Inadequate housing threatens the safety and well-being of children and is frequently at the root of child welfare involvement, out-of-home placement, and reunification delays among low-income families. While past research clearly demonstrates these links, none has looked empirically at the impact of housing vouchers on reducing child welfare involvement and improving housing outcomes. We are conducting a study that investigates this impact as seen in the Family Unification Program (FUP), which provides housing vouchers for child welfare involved families.  The goal of this evaluation is to study the impact of FUP vouchers on child welfare outcomes including prevention of removal, time to reunification, and subsequent maltreatment.  This paper reports on impacts of the FUP program in San Diego, Portland, and Salt Lake County, three of the four sites in our study.

We use quasi-experimental designs to estimate the impacts of FUP vouchers on child welfare outcomes in each site.  In San Diego, we use a waitlist to form a comparison group. From July 2010 to June 2013, more than 240 families were referred by their caseworkers for the 100 FUP vouchers available. In Portland and Salt Lake County, we use propensity score matching to identify comparable families in the child welfare system. We conduct both an Intent-to-Treat analysis and a Treatment-on-the-Treated analysis to determine the impact of FUP vouchers on child welfare outcomes including prevention of child removal, time to reunification, time to case closure, and subsequent maltreatment. Given the subjective judgment of the referral process, we account for possible caseworker-specific effects (Doyle, 2007).  For families that receive FUP vouchers, we also examine whether they maintain their housing.

San Diego has a low lease-up rate compared with Portland and Salt Lake County, probably reflecting site differences in the provision of assistance to help families through the application process and in the search for housing. We analyze potential predictors for successful lease-up including the family’s original housing situation and potential caseworker-specific effects.  In San Diego, due to the low lease-up rate for the first 100 families, 48% of those on the waitlist were eventually referred to the Housing Commission for vouchers.  Following Angrist, Imbens, and Rubin (2006), we instrument for the original group assignment to account for the families on the waitlist that get referred to FUP.