Association for Public Policy Analysis & Management

The most rigorous research on the causal effects of assisted housing on children’s outcomes finds no such effects.   In this paper, we use longitudinal data from the Panel Study of Income Dynamics, enriched with Census, American Community Survey and U.S. Department of Housing and Urban Development administrative data, to unpack these nil effects.  The analysis sample includes 194 children ( age = 6.2 years) living in assisted housing in 1995 or later who were 13-17 years old in 2002 or 2007, and an unassisted comparison group of 215 children who were income-eligible for, but never received, housing assistance.  We first estimate the effects of living in assisted housing during childhood on adolescent cognitive, behavior, and health outcomes, addressing selection using propensity score matching and instrumental variables.  Finding no effects on these outcomes at the mean, we then estimate quantile regressions and uncover heterogeneity in cognitive and behavior outcomes.  Assisted housing appears to provide an added boost for children with the best cognitive performance and fewest behavior problems but has just the opposite effect on children with the lowest cognitive scores and most behavior problems.  Further tests indicate that these differences are not explained either by neighborhood effects or housing quality effects.  A potentially fruitful avenue for future research into the heterogeneity of adolescent outcomes is differences in how parents take advantage of the housing affordability that is provided by assisted housing to benefit their children.