Did Subsidized Housing Ensure Healthy Development Environments for All Youth? an Application of Sequence Analysis

Background and Purpose:

Social workers routinely provide services to residents of site-based, subsidized housing. The housing literature includes work on child development (Leventhal and Newman, 2010); on-site supportive services (e.g., Popkin, Cunningham and Burt, 2005; Bratt, 2008); racial segregation and desegregation (e.g., Coulibaly, Popkin et al., 2003; Goetz, 2004; Polikoff, 2006); and comprehensive revitalization initiatives (e.g., Buron, 2004). Others examined how neighborhood conditions of site-based subsidized housing and compare to those in the private rental sector (e.g., Freeman, 2004; McClure, 2006, 2008; Talen and Koschinsky, 2011). Public housing is typically in neighborhoods with higher poverty than private low-rent apartments, or newer generations of site-based assisted developments (Galster, 2013). These studies have not addressed how families in public housing fare geographically in terms of long-term residential mobility compared to similar families who never lived in subsidized housing. Did residence in assisted housing result in distinctive neighborhood poverty trajectories?

Methods:

Design. We use quasi-experimental design to compare three matched samples of low income families to assess the impact of different types of low-income housing. The neighborhood quality outcome is generated using sequence analysis, a way to classify a population by how a categorical variable changes over time.

Participants, Sampling Methods, Data Collection. We use secondary data from the nationally representative Panel Survey of Income Dynamics (PSID) to study low-income households formed after welfare reform from 1988 to 1992. We merged the PSID with Census tract data to examine these households' neighborhood characteristics through the year 2009 when they were likely raising children.

Measures. Using data from PSID, we define two levels of policy treatment variable and control: (1) started in public housing (n=42); (2) started in privately, subsidized housing (n=58); and (3) the controls that lived solely in private, unsubsidized housing (n=661). The two treatment groups were matched to controls using individual and neighborhood characteristics. The dependent variable is the sequence category membership derived from quintile neighborhood poverty (e.g. 0-19.9%, 20-39.9%, etc.).

Analytic Approach. We convert quintile neighborhood poverty into a single character value called a sequence, classify them using cluster analysis, and rank the clusters by whether the trajectory “improves” (e.g., families move to lower poverty neighborhoods over a 20 year period). The outcomes are assessed using multinomial logit.

Results:

We identify six clusters: 1) improving low to very low (7%); 2) improving high to moderate (29%); 3) inconsistent high to moderate (16%); 4) consistently high (8%); 5) inconsistent very high to high (23%); and 6) consistently very high (17%). Our results indicate that a spell of residence in public housing leads to longer durations of residence in poorer neighborhoods, but there are no statistically significant differences in cluster membership.

Conclusions and Implications:

These results provide mixed implications for social policy. Public housing did not trap all low-income families as some have portrayed. However, other housing assistance programs did no better than traditional public housing in promoting mobility. Social workers need to advocate for all low income families to ensure healthy development for all youth.