Health, Neighborhood Context, and Mobility

Neighborhood context may be an upstream cause of economic mobility, yet few neighborhood studies are experimental, weakening causal inference and limiting policy translation. Our study proposes to analyze newly- available data from a large, federal government initiated social experiment of voluntary neighborhood relocation using housing vouchers in 5 US cities (the Moving to Opportunity, MTO, Study). Our project will test whether, how, and among whom random assignment of an offer to move to a lower-poverty neighborhood influenced the economic mobility of 4600 low-income families over a 15 year period. The goal of MTO was to interrupt the cascading effects of neighborhood poverty for minority low-income families, by increasing area- based access to opportunities, and thereby promote economic mobility. Although the MTO voucher intervention did not consistently affect economic outcomes, it did profoundly affect health. Unfortunately, health researchers have had limited access to MTO data, compromising the scientific payoff of the $70+ million investment. So as of yet, we have no clear understanding of why results might differ across outcomes, groups, or what mechanisms are at play, since few studies have tested mediation or moderation. Since health is one of the most important barriers to economic self-sufficiency among low-income parents, it is conceivable that economic and employment gains were concentrated among families who first experienced health improvements. However testing whether factors that occurred after random assignment are precursors for other outcomes is methodologically challenging. Fortunately, recent methodological developments in the epidemiology discipline offer cogent solutions to model this complexity and bound any bias. We propose secondary data analyses with newly available 15 year follow-up data from this experiment, to test whether health influenced the economic effects of the randomized MTO housing voucher treatment. Paired with this experimental design, we propose to apply innovative causal methods and machine learning techniques for assessing mediation and effect modification, to overcome limitations and potential bias of traditional approaches, and strengthen causal inference. Our R01 project builds on a productive, interdisciplinary team, experienced with the MTO data, and draws on a barriers-to-employment framework. We propose 4 aims to determine: whether the effect of MTO on economic outcomes was modified by baseline health vulnerability; whether MTO improved parental economic outcomes, if they or their children experienced health gains; whether MTO improved children?s employment and education, if they experienced health gains; whether MTO improved economic outcomes, if families moved to neighborhoods with fewer spatial barriers to employment. Our findings have direct policy relevance since they address a key policy question in housing voucher implementation: incorporating elements from non-housing sectors (e.g., health, employment, education) to promote moves to higher opportunity neighborhoods, to ultimately improve outcomes for low income families.