Association for Public Policy Analysis & Management

Health is one of the most important barriers to employment among low-income families. Without addressing health and other documented barriers to work, policies intended to impact employment may not be reaching their full potential. In this study, we leverage newly-available long-term follow up data from the Moving to Opportunity (MTO) housing voucher experiment, to document low-income mothers’ health barriers, estimate the impact of health barriers on employment and welfare-to-work transitions, and estimate whether the well-known positive impact of MTO on health can lead to better employment outcomes among participants at the 15-year follow up. As policymakers consider new work requirements to maintain eligibility for social programs, it is essential to understand how health barriers to work affect program participation.

The MTO study randomly assigned 4248 low-income families living in public housing in one of 5 cities, to receive one of three types of housing assistance: (1) a housing voucher to move from public housing to subsidize rent in low poverty neighborhoods (<10% of tract in poverty) along with housing counseling; (2) a housing voucher to move and subsidize rent in any type of neighborhood; (3) continued eligibility to live in public housing (in-place controls). In addition to baseline data (1994-1998), two follow ups were conducted in 2001-02 and 2008-10, measuring employment, earnings, income, program participation, as well as health (mental health, physical health, substance use). The MTO sample was comprised of low-income single-mother headed households and their children. At baseline, 72% of mothers were not working, and 75% were receiving welfare (AFDC or TANF). Prior MTO analyses have analyzed the main effects of the housing voucher trial to understand whether offering voucher-based housing assistance (compared to remaining in public housing) can promote employment, with mostly null results. However the pathway from housing to employment is complicated by high prevalence of barriers to work. Prior research has not incorporated the barriers to work framework to understand if more complicated patterns emerged.

The specific aim of our study is to test whether the MTO housing voucher experiment improved mothers’ employment outcomes, only if they could overcome health-related barriers to work first. We hypothesize that the MTO housing voucher treatment was only effective for promoting employment and welfare-to-work transitions, if families did not have barriers to work that could not be overcome via housing vouchers.

This study uses novel Controlled Direct Effect (CDE) causal mediation methods to test our hypotheses to accommodate mediators that are simultaneously moderators, (e.g. barriers to work). Our preliminary analyses support our hypotheses that health as a barrier to employment, was a modifier of the MTO housing policy, on economic outcomes; MTO caused health improvements, and MTO improved economic outcomes for healthy participants, but adversely affected economic outcomes of unhealthy participants. Given the experimental design is such a strong design for causal inference, and since the housing voucher program matches onto existing largest US federal housing policy, this housing policy provides an important window into how to change barriers to work to promote employment.