Association for Public Policy Analysis & Management

In this proposed research, we investigate the extent to which lower educational attainment is a mechanism for the intergenerational transmission of food insecurity. A major path out of childhood disadvantage is the acquisition of human capital in young adulthood that allows for self-sufficiency in adulthood. To the extent that such human capital investments are particularly difficult to access for young adults emerging from food-insecure homes, we are concerned that lower educational attainment may be one mechanism by which food insecurity is transmitted from one generation to the next.

 

Using the Panel Study of Income Dynamics (PSID), we will create a panel data set linking individual data from the Child Development Supplement (containing child food security measured several times) to the same individuals in the Transition to Adulthood Supplement (containing detailed educational information) and the core PSID (containing their food security as adults). We will analyze these data by capitalizing on recent work by Flores and Flores-Lagunes (2009, 2010) that separately identifies direct and indirect effects of a treatment variable via a mechanism variable on a given outcome. This will be the first paper to apply their methods (for point-identification and partial-identification) in an intergenerational transmission context. We are, of course, also concerned about the possible endogeneity of childhood food insecurity (i.e., food insecurity in the parents’ home) in a model of adult food security, since omitted variables may cause intergenerationally correlated error terms. While the PSID is arguably rich enough to allow us to measure many of these omitted variables (as we need to do for the previously mentioned models), we also propose a more traditional IV approach in which Supplemental Nutrition Assistance Program rules that vary by state are used as instruments for food security during childhood.

 

Our findings will deepen the existing understanding of the mechanisms for intergenerational transmission of food insecurity, and in particular the role of education as one possible mechanism. This has implications for food and nutrition policy. Current food assistance programs are primarily targeted toward children and families, while young adults making crucial educational decisions are largely excluded from food assistance as they are subject to strict time limits or work requirements. If we find that pre-existing food insecurity per se is leading to less educational investment, this would have implications for both the importance of developing programs that improve child food security as well as those that maintain food access during years with potential educational investments. To the extent that we find lower educational investment driven by childhood food insecurity, we may inadvertently raise the need for food assistance later on by failing to address it in the early years, thus raising program costs. Our findings thus have implications both for the appropriate age targeting of SNAP and for efficiency in budgeting to prevent long-term SNAP dependence due to educational underinvestment.