Panel Paper: The Longer Long-Term Impact of Head Start: Intergenerational Transmission of Program Effects

Thursday, November 6, 2014 : 10:55 AM
Jemez (Convention Center)

*Names in bold indicate Presenter

Andrew Barr and Chloe Gibbs, University of Virginia
Head Start began in 1965 as a federal program targeting low-income, preschool-aged children and their families for comprehensive education, health, and parenting services. The program continues, serving approximately 950,000 children at a cost of nearly eight billion dollars per year. While findings from the recent National Head Start Impact Study—the only experimental evaluation of the program—have reignited debate over the program’s effectiveness, these conversations have existed since the program’s inception. Notably, one feature of those debates has been disagreement over what the intended outcomes of Head Start participation are. In the literature and in the policy discourse, there is considerably less focus on the impact of Head Start on the participant’s family and parents, and virtually no attention on the potential impact of program participation on that child’s future family.

 

Given quasi-experimental evidence of Head Start effects on long-term outcomes for participants (Deming 2009, Garces, Currie & Thomas 2002, Ludwig & Miller 2007), it is plausible that Head Start participation effects may be transmitted across generations in the form of improved human capital for participants’ children. In this paper, we explore the impact of Head Start enrollment in one’s childhood on a number of medium-term outcomes for that individual’s children. While there is evidence demonstrating the effects of Head Start on improving childhood development, reducing crime and teenage pregnancy, and increasing educational attainment, we know very little about the degree to which these positive effects carry over to the next generation of individuals.

 

We exploit two sources of exogenous variation to identify the intergenerational impact of Head Start participation. First, we leverage sibling comparisons to isolate the effects of an individual’s Head Start attendance on their own children as compared to their non-participating sibling’s children. We use data from the National Longitudinal Survey of Youth 1979 (NLSY79) for information on the first generation’s Head Start status, and we match that data to the NLSY79 Children and Young Adults (CNLSY) data which follows the biological children of the NLSY79 survey respondents to look at outcomes for the second generation.

 

In addition, we combine detailed information on the location of birth of individuals in the NLSY79, born from 1957 to 1964, with data on the rollout of the Head Start program during the 1960s from the National Archives and Records Administration, to estimate the effect of the availability of a Head Start program on these intergenerational outcomes. The plausibly random nature of the Head Start rollout, as well as variation in exposure by birth cohort, creates a natural experiment to compare similar individuals in the first generation with differing likelihood of participation based on availability.

 

Our preliminary results suggest that Head Start participation in the first generation translates into improved longer-term outcomes for the second generation, in the form of increased educational attainment and decreased engagement in risky behaviors. These results, if they sustain multiple robustness checks, would indicate that we understate the cost-effectiveness of Head Start and other preschool interventions when we estimate impacts on the participants alone.