Association for Public Policy Analysis & Management

Over the last decade, a large body of work studying the roll out of the safety net – with a focus in particular on programs which were created or greatly expanded during the War on Poverty – has given us evidence about short-run and particularly long-run effects positive effects of these programs. During the mid or late 1960s, the Food Stamp Program, Medicaid, Medicare, the School Breakfast Program, job training programs, and Title 1 were all created or made permanent. Other programs such as the National School Lunch program and AFDC were expanded or altered. The Office of Economic Opportunity made grants to family planning clinics, Head Start programs, and community health centers. And this interacted with civil rights legislation requiring recipients of federal funds not to discriminate. Finally, several important parts of today’s safety net followed in the early 1970s-Supplemental Security Income, the Earned Income Tax Credit, and WIC. This emerging literature focuses on the roll out of many of these program, comparing places where the programs were implemented early to those where they were implemented later, controlling for national shocks and time-invariant differences across locations. We take advantage of variation from this existing literature and combine it with rich administrative data on earnings and use of SSI and SSDI benefits – the Continuous Work History Sample. Thus, our key independent variable is the share of time an adult was exposed to the food stamp program from birth through age 5, exposed to the Head Start program at ages 3 to 6, or their interaction. Our key dependent variables is a measure of earnings from the administrative records. These data are taken from earnings histories of a one percent sample Social Security Numbers ever issued. Restricting ourselves to natives, born between 1955 and 1980 where we can link their place of birth to the data on the rollout of SNAP and Head Start, we have a sample of nearly 1 million individuals. We model annual earnings as a function of dummy variables for year of birth, calendar year, county of birth, for being white, and stratify by gender. For women, living in a county where Food Stamps were available for the entire time from conception to age 5 leads to an increase in total FICA earnings of around 3% at age 34 and a reduction in participation in SSI while on SSDI of about 6%. Effects for men are more varied. The point estimates for effects of Head Start are positive but findings are more mixed. Estimates of the interaction suggest having been exposed to both may be valuable. These findings suggests important positive long run effects of the Food Stamp program and important program interactions.