Head Start was introduced to the United States over 40 years ago as part of President Lyndon Johnson’s War on Poverty. The program provides low-income, pre-school-aged children and their families with schooling, nutrition, health, and social welfare services during a crucial time of human development. The first study of Head Start’s impacts, published in 1966, just a year after the program began, argued that any benefits to children fade quickly. Skepticism about the lasting benefits of Head Start—now serving close to 1 million children at the cost of $7 billion per year—has persisted ever since.

In a recent paper published in the Annals of the New York Academy of Sciences, Jens Ludwig, the McCormick Foundation Professor of Social Service Administration, Law, and Public Policy in the School of Social Service Administration and the Harris School (with Deborah A. Phillips of Georgetown University) argues that the most appropriate framework for judging the value of Head Start is through the lens of benefit-cost analysis. Ludwig and Phillips determine that Head Start does appear to generate benefits to participants and society as a whole that are large enough to justify the program’s costs.

“These findings counter the view that only intensive and expensive early childhood interventions can generate long-term benefits,” the authors write, “and they run counter to the perception that Head Start has been a failure from its inception.”

Much of the current policy discussion about Head Start has focused on its relatively modest short-term impact on test scores and the “fade out” of these academic benefits. Head Start has been additionally criticized for generating smaller long-term effects than more expensive interventions, and, more generally, for not doing enough to eliminate achievement test score disparities either between minority and white children or between poor and wealthier children. But Ludwig and Phillips note that under the standard benefit-cost framework, the right standard for success is that programs generate net benefits, not miraculous benefits.

In assessing the benefits and costs of Head Start, the authors look at three studies showing that the program—from the 1960s through the 1980s—generates lasting gains in health, educational attainment, and possibly even reductions in criminal activity among participants. These findings come from separate studies, one by Janet Currie of Columbia University and a second by David Deming of Harvard University, that compare the experiences of children who participated in Head Start with siblings within the same family who did not. A third study, by Ludwig and Doug Miller of the University of California-Davis (published in the Quarterly Journal of Economics), exploits a sharp, persistent discontinuity in Head Start funding rates across counties (resulting from the way the program was initially implemented in 1965) and compares factors like mortality rates and years of school completed.

Taken together, these studies show that Head Start generates lasting improvements in many areas including educational attainment despite the fact that impacts on achievement test scores seem to fade out, particularly among African American children. In addition, Head Start—as operated in the 1960s through the 1980s—had a benefit-cost ratio that may be as large as 7-to-1. These studies also lead to speculation that program effects on children’s socio-emotional and behavioral skills may have been the key behind the observed long-term benefits.

The more difficult task is to determine the long-term benefits of Head Start for more recent cohorts, since all that can be observed at this point are short-term outcomes. Fortunately, the federal government has recently sponsored a formal randomized experimental evaluation, which shows that Head Start seems to increase most reading and math tests by around 0.1 to 0.2 standard deviations.

Determining what these short-term impacts imply for long-term outcomes necessarily requires some extrapolation. Ludwig and Phillips’s best guess is that even small effects might generate lasting benefits that are large enough to justify program costs, in part because Head Start had short-term test score results of around this same magnitude for the program’s earlier cohorts of children.

The authors note that the development of better methods for mapping short-term impacts into long-term benefits remains a key priority for the early education field. For that purpose, the National Academy of Sciences recently formed a Committee on Strengthening Benefit-Cost Methodology for the Evaluation of Early Childhood Interventions, funded by the MacArthur Foundation, to which Ludwig has been appointed a member.

Ludwig and Phillips conclude their paper by considering some of the recent suggestions that have been made to improve early education. Proposals to strengthen Head Start include making it more like state pre-kindergarten programs, which are more academically oriented and require teachers to hold a four-year degree. Others have recommended redistributing resources and diverting funding from Head Start to the state pre-K programs.

But, Ludwig and Philips caution that fundamentally shifting Head Start services would create uncertainty about its benefits. While funding changes could generate better academic outcomes, the effects on noncognitive skills remain unknown. “Policy actions that would shift or withdraw resources from Head Start are therefore risky,” they conclude, with uncertain upsides and some downside risk.