This article is part of Promoting Social Mobility, a forum on using early intervention to reduce inequality.
There is a strong argument that the roots of inequality are in early childhood and therefore we could use a major shift in social policy toward early intervention.
I agree with James Heckman that spending on early childhood education has a high rate of return, and that the benefits of a large increase in spending on early childhood education would most likely exceed the costs. In a perfect world, we would have high quality programs available to all children. In reality, since resources are scarce, we have to make hard decisions about the best use of public funds. A key tradeoff is whether to spend money on increasing program quality through policies such as increasing the education requirements (and thus salaries) for teachers, or on expanding existing offerings to reach more children. I worry that researchers and policymakers have become too focused on quality improvement, when all the evidence suggests that expanding access for disadvantaged children yields larger returns on investment.
The long-run benefits of early childhood education are greatest for children who need the most help and when such education replaces an environment of severe deprivation. In the Perry Preschool Project, children were selected based on low levels of parental education and baseline IQ. Only 17 percent of parents had graduated from high school. All children in the study had an IQ between 70 and 85, which the state of Michigan classified at the time as “borderline educable mentally impaired.” About half of these children received weekday classes and a weekly home visit for a total of roughly fifteen months over two years, and half received nothing. As Heckman notes, by age 40 the children in the Perry Project’s treatment group attained higher levels of education and earnings, committed fewer crimes, and were less likely to go on welfare than those in the control group.
Contrast this with more recent evidence such as the results of the Head Start Impact Study (HSIS), which found modest impacts on test scores—impacts that faded away by first grade. Explanations for the disappointing effect of Head Start relative to that of a “model” program such as Perry have focused on deficiencies in Head Start, but they could equally focus on improvements over time in the availability of alternative arrangements. In the HSIS about 50 percent of four-year-olds and 86 percent of three-year-olds in the control group attended at least a year of center-based care (often in another Head Start center) prior to entering kindergarten. Half of the three-year-olds in the control group simply waited a year and enrolled in Head Start at age four.
Programs vary in quality, but any is better than none.
The HSIS did not find that Head Start has no impact on test scores at all, just that it does not increase test scores any more (or less) than other center-based programs in a nationally representative sample. This is a crucial point that is often lost in the debate over the effectiveness of Head Start. As the quality of available alternatives increases, the impact of any program (including Head Start) shrinks. In an environment with many good programs, it is hard to stand out from the pack.
The best available evidence suggests that access to any center-based program of reasonable quality is far more important than differences in quality across programs. This does not mean that all programs are equivalent or that quality does not matter. But scarce public resources should be spent first on efforts to reach 100 percent coverage for disadvantaged children.
This could be achieved through a significant expansion of Head Start, which due to funding constraints currently serves only about 60 percent of poor children. Another approach is to offer a means-tested cash grant to families seeking to enroll their children in a preschool program of their choice, similar to the way the Pell Grant program operates for postsecondary education. In either case steps must be taken to ensure that quality does not suffer, and policymakers should aggressively seek to improve quality through cost-neutral policies such as increased accountability and monitoring.
While spending on quality improvement may be a worthwhile use of taxpayer dollars, it is unlikely to make a serious dent in economic inequality. Our best hope for meaningful “predistribution” is to ensure that no poor child is denied access to the great gift of a rich and stimulating early environment.
To comment on this forum, click here to return to the lead article by James J. Heckman.
David Deming is Assistant Professor of Education and Economics at the Harvard Graduate School of Education.
Promoting Social Mobility, a forum with James J. Heckman, Mike Rose, Robin West, Charles Murray, Carol S. Dweck, and others.