Skip to content

Is there a market failure in early child investment?

[This article has been published in Restoring America to highlight how increased federal spending is rarely a solution to social issues.]

During a recent Stanford event, Cecilia Rouse, chair of the Council of Economic Advisors (CEA), made the economic case for the administration’s push for higher government spending on young children.

Under the “Build Back Better” plan, federal taxpayers would finance universal pre‑K for 3 and 4‑year‑olds and extensive childcare subsidies for most families with kids. Rouse justified these programs as means of correcting market failures: if childcare and pre‑K enrollments were left to individual families, she said, parents would underspend on young children’s care and development, relative to what is socially optimal.

According to Rouse, this market failure arises partly because families ignore the broader societal benefits of high‐quality childcare and early childhood programs when deciding how much childcare or early education to purchase, such as any positive externalities associated with lower crime or future workers being more productive, for example.

Taken together with her other arguments for market failure, she believes this justifies the huge universal pre‑K and childcare subsidies proposed. Yet the evidence for this particular theory is weak, as is the idea that the administration’s mooted policies would deliver the large benefits assumed.

First, commonly‐cited studies that find large social benefits of pre‑K education are typically targeted to severely deprived children and consist of small and resource‐intensive programs that were sometimes implemented more than fifty years ago.

The administration’s Treasury report on childcare, for example, boldly states that “evidence suggests that children who participate in organized early childhood programs are less likely to commit crimes later in life.” If true, the social costs of crime would mean this would be a clear additional benefit to society as a whole.

The cited source of the claim is a report on the Perry Preschool Project, a randomized control trial that incorporated just 123 disadvantaged Black American children in the early 1960s, where the treated kids were exposed to part‐day preschool on weekdays for one or two years . Perry researchers have been criticized for using non‐standard significance levels and for using a non‐representative sample, both of children generally and potentially of disadvantaged children specifically.

To base the case for universal pre‑K for all 3 and 4‑year‐olds and hundreds of billions of dollars in childcare subsidies today on the speculative extrapolation from the results of such a small, targeted program run six decades ago is ludicrous, especially because this study does not chime with the experience of more comprehensive evaluations of recent research on large-scale pre-K programs.

For example, a comprehensive examination of Head Start, a federal preschool program for poor children, found benefits to the children themselves, but no significant effects on criminality that might benefit society. And, following the introduction of full-day kindergarten for five-year-olds and universal childcare for four-year-olds in Quebec, researchers found “striking evidence that children are worse off in a variety of behavioral and health dimensions, ranging from aggression to motor‐social skills to illness” in the short term. A follow‐up study found a “sharp and contemporary increase in criminal behavior among the cohorts exposed to the Quebec program” in the long term. This suggests, on crime at least, that universal provision of pre‐school might bring net social costs, not benefits.

There is therefore a disconnect between Rouse’s argument and the empirical evidence. It’s plausible that certain programs could improve certain societal outcomes. But Rouse’s argument for federal programs assumes they will.

Even aside from the crime issue, it would be challenging for any government to accurately determine what the full tally of social costs and benefits would be from subsidizing childcare. The programs could lead to higher parental labor market participation, for example, but this could undermine child development (and social outcomes) for those whose parents would otherwise devote extensive time to engaging their children.

Research on existing US child care subsidies finds that children exposed to subsidized child care provided through the Child Care Development Fund experienced lower math and reading scores and an increase in behavioral problems as they were beginning kindergarten. And a recent, randomized control trial (RCT) of Tennessee’s statewide pre‑K program found that children assigned to the program “had lower state achievement test scores in third through sixth grades than control children, with the strongest negative effects in sixth grade.” Negative effects were also found for disciplinary infractions and attendance.

Rouse and the Biden administration seem to just assume their policies would replicate the impacts of the small‐scale studies in delivering the benefits of “high‐quality” care for all. Yet as these other empirical analyzes prove, delivering care or early education through government to a diverse population, at scale, is difficult.

Indeed, it’s not even clear governments really know what constitutes “high‐quality” care. For example, a recent Department of Education (DOE) study that attempted to validate state program quality rating systems found that early learning and development programs that were rated highly by states were not high quality overall, and that “children attending higher‐rated programs did not have better developmental outcomes than children attending lower‐rated ones.”

Meanwhile, a 2017 study of Oregon’s quality measurement framework found that it was “very difficult” to identify specific quality measurement standards that predicted in‐classroom observed quality. According to researchers, Oregon’s quality measurement framework did not successfully differentiate “between programs rated 1 vs 2, or between programs rated 3 vs 4 or 5, or between programs rated 5 vs those rated 3 or 4.” While Oregon’s quality measurement standards have since been revised, measurement challenges in the space remain.

For these and other reasons, this particular argument of Rouse’s to justify the administration’s early childhood policies should be filed under: “speculative.” Even in its most defensible form, it might be used to support very targeted policies at a narrow range of children from disadvantaged backgrounds. This rationale certainly does not justify the huge federal expansion in childcare and pre‑K that Build Back Better would produce.

This article originally appeared in the Cato at Liberty blog and is reprinted with kind permission from the Cato Institute.


Leave a Reply

Your email address will not be published.