Leave It to the Private Sector

Studies certainly show that more than just genetics affects children's success. Research by David Armor, for instance, reveals that factors such as infant nutrition, cognitive stimulation, and the number of children in a family significantly affect a child's IQ. Even Bell Curve coauthor Charles Murray admits that "maybe we can move children from far below average intellectually to somewhat less below average." His concern is that "nobody claims that any project anywhere has proved anything more than that."

Which brings us to the central question: What can be done to optimize the outlook for children who "by accident of birth" do not have sufficient access to crucial resources?

Heckman relies primarily on two efforts — the Perry Preschool and Abecedarian programs — to illustrate that early childhood interventions can have lasting, positive effects. But are the effects meaningful in an absolute sense, rather than just in comparison to control groups, and can they be replicated on a large scale?

The long-term Perry results are decent, but not great. As Heckman reports, at 40 years of age, 29 percent of the Perry treatment group earned at least $2,000 monthly in 2004 dollars. That still-small percentage beat the control group, but $2,000 monthly — $24,000 a year — fell well short of 2004's nearly $34,000 per-capita income. Similarly, 29 percent of those treated had never been on welfare as an adult, but that means 71 percent had.

Then there's Perry's minute size and appreciable cost: Only 58 people were treated, getting 2.5 hours of preschool each weekday and a 90-minute weekly home visit by a teacher. The estimated cost per student in 2012 dollars was $12,506.

Abecedarian involved 111 subjects, 57 of whom were treated. The services started at infancy, addressed dietary and hygiene needs, and provided year-round, full-day preschool.

Abecedarian's effects as subjects hit 30 years of age were recently assessed, and outstripped Perry's. For instance, 23 percent of the treatment group graduated from a four-year college, below the national rate of 32 percent for 25-29-year-olds, but not bad. A calculation of household income put the treatment-group average at middle-class, but that was based on self-reported data and included welfare benefits. On the flip side, 27 percent had been convicted of a crime, well above the likely — though admittedly hard to calculate — national rate of roughly 5 percent.

The price tag? $17,731 per person in 2012 dollars, or 47 percent more than the average current expenditure per pupil in K-12 public schools.

Perry and Abecedarian had positive effects, but the effects were often relatively small, and questions of validity remain.

Perhaps more important, even these minor effects probably can't be replicated on a large scale, at least through government action. The record of the largest-scale early childhood program — Head Start — is informative, as is a recent attempt to bring class-size reduction to all of California.

Head Start is the federal government's primary early childhood program, with a budget of almost $8 billion. According to its most recent assessment by the Department of Health and Human Services, it has almost no lasting, positive cognitive effects, and its few, persisting social-emotional impacts are mixed positive and negative. It also suffers from widespread management problems, with federal officials struggling to keep tabs on providers and hesitant to dock poor performers. What seems to have kept it alive is advocacy by providers and widespread support for its mission.

California's class-size reduction illustrates the huge constraints on taking resource-intensive programs to scale. Inspired by the successful Tennessee STAR experiment, California undertook statewide class-size reduction in the 1990s. The effort failed, producing no conclusive achievement gains while creating a major shortage of qualified teachers. California simply couldn't staff all the new rooms.

National or statewide efforts to multiply microscopic yet dense early childhood programs would almost certainly be crippled by similar resource shortages. As a result, and because the positive effects of Perry and Abecedarian aren't overwhelming, the state and federal governments shouldn't try to recreate them. Instead philanthropists, charitable organizations, and for-profit businesses should further study — and expand — early childhood interventions.

Heckman suggests that private action should "augment" public, but private entities alone should do this because, unlike government programs, when voluntarily funded efforts fail people can easily remove their assistance. Decentralization would also minimize reliance on efficiency-killing bureaucracies and escape the very real political problem of agencies becoming advocates for — rather than supervisors of — the providers they are funding.

Clearly no one's destiny is completely locked in their genes, and early childhood interventions can have positive effects. Much less clear is how to implement truly powerful interventions for everyone who can benefit from them.