By KENNETH A. DODGE
The most intriguing innovation of the century in public financing of social services is suddenly at risk of being flushed down the drain.
“Pay for success” programs, otherwise known as social impact bonds, are a way for government to secure private investment in social programs and to minimize risk because payback with interest to the investor occurs only upon proof that the program achieves successful outcomes. The first Pay for Success contract, for juvenile delinquency diversion at Rikers Island, failed to show success, and the government did not pay investors for the program.
The second such contract, for prekindergarten in Utah, evened the score as the state government recently began to pay back investors. But according to a recent article in The New York Times, critics have argued that the way success is being measured in Utah may be flawed — that children may not actually be helped and that the government may have made a mistake.
I suggest that we not throw the baby out with the bath water here. There are numerous overlooked benefits of the Pay for Success innovation that we should emulate.
The measurement of success in the Utah contract was based on the outcomes of a group of 4-year-old children who had scored very poorly on a vocabulary test before they received free prekindergarten and presumably would have been at elevated risk for later placement into costly special education. By the terms of the contract, the investors receive financial compensation for every one of these children who progresses through elementary school without being placed in special education. If fewer than 50 percent are ultimately so placed, the investors will be paid back in full and could earn a positive rate of return on their investment.
Critics argue that no one really knows what proportion of these children would have been placed into special education had the program not been offered, and perhaps the payback algorithm is tilted in favor of the investors. Scientists will agree that the counterfactual rate is indeed not known for this group of Utah children. Only a randomized controlled trial could provide a more rigorous way to measure impact, but that option was unfeasible in this case. So no one knows for sure, and the parties in the contract generated their best estimate. I believe the best available science indicates their estimate is reasonable.
Rather than be cynical here about investors who are trying to perform a social good, I point toward a growing body of rigorous science that shows that high-quality prekindergarten can lead not only to decreased use of costly special education services but can also improve later educational performance, decrease costly grade retention and improve life skills that could bring a wide array of positive benefits. These unmeasured benefits could make the State of Utah look like a shrewd contractor.
Of course, we cannot know for sure in this case. A bonus is that the contract limits Utah’s risk by ensuring that it will not double-pay for a child’s prekindergarten and special education expenses.
Other benefits become obvious if one contemplates Utah’s alternative, which is what is happening elsewhere. In other locales, prekindergarten programs are being carried out and financed by governments without any effort to measure long-term outcomes.
So the Utah case is a step in the right direction because it shines light on outcomes rather than merely documenting implementation. Children are being tracked at least through sixth grade, which will lead officials to improve their programs if and when improvements are needed. This is a huge advance for evidence-based policy-making.
The Utah program provides an early example for encouraging innovation by measuring outcomes and shifting government’s risk on to private investors. It affords a win-win-win opportunity for taxpayers, private investors and children.
With any innovation, early steps often seem crude and inexact. That may be the case in Utah with the lack of a precise counterfactual to measure impact. But try to remember the cumbersome first personal computer or the first cellphone. I expect the measurement of impact in Pay for Success contracts will keep improving, and I hope that each future contract will look more like a sleek computer tablet. I also hope that public policy makers will not shy away from future Pay for Success ventures because they are unsure about the prototype. We should celebrate rather than castigate Utah’s pioneers.
Kenneth A. Dodge, Ph.D., is the William McDougall Professor of Public Policy and Director of the Center for Child and Family Policy at Duke University in Durham, N.C.