Recently the New York Times reported that Scott Stringer, the Manhattan borough president, is proposing a social impact bond to address early education. Today the NY Daily News published an op-ed by Stringer with high-level outline of his proposal. Here are the details of Stringer’s proposal:
Using a Social Impact Bond structure, an investor could fund 1,000 Early Head Start slots for New York City children. If the program helps to foster thriving, successful students — as defined by metrics the city would establish in collaboration with teachers, principals, community organizations and other stakeholders — the investor will profit and the city will save money in the long-term.
Early Head Start is a great program and social impact bonds are a hot innovation. But this does not imply that they are right for each other. Here are some questions that are important to understand whether social impact bond are the most appropriate financing mechanism for Early Head Start.
1. Does the proposed program produce results?
Yes. Early Head Start is a well-evaluated program that has shown results. Some evaluations can be found here: http://ccf.tc.columbia.edu/earlycare01.html
2. Would results-based financing improve program outcomes?
The answer to this is unclear, although better outcomes are certainly possible. If the city can improve the social outcomes of this program by conditioning the payment on program outcomes rather than on adherence to the program model, then there is a policy rationale to fund existing Early Head Start through a social impact bond.
The other possibility is that the program is underfunded. In this case, a social impact bond may introduce additional funding to the program. Even if social impact bonds produce the same outcomes as have been seen in the evaluation, the ability to expand coverage through SIBs would create a policy rationale for the program.
2. Do improved outcomes create lower costs for taxpayers?I believe this is not the case for Early Head Start. Several early childhood development programs create significant social benefits. But those benefits do not translate to reductions in government costs in the short run. Over the long run, a person enrolled in an early childhood development program may commit less crime, attend better schools, become a happier and more productive citizen, and so on. But the government – and taxpayers – would not see these benefits in 3-4 years after the program.
If this program improves outcomes, but does not create cost savings, then the government may still have a policy rationale for funding it through a social impact bond if other funding is unavailable. However, the premium the government would have to pay for this program in a social impact bond (to attract investors) will have to be weighed against the potential for improved outcomes that the incentive-based program creates.
Michael Belinsky 