Social Impact Bonds - Approach With Caution (Part One)
Recently a flurry of media (see the New York Times, National Post and Huffington Post, for example) has touted Social Impact Bonds (SIBs) as an innovative way of unlocking new, private sources of finance for effective social programs at no risk to government. While this description may hold true, SIBs are not a panacea: there are shortcomings to the SIB model that will limit its effectiveness to specific situations.
The basic concept of an SIB is that a service provider, usually an NGO, will raise money from private investors in order to fund a social program—let’s call it the SIB program—that will result in measurable positive outcomes. If the SIB program is effective it will change the behavior of the people it is serving—they will eat healthier, stay out of trouble, etc. This will reduce the future costs of government social services—emergency rooms, police etc.
Since the government will be saving money in the future due to the SIB program, it is willing to sign on to repay the private investors based on positive program outcomes. If the program does not achieve positive outcomes—people continue to eat McDonalds, re-join a gang after leaving jail, etc.—the government does not have to pay the private investors. So theoretically the government will capture the rewards of a successful program without holding the risk of an unsuccessful program. A good example—and the first ever Social Impact Bond—is recidivism: rehabilitating prisoners can reduce future costs to the justice system.
My doubts about SIBs were initially fairly simple, largely around the financial cost of these programs from a societal perspective: why would government pay a private investor if it is fairly certain that a program will result in real cash savings? Why would it pay a higher rate of interest to private investors than it usually pays for its government bond financing?
As I researched the issue, my questions around SIBs increased.
- Innovation: Isn’t there a mismatch between private investors who have strong incentives for low-risk returns (i.e. existing, proven programs) and the oft-stated goal of SIBs to foster innovative new programs?
- Transaction costs: Won’t the transaction costs of delivering SIBs outweigh the financial benefits to society/service providers?
- Crowding out: Will SIBs result in a narrow focus on specific programs and leave other programs with more complexity or a longer-term horizon underfunded?
- Measurement and causation: Will social programs be able to clear the significant hurdle of demonstrating that they have caused a positive change?
These are significant questions, and by no means make up a comprehensive list.
Fortunately there are a bunch of smart people that have been working through these issues for years, if not in practice at least in theory. There is no final answer to the question of whether SIBs as a general asset class will pay off: as usual, it turns out the devil is in the details. I’ll address the questions one at a time.
Financial Cost and Financing Innovation
If the government ends up paying for a successful SIB program financed by private investors, it will be paying a much higher interest rate (estimated 7-15%) than it normally pays for government financing (1% cost of collecting taxes or 2-5% cost of government bonds). So why would a government sign on to an SIB?
At least three reasons, beyond the obvious one of not having to pay for an unsuccessful program: government deficits, a focus on value for programs delivered, and the need to encourage more innovative approaches to complex social problems. By avoiding up-front funding for a program, the government delays payment, an attractive option for cash-strapped governments. The focus on value for programs delivered is a positive trend in the social sector, and one of the possible good things that will come out of SIBs is an increased government focus on all program outcomes.
As far as encouraging more innovative approaches to social problems, the impact of SIBs will be limited by the financing behind them. As things stand now, I do not believe the average private investor is willing to invest in an unproven, innovative social program—this will fall upon the foundations and individuals who are already comfortable with this mix of risk and mission. Unless governments find creative ways of encouraging private investment in social impact-driven investments—like the recently-announced Social Venture Capital Trust in the UK—the scale of SIBs will be limited.
So I have serious doubts about the types of programs private investors will finance. If they only end up financing the larger, more proven programs, as I suspect, what is the point? I have no problem with the government paying back people who are risking money on an innovative program outside of the government’s current programmatic scope. But I do have a problem with the government paying out to private investors for a program that is already proving successful outcomes.
SIBs are similar in structure to public-private infrastructure partnerships (P3s), which have been shown to end up costing governments around 16% more over the long run due to the higher interest rates as well as significant transaction costs. To overcome transaction costs there must be a level of scale. Economical P3s have been shown to require around a $40 million deal size, which is a greater amount than many social programs.
Two solutions exist, beyond eating the extra cost while SIBs mature as an asset class: pool similar programs, or pool investments. Both of these options have considerable logistical challenges and may themselves lead to higher transaction costs or program execution challenges.
My second post will take a closer look at crowding out and measurement/causation challenges, as well as wrap it up with a conclusion.
Photo credit: http://www.flickr.com/photos/alexandredebie/8175955121/