The Obama Administration has appropriately placed a high priority on driving better performance in both the public and nonprofit sectors as they seek to tackle serious social and economic challenges. This includes a heightened emphasis on research, evaluation and funding “programs that work.” All of this is certainly welcome news. Unfortunately, as part of this effort, the Administration has embraced a new idea that I fear could take us in the wrong direction.
In the FY 2012 budget, President Obama has proposed $100 million to fund co-called Social Impact Bonds. The purpose of the Social Impact Bond is to create a new capital market that will encourage private, profit-motivated investors to fund social programs that “work.” Essentially, an investor would front the money to pay for a particular program. If and when the program achieved the designated benchmarks, the government would pay the investor for the cost of the program plus enough to cover their risk and earn a profit. The theory is to create a market where investors can make money by investing in effective social programs.
Sounds good right? The government only pays if the program works. Taxpayer money is no longer wasted on ineffective programs. Private sector discipline and oversight is brought to bear on the nonprofit sector! But wait! Is this really going to work? Is creating a Wall Street-like investment market the best way to strengthen nonprofit performance? Judging from the performance of Wall Street in recent years (have you read The Big Short?), one has to at least ask the question!
While I understand the appeal of this idea – and would welcome new sources of funding for high performing social programs – we need to carefully examine the implications of this approach to funding non-profit social service programs. I see many serious questions that advocates of these bonds need to answer before tax dollars are invested.
First, Social Impact Bonds are likely to encourage functional specialization and silo-thinking. Money will flow to programs that can hit a single key benchmark so programs will be designed to do just one thing – achieve that benchmark. This will take the idea of “teaching to the test” to an entirely new level. Programs with multiple positive social impacts will be undervalued (like housing which provides economic, educational, health, public safety, and quality of life benefits) while programs with high externalities will be overvalued just as they are in the private economy (i.e. programs that improve outcomes in some areas at the expense of others.) What makes this all the more puzzling is that in other programs, the Obama Administration is promoting comprehensive, placed based approaches that seek multiple quantitative and qualitative outcomes through multiple interventions. I don’t understand why they would support a new funding scheme that drives in the opposite direction.
Second, will these bonds discourage collaboration because this funding approach places much more emphasis on which nonprofit should “get the credit” for the “success? If a student’s test scores improve, who should get the credit and therefore the money? The school? The tutor? The afterschool program? The social worker who helped the student deal with early life traumas? The parents? The next door neighbor who helped with math homework? The Little League that gave her an opportunity to grow and mature and have fun? The library where she did her homework and used a computer? How much time and money do we want to spend sorting through all the possible reasons and their relative impact on the child so we can sort out who gets how much money? Does that create a collaborative culture?
A third potential problem with this approach is that it may over-estimate our ability to identify the correct metrics and to measure them correctly. Long term data is very difficult to secure and causal relationships can be very hard to discern. In our quest for market clarity, we are certain to over simplify and choose metrics that are easy to collect, count and standardize, even if they don’t tell us the full story or even an accurate story. This is particularly true when attacking complex, systemic problems such as educational and health disparities, environmental justice, or land use development.
A fourth problem is likely to be cream skimming. Organizations will have a powerful financial incentive to cherry pick the best clients that maximize their chances for hitting performance benchmarks. Supporters say that they can guard against this, but history shows that powerful financial incentives work – they will motivate performance but they will also motivate people to “screen” clients before enrolling them in programs. And now we will have powerful investors pushing in this direction!
Fifth, advocates say these bonds will promote innovation, but I think it is much more likely that investment dollars will flow to safe, well-known, programs. New, untested ideas, will have a very hard time attracting investors – and those who are attracted may seek returns that taxpayers cannot afford.
I also wonder whether a Bond Market can really think about long term solutions. Will investors be willing to wait 5, 10 or even 20 years to see transformative impact? Or will they only be interested in programs that can achieve benchmarks within 1 or 2 years. And if we are looking at long term impacts, do we have the ability to measure impact and causation sufficiently well to ensure that the right programs are getting paid? And how would we control for macro economic impacts that may cloud our ability to see the impact of individual programs?
Finally, a Social Impact Bond Market will undoubtedly become a massively complicated system as investors seek to bundle investments, guard against losses, shift risk to other parties, scale up, and otherwise replicate traditional investment markets. Lawyers, accountants, advisors and intermediaries will be needed and they will all need to be paid (handsomely, no doubt.) A program designed to save taxpayer money could easily end up costing far more. Remember, this program will leverage private capital, but at the end of the day, the taxpayer must pay for all the costs plus the profit or return. There is no free lunch and if investors start to lose money the bond market will dry up very quickly.
Everyone wants to fund “what works” and no one wants to pay for programs that don’t achieve real outcomes. But the world is complicated and it does not help to pretend otherwise. Positive social outcomes are usually the result of multiple interventions, programs and causes – including some that operate at the macroeconomic level that is far beyond the scope of a single nonprofit program. And performance metrics can only capture so much. When a new playground opens it may significantly improve the quality of life for children and families nearby. But will a new playground translate into measurable and statistically significant reductions in obesity? Or increases in family incomes? Or improved educational attainment? Probably not. Does that mean we should stop "wasting" money on playgrounds because the "don't work?"
We absolutely need to fund programs that work. But my concern is that program evaluation is too important for us to dumb it down into numerical measures that Wall Street investors can understand but don’t tell us the true story of what is happening in our communities. Instead, we need to build capacity in the nonprofit sector to conduct robust and meaningful program evaluation and to take advantage of new information technology that can improve our understanding of program impacts. And we need to fully fund those activities. I fully support public-private partnerships and programs that leverage private investment (MACDC has filed its own legislation designed to do just that) but we have to be very careful how we design these programs less we suffer serious unintended consequences.