By Ian Galloway
Editor’s Note: This blog is one of a series that offers in which the authors respond in the areas of their expertise to the Brookings paper “The potential and limitations of impact bonds: Lessons from the first five years of experience worldwide.”
I recently read Joe Nocera’s column on Congress’s market-based solution to acid rain in the late 1980s with great interest. That novel policy, known as cap-and-trade, created a financial incentive for power companies to curb their production of sulfur dioxide, ultimately leading to a 76 percent reduction in emissions since the legislation passed in 1990. “The brilliance of the scheme,” according to Nocera, was “that while [cap-and-trade] set emissions targets, it did not tell power companies how to meet those targets, allowing them a great deal of flexibility.” That flexibility led to industry-wide innovation and, ultimately, a solution at scale: the all-but elimination of acid rain in the United States.
I was reminded of this success while reviewing Emily Gustafsson-Wright, Sophie Gardiner, and Vidya Putcha’s compelling new review of the global impact bond landscape. Gustafsson-Wright and her colleagues ably capture the current state of the industry—alternately known as pay for success financing or social impact bonds—and the important implications of shifting social sector resources towards an outcomes-based funding projects, like impact bonds. In particular, I was drawn to the study’s treatment of several common claims about impact bonds—10 in total, ranging from their net effect on overall giving patterns to their ability to reduce risk for government and sustain impact over time. But two claims stood out for me, in particular: the ability of impact bonds to 1) achieve scale in the social sector; and 2) foster innovation in service delivery. Of the 10 claims, only these two were found to be largely inaccurate.
I was struck by this. Like a lot of people, I had long assumed that impact bonds would follow the same trajectory as cap-and-trade, achieving scale through innovation. But as I reflected on that view in the context of the study’s findings, I’ve come to the conclusion that something has gone wrong here and that there is a real difference between how cap-and-trade was used to solve a difficult technical challenge and how impact bonds are being used to solve difficult challenges in the social sector.
Take the Rikers Island recidivism reduction Social Impact Bond (Rikers SIB), for instance. The Rikers SIB is famous for both being the first impact bond in the United States and for being, as of July 2, the first impact bond to be discontinued for lack of results. While critics of impact bonds have pointed to the Rikers SIB as evidence ofgovernment funded failure, the opposite is actually true. The Rikers SIB worked as designed. A program was tested, it failed, and the private sector paid the price for having invested in an ineffective intervention. New York City paid nothing. But why did the program fail? Several reasons have been offered: its effectiveness had only been tested in older prisoners, not younger ones like those treated in the Rikers SIB project; there was well-documented turmoil at the prison outside of the service provider’s control; and the provider faced unexpected budget cuts related to enrollment shortfalls (Eduardo Porter’s column, also in The New York Times, does a great job summarizing these challenges and their policy implications).
But I think there’s another culprit here: fidelity to the model. A best practice in clinical care, I’m convinced that fidelity, in this case, may have doomed the project overall. “All they were testing is whether Moral Reconation Therapy [the model] by itself would make a difference, not whether something you could do in a jail would make a difference,”according to the program’s director Elizabeth Gaynes. “Even if we could have raised money to do other stuff, we were not allowed to because we were testing Moral Reconation Therapy alone.” Fidelity to the model, in other words, trumped the outcome the model was designed to produce.
This is the crux of the issue and, in my view, where many impact bonds have veered off course. Cap-and-trade succeeded because of a two-pronged strategy: set an emissions goal and give power producers the flexibility to innovate their way to the finish line. The Rikers SIB was different. It set a goal (reducing recidivism) but insisted on a single pathway to achieving it. That isn’t a recipe for either innovation or scale. And the Rikers SIB is by no means unique in this. A constellation of factors, including public sector reluctance to relinquish control over program selection, investors’ unwillingness to invest in untested interventions, and the rigidity of the evaluation process likely all contributed to why Gustafsson-Wright and her colleagues found so little innovation or scale in the field today.
The Rikers SIB was an important step towards a more sustainable impact bond industry. After all, the performance feedback loop worked and a program was held accountable for real results. I hope, however, that we learn the right lesson from Rikers and start innovating, not replicating, our way to scale.
Ian Galloway is a senior community development research associate and field manager at the Federal Reserve Bank of San Francisco. His views are his own and may not reflect those of the Federal Reserve Bank of San Francisco or the Federal Reserve System.