By Robert Milburn
Just two weeks ago, it was announced that the U.S.’s first social-impact bond — a fixed-income like investment used to finance social good — failed to produce the desired financial and philanthropic returns. As a result, the five year investment was scrapped after three years.
The laudable $7.2 million deal, inked in August 2012, aimed to reduce youth recidivism at New York City’s Rikers Island prison by as much as 20%. Had the program hit this target after five years, the New York City government would have paid out a 4.4% annual return to investor Goldman Sachs. At the same time, it cost New York City nothing to sign off on the intervention, since Goldman Sachs was providing all the upfront capital and returns were to be paid out of savings generated from not sending these individuals back to jail.
Instead, Goldman Sachs took a $1.2 million hit, after former New York City mayor Michael Bloomberg’s family foundation paid out $6 million in “first-loss” capital.
“We’re all disappointed that we weren’t able to have a positive impact on the young people at Rikers, but ultimately the social-impact bond model proved successful by allowing the City to test and collect data on a new approach to recidivism at no cost to taxpayers,” says Andrea Phillips, vice president of Goldman Sachs’ Urban Investment Group, which made the investment. Nearly half of the youth discharged from Rikers Island each year return within 12 months.
Juvenile justice experts are currently crunching the numbers to see what went wrong on the deal, but Penta called up Phillips to learn lessons from the Rikers Island experiment, and how this knowledge might alter the fledgling social-impact bond industry, currently consisting of just seven deals and $80 million invested to date.
Goldman Sachs has invested in four of the seven impact bonds and, including the failed Rikers deal, the bank has committed $30 million to the nascent field. Goldman’s wealthy, philanthropic-minded clients are also in on the action. Such clients have invested in two social-impact bonds. One is a $17 million deal to improve pre-kindergarten literacy in Chicago; the other, similar to the Rikers investment, provides $19.4 million in funding for the Massachusetts-based nonprofit Roca to tackle youth recidivism to the state’s prisons.
So what’s changed with the new generation of impact bonds? A few things. The Rikers program worked successfully at other prisons, but not when imported to New York City’s largest jail. That’s focusing minds. The newest impact bonds are scaling up existing and successful programs already working in a particular community, Phillips says. The Massachusetts bond, for example, engages nonprofit Roca, which has been working in the Greater Boston area for 27 years and has an independently-verified track record of reducing recidivism in Boston’s prisons by 33% over one study’s three year period.
Also, the Rikers deal was defined in a very “binary way,” Phillips says, with one payout at the end of the intervention. That’s different from the way the next wave of deals are being structured. Take Goldman’s $7 million investment in Salt Lake City, Utah, to improve early childhood education in underprivileged neighborhoods. The impact bond funds a preschool program that delivers a targeted curriculum to 3 and 4 year olds. A total of 750 students were given a standardized test to assess whether they are likely to use special education or remedial services. A payment will go to co-investors Goldman Sachs and J.B. Pritzker, second-generation offspring of the family that founded the Hyatt Hotels, for each of the Salt Lake City students identified as at risk but that don’t use special education for each year in kindergarten through 6th grade. Getting rigorous data points early on and for the life of the investment, to quantify results and reassure investors, is also a key feature of the more recent bonds. The Salt Lake City bond runs for 12 years; Massachusetts, 6 years; so investors will be regularly fed benchmark data to reassure them the intervention is on-track.
“In each of these new deals, we are consistently looking to learn from prior investments to make stronger structures and stronger investments,” Phillips says. Some smart money thinks all the lessons learned and data gathered will materially move the needle. The Rockefeller Foundation estimates that the total market for U.S. impact bonds will quintuple to $500 million this year, based on the federal government money already flowing to states to finance similar deals. So the value of the Riker’s Island bond – the intelligence it passed on – could not have come at a better time. We usually learn more from failure than we do from success.