Coming to America: social impact bonds make headway | The Guardian

By Ucilia Wang

Just over three years ago, a novel approach to use private dollars to improve public social programs and save taxpayer money in the United Kingdom prompted this question:

Could the same model work in the United States?

Staff from the Laura and John Arnold Foundation in Houston set out to investigate nearly three years ago and have since put up $7.7m dollars total for two such programs – both announced in the past two months – that use so-called "social impact bonds" to reshape how public agencies evaluate and pay for social services.

social impact bond is loosely defined as a public-private investment that is expected to generate a financial return for investors while at the same time improving a public service. They're a relatively new import to America: Massachusetts, for instance, announced an $18m social impact bond last month to fund an outreach and job training program for young men on probation or recently released from jail. And in December, Bank of America announced it had completed a $13.5m bond for a similar program in the state of New York.

"The bond lowers the financial and political risk for the government to try new things," said Josh McGee, vice president of public accountability at the Laura and John Arnold Foundation, a nonprofit that supports novel approaches to fixing social problems. "It's also a way to encourage more experimentation and innovation outside of the government."

Here's how the bonds work: private investors pay the upfront costs of running a program that governments typically hire nonprofits to operate. The government pays up only if the program demonstrates results, such as lowering recidivism rates among the formerly incarcerated. The better the results, the more the government will pay – and the more likely the investors will not only recoup their initial investments, but also make a profit.

If the prison program were successful, for instance, the government would save money by not having to pay for trying and imprisoning repeat offenders – and could, therefore, pay more of that savings to investors. Meanwhile, the nonprofit running the program could use its success to win more contracts. And, of course, putting former inmates to work – and preventing more crimes – also would benefit their communities.

For the current crop of social impact investors, this new financing model is mostly a way to support programs that align with their missions, rather than to make money. If the programs fall short, investors could very well lose all of their money.

For that reason, most social-impact-bond investors so far are nonprofits, which are accustomed to spending money for social good instead of for profit, and philanthropists.

Some are skeptical that businesses will buy them. Companies may be willing to accept lower returns to support worthwhile causes, McGee said, but they're unlikely to accept the higher risk of losing their entire investment.

"Truthfully, there is a dearth of evidence on what works and what doesn't work for social programs," he said. "So that makes it more difficult for traditional investors to get involved."

Variations on a theme

Even among the few social impact bonds being set up in the US today, the structures vary.

The $18m social impact bond for the Massachusetts project, for example, includes a $9m loan from Goldman Sachs' own social impact fund, which is in turn funded by a handful of other organizations. In essence, this bond acts like a collection of loans and grants.

Meanwhile, the bond put together by Bank of America more closely resembles a traditional securities offering. BoA worked with New York state officials and a few nonprofit groups to package the $13.5m bond, which it then sold privately to more than 40 foundation and individual investors.

For these investors, the point is to make their money back so they can reinvest it in other social projects; they don't necessarily expect to make money.

"What we were seeking to do was to create something that looks more like things our clients would buy," said Liam O'Neil, head of the markets group at the bank. "But this isn't something that people are doing to get rich."

The bank will likely offer another social impact bond this year, he said.

The bond for the New York program is set to mature in five and a half years, with returns capped at 12.5% per year. Massachusetts' program will last for seven years.

How to pick a social impact bond

Considering signing up? Two issues are particularly important for investors to consider when it comes to choosing social impact bonds: the track record of the nonprofit that runs the social services and the metrics in place to measure success.

McGee notes that the two programs in Massachusetts and New York will employ a tracking method often used to demonstrate efficacy in medical trials, instead of the quicker – but less detailed – reviewing process conventionally used by government agencies.

The efficacy tracking method involves creating a control group and a treatment group. Using the example of formerly incarcerated people, the control group will be made up of former inmates who aren't enrolled in the program. The treatment group will be made up of former inmates who are enrolled in the program.

An independent company will monitor the progress of these two groups and see if the reoffending rates drop for those in the treatment group for the duration of the program. The state will pay investors only if the re-offending rates fall for the treatment group, demonstrating a proven track record.

If that happens, the amount of the payments will depend on the difference in the number of days the two groups spend in prison during the term of the program.

Critics often point to government-funded programs as being both inefficient and overly expensive. The new model of using social impact bonds is a worthwhile experiment that links results more closely to spending, a metric that is more commonly found in the business world. If proven successful, it will be one import that is here to stay.