Goldman Sachs paid to expand pre-K in Utah. It worked.

October 19, 2015
Author: Libby Nelson
http://www.vox.com/2015/10/19/9568639/social-impact-bonds-utah

There's something for just about everyone to feel good about in Salt Lake's pre-kindergarten program.

The state, with money from Goldman Sachs, expanded the program to serve more children from low-income families. In the end, Utah saved money, investors made money, and nearly every student who participated ended up not needing special education in kindergarten.

The Utah pre-K experiment is worth paying attention to for two reasons. It's another piece of evidence that good-quality early childhood education can save money in the long term. But it's also the result for a new, mostly untested way of paying for social programs known as a social impact bond: Private entities put up the money upfront, and the government only pays if the program gets the results it wants.

Social impact bonds are a relatively new idea, but they're spreading quickly. About 40 projects are underway, seven in the US. The Utah pre-K program is, so far, the first American success.

How the Utah pre-K program worked

The pre-kindergarten program, which was free to low-income children in the Granite School District in Salt Lake County, had already proven to be good at getting disadvantaged kids ready to start school. A three-year study of the program looked at 737 kids who participated between 2006 and 2011. More than 200 were expected to end up needing special education in elementary school; only 11 of them did.

But the half-day program, which cost $1,500 per student, had a long waiting list. In 2013, Goldman Sachs and the Pritzker Family Foundation put up $7 million to expand a pre-K program run by the United Way of Salt Lake.

The deal was that if the program delivered the results it promised and kept kids from needing special education, the state would pay most of the savings — about $2,600 per student — to the investors. If the program failed, the nonprofit Pritzker Foundation would help cover Goldman's losses.

The loan allowed Utah to expand the program to 595 more 3- and 4-year-olds last year, about half of the waiting list. All of them were from low-income families, and 110 were expected to need special education in elementary school. But this year, when students were tested in kindergarten, only one of them did.

That saved the state $281,550, according to the United Way. Investors will get a payment of $267,473, with additional payouts to follow if the students continue not to need special education in grade school. And the state now has proof that spending more on early childhood education still saves money, because a year of pre-K is cheaper than a year of special education.

This is the first time a social impact bond has worked in the US

It would have been even more cost-effective for Utah to expand pre-K itself. That way, the state could have kept all of the savings rather than paying most of the money to Goldman Sachs. But it also would have been riskier; if the program hadn't worked, the state would have spent the money on pre-K with nothing to show for it.

That's why social impact bonds are so appealing to policymakers. The bonds offer many of the benefits of new policies with none of the risk.

It can also provide a way for states and local governments to try things that might be politically difficult, said Allen Grossman, a professor of management at Harvard Business School who has researched social impact bonds.

"This has the other and very important benefit of removing it from all the vagaries and sclerotic dimensions of the bureaucracy," he said.

And so social impact bonds are spreading. In the past five years, the method has started experiments with new policies to lower unemployment for migrants in Brussels, find homes for children in need of adoption in the United Kingdomreduce recidivism for young men who served prison terms in Massachusetts, and help children in Saskatchewan avoid foster care.

But while the Utah program is a success, other social impact bonds have faltered. Either they haven't delivered the results they promised, or they weren't given a chance to play out fully.

The first social impact bond project started in Peterborough, in the UK, in 2010 and was meant to reduce recidivism among prisoners. Early results were promising but not spectacular: The program was effective enough to continue, with repayment expected in 2016, but it wasn't effective enough that investors were repaid immediately. Then the UK changed its policies for dealing with prisoners and recidivism, and the project was cut short.

In the US, meanwhile, Goldman Sachs and Bloomberg Philanthropies put up the money for a program meant to reduce recidivism for juveniles held at New York's Rikers Island prison. But it didn't work — an evaluation found no difference between people participating in the program and a control group — so the city killed the program.

Supporters of social impact bonds tried to portray this as good news: The city hadn't wasted any money on an inefficient program, and once the program was found not to be effective, it was discontinued quickly. But the news out of Utah is much better.