A new way to attack social problems: ‘pay for performance’ projects

The cornerstone of criminal justice reform is the belief that offenders leaving prison could be prevented from committing new crimes and getting locked up all over again, if only government could find the right social service organization to provide the right programming. Crime would drop, some prisons could close and taxpayers would save money.

First, though, officials have to identify rehabilitation programs that work, and that means evaluating claims and evidence offered by competing providers, and perhaps making so many wrong choices before landing on the right one that the effort hardly seems worth it. Even elected officials and high-ranking bureaucrats who believe in criminal justice reform are skittish about trying something new, so they often give in to their colleagues who prefer costly and unsuccessful but comfortably familiar policies on sentencing, imprisonment and parole.

But what if someone else agrees to take all the risk? What if some outsider — a nonprofit service provider, let’s say, or a charitable foundation, or maybe even a commercial bank — raises the funds, runs the program, produces the results, then gets reimbursed with public money only after presenting verified proof of success?

Later this year, analysts will publish results of an experiment along those lines begun in 2010 at Peterborough Prison outside London. The social impact bond project, as this kind of financing and problem-solving innovation is often called, uses money put up by investors and managed by a nonprofit group, which contracts with another organization to provide recently released inmates with mentoring and other services intended to break the cycle of re-offending.

If an independent evaluator confirms that the program “worked,” as defined by agreed-upon criteria for decreasing new convictions — and preliminary analyses are encouraging — the British government will repay the investors’ capital plus an agreed-upon premium. If the success targets aren’t met, the investors eat the costs and the taxpayers owe nothing.

Although the Peterborough experiment is still underway, it has become the talk of the financial and nonprofit worlds, and now governments, nonprofits and financiers are clamoring for a piece of the social impact bond action. Michael Bloomberg, while mayor of New York, launched a project in 2012 to reduce criminal recidivism of young adults held at Rikers Island, with financing and administration handled by Bloomberg Philanthropies and Goldman Sachs. Massachusetts followed with a similar project.

But social impact financing need not be restricted to recidivism programs. Utah is doing one to fund preschool. Fresno is working on one for asthma prevention. Supervisors in Santa Clara County are considering social impact bonds to address mental illness and homelessness.

Last year, two state senators, Curren Price (D-Los Angeles) and Ted Lieu (D-Torrance), introduced a bill to establish a pilot program for social impact bonds (more accurately called “pay for performance” or “pay for success” financing because no actual bonds are sold), managed out of the governor’s office. Price has since left to join the Los Angeles City Council, but Lieu is pressing forward with SB 593, and it comes before the Appropriations Committee on Tuesday. If passed and signed into law, it would require California to propose three pay-for-success projects each year for five years.

As some proponents see it, those projects would allow social activists and philanthropists to be more effective with their donations. Others see a chance to put private venture capital and market discipline and innovation to work solving social problems that government so far has found intractable. In this view, Silicon Valley might pay to end homelessness. Wall Street might pony up to treat the mentally ill. For a profit.

What could possibly go wrong?

Plenty, of course. The United States is no stranger to market fads, and the words “bubble” and “subprime mortgage” have taken on new and very unhappy connotations over the last decade. Besides, even without the lure of private capital, demand for rigorous measurement under programs such as No Child Left Behind has brought us phenomena such as cheating principals, who now have an incentive to change numbers or expel low-performing students to present their evaluators with better test scores. Add the profit motive, and the pressure to doctor results could become immense.

In Peterborough, for example, someone just might hit upon the idea of loading the ex-offenders onto a ship headed for the mid-Atlantic and not turning it around until the evaluation period has ended, lower crime numbers are in, and investors have pocketed their public money. Everything would look right on paper, but recidivism would remain unsolved as offenders return to port, ready to resume their mischievous ways.

Still, being able to point out pitfalls in a new approach isn’t the same as demonstrating that the old, failing ways are best. If they were smart, the Peterborough experimenters shaped their project and its rules to prevent anything like the Atlantic scheme, and if they didn’t, they will learn from their mistakes and do better next time. Likewise, those who craft California programs just might do some good if they can watch and learn from others, avoid a conveyor-belt-like approach and help politicians and bureaucrats recognize and overcome incentives to fudge numbers and fake success.

The Price-Lieu bill leaves much of the details to the governor, whose Office of Planning and Research would draft regulations. It is worth a try, and lawmakers should move the bill along. In the meantime, some California cities and counties are going ahead with their own programs, and Sacramento should plan to assist them with data and expertise — and to learn from their experience.