First Social Impact Bond Fails to Meet Halfway Mark Performance Target

Print Share on LinkedIn More


Social Impact Bond Misses Target

August 8, 2014;National Union of Public and General Employees

The first Social Impact Bond project in the world, the Peterborough prison effort to reduce recidivism by 10 percent, has encountered a problem: It didn’t meet the 10 percent target for the first cohort of released prisoners and therefore failed to qualify for an early payout for investors. In fact, according to this report from the National Union of Public and General Employees, “the data released this week shows the scheme will come nowhere near achieving what proponents claimed it would when the project was launched.” (“Nowhere near,” it appears, actually was a mere 8.4 percent reduction in recidivism.)

This was an early payout possibility for the investors in the SIB, and they missed it, but investors have the ability to get some payout by the end of the project. While the payout for the first cohort would have been triggered by a 10 percent reduction in recidivism, the payout target for the combined first and second cohorts of released prisoners is only a 7.5 percent recidivism reduction (compared to a national comparison group with characteristics similar to the Peterborough prison short-sentence population). The sponsors of the Peterborough bond apparently quickly emailed a statement that the project was “on track” to meet its targets for a payout to investors by 2016 for the first and second cohorts.

Third Sector’s Stephen Cook notes that the third cohort of the Peterborough SIB was cancelled by the government, as the program is being transformed into a national program under the Ministry of Justice called Transforming Rehabilitation. The new national program is still measurement- and results-oriented like the Peterborough SIB, but as a national program, it will be restructured to be roughly equivalent to what American nonprofits experience as performance-based contracting; participating nonprofits implementing the anti-recidivism projects will be compensated when they reach predetermined performance benchmarks. The process is sort of like a SIB but, as Cook notes, but “without any social investors,” or as we would note, without any private investors anticipating a profit from the program.

Like the reaction of SIB promoters to the news that the third Peterborough cohort had been cancelled, the SIB community is publicly talking up SIBs despite the project’s falling short at its first payment point. But Cook senses something else coming from the SIB promoters’ bleachers. “The overall impression is that everyone concerned is determined to talk up the Peterborough SIB in public, despite qualified success and difficulties over statistics and measurement,” he writes. “In private, some of those involved are much more downbeat about the prospects for replicating what might prove to be a one-off experiment.”

For some observers, the problem posed by SIBs like the Peterborough project is more than the likelihood that they might be simply rather expensive one-time experiments. The fact that the UK government was able to turn the project into a program—minus having to pay a premium to private investors—goes to a point that critics have raised before. If a governmental agency and related nonprofits are going to go through what was needed to structure the Peterborough project, which the National Union of Public and General Employees says took 29 months, including 11 months just negotiating the measures by which success would be determined and the investors paid, why not work on a more broad-based program? That’s time that might be better spent on public policy lobbying to create a more widely available program rather than designing the arcane structure of a SIB for the benefit of private investors tied to one project.

Cook suggests that the biggest question about SIBs might be “whether they will ever fulfill the initial intention of attracting a wide range of investors—not just forward-thinking foundations willing to take risks for the sake of innovation.” In the Peterborough case, 17 charitable foundations invested the £5m to work with the Peterborough prisoners. Charitable foundations can obviously accept little or no financial return on their investments in social programs; they do that through a concept called “grants.” Private investors want a return on their investment—called “profit.”

In New York, the anti-recidivism SIB that has Goldman Sachs as its primary investor is structured to give Goldman a return of more than 20 percent if performance targets are met, and even that wasn’t enough for Goldman. The massive Wall Street firm’s SIB is also the beneficiary of a 75 percent guarantee from Bloomberg Philanthropies. According to the National Union article, the response of the SIB community to the Peterborough outcomes is not one of “acknowledging the flaws in the privatization schemes…[but instead] proposing [that] governments provide incentives or guarantees to encourage people to invest in” SIBs, that is, finding ways to further subsidize SIBs.

Therein lies the real problem. How much are SIBs aimed at solving social problems versus finding new opportunities for private investors to profit from the largesse of taxpayers?—Rick Cohen

  • Lars Boggild

    I’d consider it disingenuous to claim that the first cohort of the Peterborough ONE Service was a failure. The 10% threshold for reducing recidivism was a trigger to motivate out-performance, providing an incentive for even larger social impacts. The 7.5% threshold should actually be considered the performance benchmark for the SIB, which the program is ultimately succeeding at. Why not just pay out at the 7.5% threshold then? To establish greater levels of statistical confidence in the program as the ultimate cause of the impact. Current results at 8.4% for the first ~1000 prisoners, give the intervention a confidence level of 90%, whereas by the second cohort, an additional ~1000 prisoners, the resulting measured performance will have a confidence level of >95%, which is the stronger standard used in social science.

    I would note that ultimately when you mention a “mere” 8.4% reduction in recidivism, you are not just attempting to malign SIBs, but ultimately passing harsh judgement on the actual work of the non-profits and charities, such as St. Giles Trust, Ormiston Trust, and others.

    Further, many of those committed to seeing positive social impact for prisoners, including some of the social investors mentioned in the article, are highly skeptical of the approach being taken by Transforming Rehabilitation, the national program you mention which caused the early termination of Peterborough.

    For example see:

  • Stratis Voutsas

    SIB’s should focus on Public Benefits and not Private Benefits. They are a Public Benefit substitute solution and not a Private Benefit solution. SIB’s are constrained since they can only recover funds from direct Council savings. They cannot recover consequential benefits beyond Council borders.

    What you need to reduce program costs is to offer a full service program, one that really works, either govt funded or private, and to monetize the total positive economic footprint of an undertaking, including its consequential external benefit (the positive externality). The funds recovered from monetizing the positive externality can be used to reduce program costs. SIB’s can work if they can tap into the external benefit. We know its there and we know how to tap into it as an add-on to a Council or SIB program.

  • Neil Wollman

    I am not quite as concerned for a few reasons. They still did better then the control group, they still will likely meet their later goal, and importantly, most SIBs will help some people who would not otherwise be helped because government dos not have the resources to do so. But for better or worse, it is moving ahead with Bills introduced into Congress (not passed yet) but is bipartisan and hearings in the fall, I believe (though an earlier haring had some doubters)

  • Richard Lightbown

    A public program based on pay for performance to non-profit service providers ignores three facts:
    first, non-profits cannot bridge services funding to PFP (the purpose of private capital in SIBs);
    second, program evaluation methods are expensive, and beyond the reach of non-profit service providers;
    third, program evaluation methods, unlike clinical trials of new medications, cannot be amortized over the next 5 million service transactions, because of constant changes in demographics and client needs, from cohort to cohort.

    The current implementation of standardized pay-for-results with Medicaid and Medicare in the US is under fire for putting service providers addressing the most needful populations in urban communities at serious dis-advantage to providers serving middle class clienteles.