Social Impact Bonds: Lessons from the Field

In piloting social impact bonds, governments have already yielded some lessons from the field.

A year after the British Ministry of Justice piloted social impact bonds to reduce the 60 percent recidivism rate for the 3,000 criminal offenders who passed through the doors of a private prison in Peterborough, UK, the innovative funding mechanism captured the imagination of many social entrepreneurs. These bonds, also known as pay-for-success contracts, promise to transform the relationship between governments, nonprofits, and funders. “Social impact bonds” became one of the top ten buzzwords of 2011. And local governments in England, Australia, Canada, and the US have started exploring these contracts. As state governments at home and abroad prepare to pilot these bonds next year, the processes have already yielded some lessons from the field.

The first lesson is that pay-for-success contacts may ultimately encompass several ways of engaging pure for-profit investors, impact investors, and foundations in producing social returns. Steve Rothschild, an Ashoka Fellow and CEO of Twin Cities RISE! has had some success in Minnesota with one form of pay-for-success contract, which he calls human capital performance bonds. In this mechanism, a state raises funds by issuing general obligation bonds, directs those funds to nonprofits that have generated positive social outcomes and created government savings, and uses cash unlocked by those savings to repay the bondholders. Because the full faith and credit of the state backs these bonds—and no state since World War I has defaulted on its bond obligations—these bonds will allow social entrepreneurs to access capital markets.  The $10 million designated toward human capital performance bonds in Governor Mark Dayton’s July 2011 budget made Minnesota the first—and so far the only—state to pass legislation on pay-for-success contracts.

Social Finance, Inc., in Massachusetts is exploring a different approach to engage capital markets. Like its sister organization in the UK, Social Finance might front working capital, raised from investors with a range of risk-return appetites, to nonprofits in return for a government promise to pay when those nonprofits achieve predetermined outcomes. Unlike the UK operation, which attracted only philanthropic investment for its Peterborough bond, Social Finance here might slice the expected revenue stream into tranches, offer the lower-risk tranches to foundations, and pitch the premium tranche to investors with the highest appetite for return. (This idea might also work for UK’s second social impact bond, which plans to target troubled families.)

The New South Wales (NSW) Government in Australia, which plans to enter into a social impact bond in late 2012 to address either recidivism or foster care, is exploring attracting for-profit investors by offering to share some of the bond’s downside risk—essentially buying the lowest tranche of its own bond.

The second lesson is that the very process of implementing social impact bonds yields valuable returns for the government that engages in it. Governments that start thinking about how much return taxpayer dollars generate in one area may take that thinking to its other investments. The process of paying a service provider based on outcomes pushes the government to introduce data monitoring and evaluation systems above and beyond the ones it may be using today. Monitoring systems that it needs to set up to track leading indicators, as well as final outcomes, may then be used to track outcomes for populations not engaged through the pay-for-performance contract. And evaluation systems that rigorously compare outcomes to counterfactual scenarios help push the conversation around performance measurement.

The social impact bond is ultimately a learning tool for a society that is still discovering best practices. If governments already knew the most efficient uses of taxpayer money, they would not need to ask the service provider to find the best solution, as pay-for-success contracts do. And if service providers delivered known or similar outcomes, then delaying payment until we observe performance would be an unnecessary burden.

Therefore, perhaps the largest lesson, and greatest success, of this young innovation has been its ability to anchor the conversation of governments, social entrepreneurs, and impact investors around measurement, metrics, and outcomes. Nick Hurd, Britain’s Minister for Civil Society, recently lamented the “massive culture of risk aversion in the public sector,” that his use of social impact bonds has revealed. Impact investors across the board are starting to hold their investments to more rigorous standards of social returns. And as governments innovate, they are rewriting the social contact with their citizens and businesses, as Westminster has done this month.

The next half-dozen pay-for-success contracts that come online in 2012-13 will expand the number of social services, types of investor, and bond structures that will, in turn, increase the size and complexity of the nascent social impact bond market. And, as that market evolves, we will come closer to that transformational promise of a new social contract in which taxpayers, entrepreneurs, and governments hold each other accountable for performance and impact.

Update: On January 18, the Commonwealth of Massachusetts issued a series of requests for proposals to use pay-for-success contracting to address youth recidivism and chronic homeless in the state.

Read the follow-up article, "Social Impact Bonds: Lessons from the Field"

Tracker Pixel for Entry


  • BY steve rothschild

    ON January 13, 2012 02:47 PM

    thank you for including mention of the Human Capital Performance Bond(HUCAP) in the article. However, there are a number of errors, some small and some material, that need to be pointed out to insure accuracy:
    In the article it is stated that “Steve Rothschild, CEO of Twin Cities RISE!  has had some success in Minnesota with one form of pay-for-success contract, which he calls human capital performance bonds.” In fact, while the HUCAP is now a law it has not yet been implemented. We expect the pilot to get underway within the month. Actually, Twin Cities RISE has had a separate pay for performance contract with the state of Minnesota since 1997. it has been hugely successful, with the state enjoying over a 600% return on its investment.Those funds are not provided by a bond but rather by state funds.
    The article states that the bond will be a general obligation bond when in fact it will not and cannot for legal reasons. Rather it is expected to be a state annual appropriation bond, which does not carry the same legal obligations of a G.O., but nevertheless is accepted in the marketplace by investors - albeit at a slightly higher interest rate.
    While savings from lower entitlement and corrections spending are likely to provide the greatest economic value, the state will also bnefit from incremental sales, income and property taxes as successful clients see their incomes increase from the intervention of high performing social enterprises like workforce programs.
    Finally, I am the Chairman of Twin Cities RISE!, not its CEO; but it was Invest In Outcomes( which developed and promoted the Human Capital Performance Bond.


    Steve Rothschild

    Founder and President, Invest in Outcomes
    Founder and Chairman , Twin Cities RISE!

  • BY Jason Burwen

    ON January 24, 2012 11:21 AM

    While I like the idea of pay-for-results, I wonder if this approach will produce perverse incentives of service organizations to go after the “best” clients in the service population. For example, if there’s an organization that works on recidivism that is in a pay-for-results contract, I can imagine they’d have a lot of pressure to avoid ever engaging clients they deem at the outset to be “too difficult.” The contracting incentives are well-intentioned, but the organizations responding to them will have many reasons to engage in selection bias and artificially enhance their metrics. What do you think?

  • BY Nicholas Gruen

    ON January 26, 2012 10:19 PM

    Thanks for the article, though, together with the optimism and enthusiasm for innovative approaches, I would have liked to have seen more awareness of the obvious risks - mentioned by the last commenter. It will be interesting to see how social impact bonds play out. They may offer substantial impact and if not perhaps at least insight. Let’s hope so.

    But the perverse incentives problems will be substantial.

  • Brian O'Shaughnessy's avatar

    BY Brian O'Shaughnessy

    ON January 28, 2012 03:02 PM

    Jason and Nicholas

    Constructed properly, a social impact bond could actually have imbedded incentives to address the highest risk population because —whether formerly incarcerated, homeless or at risk youth—the highest risk population costs the most.  Avoiding problems for one high risk individual can recoup the costs associated with 10 lower risk clients.  While not every governmental entity captures such detailed cost information, they should and SIBs may provide the motivation to do so in the future.

    This is in fact what has happened in the UK where short term offenders have an extraordinary recidivism rate.  Ultimately, much will depend upon the rigor with which the SIB intervention and terms are designed.  For the integrity of the underlying investment and the efficacy of the SIB concept, we should hope that the first SIBs or pay-for-success programs are properly designed to address the “problem” and not general ideas of cost.  Viewed as a political tool to reduce budget deficits, SIBs will fail.  Seen as a tool to provide social services that truly address an issue they have unlimited potential.

  • great post.

  • BY Thien Nguyen-Trung

    ON February 29, 2012 06:08 AM

    Michael, thanks for these lessons from the field!

    I saw a very similar article like this (actualy several) and still hold the same opinion:

    I suppose I should just get in line with all the others in the industry who are waiting finally, finally to hear about another case study than the UK prison example of social impact bonds being used.

    I think here are some areas where we are not talking more openly about the real possibilities:

    1) The big thing is the concept of “money for impact”

    I think really what gets a bit downplayed in the many articles on (the same case of) SIBs is that it is rare to formally tie the performance of a social organizations (“impact”) to certain outcomes on a financial instrument. If we want to get excited about impact investing, we should get particularly interested in financial instrument innovation that will map lenders’ expectations to metrics which make sense for social organizations that are doing the borrowing. Instead, what we have nowadays is an attempt to “find” suitable organizations that fit the lenders expectations of firms from which they can make some money.

    Thus, I would like to see more companies, together with impact funders, experimenting with instruments that are designed on a customized basis around what can be realistically expected from a social organization and try to set up realistic milestones. Most likely, this process will take considerable adjustment along the way to figure out a good balance so that borrowing organizations can be held to high impact expectations without breaking their backs a the same time.

    And agreed with the readers above that obviously this cannot be used for all kinds of organizations to avoid odd incentives.

    1) From Government to Companies

    One interesting but yet-to-be-tested idea is how we can issue these bonds from social enterprises instead of governments. The amounts would be clearly smaller given limited financial capacity, but again, if a company IS sustainably spinning off some cash or making revenue, perhaps we can find a way to let them access a new way of raising growth capital through new instruments in between philanthropic grants and pure-play bonds.

    I cannot wait for the first results to come in from impact investors trying this out with their star social enterprises.

    Thien Nguyen-Trug

  • BY Rick Brush

    ON April 2, 2012 10:55 AM

    UC Berkeley and Collective Health recently completed a white paper on the use of social impact bonds in health, commissioned by the California Endowment. The paper is linked here:

Leave a Comment


Please enter the word you see in the image below:


SSIR reserves the right to remove comments it deems offensive or inappropriate.