Can Social-Impact Bonds Really Have Big Impact?

In the past year, social-impact bonds were unveiled in California, New York, Ohio, and elsewhere, and the buzz ranged from praise — “They will revolutionize the way government provides social services, unleashing private capital for social good” — to criticism — “They are, bottom line, making money on the backs of the poor.”

Social-impact bonds seem like a simple concept: Donors and other investors provide money to a nonprofit that has a great way to solve a problem, like the number of criminals who land back in prison or the number of low-skilled people who can get good jobs. If the project works and saves the government money by reducing the number of Americans behind bars or on welfare, then the government pays the donors back for their investments. And if the project does better than expected, donors might get an extra payment beyond their initial investment.

As the leader of a foundation that has pushed to transform the way nonprofits are financed, I am happy to see real attempts to provide incentives to prevent social problems rather than pay for their cures, but I know how difficult these new efforts are to pull off.

While the F.B. Heron Foundation made a commitment to an impact bond (it did not ultimately go forward) and is open to considering other such proposals, I am skeptical about this financial tool in the way that any investor should be skeptical about any financial technology.

It can be used adeptly or ineptly. It’s not the tool, it’s what it does. Like loans, contracts, and equity shares, social-impact bonds are pretty much morally neutral.

Is it the right tool for the job? Is it a good deal, not only for the investor but for all parties to the transaction? (A particularly important question in the nonprofit world.) Is it affordable, given its life cycle? Are the parties competent? Who takes the risks? If it fails, who gets hurt?

Social-impact bonds aren’t new. One way to remove the rouge of innovation from the wizened face of these financial instruments is to refer to them as “performance contracts.”

Versions of pay-for-success have been used in many states for years. For example, they bear some similarity to the shared-energy savings contracts of the ’70s and ’80s, in which a company fronted the cost of energy-saving equipment (like fluorescent lighting, a new boiler, or timer switches) in a building at no up-front cost to the owner, then split the subsequent “savings” (the value of the reduction in energy costs) with the owner until the original investment and a return were realized.

These energy-performance contracts sounded straightforward to monitor, but actually even the simplest was complicated.

And even before the equipment was installed, complexity entered the picture. Is it the right equipment? What if it breaks down? What if the installer or the manufacturer goes bankrupt? And monitoring “savings” made installation seem straightforward despite variables like energy costs and weather and operating schedules and demand. We learned that the best applications were the simplest and most reliably predictable.

As I see it now, this rule doubtless needs to be applied to social-impact bonds as well.

The Achilles’ heels of the bonds are their complexity and transaction costs. People are infinitely more complex than boilers or valves, and the bonds invite as many as five — or even more — parties to a transaction in addition to the energy contractors. Those parties can include, say, ex-prisoners or low-skilled workers, the nonprofit that provides a service, a government agency, an intermediary, an evaluator, and the investors.

If the market found shared-energy savings complicated, imagine something as complex as prison recidivism: What is reoffending: the same offense? A lesser one? Where? For how long? Who decides? Who adjudicates if there’s a disagreement among the parties? What if something unanticipated happens? What if the main parties go out of business?

Savings are reductions in expenses, which in the absence of positive cash flow don’t translate into payments to anyone. What is the threshold for financial savings? With the brainpower involved in social-impact bonds, these questions are undoubtedly answerable, but does that largely technical exercise add value?

I fear that even skilled nonprofits and intermediaries will have trouble translating great ideas into contracts that really provide the right incentives. When you look at how the parties do the accounting, how they measure “savings,” how they decide what the return should be and to whom at various break points, it gets inelegant, to say the least.

Some deals that we have seen don’t balance the interests of the parties well. Nonprofits often take more risk than they should, and investors take less.

What’s more, nonprofits build their efforts around one transaction rather than focusing on how to strengthen their services over all. The government gets a lower return on investment, typically to court investors by offering, for example, guarantees. The group handling the transaction is often left with highly complex and difficult negotiating, monitoring, and organizing tasks for very little money.

There is also potential for perverse incentives. After all, the burgeoning social-impact bond “marketplace” depends on an ever-expanding supply of incarcerated people or other ripe social and financial opportunities. Nobody wants to end up in perpetual symbiosis with the Corrections Corporation of America to assure the market is expanding.

That is not anyone’s intention, of course.

Given that a simpler version of these arrangements might be less expensive, perhaps governments will just do this work themselves using the expertise developed in this process. (One of the best examples of an impact-bond-like arrangement in the country, in Minnesota, has done just that, for years.) Then more of the savings to government could go to serve other social needs rather than to investors or transaction costs.

In world nonprofit finance, where real adherence to mission demands that we look for ways to make ourselves superfluous, at least in health and social services, the big win will be that prison populations get so low that social-impact bonds aren’t needed at all and the market dries up completely. That’s victory for everyone.

Perhaps another reason not to see social-impact bonds as a panacea is this: They tend to focus financial resources on remedies that have the most reliable and short-term savings.

Thus, they work most straightforwardly for problems that have become so bad that the cost to help the incarcerated, the seriously ill, and others in need is the highest, the remedy closest to the incurred costs, and the financial savings high and relatively easy to measure. That’s fine unless it means that we trade this low-hanging fruit for much more desirable but tough-to-measure long-term social investments.

That could well happen if the success of social-impact bonds trains the capital markets and social investors to expect maximum, easily measurable financial return on efforts to solve all social problems, starving some of the best programs because they’re longer-term and have less circumscribed social return.

The real bulls-eye is for youthful offenders never to enter a prison, that they be born healthy and grow up in happy, stable families with a working parent.

It’s important that universal prenatal care, great preschool, and school, college and trade-school scholarships, not to mention living-wage jobs for parents, get the lion’s share of investments and attention, even though these investments may not “pay off” with high margins for investors and may be prohibitively expensive to track. In the words of a major investor, we don’t want social-impact bonds — or social investment in general to be “stuck in prison.”