What to Make of Social Impact Bonds

Published Friday, March 4, 2016 | by Barbara Ray

Quick Facts

Issue Area
Document Type
Are SIBs as revolutionary as some argue? Or are they just another way for big banks and their investors to make money off the backs of the poor?

On Nov. 3, 2015, The New York Times broke a story that some in liberal social circles had been secretly wishing to see. The initial findings from a much-heralded social impact bond in Utah — a new way to finance social welfare programs involving private funders like Goldman Sachs — were suspect. The results were likely significantly overstated, The Times reported, in claiming that 109 out of 110 children avoided special education because of a preschool program.

“Aha,” said critics, here is evidence that the bankers are fudging the data to ensure they get paid. Former Assistant U.S. Education Secretary Diane Ravitch called the initiative an outright “threat,” telling readers of her blog to phone lawmakers “at once to stop this money-making scheme.”

You can’t blame them for being suspicious. Highly arcane financial products ultimately plunged the United States into a devastating recession and threw millions of people out of their homes. In some ways, this smelled like déjà vu all over again.

Begun in the United Kingdom in 2010, social impact bonds (SIBs) are indeed arcane. Akin to seed capital for social programs, they tap private dollars to fund government programs such as homeless services or prison recidivism programs. If those initiatives deliver as promised, the investor — typically a large bank — gets paid back by the government, with interest. The bonus for taxpayers is that if the programs do not work as promised — if they fail, for example, to reduce the number of needed beds in a homeless shelter by a set amount within a certain time frame — the government owes the investor nothing. In that case, the investor has risked money and lost. Should the program meet its marks, the government wins by saving the longer-term costs from fewer prisoners.

It seems straightforward enough, but between the start and finish are layers of complicated details involving expensive intermediaries, detailed financial stipulations, and a new type of "quant" who evaluates the program’s progress.

Despite their complexity, the excitement over SIBs is palpable. Governing magazine listed SIBs as one of the top legislative issues to watch in 2016, large nonprofits are jumping on the bandwagon, intermediaries who help put together the deals are multiplying, and stories are gaining traction in the news media.

Banks like them because SIBs help the institutions meet their Community Reinvestment Act requirements to provide credit to low-income neighborhoods, and also allow banks to exercise corporate social responsibility. Politicians like them because they inject accountability into social programs. Bipartisan (!) bills are moving through Congress to set aside more funds for SIBs, and the Obama administration’s Social Innovation Fund has dedicated a SIB branch to its focus on evidence-based community development. Nonprofits like SIBs because they provide enough capital to fully fund a program and allow service providers to expand the most promising programs. SIBs, in short, are hot.

But not so fast, say critics. Why do we need private investors earning a payout? Shouldn’t the government be funding good social programs anyway? Aren’t SIBs just a cloak of respectability for big banks after the bruising they took over the federal bailouts? Others, tired of the vaunted private sector being treated as a savior, argue that we don’t need more business acumen to improve social programs, just more political will to fund them. Still others see layers of additional costs, from paying for intermediaries to the legal fees needed to get deals done.

So do they have a case? Are SIBs as revolutionary as some argue, or just another way for big banks and their investors to make money off the backs of the poor?

The Emergence of SIBs

The first U.S. SIB was launched in New York City in 2012 with the aim of cutting recidivism among young men leaving Rikers Island jail. Because of several hitches when rolling out the program, the Rikers SIB was ultimately cut short. Currently, seven SIBs are operating in the United States, including early childhood education programs (in Chicago and Utah), supportive housing for the homeless (in Santa Clara County, California, and Massachusetts), rapid housing for homeless mothers to reduce foster care placement of their children (in Cuyahoga County, OH), young adult recidivism programs in three cities in Massachusetts, and a New York State program to increase employment and reduce recidivism among ex-offenders. Dozens more are in the pipeline.

SIBs emerged amid two overarching trends — the push for more accountability and evidence-based decision-making in government, and a general fatigue with social welfare programs that never seem to work. To some, SIBs are the shot of innovation the social sector needs, aligning incentives to nudge the welfare sector beyond good intentions or political ideology. To others, the deals elevate “what works” and let the market winnow out programs that do not. According to two former Office of Management and Budget directors, Peter Orszag and Jim Nussle, less than 1 percent of government spending is backed by evidence of its effectiveness, which makes elevating the value of research and evidence sound like a good idea.

“The real innovation here,” says Ian Galloway, senior research associate in community development at the San Francisco Federal Reserve Bank, “is performance-based contracting” — getting government to pay for impact, not inputs and outputs like the number served or the number of houses built.

“It’s a hugely disruptive idea,” he says. “It upends everything: the role that evidence plays, the role of nonprofits, private-sector investors. It’s all new.” As such, it is bound to shake up the “social-sector industrial complex,” he says, with its political patronage and winners in the game of social-program funding. “It’s not an accident that we generally don’t know what works.”

Others, while waiting to see how things shake out, agree. “Its focus on evidence has the potential to transform the way government procures goods and services,” says Gordon Berlin, president of MDRC, which was the research firm serving as the intermediary in the Rikers Island SIB. (Berlin is writing a report on the Rikers experience, “Learning from Experience: A Guide to Social Impact Bonds,” due out in February.)

SIBs also bring the right people to the table, raising awareness of effective programs. “For us to be successful,” says Brad Dudding, chief operating officer of the Center for Employment Opportunities (CEO), a nonprofit focused on curbing recidivism, “we need government champions. SIBs are a surefire way to get those champions to come to the table.” CEO is the service provider in a SIB with Bank of America Merrill Lynch private banking clients that began in 2014. The goal was to reduce recidivism and increase employment among the ex-offenders in Rochester and New York City.

SIBs may also be a cutting-edge way to fund additional social programs. In a “budget-neutral” environment, where for every program added, another program must be shaved, “federal policymakers can create a virtuous cycle by funding those programs [that save money] … which in turn makes room for … new or improved social programs,” Ron Haskins, a Brookings Institution scholar and former White House adviser, writes in his 2014 book, "Show Me the Evidence."

Similarly, as George Overholser, CEO of Third Sector Capital Partners, wrote in the Stanford Social Innovation Review, SIBs also could create a recyclable pot of money for social programs when the initial investment is returned with interest to be reused in the next SIB. SIBs can also encourage greater collaboration across government agencies because they target outcomes for populations whose issues cross departmental silos.

SIBs can further help ensure the sustainability of programs. Relying on one funding source, whether a foundation or government, is rarely a good idea. Foundations move on to other pursuits, while governments cut funding with new administrations or economic downturns, often leaving programs in the cold. With SIBs, funders can be a mix of banks and philanthropies with different risk thresholds.

SIBs have one other important advantage: the upfront capital can help innovative programs go to scale, serving a much wider need than typical one-off programs. SIBs essentially create a handoff between philanthropy, which provides early capital for risky bets, and government, which funds those programs at scale.

Even More Ambitious: SIBs in Community Development and Health

SIBs in the United States are designed for programs with a fairly direct path from intervention to impact, such as expanding preschool for at-risk children to avert later placement in costly special education. But Galloway and others think they could be used in more complicated arenas like a neighborhood’s effect on health.

Improved health is certainly a valuable outcome both for society and the health care system. Identifying programs that address the upstream determinants of health, such as neighborhood factors or poor housing, can go far in reducing downstream costs from hospital visits, ER visits, and other health issues.

SIBs could help create a quasi-market for community health and its trickle-down cousin, individual health. But because improving health is a multifaceted affair, “You need a neighborhood scale of intervention to move the dial,” says Galloway.

In such a SIB, he envisions a master contractor with subcontractors, akin to the master contractor on a home renovation. The many small nonprofits that provide the basketball games or swimming lessons or other programming that ultimately contributes to weight loss and better health are too small to play in the SIB field. Instead, the private sector could fund a large community development nonprofit, which then contracts with smaller nonprofits. The master contractor is held accountable for long-term health outcomes in the community.

“We need a SIB contract that acknowledges the scale needed to tackle these big, interrelated challenges,” says Galloway.

Approaches such as these, however, need “patient” capital — investors willing to sink their money into deals that may not pay out for 10 or more years. For the time being, most SIBs are short-term deals with payouts in the realm of three to five years. The deals also need to solve the hairy problem of benefits going into the wrong pocket; investors put money into improving housing, but the returns go to the health care system.

In the interim, several SIBs are addressing an important issue — chronic homelessness — that spans health and community development, albeit in a more linear way. In Santa Clara County, California, Project Welcome Home is funded by a SIB. Because of their precarious situation, the chronically homeless frequently end up in the most expensive forms of care, such as the emergency room or in jail. Chronic homelessness is estimated to cost the county on average $83,000 per year per homeless person in public services. The “housing first” movement finds that rather than requiring people who are homeless to address their substance use or other health issues before they are given a place to stay — as was the traditional model of addressing homelessness — it is more cost-effective to get them into stable housing first. As such, Welcome Home will provide community-based clinical services and permanent supportive housing to 150 to 200 chronically homeless individuals who are frequent users of the county’s emergency rooms, acute mental health facilities, and jail.

Although this SIB is a more traditional structure, its funding is unique in its broad blend of funders investing with different rates of return (reflecting different levels of risk). Private and philanthropic funders include the Sobrato Family Foundation, the California Endowment, the Health Trust, The Reinvestment Fund, Corporation for Supportive Housing, the James Irvine Foundation, Google.org, and the Laura and John Arnold Foundation.

This broad set of investors spreads the risk and as such may in the future attract bigger pools of capital that would be able to tackle more complex problems, like addressing the social determinants of health.

So What’s Not to Like?

Whatever the challenge to SIBs, on their surface, they seem almost too good to be true — innovative, able to bring new money to the table, evidence-based, taxpayer-saving, and able to change how governments procure social programs.

But others warn that the excitement is getting ahead of itself. “A number of structural challenges will need to be overcome if this new financial instrument is to gain widespread use,” says MDRC’s Gordon Berlin. In his list of to-do’s: finding the right balance between risk and reward for investors and payers, guaranteeing accountability by tying payment to reliable evidence, injecting more flexibility into the contract terms, and ensuring that effective programs are sustained.

The devil, in other words, is in the details. Putting a SIB together is a Job-like test in patience. As one investor told me, it’s the Wild West with a Tower of Babel thrown in. No one speaks the same language and everyone is learning by doing. Investors think of risk and due diligence, nonprofits think of service delivery. There are no templates for the roles, metrics, or accounting standards. Intermediaries are hired to herd cats, translate between the groups, and ensure the program is running as promised. The dealmakers hire evaluators to evaluate the evaluators. Settling on the trigger points to payout is like setting monetary policy at Bretton Woods.

Even if both sides can agree to likely benchmarks of success, investors face the crazy notion that a 9.5 percent reduction in, say, recidivism, is “failure,” while a 10 percent reduction is success. “Is the difference between 8.5 percent and 8.4 percent substantially different from the difference that occurs between 8.4 percent and 8.3 percent? No, but the return drops off a cliff,” Andrea Phillips of Goldman Sachs told MDRC.

Just getting to that evidence benchmark is another hitch.

The Utah SIB is a case in point. There, Goldman Sachs and the M.K. Pritzker Family Foundation funded the expansion of a preschool program designed to bolster services for at-risk children to avert their later placement in expensive special education classes. Goldman Sachs provided loans of $4.6 million and Pritzker provided $2.4 million as the junior loan, reducing the risk for Goldman. Based on past research, the banks and the state government, like in all SIBs, hashed out the terms of the deal, including when to pay and how much.

By early accounting, the program was a phenomenal success. Too phenomenal in fact.

Of the 110 children deemed at risk for needing special education, 109 avoided it because of the program, according to the evaluation, saving the government $281,550, based on a special education costs of $2,607 per child.

Clive Belfield, an economics professor at Queens College in New York who studies early childhood education, was skeptical. As he told The New York Times, “They [the program] seem to have either performed a miracle, or these kids weren’t in line for special education in the first place.” If they weren’t in line in the first place, then the reduction in numbers needing special education — the “success” of the program — would have been inflated.

Yet Goldman Sachs will receive 95 percent of the savings, or approximately $250,000, in its first of several hoped-for payouts as children bypass remedial or special education through Grade 6. After the first payout, the repayments drop to 40 percent of the cost savings, or $1,040 per child per year, for every year the child does not use those services.

The problem was that there was no counterfactual in the impact study, asking what would have happened if the kids had not had the program. How many of the 110 children would have actually needed special education without the program? The payout seemed to assume that 100 percent would have required it. And that seems doubtful. Therefore, can the government be sure they were paying for real results?

The method of evaluating success is at the heart of a big concern among skeptics. How do you ensure that the game is not rigged? Any banker will want to do due diligence before investing, but social science research is hardly airtight. For every issue, whether homelessness or asthma, multiple programs have been designed to address the problem. And there are multiple studies of those approaches, some more rigorous than others. While random control studies are the gold standard in measuring true impacts, they are expensive and not always ethical, given that the control group does not receive the service. Researchers, therefore, have devised other methods to test effectiveness, but these studies can introduce a lot of mushiness in the findings — as well as sometimes much larger effects than a random control study would find. As such, investors may lean toward the less rigorous approaches when documenting “success” because private investors want to do a deal that increases the guarantee of a payout. That is, they want to lower risk if at all possible.

Whose Risk? Whose Reward?

In many ways, the level of risk is at the heart of the problem. SIBs are designed to be imbued with risk. Yet private investors want to reduce their exposure to loss.

“Social programs are starved for dollars, and the hope is that the private sector is a source of that capital,” says Berlin. “But those investors have to worry about the risk they’re taking. That’s one of the structural impediments to any deal. If you demand high-quality research, you raise the likelihood — risk from an investor’s point of view — of finding no effect.”

The Government Accountability Office agreed. “In practice, investors told us they prefer to back programs that already have a rigorous evidence base because these programs have a known likelihood for success.”

More to the point, critics contend, if these programs are known to work, why do we need Goldman Sachs profiting from them? Shouldn’t the government be funding success anyway? Or as the late journalist Rick Cohen, writing in Nonprofit Quarterly, asked, “Do voters need the validation that comes from having Wall Street bankers bless the projects?”

To CEO’s Brad Dudding, that misses the point. The role of government is not going away, he says. “We think it’s an interesting way to use new forms of capital to innovate and fund what we do, but we see it as a means to an end. A discussion needs to be had about the future role of government in the social sector because needs will continue to grow, but resources will be less. So the social sector has to get better at what they do and do it at a lower cost. I think this funding innovation makes the sector think about that.”

Specifically, Dudding says, SIBs are helping to build the evidence base of what works. There’s a staircase to building evidence that everyone needs to be on, says Dudding. “The SIB is getting us to the north star about what our responsibility is, given what we’re facing.”

During the climb up the evidence staircase, investors may still look to the tried and true programs, reducing their risk. But Deborah Kasemeyer, senior vice president and community reinvestment act officer at Northern Trust Corporation, a lead investor in a Chicago SIB, hopes there will eventually be tiers of evidence and risk, and different payouts accordingly. “Over time, the hope is that transactions that present themselves as lower risk and less reward would get lower rates of return. The ones that are higher risk would have higher rates or layer in more complicated capital stacks.” That way, she says, there will be room for everyone and all kinds of interventions.

Others point to untallied costs, including for the intermediaries who shepherd the deals through. These costs can tack on several million to a program’s operating costs.

According to McKinsey & Company, “SIBs are a more expensive way to finance the scaling up for preventive programs than if the government simply went to service providers and paid them to expand an intervention to more constituents.” Cash-strapped state and local governments that feel they cannot fund prevention services directly can be lured by SIBs, which requires no cash outlays upfront, even though such SIBs are more costly in the long run.

Can SIBs Survive?

Given these hurdles, are SIBs doomed? “If I had to predict,” says the Fed’s Ian Galloway, “I’d say it’s going to fade because no one wants that level of disruption. But that’s absolutely tragic for so many reasons. I can live with an 80 percent solution with SIBs,” he says, “because what we have now is clearly inadequate. Anything that moves us in the direction of measuring impact and paying for outcomes is a huge win over the status quo.”

Many agree that we’re at a fork in the road. But then, they say, we’ve been here before with other complex products, from the 30-year mortgage to the Low Income Housing Tax Credit.

“It’s a very new investment vehicle and it will take time to gain experience and to understand the nature of the risk and the kinds of due diligence that is required,” says Berlin. “And as we gain that experience, we’ll be better positioned to judge this new instrument’s potential. But there’s a lot to be done.”

The beautiful theory, in other words, is butting up against the always-ugly realities.