Skip to Main Content

Paying for Success

State and local governments are trying a new financing model for social programs

Econ Focus
First Quarter 2018
brick path
Pay for success is helping fund projects like the green roof at the Fort Reno Reservoir in Washington, D.C.

Since 1990, the federal government has conducted randomized controlled trials of 11 large social programs, totaling more than $10 billion in spending per year. Ten of those programs were found to have "weak or no positive effects" overall. Many other programs are never evaluated, a state of affairs that has led some critics to deem government spending on social programs "a triumph of hope over evidence."

In recent years, however, governments at all levels have made increasing use of data and rigorous evaluations to assess programs, a practice generally known as "evidence-based policymaking." Cities and states in particular have begun using a new financing model known as "pay for success," or PFS, which has links to evidence-based policymaking. In this model, private investors provide the upfront payments for a social service such as job training or supportive housing, and the government repays them only if the service achieves predefined outcomes. (PFS financing is also referred to as "social impact bonds.") Since the first PFS project in the United States was launched in New York City in 2012, an additional 19 projectsOffsite have officially gotten underway, including two in the Fifth District. (The first project in the world ran in England between 2010 and 2015.) More than 50 U.S. projects are in some stage of development.

Pay for success has attracted bipartisan support as well as the attention of many community development practitioners, including at the Fed. "This is a new way to apply community finance to chronic social issues," says Jennifer Giovannitti, a regional community development manager at the Richmond Fed. "We can bring in new thinking and new efficiencies." Still, some people in the nonprofit community are concerned that focusing too much on pay for success could siphon resources away from social issues that aren't a good fit for the model.

How Do You Pay for Success?

In a PFS contract, an investor or group of investors gives a nonprofit service provider the money to deliver its service for a set amount of time. Over the course of the project, a third party assesses the program's results, and as predetermined milestones are achieved, a payor (typically but not always a government agency) makes "success payments" to the investors. If all goes well, at the conclusion of the project the investors have been paid back, potentially with some interest, and the service has proven cost-effective enough for the government to continue and possibly expand it.

Our Related Research

"Urban Transportation and Inter-Jurisdictional Competition," Working Paper No 17-10, October 2017.

"Rethinking Detroit," Working Paper No 17-04, February 2017.

"From Stylized to Quantitative Spatial Models of Cities," Economic Quarterly, Third Quarter 2016.

"Understanding Urban Decline," 2016 Annual Report.

The deals aren't easy to put together, Giovannitti notes. "It takes a lot of capacity on the ground. A lot of players have to be involved." In addition to the investors, the service provider, and the payor, those players include legal counsel, a third-party project manager — sometimes several project managers — to help structure and oversee the deal, and an independent evaluator to assess the results.

David Hunn is the president and CEO of the SkillSource Group, which administers federal and state funding for workforce development efforts in Northern Virginia. SkillSource just launched a PFS project targeting young adults who have been involved with the foster care or justice systems; success payments will depend on the youths' employment and education outcomes after leaving the program. "Planning the project and assessing its feasibility required a whole new degree of rigor," he says. "Most of us in the local workforce boards don't have that expertise — we needed an outside expert to help us move the ball forward."

SkillSource worked with Third Sector Capital Partners, one of several nonprofits in the United States dedicated to PFS projects. There are also three academic centers, including the Government Performance Lab at Harvard University, the Sorensen Impact Center at the University of Utah, and the Pay for Success Lab at the University of Virginia (UVA), where Giovannitti is on the advisory board. (See "Growing the Pipeline of Pay-for-Success Projects," Richmond Fed Community Practice Papers, February 2018.) "Our job is part education, part research, and part analysis," says Josh Ogburn, the director of the UVA lab. "People have heard about the concept, but they don't know the ins and outs or how to get started. So we help communities develop a project idea and connect them with other advisers."

Investing in Success

In 2016, Denver launched a five-year, $8.6 million project to provide permanent housing and other support services to 250 chronically homeless individuals. The goal is to reduce the amount of time they spend in jail, detox centers, and emergency rooms — services that typically cost the city $7 million annually, or $28,000 per person. A group of private foundations put up the $8.6 million; at the end of 2017, the city made its first success payment of $188,000 based on initial reductions in jail time.

If the program weren't meeting its benchmarks, the city wouldn't have to pay anything, and the program could be discontinued. That's what happened to the first U.S. PFS project, which was intended to lower recidivism among juvenile offenders at Rikers Island in New York City. Three years after it launched in 2012, it had failed to meet the minimum goal of reducing recidivism by 8.5 percent, and the Goldman Sachs Urban Investment Group exercised its option to cancel the project. The New York City Department of Corrections didn't make any payments to Goldman, which by that point had invested $7.2 million of a pledged $9.6 million. (Goldman didn't actually lose that total amount, as Bloomberg Philanthropies had guaranteed three-quarters of the investment. If the project has succeeded, Goldman would have received the entire return.)

The potential return to investors in a PFS project varies considerably. Goldman estimated it would earn a return of between 11 percent and 22 percent, depending on if the program met or exceeded its performance goals. The estimated return to the Denver investors, however, assuming the project is successful, is just 3.5 percent. But the investors in a PFS project aren't necessarily looking to make a lot of money; they're what are known in the nonprofit world as "impact investors" rather than"return investors." Goldman's Urban Investment Group, for example, is committed to "double bottom line" investing, which emphasizes both financial and social returns. And about half of PFS investors so far have been philanthropies or foundations rather than banks or investment funds.

Enthusiasm for Evidence

Technological changes in recent decades have made it easier and cheaper to collect, link, and analyze data, and have contributed to a bipartisan push for evidence-based policymak¬≠ing. Both the George W. Bush and Obama administrations increased funding for data collection and tried to incentivize federal agencies to use more of it. In 2016, Congress passed a bipartisan bill creating the Commission for Evidence-Based Policymaking, which was charged with assessing how the government could make better use of data. The commission issued its final report in September 2017.

In line with the greater focus on evidence, the federal government has supported state and city efforts to develop PFS projects. In 2014, the Workforce Innovation and Opportunity Act, which revamped workforce development, allowed local workforce boards to set aside 10 percent of their funding for PFS programs. (The SkillSource project is the first one to use that funding.) Also in 2014, the Social Innovation Fund (SIF), which was established by the Obama administration in 2009, began awarding "technical assistance" grants to help nonprofits and local governments conduct feasibility studies and gather data. The fund gave out nearly $17 million for PFS projects. The SIF was defunded in 2017, but the budget deal that passed in February 2018 included $100 million in competitive grant money to help cities and states develop PFS projects.

Pay for success also has attracted the attention of community development practitioners at the Fed. Several regional Reserve Banks, including the Richmond Fed, have hosted events for community groups and other stakeholders interested in setting up a PFS project; the San Francisco Fed devoted an entire issue of its journal Community Development Investment Review to the topic in 2013. "We're investment oriented," says Giovannitti. "How can investment move the needle on social issues?"

The Fed's community development function grew out of its role assessing banks' compliance with the Community Reinvestment Act (CRA), which encourages banks to meet the credit needs of their local communities, including low- and moderate-income communities. In the future, it's possible banks could earn CRA credit by investing in PFS projects, as one of the San Francisco Fed articles explored, although regulators have not yet given banks any specific signals regarding how or if PFS financing might satisfy CRA requirements. Still, the potential is there, says Giovannitti. "Pay for success is a natural fit for some of the issues banks are likely to be interested in, such as affordable housing and workforce development."

The Limits of Pay for Success

Not every issue can be addressed with a PFS project; the model requires readily available data, clearly measurable outcomes, and defined cost savings within a reasonable time frame. Recidivism is a good fit, for example, because "the outcome is straightforward," explains Ogburn. "Did the person go back to jail or not? It's easy to verify, and everyone can agree on the definition." Certain objectives in education however, may be more challenging. "The outcomes are harder to quantify because the social benefits and fiscal value accrue further in the future." (There are PFS projects in development targeted toward early childhood education, with short-term metrics such as kindergarten readiness.)

In addition, the number of service providers with the capacity and expertise necessary for the rigorous data collection required for PFS is relatively small, creating the potential for the same few high-performing providers to receive the majority of PFS funding. Some observers are concerned that "rather than motivating the rest of the pack to 'lift' their game and demonstrate effectiveness, the inability of these other organizations to raise PFS funding could hamper their ability to deliver social services," as V. Kasturi Rangan and Lisa Chase of Harvard University wrote in a 2015 article in the Stanford Social Innovation Review. Other concerns are that the focus on clearly measurable — and successful — outcomes will lead governments and service providers to focus on the populations most likely to succeed to the detriment of those who are harder to serve.

In addition, while a major selling point of pay for success is saving the government money, some projects could end up costing more than they would have under a traditional contract. In 2013, for example, Maryland's Department of Legislative Services concluded that the cost of designing a program and negotiating a PFS contract would probably exceed the pilot program's projected benefits. Even if the long-term savings are potentially large, PFS projects could be a hard sell to the many states facing immediate budget shortfalls.

Detractors of PFS point to the failure of the Rikers Island project as proof that the model doesn't always work — or at the very least is overhyped. But others say that's the wrong conclusion. "PFS is a financing model, not an intervention," wrote Paula Lantz and Samantha Iovan of the University of Michigan in a 2017 article. "The 'does it work' question should be focused on the quality and impact of the interventions selected for a PFS performance-based contract, not the model itself."

Pay for Success in the Fifth District

Several cities and states in the Fifth District are moving ahead with PFS projects. In South Carolina, a group of investors including the Boeing Company, the Duke Endowment, and the BlueCross BlueShield of South Carolina Foundation have put up $17 million to expand the Nurse-Family Partnership, a program that pairs nurses with first-time, low-income mothers. (The investors plan to reinvest their success payments in the program.) Multiple studies of the partnership in other states have found that it reduces preterm births, that children are more likely to be vaccinated and less likely to visit the emergency room, and that the mothers wait longer before having a second child — all of which potentially reduce spending on safety-net programs. The PFS project will enable South Carolina's Nurse-Family Partnership to roughly triple its reach to 4,400 families. A similar program is undergoing a feasibility study in Virginia.

Another program targeting the health needs of children is being developed in Richmond, which consistently ranks as one of the worst cities in the country for asthma sufferers, according to the Asthma and Allergy Foundation of America. Asthma is a leading cause of missed school days, emergency room visits, and hospitalizations for children; low-income children, who typically have less access to health care and more exposure to environmental pollutants, are at greater risk. In May 2017, the Richmond City Health District was awarded a $350,000 grant to determine the feasibility of creating a PFS program. Baltimore is also in the planning stages of a program targeting childhood asthma.

In a very different vein, Washington, D.C.'s Water and Sewer Authority is using pay for success to finance improvements to its stormwater runoff system. Goldman Sachs and the Calvert Foundation purchased a $25 million bond issue, which DC Water will repay only if the new infrastructure reduces runoff by a certain amount.

Other projects being discussed in the Fifth District include expanding the scope of Baltimore's Meals on Wheels program, which delivers meals to older adults, to include safety checks and case management; training emergency personnel in Greenville and Oconee counties in South Carolina to provide primary and preventive health care to people without other access to health care; and, akin to the project in Denver, creating supportive housing in Richmond.

As a model in its infancy, the evidence on PFS is minimal. But supporters are optimistic that at the very least, the effort will increase policymakers' reliance on evidence rather than on good intentions. "It's a much more rigorous process than the status quo," says Ogburn. "Entering into one of these projects reorients everyone around an outcome-based mindset."


Giovannitti, Jen, and Joshua Ogburn. "Growing the Pipeline of Pay-for-Success Projects." Federal Reserve Bank of Richmond Community Practice Papers, February 2018.

Federal Reserve Bank of San Francisco. "Pay for Success Financing.Community Development Investment Review, April 2013, vol. 9, no. 1.

Lantz, Paula, and Samantha Iovan. "When Does Pay-for-Success Make Sense?Stanford Social Innovation Review, Dec. 12, 2017.

Liebman, Jeffrey B. "Using Data to More Rapidly Address Difficult U.S. Social Problems.Annals of the American Academy of Political and Social Science, January 2018, vol. 675, no. 1, pp. 166-181.

Subscribe to Econ Focus

Receive an email notification when Econ Focus is posted online.

Subscribe to Econ Focus

By submitting this form you agree to the Bank's Terms & Conditions and Privacy Notice.

Phone Icon Contact Us

David A. Price (804) 697-8018