How to Design Contracts that Deliver Results

The “pay for success” movement in the non-profit world is starting to take hold at the federal, state, and local levels. But a prerequisite is having some way of measuring success—and ensuring that funding models encourage it.

The Urban Institute has launched a new web resource to explain various forms of performance-based contracting aimed at delivering targeted, high-impact preventative social services where an intervention at an early stage could reduce the need for higher-cost services in the future.

Pay for success funding systems can take many different forms and already operate in different policy arenas. “They include value-based payments to hospitals and nursing homes, performance-based contracts with workforce providers, merit-based pay in schools, and performance-based payments to colleges and universities,” according to Patrick Lester, director of the Social Innovation Research Center at the IBM Center for the Business of Government. These involve many billions of dollars in annual public funding.

Two of the most prominent of these outcome-based funding systems are social impact bonds and performance-based contracting. A new IBM Center report by Lester offers a guide to understanding and selecting the best approach—through performance-based contracting or social impact bonds—depending on the situation at hand.

Social Impact Bonds

In 2010, the British government launched an innovative funding scheme, which it called Social Impact Bonds, where private sector investors committed funding upfront to pay for improved social outcomes that result in public sector savings. The investors were repaid by the government only when the outcomes were determined to have been achieved. This funding scheme has attracted substantial attention in the United States where it and many variations are being piloted, as noted in a report by the Government Accountability Office.

Like performance-based contracts, social impact bonds tie payments to a provider’s performance in delivering outcomes. However, social impact bonds tie payments to meeting pre-determined performance goals. Payments are not made until results are achieved. This creates substantial risks for providers, but it has attracted bipartisan political attention, and the attention of various non-profits, universities, and foundations.

In some situations, social impact bonds can hold a number of advantages over performance-based contracts:

• They provide the ability to measure the impact of services. Lester notes: “One of the greatest challenges faced by outcomes-oriented policies such as performance-based contracts is that improved outcomes do not guarantee im­proved impacts. Outcomes, like placing a child in a home, can reflect not only the work of a provider, but also other factors such the availability of foster parents, variations in the economy or local poverty rates, and potential perverse effects like cream skimming, tunnel vision, gaming, and fraud.” Measuring impact—something the provider actually does to make a difference—can sort out these effects, says Lester, “but it requires a rigorous evaluation.” While evaluations of outcomes-based funding mechanisms are relatively rare for many programs, Lester says they are common for social impact bonds.

• They are more suitable as demonstration projects. Performance-based contracts are usually adopted at a system-wide level, usually the result of restructuring funding streams that already exist. However, social impact bonds are more often created as demonstration projects to prove the value of a new initiative, such as preventive services.

• They are less likely to generate political opposition. Lester observes that systemic changes like performance-based contracts may generate opposition from existing service providers comfortable with established funding mechanisms. While social impact bonds may also draw opposition, the scale of smaller demonstration projects may generate less pushback than system-wide changes.

Performance-Based Contracts

Performance-based contracting’s early roots were in weapon system development in the military. However, states and localities have developed similar forms of performance-based contracts to create financial incentives or penalties for providers to meet pre-defined performance benchmark targets in social services. This approach differs from traditional fee-for-service contracts that use fixed payment rates for services provided. In the case studies in Lester’s report, the performance incentives are to place children in permanent homes with foster families instead of housing them in group homes.

Performance-based contracting has been seen as problematic in harder-to-define social services because program evaluations often found no relationship between the metrics used and the results achieved. However, the author notes: “there is reason to believe that the success of previous pay-for-performance efforts has had less to do with the concept and more to do with execution.”

While social impact bonds may be ideally suited as demonstration projects, performance-based contracts may be better suited to taking a policy initiative to scale. According to Lester, some advantages include:

• They are straight-forward and easy to understand. “While social impact bonds commonly require complex contractual arrangements among funders, service providers, independent evaluators, and government agencies, performance-based contracts are usually much simpler, with payments determined by law or regulation,” writes Lester.

• Upfront, third-party financing is not required. Outcomes-based funding systems typically do not rely upon external third-party funding arrangements, which are a key element of most social impact bonds. Third-party investors are not needed to cover provider costs until performance metrics are met and payments are made. Instead, performance-based contracts commonly combine upfront base payments with bonuses or penalties assessed after the fact—and the government funder bears greater risk.

• There’s flexibility allowed in changing service delivery approaches. Most social impact bond projects spell out in precise terms how providers will deliver their programmatic interventions. Lester says: “In theory, fidelity to specified evidence-based models provides added pro­tection for investors.” However, this does not give providers enough flexibility to adjust the proposed intervention to local conditions or changing circumstances and they can fail. Lester says: “Most performance-based contracts, by contrast, specify only the outcomes to be achieved, leaving providers freedom to decide how to meet them.”

• They can be scaled to serve larger populations. Lester writes: “While social impact bonds have drawn substantial interest, they may prove difficult to scale. Due to their high-risk profiles and relatively low rates of return, at least some experts are questioning whether they will be able to draw enough private investment capital. Moreover, current methods for determining success—such as randomized, controlled trials—may prove too impractical and expensive once they expand beyond the demonstration stage.” Performance-based contracts, on the other hand, have demonstrated their ability to scale in many different policy arenas and areas of the country.

“Taken together,” Lester concludes, “performance-based contracts and social impact bonds can be highly synergis­tic. While social impact bonds have many features such as rigorous impact evaluations and cost-benefit studies that make them well-suited as demonstration projects, performance-based contracts may be easier to scale. . . . Where performance-based contracts do not already exist, social impact bonds could provide proof-of-concept before rolling out a pay-for-performance system more widely. Where such contracts exist already, social impact bonds could be used to assess their impact and test potential improvements.”