December 2, Submitted by: Social Finance, Inc. 77 Summer St, 2 nd Fl Boston, MA 02110

Transcription

1 The PFS Incentive Fund RFI Response Strategies to Accelerate the Testing and Adoption of Pay for Success (PFS) Financing Models Docket ID TREAS-DO December 2, 2013 Submitted by: Social Finance, Inc. 77 Summer St, 2 nd Fl Boston, MA Jane Hughes Director of Knowledge Management jhughes@socialfinanceus.org Lara Metcalf Managing Director lmetcalf@socialfinanceus.org

2 Executive Summary The President s FY2014 Budget proposes to employ the $300 million Pay for Success (PFS) Incentive Fund to encourage innovation and accelerate the use of evidence-based approaches by lowering the risk associated with initial investments. As a social innovation financing intermediary that has engaged with many service providers on various state- and municipally-driven PFS procurement processes, Social Finance is pleased to submit this response to the Treasury s Request for Information. The Incentive Fund may play an important role in catalyzing this market. Support may be divided into three categories outcomes payments, credit enhancement, and capacitybuilding, as follows: First, using the Fund as a supplemental outcomes payor would help to catalyze the sector by seeding demonstration projects and by reflecting the full range of public-sector benefits accruing from PFS projects, e.g. through federal Medicaid dollars. Next, financing support through the Fund is also conceptually valid and may prove less cumbersome to implement than including the federal government as an outcomes payor. Federal officials have experience in providing credit enhancements (e.g. guarantees through the Small Business Administration), and the power of a federal guarantee for all or some of a program would greatly enhance the attractiveness of the transaction to potential investors. In particular, federal funds may be used to allay appropriations risk, which is otherwise daunting to many investors. Finally, capacity-building among state and local governments is another potential use of Incentive Fund dollars. Many governments lack the administrative data and technical expertise that are fundamental to the building of evidence-based programs. Investing in initiatives to develop data collection, sharing, and matching across various agencies and levels of government would be a major step forward for the PFS sector. Similarly, providing funds to support the work of local governments in developing PFS projects would be a useful exercise. Overall, monies under the Fund should be directed to the most promising program areas and projects, which include chronic illness, early childhood education, maternal and infant health, and recidivism. We believe that early PFS contracts should look to attract impact investors, i.e. those who actively seek opportunities that reflect their values and generate rewards for society as well as themselves. Accordingly, early PFS contracts should look to attract this type of capital by employing proven, rigorously tested interventions delivered by service providers with a high-quality evidence base of successful outcomes. Perhaps most important, at this stage each PFS project is bespoke. While the development of norms and standards is a worthy goal, the market is not yet this far developed; ongoing experimentation and on-the-ground experience are needed in order to determine standards and best practices going forward. 2 P a g e

3 Question 1: Instead of focusing on particular programs, the budget language proposing the Fund is broad in scope. What agencies and/or program areas are best suited for the Fund and why? What level of evidence exists in these areas about interventions that work? What is the threshold of evidence that a program should have in order to merit consideration for a PFS approach? What other factors should be considered in setting resource priorities for the Fund? Program areas best suited for PFS The President s FY2014 Budget proposes to employ the $300 million Pay for Success (PFS) Incentive Fund to encourage innovation and accelerate the use of evidence-based approaches by lowering the risk associated with initial investments. 1 Social Finance is a social innovation financing intermediary that has engaged with many service providers on various state- and municipally-driven PFS procurement processes including those in Colorado, Denver, the District of Columbia, Illinois, Massachusetts, Michigan, New York, and South Carolina. We are committed to identifying PFS-suitable programmatic areas and interventions. Core to our mission at Social Finance is the development of PFS-based financial innovations, such as Social Impact Bonds (SIBs). PFS projects have the potential to enhance taxpayer efficiency and public accountability through their emphasis on outcomes. However, they are not a panacea; not every social ill can be addressed through PFS-based programming. Partners including governments, service providers and market intermediaries must undertake stringent evaluation of various success criteria before deciding whether or not to launch a PFS project. The points below detail general criteria for assessing the viability of a PFS project: First, do no harm: Those initiating a PFS project must determine, first and foremost, that their work will not cause any negative, unintended consequences to the population that they hope to serve. For example, could excluding some needy individuals from services in order to provide a control group cause harm? Could family restoration services actually prove counterproductive? If the answer is yes, then the project must be redesigned or abandoned. Do the right service providers exist? A successful PFS project rests upon the ability of dedicated and skilled service providers to deliver social interventions on a larger scale. These organizations must have a ready supply of knowledgeable personnel with roots in the community and must be able to deploy proven social interventions. Is the financial case convincing? Rational investors will expect to see a solid financial case for a PFS program; there must be a strong investment thesis that the 1 "The Budget for Fiscal Year 2014: Department of the Treasury," The White House, 3 P a g e

4 expected payout from the program is greater than the cost of providing it. In other words, projected returns should exceed the cost of setting up, running, and paying for the interventions financed through a PFS contract. This assessment is part of our due diligence process at Social Finance. Are the outcomes measurable and verifiable? Evaluators must be able to measure outcomes using clear, objective data. Outcomes should be evaluated against a counterfactual either by projecting historical data forward or by establishing a control group in order to demonstrate that benefits arose as a result of the intervention and not an external factor. At Social Finance, we believe that the most promising areas for PFS contracts are those that include three key elements: Evidence-based intervention: Programs have been evaluated by independent parties and found to be effective; Sufficient net savings within time horizon: Results should be realized within a time frame that suits the needs of investors; Replicable and scalable: The interventions can be expanded to serve a larger population. Service providers exist in the sector with the capability to scale up. Figure 1: Promising Applications for SIBs 4 P a g e

5 PFS Evidence Base While some service providers have a solid, consistent evidence base that demonstrates successful outcomes, many others do not collect and assess data with the same rigor. Social Finance has sought to identify interventions and organizations that are supported by high quality evidence derived from randomized control trials (RCTs) and other robust experimental evaluation methods over a significant period of time. In the past year, Social Finance completed a Non-profit Service Provider Analysis, which sought to identify top-tier SIB-ready service providers culled from 10 different evidence-based registries (e.g. the Coalition for Evidence-Based Policy s Social Programs that Work and the Washington State Institute for Public Policy cost effectiveness studies). Preliminary analysis suggests that only providers employ SIB-ready interventions based on our research into their outcomes, evidence base, reputation, scope of operations, scale, and leadership. Of these, perhaps were Tier 1 or Tier 2 organizations, largely due to an evidence base supported by at least one RCT or multiple non-rct experimental studies. Further research and analysis may indicate a broader range of SIB-ready interventions, especially as social service organizations focus more attention on evidence-based work. The following provide examples of such high-quality interventions: Center for Employment Opportunities (CEO): CEO provides life skills, transitional jobs, job placement support and post-placement services to individuals recently released from prison in order to decrease recidivism and enable these persons to become productive members of society. MDRC, an independent social policy research organization, conducted a 3-year randomized control trial of CEO s services and concluded that the intervention significantly reduces recidivism at every level of the criminal justice system including arrests, convictions, and incarcerations. According to the study, CEO s impact only increased with the more disadvantaged or higher risk individuals. 2 Multisystemic Therapy (MST): MST is an intensive, family-focused and community-based treatment program for (often justice-involved) youth that addresses an array of systems in an adolescent s life. MST aims to help caregivers manage and nurture challenging adolescents more effectively. MST has been evaluated in eight RCTs involving over 700 violent and/or substance abusing juvenile offenders. MST evaluations demonstrated a 25-70% reduction in longterm re-arrest rates, a 47-64% reduction in out-of-home placements, decreased substance abuse, and better functioning families for male and female youths ranging in age from 12 to 17 and across ethnic groups. Additionally, follow-up 2 Megan Milenky, Cindy Redcross, and Timothy Rudd, "More Than a Job: Final Results from the Evaluation of the Center for Employment Opportunities (CEO) Transitional Jobs Program," MDRC: Building Knowledge to Improve Social Policy, V. 5 P a g e

6 studies (such as that of the Missouri Delinquency Project) have further bolstered the evidence of MST s long-run impact. 3 Nurse-Family Partnership (NFP): NFP pairs expectant first-time, Medicaideligible mothers with a nurse who provides home visits from early in pregnancy until the child s second birthday. The NFP service delivery model is a product of almost four decades in the research laboratory and over a decade of field implementation. Multiple (30+) NFP evaluations including three well-designed RCTs dating from 1977, 1988, and 1994 with different populations and geographies have demonstrated that the program achieves significant, sustained outcomes for high-risk families. Across a range of outcomes, the program s impact is more pronounced for higher-risk mothers who are living in communities of concentrated social disadvantage. NFP benefits include but are not limited to a 16% reduction in preterm first births, a 24% reduction in very closely spaced births (within 15 months postpartum) and a 31% reduction in child maltreatment through age PFS Evidence Threshold When considering a threshold of evidence for programs meriting PFS financing, the relatively nascent state of the market must be considered. As of yet, there is no single, proven approach to structuring PFS projects. For the foreseeable future, transactions will be done in a bespoke fashion, varying greatly depending on the policy priorities of the procuring government and the available evidence base of various service providers. Social Finance is dedicated to drawing private sector capital into the PFS space because we believe this ultimately represents a larger and more sustainable pool of funds than philanthropic dollars. In order to create a capital structure with significant funding from mainstream impact investors, PFS interventions will likely need an evidence base that is convincing and solid enough to attract investors. These investors will be interested in scaling what works, rather than developing new and innovative service platforms. However, for various reasons governments may be willing to consider interventions with weaker bases of evidence when procuring for a PFS contract. For example, a government may be interested in funding promising but novel interventions yet to undergo independent evaluation. Or, it may want to run two evidence-based interventions 3 "Evidence-Based Treatment, MST Proven Results from MST Services," Multisystemic Therapy, "Multisystemic Therapy: Research at a Glance," Multisystemic Therapy: Breaking the Cycle of Criminal Behavior, 4 Findings from these randomized controlled trials in Denver, Elmira, and Memphis have consistently demonstrated that NFP improves prenatal health and birth outcomes, child health and development, pregnancy spacings and economic self-sufficiency of families. See including: Kitzman et al., Journal of the American Medical Association, 278:8, , 1977; Olds et al., Pediatrics, 77:1, 16-28, 1986; Olds et al., Pediatrics, 110:3, , P a g e

7 concurrently, examining whether this produces amplified impact for a targeted outcome. Governments may be willing to test out less proven interventions/designs for experimental purposes; in this case, PFS financing based on philanthropic rather than mainstream investors could provide a more accommodating financing method as compared to typical budgeting procedures. It should be noted, however, that using PFS to finance unproven interventions is risky. Such a project would be less likely to attract private capital, making it more dependent on philanthropic funding until the intervention develops a solid evidence base. In summary, the threshold of evidence required for a PFS project will depend heavily on the target investors, a government s preferred cost of capital, and public sector policy objectives. Other Factors to Consider A $300 million federal PFS fund could go a long way toward catalyzing PFS contracts across a wide array of governmental jurisdictions and programmatic areas. Importantly, Social Finance believes that the true innovation of PFS is channeling a substantial, sustainable pool of private capital towards interventions that are designed to generate positive social outcomes for beneficiaries, create government/taxpayer benefits, and produce financial returns for investors. Indeed, evidence suggests that many investors are actively seeking opportunities that reflect their values and generate rewards for society as well as themselves. 5 Consequently, we believe that early PFS contracts should look to attract this type of capital by employing proven, rigorously tested interventions. If the PFS market takes flight, philanthropic dollars can be directed elsewhere. For example, these dollars can go towards running demonstration projects or evaluations for promising but unproven interventions that may one day be good candidates for PFS financing. Question 2: The budget proposal encourages maximizing the leverage of Federal funds by engaging intermediaries, including state, local and tribal governments. What other kinds of groups should be considered as intermediaries? Are there other organizational constructs that should be considered? The ability to demonstrate whether a PFS intervention produces the desired results is the backbone of the model. How can the Federal government encourage the adoption of low-cost yet rigorous outcome measures? What are some of the barriers to using administrative data in a PFS scenario, and how might they be addressed? 5 In its 2013 Insights on Wealth and Worth, U.S. Trust found that six in 10 wealthy individuals feel that they can influence society through their investments, and 46% would be willing to accept a lower return from investments that generated a greater positive impact on society. The full report is available at: Wealth-and-Worth-2013.pdf. 7 P a g e

8 The Role of Intermediaries in PFS Contracting Intermediaries such as Social Finance have played an integral role in developing the US PFS market. Indeed, social impact financing is not easy. Intermediaries can take on the tough and complicated work of creating a structure that will optimize incentives for service providers, offer a compelling investment opportunity to draw in new pools of capital, and promote beneficial use of scarce resources for governments and taxpayers. An intermediary is a party that facilitates the structuring and negotiation of a PFS contract between other stakeholders in a transaction such as the government, service providers, and investors. The UK s SIB Knowledge Box explains: Intermediaries have played a pivotal role in the development of social impact bonds, most notably Social Finance that originally developed the model and implemented it with the Ministry of Justice as the first social impact bond in Peterborough. A social impact bond could, theoretically, be developed without an intermediary. It could also involve more than one, playing different roles. 6 Intermediaries may offer a variety of services, including: Identify and conduct due diligence on potential service providers and interventions Conduct financial modeling and structure the PFS program Establish outcome metrics and evaluation methodology with all partners Negotiate PFS contracts to align incentives among stakeholders Develop the financing vehicle and raise investor capital Manage performance and spearhead mid-course corrections as needed Work with independent evaluators to assess performance outcomes Educate the market and serve as a knowledge resource for others A 2012 McKinsey report spells out the knowledge base and skill set necessary for an intermediary operating in this space: Deep knowledge of the relevant social issue Financing expertise: how to conduct due diligence, raise capital from investors, and structure deals Project management competence Collaborative and consensus-building skills 7 6 "The Knowledge Box," Cabinet Office: Centre for Social Impact Bonds, 18 Apr. 2013, 7 Laura Callanan, Jonathan Law, and Lenny Mendonca, "From Potential to Action: Bringing Social Impact Bonds to the U.S.," Mickinsey & Company, May 2012, Innovation/McKinsey_Social_Impact_Bonds_Report.pdf. 8 P a g e

9 Including an intermediary in a PFS-based project confers a number of advantages. The intermediary is able to facilitate negotiations by acting as a neutral broker among all stakeholders and by translating among partners who may not normally work together. An intermediary may contribute skills and experience that the other parties need but do not yet have in areas such as financial analysis, social programming, and negotiating a multiparty deal. This may significantly shorten the development time for a project. If the intermediary develops a strong track record of delivering results, investors may commit to a PFS project because they have confidence in the intermediary, although they may not yet have that same confidence in the service providers and/or interventions being delivered. 8 Including an intermediary may add to the costs of a transaction, but a high-quality intermediary should add value in terms of deal efficiency that more than offset those costs. Additionally, intermediaries may be willing to take a portion of their fees in the form of project-related success payments. This means that intermediaries assume some of the risk around performance outcomes, further aligning the incentives of all key stakeholders. Also, future intermediary organizations will, for the foreseeable future, likely remain nonprofits with access to philanthropic support that subsidizes some of the early-mover costs involved in developing PFS transactions. Intermediary Constructs This RFI response contains multiple references to the fluid nature of the US PFS market, and the intermediary role is not unique in this regard. Procuring governments may have differing or ambivalent views towards intermediary participation, giving rise to multiple constructs: No intermediary: The UK Cabinet Office Social Outcomes Fund does not require any intermediary participation in its PFS contracts. Additionally, some US state governments have shown interest in direct contracting with service providers, bypassing an intermediary. Single, full-service intermediary: This construct has the benefit of limiting parties to the deal, while also having in place a knowledgeable, non-partisan organization touching all aspects of the PFS transaction. There are currently only two US organizations specifically constituted to fulfill this full-service role: Social Finance and Third Sector Capital Partners. However, organizations in the evaluation, philanthropy, community development financial institution, and other sectors may be looking to enter this space as well. Intermediary organizations need not be exclusively focused on PFS development, but intermediary growth will be essential to developing the PFS pipeline. Joint intermediation: This approach has the benefit of specialization, in which 8 "The Knowledge Box," Cabinet Office. 9 P a g e

10 one party could focus exclusively on the financial side of the PFS contract, while another could have programmatic responsibility, ensuring the intervention is run with fidelity-to-model and outcomes are achieved. Social Finance and the NFP National Service Office (NSO), to cite one example, are looking to structure PFS contracts with this dual governance structure. As the financial intermediary, Social Finance would be responsible for contract negotiations, financial modeling, developing a fundraising strategy, raising money, and managing investor relations. As the program intermediary, the NSO would shoulder primary responsibility for identifying service providers, designing the program, and managing the project/making course corrections (see Figure 2 below). Figure 2: Example of Co-Intermediary Division of Responsibilities NSO Lead Responsibilities SF Lead Responsibilities Shared Responsibilities Program design / Fund-raising strategies PFS contract project development and implementation negotiation, project Service provider selection, Investor relations and management, and contracting, and ongoing reporting stakeholder management Financial modeling management Performance Capital market relations Modeling assumptions measurement design management Investor disclosure data / Program management documentation Evaluation design in conjunction with the State PFS Outcome Measures and Administrative Data The ability to properly demonstrate causation and measure outcomes is foundational to PFS contracts. The entire construct is premised on the concept that taxpayers should only pay for verifiable outcomes achieved. There is, however, some tension between the most rigorous outcomes measurement instrument (randomized controlled trials, or RCTs) and time/money constraints. RCTs are expensive and time-consuming compared to other outcomes measurement techniques, as they add operational complexities and additional requirements with regard to population and control groups. Thus it could be argued that top tier service providers like NFP which have already proven their effectiveness through multiple RCTs do not require additional RCTs with each new PFS contract. Additionally, researchers are working to design outcomes measurement techniques that place fewer demands on resources. 9 Different statistical methods may offer viable results, 9 Jon Baron, "Rigorous Program Evaluations on a Budget: How Low-Cost Randomized Controlled Trials Are Possible in Many Areas of Social Policy," Coalition for Evidence-Based Policy, March-2012.pdf, 3: Under certain conditions, low-cost ($50,000-$300,000) RCTs, the most rigorous evaluation method available to-date, may be carried out in a variety of policy areas. In traditional RCTs, 10 P a g e

11 while the use of a lower-cost counterfactual may also be acceptable with proper statistical analysis. Indeed, one potential use for the Fund may be to support projects where RCTs are not included in the outcomes measurement, especially if RCT evidence for the intervention has already been established. The development of lower-cost, less demanding but still rigorous outcomes measurement techniques would streamline and stimulate PFS development. According to the Harvard Kennedy School SIB Lab, historical baselines determined through the use of administrative data are often necessary to determine value from improved outcomes. These baselines are essential for intermediaries to model public sector benefits resulting from scaled interventions. 10 To date, PFS outcomes have been measured for a period of at least five years (and payments may be made on outcomes a decade or more beyond the contract s length), so accurate and comprehensive administrative data is essential even at the due diligence and contract negotiation stages. This data only becomes increasingly important as the parties develop an adequate evaluation methodology. For many PFS projects, outcomes measurement may require data to be matched across data sets that reside in multiple government agencies. For example, a list of recently released ex-offenders may be matched to data sets regarding subsequent convictions/sentences or reported income records in an unemployment insurance system. In order for government agencies to properly match administrative data sets, lawyers often establish inter-agency Memorandums of Understanding (MOUs), and adequate information technology infrastructure and staff must be in place. If this infrastructure does not exist, PFS contract stakeholders may need to procure the assistance of outside data consultants (local universities can also assist in this endeavor). A high degree of security must be insured for individual data used, while also allowing the intermediary and evaluators to perform their roles. There are examples of extant administrative data systems that meet the requirements necessary to conduct a rigorous, low cost RCT. For example: A $225, ,000 RCT was carried out to determine the impact of the Positive Parenting Program (Triple P) 11 in 18 rural and urban counties in South Carolina. This study found reductions in child maltreatment, out-of-home placements and the rate of hospitalizations for child maltreatment injuries. Low-cost, yet rigorous outcomes could be obtained using routinely collected data (e.g. child maltreatment records), rather than necessitating new collection techniques such as data collection comprises the largest share of costs. However, by using administrative data already gathered by public agencies in order to measure treatment and control group outcomes, collection and study costs may be greatly reduced, while still producing valid evidence that is of policy and practical experience. 10 Jeffrey Liebman and Alina Sellman, "Social Impact Bonds: A Guide for State and Local Governments," Harvard Kennedy School Social Impact Bond Technical Assistance Lab, June 2013, 11 For more information on Triple P, see 11 P a g e

12 surveys. 12 As part of a New York City Teacher incentive program, an RCT was conducted between 2008 and 2010, which measured a sample of 396 low-performing schools. If the treatment group produced positive impacts in student achievement, attendance, graduation rate, behavior or GPA, then teachers would be awarded a bonus of up to $3000. The $50,000 RCT used administrative data, such as state test scores, to determine improvement, rather than designing a new, costly system to do so. Through this evaluation, the program was determined to have produced no statistically significant effects, leading the city to cancel the initiative in order to direct resources elsewhere. 13 As Social Finance, Center for Employment Opportunities and New York State have worked diligently to structure a PFS-contract related to prisoner recidivism and employment, the process has been smoothed by the State s robust administrative data infrastructure. However, this is not true nationwide. Many other states have different data centers and face both technological and political issues when attempting to match data sets. Even in well-matched systems, timeliness of data retrieval is an issue and delays are common. As mentioned above, these shortcomings may necessitate the hiring of outside contractors in order to integrate data sets within agencies and across silos. However, this may create new issues around privacy (e.g. de-identification of data, application for data access, etc.). Social Finance has experienced these issues while trying to obtain and analyze patient records in a Fresno, CA asthma-related demonstration project (dealing with data from insurance providers, as opposed to governments). These concerns may be especially relevant in the likely event that health-focused PFS contracts are pursued in the future. Enabling administrative data systems to actively integrate and communicate across silos would be a major step forward for PFS contracting. A federal PFS Incentive Fund could earmark money specifically for this purpose. Support through the PFS Incentive Fund could take different forms. For example, the federal Fund could target administrative data integration at the state level using a broad net. (This action would have the advantage of positive spillover effects beyond PFS contracting). The Fund could also support initiatives within government to augment systems for data collection and analysis, to ensure that data is high quality, in a digitized format, and collected in a consistent manner. Alternatively, monies could be doled out to improve data collection and analysis specifically for individual transactions in jurisdictions procuring for PFS. In scalpel-like fashion these funds would be deployed to ensure successful due diligence and evaluation for PFS projects in active development. Questions 3: Outcome payments and financing support (e.g. credit enhancement, loans 12 Baron, Rigorous Program Evaluations on a Budget, Ibid P a g e

13 or advances) are two forms of assistance meant to complement one another in stimulating PFS approaches. What criteria should be used to decide how to split the Fund between these two forms of assistance? Should a certain proportion of the fund go toward outcome payments versus financing support, such as 50/50, 30/70, etc.? Question 5: Among the possible forms of financing support, would credit enhancements, loans or advances be most helpful? What role would financing support play in the overall structure of a PFS structure? We have chosen to respond to these questions together under the broad umbrella of how to optimize outcome payments and financing support through the federal Fund. In order to catalyze the market, the federal Fund could provide either supplemental outcome payments or financing support for early PFS transactions. Some examples and the issues surrounding these forms of assistance are covered below. Supplemental Outcome Payments Using the Fund as a supplemental payor of outcomes for PFS financings would help to catalyze the sector, facilitating transactions that may not otherwise get off the ground. Because the federal government would be a significant beneficiary of many PFS projects either directly or indirectly federal Medicaid dollars is but one example of this potential benefit there is a strong conceptual argument for including it as a payor. In fact, the federal government has already begun to implement this approach. In September 2013, the US Department of Labor awarded nearly $24 million in PFS grants to New York and Massachusetts, monies earmarked towards payments for employment and recidivism outcomes among formerly incarcerated individuals. The use of Fund monies for outcome payments would facilitate the development of PFS programs by seeding demonstration projects and by reflecting the full range of publicsector benefits accruing from PFS projects, as follows: Seeding demonstration projects: The Fund could be used to help provide outcome payments for new PFS applications, in order to incentivize state and local PFS development. Because the PFS approach represents a new way of doing business in the provision of social services, some governments may be reluctant to take up PFS on their own. For example, after Goldman Sachs, J.B. Pritzker, and the United Way of Salt Lake City crafted the first US SIB demonstration project related to early education (see our response to Question 6 for more detail), the Utah State Legislature narrowly missed passing the Results-Based Early Education Act, which would have enabled the state to act as outcomes payor. Consequently, the United Way agreed to provide outcome payments of up to $1 million for the first year of the project. The federal Fund has the potential to play a similar role, seeding the space with demonstration projects to help build a track record for the market. 13 P a g e

14 Outcome payments from the Fund could also incentivize PFS launches by helping to overcome the issue of silos across different government levels and agencies. The costs related to social challenges are unevenly distributed across various public-sector levels (federal, state, local) and agencies (e.g. corrections, criminal justice, health). Similarly, the benefits associated with a successful PFS program that addresses these challenges are also unevenly distributed. As a recent University of Pennsylvania study notes, Bureaucracies are necessary to public administration; they are how government gets things done. But dysfunction arises when threatening social problems or national crises require simultaneously addressing the needs of the whole person but the boundaries between bureaucracies are too rigid to permit coordination and collaboration. 14 This picture greatly complicates the economics for many contracts that would otherwise be a good fit for a PFS. Using the Fund to top up outcome payments would incentivize the appropriate public-sector groups to work together on multidisciplinary PFS programs that bridge gaps across traditional dividing lines. Encompassing holistic public-sector benefits: At this point, procuring states are inclined to structure outcome payments based on strict budgetary savings that accrue over a relatively short timeframe, which makes political sense. However, a federal government Fund can and should take a broader view, accounting for value in addition to budgetary savings reaped through Medicaid and other programs. PFS projects may create social value (e.g. better quality of life, lower crime rate) that does not accrue directly to state or federal budgets, as well as cost savings that materialize beyond the duration of the investment. An excellent example of this type of project is early childhood education programs, which include a range of comprehensive supports for at-risk children and their families from the prenatal period through the first five years of life in order to give children in poverty the best chance for success in school and in life. Over 20 years ago, studies such as the Perry Pre-School Study and The Carolina Abecedarian Project, articulated the significant impact of effective early childhood programming on the long term outcomes of children. Organizations like the Ounce of Prevention Fund support programs built upon this solid foundation of research and evaluation to narrow the academic achievement gap for children in poverty. Today, the Ounce is engaged in multiple research studies to prove the value of its early education programs for at-risk children in the current climate. The multi-year cross-sector impact of early education programs, like those provided by the Ounce, is apparent for individuals and communities that benefit from this type of programming Dennis Culhane, John Fantuzzo, and Heather Rouse, Actionable Intelligence for Social Policy: Using Integrated Data Systems to Achieve a More Effective, Efficient, and Ethical Government (University of Pennsylvania, November 2013), "Early Childhood In The News," The Ounce of Prevention Fund - Giving Children in Poverty the Best Chance for Success, 14 P a g e

15 However, much of this value is long-term in nature and not easily quantifiable. Without these benefits, some evidence-based interventions may not generate the proper economics for PFS transactions. In this context, the federal Fund could fulfill an important, supplementary role. Fund monies would incentivize states to launch PFS projects with the potential to deliver more holistic public-sector benefits, encompassing both long-term budgetary relief and societal value. However, despite these potential benefits from federal participation, designing and implementing a complex contract with multiple outcome payors can be cumbersome. Traditional federal grant guidelines and rules may not be easily adaptable to pay-forsuccess contracting, and federal involvement could also add indirect expenses and time to complete the transaction. As a general rule, mainstream impact investors prefer simplicity; a more complex contractual arrangement is by definition less streamlined, less attractive to investors, and less replicable. Thus it is important to ensure that federal participation through the Fund be as straightforward, streamlined, and simple as possible. Financing Support Using the Fund to provide financing support, including credit enhancements and/or subordinated loans, could also help to drive the PFS market forward. 16 Outcomes payments largely serve to catalyze state and local government participation in PFS programs, while financing support centers around catalyzing investor participation in these programs. The federal government is already involved in providing similar types of financing support, if not in the PFS sphere. One Small Business Administration (SBA) program guarantees up to 85% of commercial loans to small businesses. Additionally, Fannie Mae and Freddie Mac have been used to enhance access and reduce costs related to home ownership for low- and middle-income Americans. Two lesser-known examples of federal activity in this realm are included below: The California FreshWorks Fund: The fund aims to increase access to healthy, affordable food in underserved communities, spur economic development, and drive innovation in healthy food retailing. The debt facility for this project totals $132.5 million, with $7.5 million acting as catalytic first-loss capital (CFLC). The California Foundation put up one-third of the CFLC, in order to increase financing for healthy food access, and opened initial negotiations that led to the JP Morgan Chase Foundation s $2.5 million CFLC contribution. Additionally, through its Community Development Financial Institutions (CDFI) fund, the US Treasury provided $2.5 million to CFLC. As a consequence of this multi-sector co-investment, it became easier to attract FreshWorks investors at both the senior 16 Financing support includes various forms of credit enhancement that reduce an investment s risks and cost of capital, such as a guarantee for all or some of an investment s principal; preferred status for private investors; and insurance against certain types of losses. 15 P a g e

16 and junior tranche levels. For individual credit facility loans, losses of up to $7.5 million will be absorbed by the CFLC in order to protect senior investors. 17 Democracy Prep Charter School: This 2008 project aimed to turn a dilapidated church vestry in Harlem into an inspiring, yet affordable space for students. The total school investment is roughly $5 million, including $300,000 in first-loss capital from Civic Builders. Civic Builders received this money as part of an $8.3 million grant from the US Department of Education credit enhancement facility, meant for Civic s charter school capital projects. As in other projects, Civic uses the education money in conjunction with its own equity investments, which the CFLC protects. 18 The PFS market would be a beneficial target for federal credit enhancement. This is particularly true for promising interventions without a long track record of experimental evidence. In a PFS project, especially at this early stage, mainstream impact investors are generally more concerned with protection of principal than with maximizing return making federal principal protection (in whole or part) a valuable lure. The Australian experience provides a good example of the value of governmentsupported credit enhancements in catalyzing the PFS market. The first Australian Social Benefit Bond (SBB) was issued in spring 2013 with a guarantee of 50-75% of investor principal. The second, launched in October 2013, featured a two-tiered credit enhancement structure. Investors in the lower-return, lower-risk tier enjoy a 100% guarantee of principal but only 10% of their return is dependent on outcomes, while investors in the higher-return, higher-risk tier put their capital at risk but also have up to 30% of their return tied to outcomes. In financing the first US PFS contract, the Goldman Sachs Urban Investment Group (UIG) provided a $9.6 million loan to support programs aiding juvenile offenders at Rikers Island correctional facility. Given the esoteric nature of this transaction, Bloomberg Philanthropies furnished a $7.2 million grant, guaranteeing a large portion of Goldman s loan (in the event that outcomes are not met) and covering some of MDRC s pilot and intermediary costs. If program outcomes are achieved, Bloomberg and MDRC will work to repurpose the grant toward future efforts. In this fashion, philanthropic capital worked to catalyze the still nascent US PFS market. Other early-stage projects seem likely to include similar forms of financing support, although probably at lower levels as time goes on. Foundations have limited capital and resources, which leaves a gap that the federal Fund may fill. Below are a few additional considerations with regard to the federal government s potential role as provider of credit enhancement for PFS financings: 17 Amit Bouri and Abhilash Mudaliar, "Catalytic First-Loss Capital," Global Impact Investing Network, Ibid P a g e

17 A PFS financing with the full faith and credit of the federal government behind some portion of the investors principal would be much more attractive to investors than a deal without such financing support, which would lower the deal s cost. However, the government should be ready for a robust response to potential inquiries about why it is supporting private either philanthropic or commercial investors (e.g. helping to catalyze a market that directs taxpayer dollars to social programs that reliably produce positive outcomes). Credit enhancement by the Fund may eliminate or greatly reduce investor demands that the state government escrow funds for outcome payments, making PFS projects more attractive to state officials. Timing is critical in the capital markets, so Fund administrators would need to develop a streamlined and straightforward process for implementing a guarantee facility. This process should be quick and fluid in order to meet the needs of investors; excessive delay in processing the facility would render it useless in the context of the capital markets. The Fund should lay out a concrete strategy, either generally or in regard to individual PFS grants made, regarding what would happen to grant monies that go unused in a successful PFS financing. In NYC, the Bloomberg funds would be rolled into future projects. Conclusion The federal Fund can catalyze the PFS market through either outcome payments or various forms of financing support. Given that the concepts and processes of certain financing support structures are familiar to federal officials; this may make financing support facilities easier to develop than outcome payments. However, financing support may be more difficult to justify politically, for it could be seen as propping up wealthy investors or well-endowed philanthropies. At this stage in the market, it would be premature to establish a fixed proportion of the Fund for outcome payments or financing support. Both are conceptually valid and have the potential to add great value to PFS financing. Initial PFS projects are still bespoke in nature rather than standardized, meaning that one may have different financing requirements than another. Over the longer term, learnings from early projects can and should be used to determine the practical advantages and disadvantages of each strategy. Question 4: Is there an optimal structure for both the timing and tiering of outcome payments? For example, should the projects allow for some degree of progress payments based upon achievement of early outcomes? Should the projects allow for bonus payments for extraordinary performance? What are the trade-offs of adapting different structures to different projects versus supporting a standardized approach? At this still early stage of PFS development a period in which key stakeholders may have little experience working together a federal Fund must retain a high degree of 17 P a g e

18 flexibility in order to stimulate growth and best practices. Currently, it is too early to determine what the optimal structure for timing and tiering of outcome payments might be; indeed, there is a distinct danger to standardizing before on-the-ground lessons have clarified the benefits and issues associated with various structures. Within these broad parameters, outcome payment structures will necessarily be customized depending on the specific characteristics of a transaction. For example, the state may target a promising yet unproven intervention or holistic combination of evidence-backed interventions that have yet to be evaluated in unison. Under these circumstances, a PFS project would depend on support from philanthropic investors, and payments may be tiered depending on the willingness of investors to take on the inherent increase in risk associated with the project. Conversely, for a single intervention like Nurse-Family Partnership which is backed by multiple longitudinal RCTs and would be more appealing to mainstream impact investors using tiers as a method of protecting investor principal may not be necessary. Ongoing experimentation is needed in order to develop norms and standards around these issues. However, some guidelines for this customization should be noted: Progress payments (i.e., providing some portion of outcome payments along the way rather than at the end of the project) could add significant value to a project by increasing investors internal rate of return (IRR). Progress payments could also facilitate programs with longer timelines to realize measurable outcomes, such as early education. Investors would be more willing to invest in longer-term programs if they received some portion of outcome payments during the course of the project. The Harvard Kennedy School SIB Lab has warned against the use of interim payments for financing the work of service providers, as follows: While it will often be appropriate to make interim performance-based payments before the end of the project, to avoid shutdown risk in the event that early performance targets are not met, these interim payments should not be used to finance service provision. Social impact bond projects should be designed with a long enough duration of services to allow for learning and mid-course corrections to occur. 19 Bonus payments for exceptional outcomes could add significant value by increasing IRR for investors and by incentivizing service providers to maximize outcomes. However, bonus payments could trigger a public backlash if they are perceived as too generous to investors. Bonus payments could also create perverse incentives if they are not carefully structured, e.g. by incentivizing service providers to cherry pick subjects who are most likely to succeed over the short term. On the other hand, bonus payments to service providers would enhance risksharing and the alignment of incentives among project partners. Most important, whatever structure offers the greatest chance of success for an individual project should be utilized. As noted above, PFS-based financing in the 19 Liebman and Sellman, "Social Impact Bonds: A Guide for State and Local Governments. 18 P a g e

19 social sector is highly customized at this stage, so flexibility is more critical to success than standardized structures. At the same time, there is value in standardization once the market has developed beyond its initial stages of experimentation and learning. Eventually, standardization in overall PFS structure will allow investors to gain greater comfort with the instrument, enabling intermediary organizations to increasingly attract private capital into PFS deals. Additionally, as more PFS contracts are negotiated, the process will be increasingly streamlined, decreasing the time involved for all parties to bring such contracts to fruition. Standardization, to the extent that it proves possible for this market, will only help this process along. Question 6: Please suggest one or more examples of promising PFS projects or programs. For each example, what are its characteristics or features that make it a good candidate for PFS? Who would be the key partners and what would be their roles? How would the activity be funded? How would risks be shared and interests aligned among the partners? What might be appropriate outcomes and metrics? Over what timeframe would outcomes be determined? Examples of promising PFS program areas and projects As noted above, a project may be a good candidate for PFS if it features a strong evidence-based intervention, sufficient cashable savings within a reasonable time frame as well as substantial social value, and a model that is both replicable and scalable. In assessing whether a program is appropriate for PFS, it is also important that the right service provider(s) exist in the given issue area and geography. In accordance with these guidelines, Social Finance has identified four issue areas that meet these criteria and offer substantial potential for a successful PFS program: chronic illness, early childhood education, maternal health, and prison recidivism. The following describes each of these issue areas, including key partners and funding models associated with promising projects. With the exception of recidivism, all of these projects are in the pilot or development phase. Chronic Illness: The US spends far more on per capita health care than any other industrialized nation, roughly 2.5x the OECD average. 20 According to the Centers for Disease Control, more than 75% of healthcare costs can be traced back to chronic illnesses, although these are generally among the most preventable and/or manageable conditions. The combination of need and the potential of preventative interventions make this area ripe for PFS financing. 21 A promising PFS project in this area is: 20 "Americans Spend over Twice as Much Per Capita on Healthcare as the Average Developed Country," Peter G. Peterson Foundation, 21 "Chronic Disease Prevention and Health Promotion," Centers for Disease Control and Prevention, 19 P a g e