Social Impact Bonds: Implementation, evaluation and monitoring

48 downloads 296 Views 2MB Size Report
Development Discussion Paper: 2016-04. Abstract. Traditional ... Social impact bonds (SIBs) are „pay-for-success‟ pr
Social Impact Bonds: Implementation, evaluation and monitoring Foroogh Nazari Chamaki Department of Banking and Finance, Eastern Mediterranean University, North Cyprus E-mail: [email protected] Glenn P. Jenkins Department of Economics, Queen‟s University, Kingston, Ontario, Canada, K7L3N6, and Department of Economics, Eastern Mediterranean University E-mail: [email protected] Development Discussion Paper: 2016-04 Abstract Traditional approaches to public policy increasingly fail to resolve social challenges, particularly in the field of criminal justice. High rates of juvenile recidivism, for example, are often linked to inequality in education and persistent, long-term unemployment—factors which, while complex, are nonetheless conducive to preventative strategies. Social impact bonds (SIBs) are „pay-for-success‟ programs that attract private-sector, upfront funding for social interventions. If the program achieves agreed targets, taxpayer funds repay the investor. If the program fails to meet agreed targets, investors take the loss. This innovative form social finance through public-private partnership (PPP) has helped spur efficiencies and improvements in the provision and outcomes of criminal justice services. However, the success of a SIB depends on careful implementation, evaluation and monitoring.

Keywords: Pay-for-success, social service, social impact bond (SIB), public-private partnership (PPP), social finance, service provider. JEL Classification: H53, A13, L33

1

INTRODUCTION Social Financing and Social Services Social financing relies on investors willing to bear higher risks and lower financial returns—a model that combines the (financial) incentives required by commercial investors with the (social) gains sought by philanthropic organizations and charitable donations. Over recent years, a number of western countries have witnessed rapid growth in the social financing of services provided by state and local governments, charities and philanthropic organizations. In the US, social financing has expanded at a faster rate than traditional commercial investments. In 2010 social financing in the US totaled $3.07 trillion (US SIF, 2010), rising to $6.57 trillion in 2014—a 76% increase on the $3.74 trillion achieved in 2012 (US SIF, 2014). Social Impact Bonds or Pay-for-Success Contracts Social impact bonds (SIBs), also known as pay-for-success (PFS) contracts, are multistakeholder mechanisms designed to finance more effective social-service interventions at lower cost to the taxpayer. A SIB is a contract between one or more public entities (at the local or state level), investors and a private service provider. The public entity specifies the social-impact goals, the investors shoulder upfront costs, and the service provider implements the project using the investor funds. Once the project is complete, a monitoring entity assesses the outcomes achieved. If these are in accordance with those stipulated in the contract, the public entity pays investors their principle investment, plus an additional return linked to specified project objectives. These additional returns are funded by savings accrued to the government by implementing the project through the SIB model (Figure 1). 2

Figure 1: Basic Model of SIB Contract

If the project does not meet agreed targets, investors are not repaid. A long-running contraction in the global economy has led to increasing demand for public services amidst ever-shrinking government finances. Essential services such as those related to criminal, particularly juvenile, justice, are often viewed as failing or under unsustainable strain. The SIB model is therefore an important tool in the provision of more efficient and more effective social services. A feasible SIB project depends on specific parameters that must be clearly detailed in the project contract. These parameters are outcomes, rate of return (ROR), service delivery, and project period and payment point. 3

The successful implementation of a SIB depends on a number of factors. First, is a SIB a good fit with the project in question? Second, what is the best policy intervention to pursue? Third, which analytical tools are best suited to the project model and specifications? Do these tools enable the development of clear formal structures appropriate to all parties? Fourth, what are the project phases and how will each be monitored? And fifth, how will it be determined if the project has achieved target outcomes (Liebman and Sellman 2013)? The SIB model is not ideal for every project or policy area. Important preconditions are: the potential for meaningful cost savings; the availability of clear and measurable outcomes; contracts properly adjusted to reflect the responsibilities of multiple stakeholders, including investors public entities and service providers; and the appointment of an effective coordinator. SIBs are most appropriate in the field of high-cost services, the provision of which is essential to the wellbeing of society as a whole. These include criminal justice as well as education, job training, healthcare, the provision of disability services and foster care. Notwithstanding the success of the SIB model, however, governments must not use SIBs as an excuse to cut the budget for social services. Policy-makers must guard against the withdrawal of government from the design and provision of essential services. If the core, founding mechanisms for the delivery of social services are lost, SIBs will not fill the gap. At the same time, the success of SIBs should not overshadow efforts to improve and develop other financial tools and mechanisms. In short, the SIB model and others are not a replacement for but a means of improving traditional social-service systems. The independent monitoring and evaluation of SIB impacts and outcomes is critical, promoting accountability and helping to keep projects on track. Close monitoring and regular evaluation

4

also provide a mechanism by which projects can be halted due to cost overruns as well as facilitating the learning process in the context of future implementation of SIB projects. Advantages of the SIB model Under a SIB model, the public sector (state or local government) pays for success—for outcomes and results that provide a reasonable cost-saving to the taxpayer. The model transfers the risk of financial losses incurred in the implementation of ineffective projects from the public to the private sector. Another significant benefit to public sector bodies is the opportunity to explore innovative solutions to intractable social problems. Innovation poses both great risks and great rewards. However, government entities with limited budgets cannot afford such risks. SIBs enable government to transfer the risk of social innovation to private sector interests with greater flexibility and resources, in exchange for the opportunity to realize a profit while engaged in an altruistic activity. As such, a great number of private sector organizations view SIBs as a winwin instrument. Finally, SIB projects improve performance and lower costs. A focus on prevention rather than remedial interventions is more efficient, resulting in more effective outcomes and higher quality services. In addition, the strict feasibility analysis and step-by-step monitoring of a SIB project facilitates the achievement of specified outcomes while controlling costs, as well as enabling rapid adjustments to adapt to new or unforeseen circumstances. SIBs benefit all parties—government, private sector entities and society as a whole.

5

LITERATURE REVIEW AND HISTORICAL DEVELOPMENT Literature Review The literature relating to social impact bonds (SIBs) can be divided into two. The first group describes the origin and history of the concepts underlying SIBs—how they emerged and why they came to be so widely used. This literature focuses on the UK, where SIBs were invented and first implemented, and bears particular testimony to the role of SIBs in a fiscal environment dominated by restricted government budgets amidst rising demand for social services. The second group of literature describes the legislative and regulatory structures governing the implementation of SIBs in the UK and elsewhere in Europe, as well as Canada and the US. This literature highlights variations in legal and regulatory structures between regions and countries, as well as the evolution of project implementation. Origin and History of SIBs The use of SIBs originated in the UK in 2007, when Prime Minister Gordon Brown asked the Council on Social Action (CoSA) “to generate initiatives through which government and other key stakeholders could develop and celebrate social action” (Strickland, 2010). CoSA was established by the Office for the Third Sector—a Cabinet Office group covering charities, voluntary communities, social activists and philanthropists—as an independent advisory group tasked with finding new ways to control the cost to government of social services. SIBs were presented as one novel solution, creating innovative partnerships between government, service providers and investors to address social issues through cost-effective projects based on concepts of early intervention and prevention.

6

CoSA began identifying feasible SIB-packed projects, examining costly services that would gain from a focus on root causes rather than consequences—areas such as criminal justice, education, homelessness, foster care, and health and mental health care. The first SIB project was launched in September 2010 at Peterborough prison in Cambridgeshire, England—a category B prison operated by Sodexo Justice Services (Disley, 2014). Under the scheme, the UK Ministry of Justice raised £5 million from charities and private investors to pilot a six-year intervention project aimed at reducing rates of recidivism in a group of 3,000 adult male former inmates. Project partners were the Ministry of Justice and the Big Lottery Fund, with Social Finance (UK) acting as private investors and bond-issuing partners (Centre for Social Impact Bonds, 2013). Investors in the Peterborough SIB are mainly philanthropists and charitable trusts. A total of £5 million in social finance was raised from 17 investors, including the Henry Smith Charity, the Barrow Cadbury Charitable Trust, the Johansson Family Foundation, the Esmée Fairbairn Foundation, the Lankelly Chase Foundation, the Friends Provident Foundation, the Panahpur Charitable Trust, the Monument Trust, the Tudor Trust and the Paul Hamlyn Foundation (Social Finance, 2010). The Big Lottery Fund provided £6.25 million in funding. Although final evaluations are not yet available as the project is ongoing, interim data indicate that ex-offenders attending Peterborough prison SIB project services have recidivism rates 8.4% lower than the control group. If these results are sustained, investors can expect to receive a return on their funds. However, investors will not receive early returns as the 8.4% reduction is not sufficiently above the minimum agreed threshold of 7.5% (Perakis, 2014). The main logic driving the Peterborough SIB project is to reduce costs. According to the UK Ministry of Justice, 60% of prisoners will reoffend within a year of their release. The annual cost

7

of crime committed by reoffenders in terms of policing, damage to property, court and legal fees, and health care is estimated at almost £10 billion. For every £1 invested in the Peterborough prison SIB, the UK government saved £10. Another SIB project for ex-offenders is run by the St Giles Trust, offering support to 3,000 shortterm former inmates with the aim of preventing reoffending. This six- to eight-year project generates £5 million to provide participant with education and training, and has resulted in a 40% decline in rates of recidivism. According to the project terms, investors will receive 7.5% returns if the rate of reoffending is reduced by 10% compared to a control group, rising to a maximum of 13%. Payments to investors are made in years six and eight of the project. The UK Ministry of Justice has announced a further ten SIB pilot projects in the criminal justice sector. Six first-round pilot projects are located in Nottingham, Birmingham, Merseyside, Perth, and London (one in Shoreditch and one in Newham). A further four second-round pilot projects are located in Greater Manchester, South Wales, the Thames Valley area and West London (an area comprising six local authorities). All ten projects have focused on young people, particularly the homeless, drug users, and the mentally or physically disabled. Most European societies expect government to provide citizens with a wide range of high quality social services. The private provision of such services is unusual and even difficult under prevailing legal structures. In such contexts, SIBs have been criticized as a process by which public responsibility for the welfare of society as a whole, is replaced by private interests working to achieve specific, narrowly defined outcomes. However, the pressure of contracting economies and restricted public finances, combined with ever-increasing demand for social services, have led many European governments to accept SIBs as an opportunity to maintain a high standard of social services. In such cases, governments have tended to overcome initial

8

political resistance and legal hurdles by promoting SIBs as an essential source of seed capital in efforts to develop innovative social programs. One of the first SIBs to be implemented in continental Europe was launched in Belgium in January 2014, by the Brussels Employment Agency and the government of the Brussels-Capital Region. The three-year project aims to reduce recidivism rates of juvenile ex-prisoners, aged between 11 and 18 (Dermine, 2013). The project was developed with the assistance of Jeffrey Liebman of Harvard‟s Mossavar Rahmani Center of Business and Government. Maryland became the first US state to adopt the SIB model, in 2010 (Mosenson, 2013). The project aimed to reduce the length of stay of children in foster care, and to reduce recidivism in juvenile justice programs through multisystemic therapy, job training, paid internships and other youth-development activities (Open Society Foundations, 2012). The first SIB contract aimed at reducing juvenile recidivism in the US was launched by the state of New York in 2012, for work with Rikers Island prison. The investor Goldman Sachs paid the intermediary $6.9 million to implement the project. If the rate of recidivism is reduced by 10% the principle will be repaid in full, with an additional $2.1 million to be paid if rates are reduced by more than 10%. However, the contract is unlike previous SIB mechanisms used in New York, as the maximum Goldman Sachs can lose has been limited to just $2.4 million (Callanan, 2012). Another, four-year, recidivism and employment SIB project was launched by New York Governor Andrew Cuomo in December 2013, with $13.5 million in investor funds (Humphries, 2014). In May 2011, the US state of Massachusetts‟ Executive Office of Administration and Finance issued a request for information (RFI) for a SIB program, announcing in August 2012 that two SIB projects were to be pursued in the areas of juvenile justice and chronic homelessness (Mass,

9

2012). The Massachusetts juvenile justice SIB was launched in January 2013 as a seven-year project involving 1,000 young men. The project raised $18 million from investors, including New Profit, Goldman Sachs, the Laura and John Arnold Foundation, the Kresge Foundation, the Boston Foundation and Living Cities. If the project meets its targets, Massachusetts will pay investors a total of $27 million (Humphries, 2014). Interest in SIBs has rapidly gained pace in the US. In 2012 President Barack Obama announced $20 million for SIB grants through the USA Employment and Training Administration and the Department of Labor (Dermine, 2013). According to a report by McKinsey entitled “From Potential to Action: Bringing Social Impact Bonds to the US,” there is significant additional scope for this form of financing (Hansbrough, 2014). In 2013 alone, Harvard‟s Kennedy School of Government offered technical assistance to authorities in seven states—Colorado, Connecticut, Massachusetts, Michigan, New York, Ohio and South Carolina—and two cities— Chicago, Illinois and Denver, Colorado (Humphries, 2014). The Success and Failure of SIBs Although there is considerable interim evidence of the success of SIB projects, a complete assessment of their full potential will depend on data to be compiled at the end of ongoing projects. Equally important, however, is the available evidence relating to unviable, failed SIBs. One such example comes from the UK. Birmingham City Council in partnership with local National Health Service (NHS) primary care trusts (PCTs) launched the Be Active program in September 2009, with the aim of providing free access to local sports centers during off-peak hours. The project succeeded in encouraging more people to participate in physical activity, with measurable improvements in health calculated in quality-adjusted life years or QALYs, including decreased rates of smoking and an increase in subjective well-being (SWB). However, because 10

achievable benefits were equal to costs, the savings required by project investors could not be generated (Marsh et al., 2011). Another example of an unviable SIB is the US-based Wyman Center‟s Teen Outreach Program (TOP). Wyman received a grant to support its teen-pregnancy prevention program—one of several organizations in 32 US states to receive such a grant. A four-year SIB program involving 8,000 female students successfully reduced the incidence of pregnancy during participants‟ school years. However, feasibility analysis revealed that the savings generated were not sufficient to cover project costs (Dao, 2012).

INSTITUTIONAL STRUCTURES FOR SIBs Social Impact Bond Contracts SIBs represent a completely new funding mechanism. The method of commissioning SIBs, their financing, the forms of contract involved and structure of partner relationships all differ from those traditionally used—even analytical concepts and methods of evaluation are different from the processes commonly employed by public bodies in the delivery of social services. The usual approach to the outside provision of social services involves government agencies seeking contractors through a competitive process, with the goal of obtaining the best value for money. Under the SIB model, a contract is drawn up between the public authority and a service provider, chosen on the basis of a proposal that seems a feasible means of delivering cost savings to government. Building a SIB is obviously a collaborative endeavor. In order to operate on a larger scale, a competitive market for SIBs will be necessary. However, the development of such a market would have implications for the size of commission required

11

to attract participants. The larger the scale, the more likely it is that public authorities will make negotiate with intermediaries rather than providers

Different SIB models Social impact bonds are not suited to every project but rather depend on certain specific criteria. SIBs are appropriate to fields such as juvenile criminal justice, homelessness and health care, however model of SIB requires a particular form of evaluation and assessment. In some models, there is no need for an intermediary other than the service provider. In these cases, not-for-profit investors and philanthropists fund the project according to a contract with the service provider, which uses this working capital to deliver services. If the targets set out in the contract between the government and the service provider are achieved, the government pays the investor on a performance basis, covering the initial investment and plus the determined rate of return (Figure 2). In this case, again, an independent evaluator is needed to assesses implemention and to certify results (Figure 3).

12

Figure 2: The Social Impact Bond Model (1)

Figure 3: The Social Impact Bond Model (2)

13

In cases where there is not a contract between the government and investors, and every transaction occurs through a service provider, the service provider contracts with the investor, using working capital to meet specified targets. After an independent evaluator certifies that those targets were met, the government pays the service provider the sum invested plus interest, as agreed in the contract between the government and service provider (Figure 4).

Figure 4: The Social Impact Bond Model (3)

In rare cases, SIBs are contracted between the national and local levels of government. In this case, local government accumulates working capital for the project and implements it through a service provider. When a specified target is met, the national government pays the local government the upfront fund plus interest, as determined in a contract between the local and state governments (Figure 5).

14

Figure 5: The Social Impact Bond Model (4)

SOCIAL IMPACT BOND CASE STUDIES Be Active The Be Active program was implemented in the city of Birmingham to offer free-of-charge access to gyms, swimming pools and exercise classes during off-peak hours, with the aim of improving health by removing barriers to access and promoting social integration (Marash et al., 2011). The program was expected to increase rates of participation in physical activity, reduce levels of smoking and increase opportunities for shared social experiences. The monetary value of an increase in physical activity and decline in smoking can be estimated in terms of the reduced cost of health care and associated increase in government savings, and in the improved productivity and health-related quality of life of participants, as measured in quality-adjusted life years (QALYs). Improvements to subjective well-being (SWB) were an expected result of the opportunity for shared experience (Marash et al., 2011).

15

The Be Active program was evaluated through a feasibility study and cost-benefit analysis (CBA), as follows. An estimated 140,000 of Birmingham‟s 1.1 million residents participated in Be Active each year, generating „profits‟ of £445.2 million. These profits include healthcare savings, QALYs and productivity gains, as well as a 5.9% increase in participants‟ life satisfaction (a measure of SWB) (Marash et al., 2011). The cost-benefit ratio of the Be Active program was calculated to be an average of £21.30 for each £1 invested. However, the amount varied according to investors‟ incurred costs and benefits. The return for the local health authority, for example, was £22.80 per £1 invested. The greatest benefits were in QALYs, while the smallest (£2.60 to £0.50 per £1 invested) related to medicinal-cost savings. In addition, the Be Active program resulted in decreased absenteeism, improved productivity and increased tax payments to the Treasury. Most significantly, the cost per QALY gained was £1,164.60—far below the UK‟s National Institute for Health and Care Excellence (NICE) value of £20,000. As such, the Be Active program represented an efficient investment of resources in a public health intervention. Feasibility analysis using the data above indicates that the profits gained here were not sufficient even to cover investors‟ principal if just cost savings (realizable and unrealizable) are considered. But if investors were willing also to pay for QALY gains, Be Active would be feasible, producing a return to investors. According to cost-benefit analysis of the program‟s first five years, just a small portion of public-sector costs were recouped through this SIB. Local health authorities reaped the benefits of reduced costs of secondary care of lung cancer, type-II diabetes and other illnesses caused by a lack of physical activity and smoking. However, the greatest gains were to individual participants who realized an improved quality of life and greater SWB,

16

as well as productivity gains stemming from reduced absenteeism of employed participants (Table 1).

Table 1: Be Active Benefits Used in Cost-Benefit Analysis Source: (Marash et al., 2011)

Costs and benefits can be considered through different dimensions: perspective (different stakeholders), timing and realizability (Table 2).

Table 2: Costs and Benefits Attributed to Stakeholder

The benefits are calculated over the lifetime of individuals for 50 years, with benefits classified in three groups: 1 to 5 years (short-term), 6 to 15 years (medium term) and 1 to 50 years (lifetime). Benefits were classified as realizable according to the following criteria, used to estimate their monetary value: 17



Healthcare cost savings. Avoided costs of medication are the only realizable healthcare savings.



Productivity gains. Increased corporate tax payments due to increased income to employers were realizable productivity gains.

All figures are calculated using 2011 prices and a 3.5% discount rate, in accordance with the Treasury‟s Green Book—guidelines for UK public-sector bodies on the appraisal of project proposals.

Cost-benefit analysis of Be Active The aggregate cost of the Be Active program is estimated at £22 million over five years, with benefits exceeding costs by a total of £445.2 million for 140,000 participants. This net benefit includes £28.7 million as cash savings, £39.2 million as productivity gains and £377.2 million as improvements in quality of life (QALY). Cost-benefit analysis over the life-time of an individual represents £3,202.70 in benefits per person for every £1 invested in the SIB, with cost per person per year of £33.80. The Be Active program therefore generated an average benefit of £21.30. The Be Active program was funded through a partnership between Birmingham City Council and local NHS primary care trusts (PCTs). Assuming that Be Active users would stay in the program for five years, the total program cost per person after discounting would be £158. The total cost of Be Active was therefore calculated as being £22 million, of which the local PCTs funded £18.7 million (85%) with the local authority providing the remaining £3.3 million (15%). The results of the Be Active program can be summarized as follows:

18

 8% increase in physical activity of at least three 30-minute sessions per week (Lyon et al., 2011; Harland, 1999; York Health Economics Consortium, 2007).  3% increase in likelihood of quitting smoking (Lyon et al., 2011).  74% increase in likelihood of participating in a shared experience (NHS of South Birmingham, 2010). As stated above, the possible health-related benefits of the program accrued to participants. For instance, the chance of suffering a stroke was reduced by 8% as a consequence of long-term physical activity; the chance of developing type-II diabetes declined 7%; quitting smoking reduces by 2% the risk of developing lung cancer and by 5% the risk of developing myocardial infarction. In addition to generating £20,000 per QALY in the avoidance of such diseases, the program also generated thousands of pounds‟ worth of savings for the NHS in treatment costs (Matrix, 2011). Furthermore, regular physical activity reduces by 29% the number of yearly visits to physicians, generating additional cost savings for the NHS. In addition to all these benefits, being an active program participant contributed to a 0.04-point increase in life satisfaction at a significance level of 0.05, on a scale of 0 to 1. The opportunity for a shared experience presented by the Be Active program therefore also has a small but significant effect on SWB (Table 3).

19

Table 3: Lifetime Benefits by Stakeholder and Type of Resource

The net present value (NPV) of the lifetime benefits per Be Active participant is calculated as being £3,202.70, including the monetary value of smoking and physical activity. The results of cost-benefit analysis therefore indicate that the Be Active program is cost-effective and represents an efficient use of public resources. These results are based on as broad a range of data as possible: rates of increase in physical activity and of reduction in smoking; likely reduction in prevalence of diseases associated with low levels of physical activity and with smoking, and associated cost savings; and gains in quality of life (Table 4).

20

Table 4: Returns by Stakeholder and Type of Resource (GBP, 2011 prices)

Feasibility analysis of SIBs The cash savings generated by a SIB-backed project must be sufficient to cover the costs of implementing and running a given service, as well as investors‟ principal and additional return payable on meeting certain targets. One critical benchmark in evaluating the feasibility of a SIB is therefore the cash value-added of achieved outcomes. In other words, whether a project‟s implementation will generate sufficient cash savings for public service commissioners, and whether the use of these savings to fund a SIB is of comparatively greater benefit than their use to fund a non-SIB project. This case study is mainly focused on evaluating the cash savings of expected outcomes, as a critical first step in assessing the feasibility of a SIB. Two additional criteria for a successful SIB are measurable outcomes and a comprehensive contract based on expected outcomes (Social Impact Bonds, A Technical Guide, Social Finance, 2011). The feasibility of the five-year Be Active SIB was assessed under two scenarios, one assuming annual real rates of return of 5% and the other of 7%. 21

Typically SIB contracts consider three types of benefits and realizable cost savings: (1) realizable cost savings not including costs of medication; (2) non-realizable cost savings; and (3) the monetary value of QALY gains. Service commissioners pay for results in areas (1) and (2) as follows: 80% of short-term cost savings for first 5 years, and 20% of medium-term cost savings from years 6 to 15. In the case of area (3), commissioners either do not pay for gains related to QALYs, or pay 80% of a QALY in the short-term (according to the NICE QALY value of £20,000) and 20% in the medium-term. Table 5 outlines two contract scenarios (A and B) under which a Be Active feasibility analysis is conducted.

Table 5: Sample Social Impact Bond Contract Feasibility Analysis

Key differences and similarities relate to: Realizable cost savings. Commissioners pay 80% of short-term and 20% of medium-term cost savings under both scenarios. Non-realizable cost savings. Commissioners pay 80% of short-term and 20% of mediumterm cost savings under both scenarios.

22

QALY gains in monetary terms. In scenario A commissioners do not pay for QALY gains; in scenario B they pay 80% of the value of QALYs gained in the short-term and 20% of the value gained in the medium-term. According to the results presented in Table 5, the realizable and non-realizable cost savings achieved by the Be Active program are not sufficient to require commissioners to repay investors. The Be Active SIB model is only feasible if commissioners are willing to pay investors a minimum of 33% of QALY gains, taken to include short-term QALY gains valued at nearly £6,600 per QALY (Figure 6).

Figure 6: Scenario A SIB Contract: Sensitivity analysis of net benefit of QALYs gain in short term. Source: (Marash et al., 2011)

The benefits achieved by the Be Active program are health-care cost savings and gains in QALYs, productivity and SWB (Tables 6 and 7). Because the greatest value is generated by outcomes related to quality-of-life benefits that reduce the burden on health services, the 23

feasibility of the SIB depends on the commissioners‟ willingness to pay for the QALY gains delivered by the program.

Table 6: Cost-Benefit Analysis of Be Active: Cost savings and productivity gains by stakeholder over time

Table 7: Cost-Benefit Analysis of Be Active

24

Wyman Center Teen Outreach Program (TOP) Wyman Center‟s Teen Outreach Program (TOP) proposed a social impact bond (SIB) to support a project to provide urban teenagers with knowledge and skills aimed at reducing the incidence of early-teen pregnancy. The program involved Master‟s level Social Work students conducting weekly 45- to 60-minute sessions with participants, in return for which students gained course credits or a fee (McBride, 2014). In addition to short-term benefits of reducing early-teen pregnancy, the project envisaged longerterm benefits in terms of participants‟ better opportunities for education and employment, with associated gains to society (lower levels of youth unemployment) and government (higher tax revenue) (Nonprofit Finance Fund, 2012). Like other SIBs, the program combined private- and philanthropic-investor interests with publicsector service commissioners to launch an intervention with measurable social benefits and costsavings for taxpayers. An evaluation of the Wyman TOP indicated that the project was an evidence-based and effective means of preventing early-teen pregnancy. At the end of the school year, female participants were found to be less likely than non-participants to have become pregnant and quit school. The Chicago public school system implemented the TOP model to reach 9,000 9th-grade students in 40 target schools each year, with the aim of improving participants‟ life skills, health-related behavior and involvement in community activities. The cost-benefit analysis of the Teen Outreach Program comprises two main components: the cost of implementing the program and the savings generated by program outcomes (Table 8). Because TOP was a short-term program, the success of the SIB was assessed in terms of the

25

most easily, directly quantifiable outcomes—medical cost-savings related to prevented early-teen pregnancies over a 5- to 10-year period.

Table 8: TOP Cost Assumptions

However, TOP also resulted in important, less easily quantifiable benefits, such as a decrease in rates of crime, increased enrolment in higher education, and higher rates of employment with an associated increase in future expected tax revenues. 26

Thomas identifies three layers of cost savings under the TOP program, based on inflationadjusted rates (Thomas, 2012). In the first year of TOP treatment (year 2), savings of $1,776 per prevented birth are realized, equivalent to avoided costs of prenatal care, delivery and postpartum care. Savings associated with infant medical care amount to $2,403 per avoided birth, while savings associated with the first four years of child medical care amount to $4,179 per avoided birth (Figure 7). Figure 7: Timeline of TOP Dollar Savings (Source: Dao, V. 2012)

The rate of reduction in teen births under TOP was calculated at 45%, given a Chicago rate of teen pregnancy of 6% (girls who become pregnant while part of the program do not receive medical care under TOP). The total bond issued for the Wyman Center Teen Outreach Program was $4 million (Table 9). The average cost per student is calculated as $964 in Year 1, $907 in Years 2-4 and $305 in Year 5. These costs include a 1% fee payable to an intermediary and 1.75% to an independent assessor.

27

Table 9: Other TOP Assumptions

Table 10: TOP Cost Analysis (Source: Dao, V. 2012)

As shown in Table 10, total savings per student are between $24 in Year 2 and $226 in Year 7. However, these savings do not accrue to one branch of government, being split between the federal and the state/local administrations. The savings/cost ratio for federal and state/local levels of government means that the total savings generated are in fact much lower than those specified

28

in project contracts. As shown in Table 11, the savings/cost ratio for Year 2 is just 3%. As a result, the savings realized are just a tiny proportion of the total project cost (Table 11). Figures 8 and 9 illustrate the problem, presenting details of annual TOP cash flow and returns. As shown in Figure 8, net cash flow in Years 1-4 is negative because of high program start-up and running costs. While cash flow becomes positive in Year 6, cumulative cash flow remains negative throughout the ten years of TOP implementation. In short, the savings generated over the life of the program do not cover its costs.

29

Table 11: TOP Saving Analysis

30

Figure 8: TOP Net and Cumulative Cash Flows

Figure 9: IRR to TOP Over Time

31

Figure 9 presents the project‟s annualized IRR on total public spending, and on federal and state/local government spending. As shown, the IRR for each investor in TOP is negative at every period of the program. These analyses establish that the savings generated by TOP are insufficient to cover program costs and the original principal of $4 million. The Wyman Teen Outreach Program therefore cannot be a feasible social impact bond.

Sensitivity Analysis The feasibility of a SIB depends on a number of assumptions that can be subject to adjustment. In the case of the Wyman Center Teen Outreach Program (TOP), savings realized are sensitive to three issues: persistence factor, decrease in participant birth rate, and the Chicago early-teen birth rate. Adjusting any of these factors will change cash flows and IRRs (Figure 10).

Figure 10: TOP Sensitivity Analysis (1)

As the rate of persistence increases, the savings increase as well. This happens because the effect of TOP lasts longer than the program, and the 53% decline in teen pregnancy risk lasts beyond one year (Table 12).

32

Table 12: TOP Sensitivity Analysis

Figures 11 and 12 show two figures of particular interest, relating to IRR at a 45% reduction in the birth rate—at 0% persistence and at a 6% Chicago teen birth rate, respectively. These two figures indicate that more savings are expected from TOP when the persistence rate and the Chicago teen birth rate are higher. But even then, the IRR for these two factors is negative. It is therefore concluded that the Wyman Center Teen Outreach Program cannot feasibly be implemented as a SIB. Figure 11: TOP Sensitivity Analysis (2)

33

Figure 12: TOP Sensitivity Analysis (3)

34

CONCLUSION The inefficiency of government programs and the inability of the public sector to finance demand for services mean that social impact bonds are a useful tool for governments. Privatesector investment in SIBs would not only help public programs work better, but would also enable programs to be scaled up in order to reach population groups not previously served. An additional advantage of SIBs is the focus on measurement of outcomes. This facilitates a partnership between local/state governments and non-profit investors, which transfers financial risk from the public sector while raising sufficient capital to expand operations on the basis of rapidly and rigorously accumulated evidence of success. This paper discussed two case studies to assess the effectiveness of the SIB model: Birmingham‟s Be Active program and the Wyman Center‟s Teen Outreach Program (TOP). In the case of Be Active, economic analysis demonstrates that the program represents a costeffective and efficient use of public resources. However, there would be few benefits to implementing the program through a SIB as only a small fraction of benefits accrue to the public-sector project commissioners, generating insufficient cost savings to repay investors. A Be Active SIB is feasible, however, if framed in terms of gains in QALYs, productivity, and secondary or long-term health benefits. The avoidance of diseases caused by a lack of physical activity or smoking would bring other savings for health services, including the cost of physician time, drugs and treatment, and medical equipment and associated amortization. The failure or success of a SIB depends on the evaluation of every single benefit, assigned a monetary value and assessed over a specific timescale. In the case of the Wyman Center‟s Teen Outreach Program, SIB financing is not feasible. In addition to structural and institutional obstacles, the program considered only short-term savings

35

in avoided medical expenses, neglecting long-term, difficult to quantify benefits associated with fewer early-teen births. Social impact bonds are an innovative tool but challenges remain in their implementation. First, the current focus on feasibility analyses should be bolstered by systematic efforts to ensure program fidelity. Second, there is a need for legislation in support of proper SIB frameworks such that project partners are required to fulfill their responsibilities even if project outcomes no longer align with partner policy (as in the case of a change of government, for example). Such legislation should also ensure a speedy procurement process, as well as providing mechanisms for a change of service provider on the grounds of poor performance or where that provider is no longer able to provide the service agreed. Third, barriers to investment must be addressed in order to enhance the attractiveness of SIBs and thereby boost their utilization. In conclusion, SIBs can provide vital finance for social and other programs anywhere in world, so long as impacts can be evaluated in monetary terms (Gustafsson-Wright, 2015). Furthermore, the structure brings flexibility in roles, structures, contracts and mechanics, enabling projects to be managed to achieve maximum benefit in minimum time and with greatest efficiency.

REFERENCES 36

[1] Azemati, H., Belinsky, M., Gillette, R., Liebman, J., Sellman, A., & Wyse, A. (2013). Social impact bonds: lessons learned so far. Community Development Investment Review, 23-33. [2] Callanan, L., & Law, J. (2012). Will Social Impact Bonds Work in the United States? Washington DC: McKinsey on Society. Retrieved from http://mckinseyonsociety.com/downloads/reports/Social-Innovation/Social-impact-bonds.pdf [3] Center for Global Development. (2014). Assessing the Impacts of Service Learning on Middle School Students: Wyman‟s Teen Outreach Program. http://csd.wustl.edu/Publications/Documents/RR14-09.pdf [4] Centre for Social Impact Bonds. (2013). Ministry of Justice: Offenders Released from Peterborough Prison. Retrieved from http://data.gov.uk/sib_knowledge_box/ministryjusticeoffenders-released-peterborough-prison [5] Dermine, T. (2014). Establishing Social Impact Bonds in Continental Europe. [6] Disley, E., Rubin, J., & Europe, R. A. N. D. (2014). Phase II Report from the Payment-byResults Social Impact Bond Pilot at HMP Peterborough. [7] Dao, V. (2012). Social Impact Bond: Feasibility Analysis of the Wyman Teen Outreach Program (Doctoral dissertation, University of Chicago). [8] Gustafsson-Wright, E., Gardiner, S., Putcha, V. (2015). The Potential and Limitations of Impact Bonds. Global Economy and Development at Brookings. [8] Hansbrough, M. (2014). Social Impact Bonds: An Explanation, an Application and Cautions. [9] Health Survey of England (1999). The Health of Ethnic Groups. The Department of Health. [10] Humphries, K. W. (2014). Not Your Older Brother‟s Bonds: The Use and Regulation of Social Impact Bonds in the United States. Law and Contemporary Problems. 76(3), 433-452. [11] Jackson, E. T. (2013). Evaluating Social Impact Bonds: Questions, Challenges, Innovations, and Possibilities in Measuring Outcomes on Investing. Community Development. 44(5), 608616. [12] Langford, A. R. (2011). Social Impact Bonds in Canada: From Theory to Implementation. [13] Liebman, J., & Sellman, A. (2013). Social Impact Bonds: A Guide for State and Local Governments. Cambridge, MA: Harvard Kennedy School Social Impact Bond Technical Assistance Lab. [14] Malcolmson, J. D. (2014). Social Impact Bonds: Cleared for Landing in British Columbia. [15] Marsh, K., Bertranou, E., & Samanta, K. (2011). Cost-Benefit Analysis and Social Impact Bond Feasibility Analysis for the Birmingham Be Active Scheme. Matrix Evidence.

37

[16] Mass. Gov: Administration and Finance. (2012). Massachusetts First State in the Nation to Announce Initial Successful Bidders Pay-for-Success Contracts. Retrieved from http://www.mass.gov/anf/press-releases/fy2013/massachusetts-first-state-in-the-nation-toannounce-ini.html [17] Matrix (2011). Return on Investment and Cost Impact for Public Health Interventions. Commissioned by NICE, Study Number 433. [18] Mosenson, S. H. (2013). New Models of Private Sector Financing of Not-for-Profit Corporations (Doctoral Dissertation, Northeastern University Boston). [19] NHS South Birmingham (2011). Leisure Card data. [20]Nonprofit Finance Fund. (2012). Pay-for-Success Initiative.” Retrieved from http://nonprofitfinancefund.org/pay-for-success (Accessed June 2, 2012). [21] Open Society Foundations (2012). Maryland opportunity compacts: Public/private partnerships in action changing lives, systems, and policies. New York: Open Society Foundations. [22] Perakis, R. (2014). Centre for Global Development. (2014). Peterborough SIB Results Are In—What Do They Mean? Retrieved from http://www.cgdev.org/blog/peterborough-sib-resultsare-%E2%80%93-what-do-they-mean [23] PSSRU (2009). Unit Costs of Health and Social Care 2009. Retrieved from http://www.pssru.ac.uk/uc/uc2009contents.htm [24] Roman, J. K., Walsh, K. A., Bieler, S., & Taxy, S. (2014). Five Steps to Pay for Success. [25] Strickland, P. (2010). Social Impact Bonds: The Pilot at Peterborough Prison. Retrieved from http://www.parliament.uk/briefing-papers/SN05758 [26] Thomas, Adam. (2012). Three strategies to prevent unintended pregnancy. Journal of policy analysis and management 00, no. 1:1-39 [27] U.S. Department of Labor. (2012). Notice of Availability of Funds and Solicitation for grant applications for Pay for Success Pilot Projects. Retrieved from http://www.doleta.gov/grants/pdf/pfs_sga_dfa_py_11_13.pdf [28] US SIF. (2010). US Social Investment Forum: Trends Report 2010. Report on Socially Responsible Investing Trends in the United States. Retrieved from http://www.ussif.org/files/Publications/10_Trends_Exec_Summary.pdf [29] US SIF. (2014). US Social Investment Forum: Trends Report 2014. Report on US Sustainable, Responsible and Impact Investing Trends. Retrieved from http://www.ussif.org/Files/Publications/SIF_Trends_14.F.ES.pdf

38