The razor’s edge: Social impact bonds and the financialization of early childhood services

ABSTRACT

In a growing number of U.S. cities, social impact bonds (SIBs) introduce an experimental strategy into the politics of fiscal constraint. With limited political willpower and public funding, some have used SIBs to leverage new support for social programs. We argue that cities that engage in SIBs walk a razor’s edge between promoting public investment and the risk of deepening financialization in the social service sector. We explore efforts in 3 cities to expand early childhood services through SIBs: Salt Lake City, Utah; Chicago, Illinois; and Greenville, South Carolina. We test the balance between promise and risk through four foci: systemic change, performance metrics, cost structure, and social equity. We show that the context of political fiscal climate and strategic policy change matters in SIBs’ justification and impact; whereas Salt Lake City and Greenville scaled investment up to the state level, Chicago merely plugged short-term local budget gaps.

Social impact bonds (SIBs) offer an experimental strategy for U.S. cities navigating the politics of fiscal constraint. These emerging financing mechanisms represent an alluring possibility: increased investment in social programs through private financing (Pequeneza, 2018). We explore how U.S. cities have maneuvered through the allure of SIBs, and their Trojan horse–like dangers, as they strive for more equitable social service delivery.

The number of SIBs has grown substantially around the world (A. Fraser, Tan, Lagarde, & Mays, 2018) since the first SIB was developed in the United Kingdom in 2010 (Disley, Rubin, Scraggs, Burrowes, & Culley, 2011). At least 17 have been implemented in the United States (Finlaw, 2017), primarily by cities, counties, and states, with steady Obama-era federal support for feasibility studies and technical assistance (Gustafsson-Wright, Gardiner, & Putcha, 2015; U.S. Government Accountability Office [GAO], 2015).

But cities risk losing more than they gain due to the high costs of setting up a SIB (Edmiston & Nicholls, 2018). In an inherent paradox, though they promise cost savings, SIB transactions are expensive, because each deal is a unique, complex negotiation (Maier, Barbetta, & Godina, 2017; Warner, 2013). Though SIBs purport to allow flexibility and innovation, they demand well-documented interventions, model fidelity, and strict evaluations to trigger accurate payouts (Berlin, 2016; Maier et al., 2017). Their reliance on performance-based management induces gamesmanship (Lowe & Wilson, 2017) and may overly skew their focus toward meeting a quantifiable result (Heinrich & Choi, 2007). But the most insidious cost of SIBs is their potential to financialize social services by marketizing the “public finance value” of their vulnerable clientele (Neyland, 2017; Sinclair, McHugh, & Roy, in press; Warner, 2015). Chiapello and Knoll (2017) show that SIBs follow a variety of “welfare conventions,” organizing objectives and stakeholders according to frameworks that prioritize financial profit, competitive exchange, and entrepreneurial innovation. Though some argue that SIBs can broaden dialogue and intervention in critical social problems by demonstrating the benefits of model programs (Jupp, 2015; Liebman, 2011), others warn that they may narrow the scope of policy conversation, reducing it to a transactional discussion of inputs and outputs which can silence broader policy goals (Lake, 2015; Tan, Fraser, McHugh, & Warner, in press; Warner, 2015).

Negotiating the razor’s edge of financialization

Cities that walk the “razor’s edge of financialization” do so at some risk to their most vulnerable citizens. By making a simple link between intervention and payout, SIBs may narrow the scope of social investment to low-cost programs with short-term returns, when more comprehensive approaches are needed (Lake, 2016). When SIBs are employed to push for systemic shifts in political contexts facing severe fiscal constraint, however, they can reveal new possibilities for longer-term investment. Cities must negotiate this razor’s edge if they are to ensure that SIBs actually enhance long-term investment, rather than simply serving as a Trojan horse of financialization.

The financialization of social policy through SIBs risks obscuring visions of greater social rights. Lake (2016) characterizes the reductive nature of financialization for urban policy: “The monetization of policy goals … transforms substantive social outcomes from the status of ends in themselves to a means for reducing government spending and producing a financial return for investors” (p. 57). Similar lessons from the financialization of physical infrastructure show how broader public objectives and positive externalities are sacrificed to narrow, shorter-term, profit-seeking objectives (Sclar, 2015). Private financing has undermined the broader characteristics of physical infrastructure: universality, access, and maximization of positive externalities (O’Neill, 2010). Today, physical infrastructure investments have become a new asset class, designed to attract private investment through public–private partnerships (O’Neill, 2017).

Financialization risks marginalization of those sectors and people most in need of services by delivering prescriptive social interventions while undermining social inclusion (Shortall & Warner, 2010) and ignoring the broader structural reforms that should be at the core of social and urban policy (Lake, 2016). SIBs claim that measuring performance will accurately assess an intervention’s value, as a social service and public expenditure. They position public expenditures as defensible, accountable transactions. But by framing social services as transactions, SIBs risk becoming static, point-in-time estimates of social need. If their returns were not so rigidly linked to specific outcomes, they could allow for a more flexible, comprehensive approach. Achieving these outcomes, however, requires prior evidence of success, fidelity, and consistent results (Maier et al., 2017).

City, county, and state governments are experimenting with SIBs to expand critical early childhood care and education (ECE) services. We will show that in networks with a broader view and a balance of power across actors in early education, government, and business, SIBs may be able to promote expanded ECE investment. But when financial metrics are privileged, market logic narrows the focus. Though financialization risks reducing urban policy to its financial, not social, outcomes (Lake, 2016), we find local configurations of actors and their power relations matter. Cities can push back against these forces of marketization, effectively “riding the wave” of fiscal austerity by adapting market logics to justify public investment and enhance social policy (Warner & Clifton, 2014). N. Fraser (2011) and Ogman (2019) have used Polanyi’s (1944) theory of a double movement to look at how progressive coalitions of service providers seek to maintain services in times of austerity by partnering with more powerful economic actors in a progressive neoliberal compromise. Progressive lower-level service providers are using economic logics to justify enhanced investments in early childhood, making the “economic case” and hoping that this will lead to an expansion in social rights to care (Warner & Prentice, 2013). In so doing, early childhood interventions walk a razor’s edge between the potential to insert broader social objectives into a market mechanism and that mechanism’s capacity to undermine broader social goals.

We use this framework to discern SIBs’ possibilities for progressive impact. We look inside the details of SIB design to assess actors, political context, metrics, costs, and social equity. Whether SIBs can concretize the promise of increased investment and broader public policy response depends critically on the configuration of actors and mechanisms in their local and regional context.

Testing the razor’s edge: SIBs in early childhood education

Early childhood care and education services are a proving ground for early SIBs in the United States. Lessons for emerging SIBs can be drawn from the experience of using economic logics to support state, regional, and local-level economic development policies that expand access and funding for child care. Economic or financial logics have been used for more than 20 years in the United States to reimagine ECE services as investments, instead of just expenditures, and thus justify additional public and private support (Bartik, 2011; Warner, 2006). These logics have nudged policy discourse in the United States and Canada toward thinking of early education and care as a social infrastructure for economic development and thus an investable social service (Warner & Prentice, 2013).

ECE has a well-documented capacity to produce important social gains. Relatively short and inexpensive interventions, such as preschool or parenting education programs, can improve children’s cognitive and social development and enhance school achievement (Heckman, Pinto, & Savelyev, 2013; Olds et al., 1997). Broader supports for working families, such as full-day childcare, reduce stress on parents and enable increased parental workforce participation and productivity (Gornick & Meyers, 2003; Halpern, 2004; Hipp, Morrissey, & Warner, 2017). Early childhood interventions can even support market returns at the regional economy level, if broadly conceived (Bartik, 2011; Warner, 2006). These economic logics reconceptualize early care and education as social infrastructure for economic development. Social infrastructure can expand social rights because it shifts ECE from a welfare service to an infrastructure and the universality that infrastructure has historically implied (Warner & Prentice, 2013).

A comprehensive approach that includes human development, the productivity of families, and the regional economy is critical because it opens possibilities for deeper reform (Morrissey & Warner, 2007). The simple interventions favored by SIBs in the United States, however, such as preschool or parent education, fall short of this approach. Despite the possibilities for broader structural change, they do not address workplace policy and comprehensive support for ECE from birth onward, such as full-day childcare for working parents. Biases toward preschool alone are found in business and economic arguments (Rolnick & Grunewald, 2003). Feminist economists and educators voice the broader concerns of parents and ECE specialists, but these have not gained the investor attention they deserve (Folbre, 2006; Halpern, 2004; Stoney, Mitchell, & Warner, 2006; Warner, 2009).

Policy coalitions matter. As SIBs emerge in the ECE arena in the United States, we must assess the extent to which state and local policy actors are building links with private finance to increase ECE investment and promote further financialization of ECE services at the same time (see also Carter, in press; J. Williams, 2018). SIBs claim to demonstrate the effectiveness and cost savings of their social interventions, but the economic benefits of ECE investments are already widely known. Does the SIB mechanism itself help to produce information that can shift the policy debate? Or does it merely serve as a Trojan horse opening another form of legitimacy for neoliberal policy intrusion into social services?

Methodology

We show, through three ECE case studies, how U.S. cities have adapted SIBs to promote social investment while simultaneously risking further financialization of the sector. ECE services in the United States have suffered from chronic underinvestment compared to other countries. ECE clients, young children and their vulnerable families, are among the most voiceless members of society. But because of ECE services’ proven economic benefits, SIBs could be suitable tools to expand investment. Three SIBs currently are operating in the ECE arena in the United States: the South Carolina Nurse–Family Partnership Pay for Success Project, the Utah High Quality Preschool Program, and the Chicago Child–Parent Center Pay for Success Project. For each case, we conducted a document review of publicly available contract and loan agreements,¹ press releases, and journalistic articles. We also conducted 11 semistructured interviews in January and February 2017,² including five for the South Carolina SIB, three for the Chicago SIB, and three for the Utah SIB. Among these interviews, one was a public agency, one was an evaluator, two were funders, two were service providers, three were technical assistance providers, and two were intermediaries. Interviewees were selected based on their involvement in the design, launch, and implementation of the SIB. Interview questions covered program design, intervention scope, and management. During coding of these interviews, themes that emerged inductively shaped the four conceptual foci that guide our analysis: systemic change, performance metrics, cost structure, and social equity. We define these constructs below and follow with short descriptions of each case, summarized in Table 1:

Table 1. SIB project description summary.

	South Carolina	Utah	Chicago
Intervention	Nurse–Family Partnership maternal and child health model	Granite School District preschool model	Child–Parent Centers preschool model
Clientele	3,200 First-time low-income mothers	3,500 Salt Lake City students	2,618 Chicago students
Geographic scope	State	City, with statewide implications	City
Investor type	Philanthropy	Commercial and philanthropy	Commercial and philanthropy
Investment amount	$17 million, plus $13 million from Medicaid	$7 million	$17 million
Outcome metrics	Birth spacing Child injury Preterm births Low-income ZIP code coverage	Special education avoidance	Special education avoidance Kindergarten readiness Third-grade literacy
Outcome payor	Government	Nonprofit and government	Government
Success payment	$7.5 million	Maximum unknown	$25–34 million
Contract duration (including success payments)	5 years	12 years	16 years

1. Systemic change: the deeper possibilities of social transformation when sustainable finance and policy change are promoted.

2. Performance metrics: the scope of outcome metrics.

3. Cost structure: the costs and returns in the transaction.

4. Social equity: the consideration of social inclusion.

South Carolina Nurse–Family Partnership Pay for Success Project

Originating from a consortium of funders supporting Nurse–Family Partnership (NFP) in Greenville, South Carolina (SC SIB, 2017), this SIB launched in 2016 to expand access to NFP home visitation services for 3,200 first-time mothers across the state (South Carolina Department of Health and Human Services [SCDHHS], 2016). NFP pairs regular nurse visits with low-income mothers to improve maternal and child health. The project has four outcome metrics (South Carolina Department of Health and Human Services, 2016). Interviewees noted that the goal is to secure sustainable public financing for NFP statewide.

This SIB is unique in the U.S. context in that it does not offer an investor return. The philanthropic consortium invested $17 million, and the state secured $13 million in Medicaid funding to build a longer-term funding stream for the program.³ The maximum success payment is $7.5 million, which will be reinvested into the program for future service provision (SCDHHS, 2016).

Utah High Quality Preschool Program

The Utah SIB began in Salt Lake City in 2013 as a coordinated effort among the United Way of Salt Lake, a statewide children’s advocacy organization called Voices for Utah Children, and Granite School District. The first SIB in ECE, this project expands Granite School District’s high-quality preschool model to 3,500 students in Salt Lake City (Gustafsson-Wright et al., 2015). Special education avoidance is the single outcome metric. Interviewees clarified that the goal is to persuade the state legislature to fund preschool (Utah State Legislature, 2017).

Goldman Sachs and the J.B. and M.K. Pritzker Family Foundation invested $7 million in this SIB. Initial outcome payments have been made, but education experts criticized them based on questionable metrics, methodology, and financial agreements, claiming that investors were overpaid (Popper, 2015).

Because of the SIB, the state ultimately passed legislation to appropriate funding for preschool. The legislation also caps investor return on future SIBs at 5% above the municipal market data general obligation bond rate (State of Utah, 2014).

Child–Parent Center Pay for Success initiative

This SIB launched in Chicago in 2014, the year after Utah’s, to increase the number of Child–Parent Center (CPC) preschool sites in the city and open 2,618 new student slots (City of Chicago, 2014a). CPC is a highly researched preschool program (Reynolds, 1997) that focuses on teacher and parent engagement. The project has three outcome metrics (City of Chicago, 2014a). Unlike the other cases, this SIB does not appear to have a broader long-term financing goal.

Goldman Sachs, the J.B. and M.K. Pritzker Family Foundation, and Northern Trust invested $17 million (City of Chicago, 2014a, 2014b, 2014c, 2014d). The Pritzker Family Foundation used a program-related investment (City of Chicago, 2014d). The maximum potential success payment is about $34 million, though the expected payment is about $25 million (City of Chicago, 2014a). An initial success payment has been made (Sanchez, 2016). The project has been criticized for using special education avoidance as a metric, under concerns that this incentivizes service reduction, and for using a low-risk model that increases the likelihood of success payments (Sanchez, 2016).

Analysis

We evaluate each case against the four foci outlined in our methodology. Chicago’s and Utah’s SIBs operate at the city or metropolitan level. The South Carolina SIB operates at the state level. Our following analysis shows how the cases fall along a continuum. We find that South Carolina is most likely to achieve systemic change by both inserting social values in the market and using market players to shift the political dialogue. Utah is next along the continuum, using market pressure to shift political investment but sacrificing social objectives to investor return. We find that Chicago ranks last because it extracts funds from the public sector in the name of expanding access to preschool. A summary of the three cases is provided in Table 2.

Table 2. Evaluation of early childcare and education social impact bonds.

	South Carolina	Utah	Chicago
Systemic change: presence of a new, sustainable public funding stream	Yes—Medicaid reimbursement	Yes—state income tax	No—funding ends after SIB
Performance metrics: breadth of metrics	Broad—four metrics	Narrow—one metric	Intermediate—three metrics
Cost structure: comparison of public investment to investor profit	Maximizes public investment—reinvests success payments	Maximizes investor profit—overestimates impact	Maximizes investor profit—overpays investors
Social equity: coverage of services to vulnerable clientele	Yes—rural coverage	Yes—low-income coverage	Yes—low-income coverage

Systemic change

Interviewees explained that the goals of both the South Carolina and Utah SIBs are to secure sustainable public funding for early childhood services through state legislation. The designers of each project committed to this goal first. Then they used the SIB to demonstrate cost savings for governments and social benefits for vulnerable clientele. Both Utah and South Carolina are politically conservative states with high child poverty rates. They are difficult contexts in which to push any expansion of ECE services.

In South Carolina, the intention of the SIB was to transition funding for NFP from local philanthropy to permanent state and federal sources. A group of philanthropies, including The Duke Endowment, Blue Cross Blue Shield Foundation, Greenville First Steps, and the Children’s Trust of South Carolina, had been funding NFP since 2008 (SC SIB, 2017), including an expansion in 2013 through federal Maternal, Infant, and Early Childhood Home Visiting program funding (B. Williams, 2013). Services were limited to a few hundred cases in the Greenville area, but the funders recognized that philanthropy could not sustain these services in the long run. They launched the SIB to scale up NFP services and demonstrate its cost savings:

It’s a much more expensive intervention than even the childcare vouchers that we provide. When the community started NFP, it was 100 clients, so we were at about $450,000. … That was sustainable in Greenville County. We could have done that forever. There were 10 partners that we could have gotten to put $50,000 in … but what we realized was we didn’t need to serve 100 clients. We needed to serve 600 clients, because we started looking at our birth rates in Greenville County alone; [there] was no way that local philanthropy was going to be able to sustain a $3 million project. Our United Way’s entire early childhood investment is $2 million. That’s for everything that they do. … Early on, we realized we want to take this program to scale, and we don’t want to be left with the bill because there’s no way we can sustain that. (personal interview, SC SIB, January 2017)

Along with upfront philanthropic investment, South Carolina’s Department of Health and Human Services secured eligibility for $13 million in state and federal Medicaid funding, which took 3 years to achieve and Obama administration support (SC SIB, 2017). If the SIB is successful, the partners plan to advocate for permanent Medicaid reimbursement or other state appropriations so that the program will have a stable, long-term funding source (SC SIB, 2017). South Carolina’s experience demonstrates the greater value added from an SIB with a long-term perspective and the relative lack of value from one used for temporary service provision:

Think about where the [SIB] makes sense. It doesn’t make sense for a known recurring cost you’re going to face forever more. It can make sense for something where you either want to try to figure how to get something to scale or something that’s a little more experimental potentially. In this case, the question was, is there a way to scale NFP while also bringing its cost structure down to a place where maybe it could become sustainably financed in some way to the Medicaid program? (personal interview, SC SIB, January 2017)

In Utah, a patchwork of public and private providers offered preschool with zero state investment before the SIB. State income tax funding for education was restricted to kindergarten to grade 12 (Utah SIB, 2017). The intention of the project partners was to secure state appropriations for high-quality preschool and alter the state’s education funding formula to include preschool. The first bill to do so failed to pass in 2013 (Bennett, 2013). Project partners used the SIB as a “proof-of-concept” to further educate legislators about the benefits of high-quality preschool in terms of cost savings and child development (Utah SIB, 2017). United Way of Salt Lake contributed $1 million for the first cohort’s success payments, on the condition that, if successful, the state would pick up the success payments for the second through fifth cohorts.

The Utah SIB effectively shifted state investment in early childhood. In 2014, the Utah legislature passed H.B. 96, the Utah School Readiness Initiative, which included an ongoing appropriation of $6 million from the general fund and created the Utah School Readiness Board. The board participates in Pay for Success transactions for the state and administers grants to public and private high-quality preschool programs (Utah State Legislature, 2014). In 2016, the legislature passed SB 101, the High-Quality School Readiness Expansion, which expands preschool slots by appropriating almost $11.7 million for 3 years, mostly using federal funds (Utah State Legislature, 2016). Currently, United Way is participating in a ballot initiative to raise taxes and use the additional revenue to fund preschool and higher education. They also are working with Salt Lake County to achieve universal access to preschool in the county (Utah SIB, 2017).

In Chicago, the goal of the SIB is more limited. A priority of Mayor Rahm Emanuel is to expand early childhood education services, and the SIB uses private investment to increase access to CPC. But after the final cohort, the city will have to find new funding to sustain the expansion or else reduce the number of slots (Chicago SIB, 2017). Unlike the other cases, Chicago shows no evidence of a broader political or fiscal goal. There are no plans to further expand preschool after the SIB ends (Sanchez, 2016). Our case studies suggest that SIBs that do not have systemic change as an explicit goal and do not deliberately seek to change mainstream public funding arrangements offer less compensation for the known risks of financialization.

Performance metrics

Performance metrics are the bedrock of SIB design. Though performance measures have helped target educational services (Boyne & Chen, 2006) and can be structured to create positive feedback to organizations (Schalock & Bonham, 2003), in SIBs the concern is that these measures may reduce flexibility and innovation (Maier et al., 2017). The selection of metrics; their threshold for success, measurement method, and evaluation rigor; and the performance of the intervention against its outcome targets are the sole factors that determine the payout in a SIB contract. Because of this narrowly defined path to success, some researchers argue that SIBs’ performance-based management scheme simplifies and distorts the “complex reality” of operating a social program, turning the intervention into a game to win successful outcomes instead of to support vulnerable clients (Lowe & Wilson, 2017). This gamesmanship can drive management toward “creaming and parking, teaching to the test, reclassifying, and falsification of data” (Lowe & Wilson, 2017, p. 986). We provide the following insights into our cases’ selection and breadth of metrics.

South Carolina has the most metrics and its evaluation is the most extensive of the three. The project’s four outcome metrics include reduction in preterm births, healthy birth spacing, reduction in child injury, and coverage of services to low-income ZIP codes. All metrics were chosen based on past evidence of the NFP model’s capacity to achieve them, to reduce the risk to NFP, and to maintain the clarity of the evaluation. The SIB’s designers considered that one or two metrics would apply too much pressure to NFP to perform within acute parameters. Five or six metrics, however, would reduce or distort each metric’s significance (SC SIB, 2017).

The Utah SIB uses a single metric: avoidance of special education in kindergarten through sixth grade for preschoolers deemed at risk of needing those services (Utah SIB, 2017). Students who score two standard deviations below the mean on the Peabody Picture Vocabulary Test are considered “at risk.” Education experts criticized use of the Peabody Picture Vocabulary Test because it is not commonly used to screen for special education and may disadvantage non-English-speaking children when administered in English (Popper, 2015).

This metric led to an overidentification of at-risk children because of the “faulty assumption that many of the children in the program would have needed special education without the preschool,” which in turn led to an overstatement of impact and an overpayment to Goldman Sachs (Popper, 2015, p. B1). In the first cohort, 595 students attended preschool, 110 were deemed at-risk, but only one actually used special education in kindergarten (United Way of Salt Lake, 2015). Goldman Sachs was paid for almost the entire cohort—an unprecedented level of impact for a preschool program (Popper, 2015). Although the SIB’s partners considered these criticisms, they argued that their evaluation was legitimate and they did not make dramatic changes to the payment structure (Utah SIB, 2017). But the criticisms reveal deficiencies in the SIB’s metrics. Repayment to investors is maximized by reducing the number of students receiving special education, something against which the U.S. Department of Education (2016) has warned.

Chicago broadens its metrics to three: avoidance of special education, kindergarten readiness, and third-grade literacy (City of Chicago, 2014a). This includes a mobility factor to account for program attrition (City of Chicago, 2014a), which the Utah SIB does not have (Utah SIB, 2017). The mobility factor ensures that only Chicago Public Schools students who remain in the school district from preschool onward will be counted toward success payments. Compared to Utah, this more sophisticated methodology uses additional criteria to circumvent potential criticism about the legitimacy of a single metric. The special education metric, however, holds by far the greatest weight in the contract. The other two are weak measures by contrast.

Cost structure

Cost structure is a key indicator of whether or not a SIB will financialize public services. Comparative analysis of the three SIBs reveals critical differences in the cost structures of each. Utah and Chicago overpaid their investors at the expense of public savings and future investment in early childhood because nearly all savings go back to investors as returns. South Carolina will reinvest its returns in the NFP intervention. Without an investor premium, its SIB has become a vehicle for sustainable financing:

It comes back … to one of the central irrationalities of these SIB projects, which is that ultimately nobody borrows money cheaper than governments. You go out and have other entities borrow money and then you repay them that with some kind of interest or something on top of it. From a financial argument, you’re basically throwing money away for whatever the difference is between the state’s potential borrowing cost and then whatever you pay out instead at the end of the project. The solution for us was, there’s no private investor premium being paid in our project. We basically threw the money changers out … to get the math to work. (personal interview, SC SIB, January 2017)

South Carolina’s cost structure demonstrates that investor profit is not the focus. NFP reduced its cost of service by 25% for the SIB by increasing client–nurse ratios in order to make its services affordable and sustainable for state Medicaid funding (SC SIB, 2017). Three metrics use both fixed and variable components to determine success, so that results can indicate a zero, intermediate, or high rate of payment. The fourth metric has a minimum threshold for any success payment (South Carolina Department of Health and Human Services, 2016). Overpayment is not possible because success payments go toward further service expansion.

In Utah, the SIB contract is not publicly available. Reports show that Goldman Sachs received a $260,000 success payment in 2015 (Popper, 2015) based on savings of $281,550 for the first cohort, calculated using the special education cost rate of $2,607 per student (United Way of Salt Lake, 2015). The payment to Goldman Sachs constitutes 95% of the savings from the program.

The SIB’s designers clearly considered cost savings for the public, but how substantial these can be when nearly all of the savings are given to the investors is unclear: “They were using the cost savings of special ed. avoidance to pay back the investors up to a certain point, but after sixth grade that money remains with the state, so in theory that helps bolster the state’s education budget as well” (personal interview, SC SIB, January 2017). Utah’s SIB is based on expected cost savings, but how close the forecasts are to reality is unsure, a problem also found in other SIB studies (Edmiston & Nicholls, 2018).

Utah’s SIB designers were concerned with attracting initial investors and compensating them for their investment risk (Utah SIB, 2017), perhaps fairly so because theirs was the first early education SIB in the nation. They used a 5% interest rate to compensate for risk, assuming that at least half of the children in SIB slots would not need special education (Popper, 2015). For the first cohort, if the program performed at a certain higher level, the investors could receive a higher interest rate. This higher rate was eliminated for the second through fifth cohorts and the potential base rate increased (Utah SIB, 2017). The maximum return for the first and second cohorts is capped at 7.26% (Gustafsson-Wright et al., 2015). For future SIBs, Utah’s new preschool legislation caps the interest rate at 5% above the municipal market general obligation data bond rate. The idea behind the cap is to relieve the state from a degree of interest rate negotiation while still enabling them to attract investors (Utah SIB, 2017) and be “good government stewards” (GAO, 2015, p. 47).

The uncertainty of the Utah SIB’s risk influenced its high payout rate. But Granite School District’s preschool program had a history of success and the SIB’s metrics were skewed to increase outcome payments. The risk to investors should be relatively low. So why should investors receive a high interest rate? Though demonstrating to the state legislature that preschool is good for children, the SIB handed over a massive amount of money to investors and has generated negative press about its usurious pricing structure (Popper, 2015).

Despite details in available documentation about the Chicago SIB, its precise costs are uncertain. In the SIB contract, success payments are expected to reach over $25 million, about $21.4 million from the Board of Education and $4.3 million from the city. But one clause caps possible board payments at $30 million (City of Chicago, 2014a). SIB documents cite a return of 5% for each investor (City of Chicago, 2014b, 2014c, 2014d), but a Brookings Institution report cites that the return could be up to 6% (Gustafsson-Wright et al., 2015).

The most surprising factor in Chicago is that investors will receive success payments for 15 years after the intervention. This has been criticized for doubling investor return (Spielman, 2014). Justification for this continued return from a single point-in-time intervention is unprecedented and not supported by research on ECE. Before its launch, five city council members voted against the SIB because of its low risk, high interest rate, and complicated structure, comparing it to Chicago’s infamous parking meter privatization scheme (Spielman, 2014). Chicago’s public school system has struggled with fiscal mismanagement for years; it dealt with a debt rating downgrade in 2015 (Gillers, 2015) and, heading into the 2017–2018 school year, faced a budget deficit in the hundreds of millions (Perez, 2017). The reasoning behind the potential $34 million SIB payout demonstrates a pattern of short-sightedness about the limitations of SIBs. Chicago opened 2,618 temporary slots without regard to long-term funding and the SIB returns constitute a substantial overpayment to investors.

Social equity

By expanding access to ECE services, SIBs may promote social inclusion and increase social equity. In all three cases, the target clientele are low-income families. In Utah, over half of the students in Granite School District qualify for free or reduced-price lunch and nearly half speak English as a second language (Utah SIB, 2017). The new CPC sites in Chicago expand services from African American families, historically those predominantly served, to Hispanic families (Sanchez, 2016). In South Carolina, the Department of Health and Human Services went a step further by pushing for the inclusion of a low-income ZIP code metric, which requires that 65% of NFP coverage go to first-time mothers living in a list of predetermined ZIP codes at the time of enrollment, many of which are in rural, high-poverty areas (South Carolina Department of Health and Human Services, 2016). NFP initially pushed back against this request because of the additional cost of rural outreach and recruitment, but the state persevered and the metric was included (SC SIB, 2017).

Exclusionary procedures, such as randomized control trials (RCTs), which many SIBs use, also raise concerns for social equity. South Carolina is the only case in our study to use an RCT. RCTs are considered the “gold standard” for evaluation. South Carolina’s Department of Health and Human Services required an RCT to ensure the credibility of the SIB’s results (SC SIB, 2017). But RCTs require a control group, meaning that several hundred mothers will receive neither NFP services nor referrals to other home visitation programs. This raised concerns for the South Carolina SIB designers:

There has been one negative thing [as] a result of Pay for Success, and that is … the idea that we would have to have a control group that we would also track, but cannot benefit from, not only NFP services but any other prenatal services. … What we have now are a group of incredibly vulnerable, first-time, low-income mothers that we found during pregnancy and know they need our help. Because we need to have some sort of objective comparative data, not only can we not help them, but we can’t tell them other places to go to get that help. … We’re looking at probably 100 families now a year that I can’t help that I should be helping. I totally get that if it wasn’t for those 100 families, we wouldn’t be able to help all these other families. I totally get that five years ago, there was no NFP and none of these families were getting help. The world is better off because we’re doing what we’re doing, but I look at those 100 families and I think, “But their kids won’t be better off.” That is probably the one thing I would say keeps me up at night about a Pay for Success study, is this need to have a control group that I really think is unnecessary. (personal interview, SC SIB, January 2017)

To address this concern, some SIBs offer ancillary services to all. South Carolina offers its project participants two additional resources beyond NFP, such as Reach Out and Read programming and free childcare vouchers, but not all families participate. The Chicago SIB includes additional resources for parent engagement through its partnership with Metropolitan Family Services.

RCTs are not the only evaluation option. Collaborative developmental evaluations, which promote ongoing learning among partners (Benjamin & Greene, 2009; Patton, 2011), would be more appropriate in network situations such as SIBs. But they do not yield clear financial metrics for private investors.

The common use of special education avoidance as a metric carries the risk of reducing services for children with disabilities. The U.S. Department of Education (2016) encouraged new SIBs to “set strong guard rails when using special education as an outcome measure” (¶ 8). Interviews and document review in the Utah and Chicago cases show that the projects are not trying to remove special education access (City of Chicago, 2014a; GAO, 2015; Utah SIB, 2017). But tying SIB payments to special education can create a financial incentive to keep services away from children who need them.

Almost universally, interviewees reported that the high-profile nature of SIBs has begun to insert ECE into policy and funding conversations in which it would not have been before. SIB stakeholders are positive about the impact of their experiments with social finance. They see SIBs as helping them serve more children and opening the door to serving many more.

Discussion: Experimenting on the razor’s edge

These cases show the challenges of walking the razor’s edge between using financial logic to expand ECE investments or to narrow the conversation to financial metrics. The logic was too narrow in Chicago and Utah and led to overpayment of investors, extracting funds from the very education programs the SIBs were designed to support. By contrast, this logic was used to recast ECE expenditures as investments in South Carolina and Utah so that political support for broader public funding could be built. What determines the difference? We offer four considerations.

The first factor is political context. The benefit of a SIB is the potential to broaden support for sustainable public funding even in conservative political environments with deep fiscal constraint, as in South Carolina and Utah. The inherent risk in these cases was more political than financial. Comparatively, Chicago is a liberal political context already funding CPC. South Carolina addressed political risk by leveraging a uniquely advantageous policy framework—Medicaid is the responsibility of government:

From a government perspective … in South Carolina, early childhood investments [are] not really seen as worthwhile roles of government. It’s not the government’s role … to tell a mother of a six-month-old to make sure she has a crib that is safely secured. … In general, politicians don’t feel like that’s a function of government, but it is government’s role to pay Medicaid costs because that’s a role that’s been assigned to government. (personal interview, SC SIB, January 2017)

By demonstrating cost savings within an accepted public expenditure, through Medicaid, South Carolina already has broadened the scope of ECE services. Although the program’s eligibility for Medicaid is limited to 5 years, the intention to make eligibility permanent is a clear goal of the SIB.

In Utah, the SIB was launched in a political atmosphere that provided no funding for preschool (Barnett, Carolan, Squires, & Clarke Brown, 2013) and delivered the lowest per pupil expenditures in the nation. Children in Utah represent a greater share of the population than in any other state, so state investment is spread thin (Utah SIB, 2017). Despite mistakes in pricing and methodology, demonstrating evidence for public savings induced legislative action to finance preschool.

The second factor is a strong public actor. Complex contracting requires clear values and sophisticated partners who can overcome information asymmetries. The South Carolina state government was a strong and creative partner that helped steer the project toward broader aims and a sustainable funding stream. Philanthropic investors were not looking to extract economic rents from the project. By contrast, Utah’s partners had to use their position to attract private finance, at great cost, to pressure state policy. Private financial interests could take advantage of information asymmetries to extract rents. When the state did come on board, it used its political power to legally limit future rent extraction. In Chicago, the same financiers as in Utah took the lead. The city exacted no special requirements and thus no broader aims were met. The rent extraction in Chicago is the most usurious—claiming success payments for up to 15 years for a single-year intervention, without a sustainable funding plan. As with the Chicago Skyway and parking public–private partnerships (Sclar, 2015), Chicago has promised public revenue to private bidders at the expense of further public investment.

The third factor is coordinated networks. In both South Carolina and Utah, coalitions of local and statewide advocates and philanthropies had supported their respective interventions long before the SIBs were implemented. These coalitions worked together with state actors to pursue long-term policy solutions and funding streams, using the SIBs as leverage. But without these networks inserting broader objectives, comprehensive approaches could be lost to SIBs’ financializing logic, as occurred in Chicago.

Interviewees argued that SIBs’ appeal is their capacity to translate a complex service into quantifiable, investable variables that can change policy. But we find that translation to be overly reductive, narrowing the scope of policy dialogue from the inherent social value of ECE to its transactional monetary value. O’Neill (2010) warns us of the danger of losing key features of infrastructure—universality, bundling, access, and positive externalities. Planners know that positive externalities are primary reasons for investing in public services (Del Bo & Florio, 2012; Sclar, 2015). When a broader perspective is taken, all of society benefits, not just the client. But SIBs are limited to short-term outcomes that can appeal to investors and clearly be attributed to cost savings. Figure 1 shows how SIBs, by narrowing their focus to the monetizable elements of an intervention, draw attention away from broader social inclusion and human development.

Figure 1. How social impact bonds narrow the focus of social interventions.

A common refrain of interviewees was that SIBs’ cost savings argument, supported by data, is the only one that is politically effective. As SIBs in ECE track metrics for cost savings, they also track evidence on child development. The evidence for cost savings serves as evidence for “what works.” But the SIBs’ interventions already were well documented. NFP operates in 42 states and is well researched (Olds et al., 1997). Child–parent centers have been operating in Chicago for 50 years and have been studied closely (Reynolds, 1997). Even the Granite School District preschool program had 3 years of data supporting it (Utah SIB, 2017). The benefits of preschool, moreover, are well-demonstrated in other states (Barnett & Ackerman, 2006). These ideas are not gambles with public money—and private finance generated through the SIB is much more expensive than public finance.

Instead, the main benefit of data and the cost savings argument is the production of localized information to change political or institutional discourse, our fourth factor. Rather than tools to implement public services, SIBs are tools to produce information about both cost savings and intervention performance, which can shift public investment. The evidence for these interventions’ social and financial benefits already existed but, as other studies show, states in conservative contexts with higher poverty levels and larger populations of people of color are more restrictive with social welfare services and funding (Hahn, Aron, Lou, Pratt, & Okoli, 2017; on state discretion, Michener, 2017). SIBs can encourage such cities and states to use information to justify social investments:

I think data and rigor about results was the number one contributor [to a change in conversation]. Initially we thought that the Pay for Success mechanism, which really transfers risk from taxpayers to investors … would be attractive to policymakers in the long run. To some it is. They still see that as valuable, but I think many people see it as a good way to test and start a new idea. The feeling is, if it works, maybe the state should just fund it directly without the interest cost of [Pay for Success]. (Utah SIB, 2017)

One of the things I like the most about [Pay for Success] is that there’s a rigorous evaluation and we all agree to abide by the results, and the payment is specifically associated with the outcome we’re looking for. You look at the overwhelming majority of what we pay for on health or education … and a lot of it is based on either inertia—nothing drives the appropriations process like inertia—or it’s based on people telling stories in front of committee. The amount of money that we appropriate that is connected to real evidence is damagingly small. (SC SIB, 2017)

Policymakers are interested in the local production of data. The performance-based management theory underlying SIB design in fact prioritizes the production of data (Lowe & Wilson, 2017). The persuasiveness of local evidence means that SIBs can influence policy conversations by introducing the importance of ECE. As conversations change, policies and public investment can expand to this underserved sector, showing that the narrowing effect of SIBs in Figure 1 can be read in the other direction, at least for Utah and South Carolina. In Utah, the economic return on special education avoidance changed state investment in early education. In South Carolina, the economic return on improved public health expanded the scope of policy and secured a long-term funding stream. The possibility to sway policymakers with higher-level authority means that SIBs have a surprising capacity to scale up public investment in social welfare from the city to the state level.

The conception of SIBs as the imposition of the market economy onto society is not uniform. N. Fraser (2011) points to the importance of the structure of coalitions in determining whether progressive neoliberal coalitions will push back and redirect the market or merely justify new forms of financial intrusion. This explains the potential for SIBs as a vehicle for social rights and protection; South Carolina pushed back against the financializing nature of SIBs by arguing for the inclusion of low-income, rural areas, despite the increased costs. Philanthropic funders acted in the public interest by foregoing the standard investor premium to sustain future investment.

The margin, or the “razor’s edge,” for this broader interpretation is thin because the SIB’s neoliberal logic privileges financial gains. Figure 2 visualizes our findings on navigating this margin. South Carolina has no premium and no profit, instead inserting social objectives into the market to move the political conversation in a conservative state. Utah ran the financial risk of a SIB with a long view of moving state politics, but the investor payout and methodology undermined its social mission by diverting public investment to private investors. Chicago demonstrates the risk of financialized logic by providing short-term benefits to investors and clientele without the balance of long-term social objectives. This SIB falls totally on the extractive side of the market.

Figure 2. The razor’s edge: Do SIBs enhance investment or deepen financialization?

Conclusion: Cautioning the use of social impact bonds

We have shown that SIBs in the United States have the capacity to overcome their internal financial logic when they include long-term objectives to build systemic change. These objectives are not written into the contracts. They are not part of the design. But they leverage the capacity of SIBs to produce localized information. This capacity is especially useful in conservative political contexts because strategic networks of social advocates can use that localized information to nudge discourse and alter policy. This capacity is not among the advertised claims about SIBs, but we argue that SIBs in the United States can be leveraged to scale up public investment to the state level by strategically targeting long-term policy change.

SIBs in the ECE sector should seek to create sustainable investment. A SIB that only pays for current costs and does not consider how to sustain investment is not worth the transaction cost or the interest rate. SIBs that overpay their investors divert funding from social services, as in Chicago and Utah. South Carolina shows that an investor premium is not necessary and that philanthropies can significantly reduce SIB costs through their traditional grant-making roles.

Grappling with fiscal and political constraints, U.S. cities are trying to find new sources of funding for critical social interventions, but they should be wary of SIBs’ marketizing framework. Although our cases show the potential for new investment, SIBs may undermine their own social objectives through their financializing metrics. The risks of high transaction costs, overpayment to investors, inflexible implementation, and loss of focus on vulnerable clients are significant. Without the benefit of scaling up public investment and shifting policy, cities invite a Trojan horse of public value into their neediest communities when they implement a SIB. We caution cities to pay attention to these risks as they launch future SIB experiments.

About the authors

Allison E. Tse is a master’s student in City & Regional Planning at Cornell University concentrating on economic development. She researches social impact bonds as well as fiscal policy and austerity in local governments in the United States. Her past work has focused on rural community and economic development in the United States.

Mildred E. Warner is a Professor in City and Regional Planning at Cornell University. Her research focuses on the economic development, social service, and environmental policies of local governments. She is an international expert on new models of service delivery and finance, especially in the context of privatization and decentralization.

Acknowledgments

An earlier version of this article was presented at the Social Finance, Impact Investing, and the Financialization of the Public Interest Conference organized by Eve Chiapello and Lisa Knoll at Hamburg University in Germany and supported by the Humboldt Foundation.

Additional information

Funding

This research was supported in part by a Luigi Einaudi Chair Innovation Grant from the Cornell Institute for European Studies and a grant from the Atkinson Center for Sustainable Futures.

Notes

1. Contract and loan agreements are listed in the References section with dates and involved parties, typically including the public payors, intermediaries, and investors (City of Chicago, 2014a, 2014b, 2014c, 2014d; SCDHHS, 2016).

2. To preserve anonymity, citations for interviews are written as SC SIB 2017 for South Carolina interviewees, Utah SIB 2017 for Utah, and Chicago SIB 2017 for Chicago.

3. Under the Affordable Care Act, preventive health programs, such as the Nurse–Family Partnership, can be funded through federal Medicaid dollars. States have wide latitude in how they use federal Medicaid dollars, so funding levels and eligibility criteria for reimbursement of services differ greatly by state. The Medicaid funding for NFP in South Carolina is the first of its kind in the country.