Virtuous profits: Pay for success arrangements and the future of recidivism reduction

Introduction

In 2012, New York City (USA) launched the Adolescent Behavioral Learning Experience program in the city’s Rikers Island jail. The program sought to reduce the 47% recidivism rate of 16- to 18-year-old inmates returning home (New York City Department of Correction, 2012). To “break the cycle” of incarceration, the young inmates, 91% of whom were Black or Latino, received an evidence-based cognitive behavioral therapy while incarcerated. An evaluation of the Adolescent Behavioral Learning Experience program showed that 87% of the 1691 16- to 18-year-olds incarcerated for more than seven days at Rikers in 2013 attended at least one session of the program (Vera Institute of Justice, 2015). A nonprofit organization implemented the evidence-based intervention, Moral Reconation Therapy (MRT), which focuses on improving social skills, decision-making, and instilling personal responsibility. The young men completed workbooks designed to address a host of individual-level risk factors, including maladaptive beliefs, attitudes and behaviors, negative relationships, and hedonism. These workbooks—serving as a proxy for counseling and psychotherapy—seek to reinforce positive behaviors, enhance positive identity formation, improve self-concept, develop frustration tolerance, and encourage “higher stages” of moral reasoning in participants.

This all seems unremarkable. The therapy at Rikers Island jail is not unusual either in kind (evidence-based and individualized) or in mode (workbooks). That is to say, MRT is one of numerous evidence-based programs (Lipsey, 2009) that address the individual-level risks and needs of incarcerated people (Hannah-Moffat, 2013; Ward and Maruna, 2007) and relies on workbooks to change the thoughts and attitudes of inmates who are returning, very often, to the same social ecologies that contributed to their offending in the first place (Davidtz et al., 2015). What is unusual, however, is that the reentry success of these young men was being wagered on by one of the largest investment firms in the world: Goldman Sachs Bank USA. The global investment banking, securities, and investment management firm paid $9.6 million upfront for the Adolescent Behavioral Learning Experience program at Rikers. Contractually, the city of New York would only be required to repay Goldman Sachs for its investment if an evaluation showed the program reduced recidivism by 8.5% or more (Roman et al., 2014). A greater than 10% reduction would have led to a profit for Goldman Sachs—in the range of $500,000 to $2.1 million. If the program did not meet the performance targets, New York City would not have saved taxpayer dollars and therefore not be obligated to repay Goldman Sachs.

In what is being called a pay for success deal, the repayment hinges on an “impact” calculation: the difference between the outcomes of program participants and those of a comparison group. This arrangement—financed through what are called Social Impact Bonds—transfers the risk from government to private interests and sutures the reentry success of young inmates in New York to corporate profits. Unfortunately for Goldman Sachs, the recidivism rates for eligible youth did not differ statistically from those of a matched historical comparison group (Vera Institute of Justice, 2015). This meant that the City of New York was not required to reimburse the investment company for the three years of programming it paid for. Put into economic language: Goldman Sachs’ investment was not profitable. Consequently, since private investors are “in charge” of funding, in August 2015 Goldman Sachs terminated the Adolescent Behavioral Learning Experience program at Rikers Island jail.

The Rikers Island jail arrangement is a paradigmatic example of the next stage in the public–private partnerships that Western Democracies use to administer and fund social programming, including crime control. Beginning in the 1990s, private foundations and governments began employing “outcomes-based funding” as a way to target “specific goals for a program or initiative” (Hernandez et al., 2012: 3). This performance-based funding arrangement, which is still in use today, tied the accomplishment of key outcomes, such as school retention or graduation rates, to funding levels. With pay for success, private capital finances traditional public sector activities, and investors stand to profit if the initiatives they fund achieve certain outcomes. Governments reimburse investors only if the program saves (future) governmental expenditure. Advocates argue that, since programmatic success depends on demonstrating measurable outcomes, pay for success will encourage the adoption of evidence-based programs and discourage the use of unproven ones.

Pay for success is underway in the United States; it is being used in a more limited capacity in the United Kingdom and Australia; and it is being considered in Canada, France, Germany, Ireland, Israel, and South Korea (Patton, 2013). At the same time, there is renewed demand for decarceration or “rightsizing” of the US correctional system, with increased attention being given to diversion and reentry programs (Gottschalk, 2014; Mears and Cochran, 2014). We therefore consider it essential to explore what pay for success contracts will mean for these programs. In this article, we first describe the pay for success way of funding, implementing, and judging recidivism reduction programs and discuss what role it will likely play in the US criminal justice reforms that seem to be afoot. We then speculate about how pay for success financing will affect the nature of recidivism reduction programing as well as the social ecology of community-based social control. In the spirit of Stanley Cohen (1985), this article offers some preliminary prognostications about how the pay for success arrangement will fundamentally alter the size, logics, and practices of US community crime control as well as the dominant discourse of “what works” in criminology. The article concludes by situating these possible changes within broader shifts in social control and security under neoliberalism.

Mass incarceration, decarceration, and reentry

With nearly one of every 100 adults in prison or jail—a rate that is 5–10 times higher than rates in Canada and Western Europe—the US penal population accounts for 25% of the world’s total inmate population, including about a third of the world’s female inmate population (Gottschalk, 2014). There is, however, evidence of declines in the growth of the US inmate population. And although the decreases are not uniform across US states, locales, and the federal government, numerous calls for decarceration are being made (Garland et al., 2014). Signs also point to US politicians’ motivation to change direction, and public opinion shows waning support for mass incarceration policies (Garland et al., 2014). We may be cautiously optimistic that there is for now a waning in the eagerness to incarcerate so many people in the US—at least when it comes to nonviolent lawbreakers and legal residents.

As part of this shift, legislators have begun to increase funding for prisoner reentry programs (Green, 2013), and most Americans are generally supportive of reentry programming (Krisberg and Marchionna, 2006). Reform efforts will certainly involve changes to reentry policy and many scholars are researching and writing about this (Green, 2013; Mears and Cochran, 2014; Miller, 2014; Steen et al., 2012). Finding funds for reentry is therefore vital. However, with tight budgets and politicians reticent to soften too much on their tough on crime stance, funding will be difficult to secure. And funded programs better not fail, since policymakers are likely to be attacked for supporting unsuccessful programs. In this current climate, the mantra of evidence-based programs has exploded. But also—and fortunately for nervous politicians—pay for success contracting promises to be a politically safe way to meet social challenges such as high rates of recidivism. Governments around the world are piloting these types of deals with private investors in a host of policy areas, including reentry (Roman et al., 2014). To provide a fuller understanding of the types and arrangements of pay for success projects, we offer here a few illustrative examples that will then give us the space to envision their possible effect on recidivism reduction.

The pay for success “solution” to social problems

Global interest in the potential of this funding scheme to reduce recidivism, youth offending, and meet other social “needs” originated in the United Kingdom (Hernandez et al., 2012). In the northeast English city of Peterborough, the first social impact bond was created in 2009 to reduce the then 60% recidivism rate among inmates housed in the Peterborough Prison. Aiming to avoid the $126,000 yearly price tag for housing one repeat offender in prison, the privately funded project’s “impact” marker was a 7.5% reduction in recidivism among nonviolent, short-term prisoners. Were it to be met, the Ministry of Justice agreed to pay out from the carceral cost savings a dividend of up to 13.5% to the bond investors (Hernandez et al., 2012). Fortunately for the investors, independent assessors from Qinetiq and the University of Leicester found an 8.4% reduction in recidivism for participants as compared to a control group (Costa, 2014). If the reduction had exceeded 10%, an immediate payment would have been triggered in 2014. Despite the apparent success of the initiative, broader reconfigurations in the administration and funding of reentry services in the UK signaled the end of the Peterborough social impact bond-funded program (Cahalane, 2014).

Considered to be the classic case, under the Peterborough Prison-style pay for success arrangement, a government agency sets a definite, measurable outcome to be achieved within a well-defined population over a predetermined period of time (Hernandez et al., 2012). Although still used relatively rarely, the funding arrangement appears to be growing in popularity. As of June 2016, there were more than 30 social impact bonds in the UK, often funding programs that aim to improve the life chances of youth not in education, not employed, and not in a training program (i.e. so-called NEET youth).

The US is piloting pay for success deals as well. In addition to its investment in the Rikers Island jail deal in New York City, in early 2014 Goldman Sachs contributed $9 million to a $21 million social impact bond initiative in the US State of Massachusetts to fund community-based organizations working to lower the rate of recidivism for high-risk young men already in the justice system (Deprez, 2014). This bond funded, among other organizations, a community-based organization named Roca. According to the state’s calculations, each participant Roca keeps out of jail saves Massachusetts $12,400 for the year. Under the terms of the deal, after Roca’s program has been operating for four years, independent evaluators will compare the Roca group’s recidivism rate against a control group. If the men enrolled in Roca spend 22% fewer days incarcerated than the control group “Massachusetts would save enough money to repay Goldman Sachs’s investment” (Deprez, 2014).

A host of social problems are being addressed with pay for success-funded programs across the US. In late 2015, the state of California launched Project Welcome Home, the state’s first pay for success initiative, to address homelessness. Cities such as Nashville, San Francisco, and Salt Lake City have followed suit. The state of Ohio recently embarked on a five-year, $5 million dollar pilot project funded by a variety of private investors and philanthropies to reduce the time children of homeless parents spend in out-of-home foster care (Forbes, 2014). Having recognized that US cities and states were brokering pay for success deals, in 2014, the US federal government proposed investing nearly $500 million in pay for success initiatives, including $300 million for an incentive fund at the Treasury Department (VanRoekel and Greenblatt, 2013). And President Obama’s 2016 budget proposal increases federal funding for new pay for success deals. Although it is unclear to what extent pay for success will grow in the UK and the rest of Europe, US investment banks like J.P. Morgan estimate that pay for success (or “impact investing”) will reach between $400 billion and $1 trillion by 2020 (Quinton, 2013). With ongoing austerity measures and budget cutbacks in the public sector, the pay for success model will likely blossom as it allows governments to fund social programs only after a successful cost-saving outcome is realized. The calls for “rightsizing” the criminal justice system are occurring at the same time as the pay for success movement begins to ascend, meaning that the two may influence each other in profound ways.

Pay for success and the meaning of “evidence-based”

Pay for success would seem to encourage the funding of evidence-based recidivism reduction programs, while injecting monetary concerns into the (already contested) concepts of “evidence-based” and “recidivism.” Under pay for success, recidivism reduction efforts would focus on changing aspects of offender’s lives that are consequential in monetary terms, especially the rate of recidivism. Given that profits are paid out of supposed carceral cost savings, under pay for success arrangements, recidivism would be defined as reincarceration in a locked facility rather than in behavioral terms or rate of arrest.

Advocates claim that pay for success will increase the amount of rehabilitative and reentry programs in the US and encourage the adoption of proven programs. Preventive and rehabilitative programs often go unfunded in the US because large up-front investments from the state are deemed too risky, since they rely on outcomes that occur well into the future (Roman, 2015). Pay for success would seem to address this problem. Pay for success deals also promise to eradicate funding for programs that do not work while increasing accountability (Rudd et al., 2013). So, there is reason to believe that evidence-based criminal justice programs will be required for adoption, and that the investors with a financial stake in a program’s success will carefully oversee the actual implementation of these proven techniques to help assure a profit. A major investment firm may well be a much stricter overseer of a nonprofit organization implementing an evidence-based program than the state bureaucrat or private contractor who currently does this job under a standard public–private partnership.

We think that pay for success may fundamentally alter the disputed meaning of “what works” in US criminal justice policy and practice. As it currently stands, audits done by outside assessors gauge whether a service provider is delivering an intervention with fidelity to the intended design (Fagan et al., 2011) and this is what often counts as “success” under many public–private partnerships. But if the metric for success truly shifts to future carceral cost savings (i.e. an impact, which proves that programs are making a difference that translates into real value for the government), how an organization achieves that savings may become markedly less important. This shift away from concern about outputs and toward financial impact outcomes will mean that the audits and checklists currently used to assess the “success” of interventions may significantly recede in importance. Implementing evidence-based programs, and adhering to program fidelity, may become only a tertiary part of how an organization lowers recidivism. This might elevate the perceived importance of community-based service providers, who may now be given more autonomy in how they go about lowering recidivism. Taken together, these shifts would further complicate our field’s contested understanding of “what works.”

To illustrate, let us imagine there is an organization, Project No Jails (PNJ) that is being funded by a pay for success arrangement to decrease the number of days its participants (young men returning home to Chicago) spend in locked custody following reentry. Like many private community-based reentry service providers, PNJ offers cognitive behavioral therapy, job placement assistance, and life skills training to the young reentrants it works with (Miller, 2014). But in addition to offering individual-level programming, PNJ connects reentrants to social service, vocational, and educational agencies, all of which research shows to be important for facilitating successful reentry (Mears and Cochran, 2014). Moreover, PNJ not only connects reentrants to a social service network to better meet basic human needs, but works against barriers that restrict economic and civic opportunities of formerly incarcerated people, which mirrors the practices of many existing community-based organizations (Goddard et al., 2015; Myers and Goddard, 2015).

In addition to this service work with reentrants, and similar to other community-based organizations across the US, PNJ has forged working relationships with a number of local police and probation officers who are sympathetic to PNJ’s work. The organization believes that these relationships with law enforcement are important because criminal justice agents are more likely to contact a PNJ counselor when they stop a PNJ participant on the street rather than issue a citation or make an arrest. This happens under the public–private partnership model as well: a number of organizations in the US advocate on the behalf of young people caught up in the system or on gang registries, work to decriminalize youthful indiscretions, and reduce the use of reactive policing practices (Goddard et al., 2015). When success is determined by gauging adherence to a proven ‘evidence-based’ program model, as it is under most current public–private crime control partnerships, organizations do not receive any credit for reducing crime by changing criminal justice reactions. Under the results-focused pay for success model, however, this would count as “success.” What this potentially means for organizations is that they will be freer to design practices with the best chance of lowering recidivism for their participants. And since technical violations of parole and minor brushes with the law account for so many returns to locked facilities, changing these reactions would seem to be a very attractive option.

Altering criminal justice reactions may become a proven, acceptable strategy under pay for success. More generally, it seems very likely that if current behaviorist evidence-based programs fail to deliver the reductions in recidivism needed to assure profits, private investors will look to other approaches. The Rikers Island jail pay for success deal’s primary intervention was the evidence-based MRT program (Rudd et al., 2013), which did not lead to the desired outcome for profitability. Since the aim is to make a profit, investors will likely look to additional ways to achieve the outcome needed to trigger the success payment.

The most expedient way to lower recidivism is, of course, to change laws, how they are enforced, and significantly reduce technical violations for reentrants (Mears and Cochran, 2014). A variant of this approach has already taken place in at least one pay for success project. The Boston-based organization Roca, which is partially funded by Goldman Sachs, claims to be not only delivering cognitive behavioral therapies, but working with local law enforcement to change the way policing and gang enforcement is done with young people returning home from locked facilities (Deprez, 2014). As its website states: “Through dialogue and action, Roca continues to promote and produce alternative, restorative policies in order to impact systemic change in how our communities address the needs of high-risk youth” (Roca, 2014). It appears that the ability to forge strong relationships with local social control agencies will ascend in importance. It is not that evidence-based program models focused on individual behavioral change will vanish; however, their importance will likely fade and their adoption will become one aspect of what makes for, after involvement in a few profitable pay for success ventures, an evidence-based organization.

If establishing relationships with social control and social service agencies ascends in importance, this may change which nonstate organizations are best positioned to compete for private investor partnerships, as local knowledge of the social ecology of services and social control will become crucial. Brokering such deals requires a host of nontechnical “soft skills” that have historically been outside of the scope of what is thought to be “the crime sciences” or “what works.” As it stands, an organization that “works” is one that can demonstrate its correct implementation of an evidence-based program, as measured by the state or a third-party evaluator working on its behalf. What matters is the extent to which organizations can be effectively colonized by “what works” ideas for reducing crime, perform them as prescribed (at least when the auditors show up), and ignore—or play-down—local solutions and services not mandated in their state contract, including their partnerships with other organizations. However, the ability to partner with other service providing agencies might become a best practice if the variable on which providers are being judged switches to future success outcomes. And this is what the persons managing the Peterborough prison pay for success deal claim: a key reason for the program’s success was its “…excellent working relationship with Peterborough prison and close links with local statutory and voluntary partners in the area” (Social Finance, 2014: 3).

Pay for success may encourage and reward organizational flexibility, rather than enforce strict adherence to program model fidelity. RAND Europe’s independent evaluation of the Peterborough social impact bond concluded that the “Flexibility of funding in the Peterborough pilot was widely considered to be innovative and central to the delivery model” (Disley and Rubin, 2014: 34). If this trend continues, locally managed interventions with flexible funding arrangements that can adapt to local conditions will likely become a best practice under pay for success financing.

But how might pay for success change our thinking about “what works”? If how an organization achieves a reduction in recidivism becomes less important, the change may be profound. Most fundamentally, we may very well move away from proven models, therapies, and frameworks—which commonly work to prevent harmful behaviors or promote positive behaviors—to proven organizations or providers as the sought-after gold standard. As it currently stands, the component that matters is the program model—the organizations that usually deliver interventions are thought to be interchangeable. But under pay for success financing, we are likely to see the provider become the key factor rather than the program model (a possible outcome made even more likely when proven programs fail to achieve their targets). Rather than being judged on whether an audit shows them to be delivering MRT or multisystemic therapy correctly, organizations will be judged on whether they can achieve, by whatever means are necessary, the prearranged impact outcomes. These outcomes save the state criminal justice costs and win investors their profits. Given this, what will be at a premium here are impactful organizations more so than scientifically proven program models. The term “evidence-based” is unlikely to go away, but we may go from talking about “evidence-based programs” that reduce recidivism to “evidence-based organizations” that consistently deliver quantifiable reductions in public spending and deliver profits to investors.

If the crucial element switches to evidence-based organizations, this could be unwelcome news for administrative criminologists. Along with face-to-face consultation and evaluations, criminologists sell proprietary assessment tools, rehabilitative program checklists, and off-the-shelf programs to different entities in the justice system. To succeed in this new arrangement, the skills necessary to run a functional organization will need to be imparted. When asked what organizational qualities his company looks for in a pay for success partner, a Goldman Sachs investor involved in the Roca arrangement explained it this way:

“It’s not that different than if you were looking at any company” to invest in, says Andrea Phillips, vice president of Goldman’s Urban Investment Group. “You’re relying on a management team. You’re relying on their human capital, their folks who work at the plant and deliver and do what they’re supposed to do.” (Deprez, 2014)

Pay for success and the social ecology of community crime control

To broaden our analysis, we now explore how this funding arrangement may shape the social ecology of community crime control. As noted famously by Stanley Cohen (1985), much of the social control system’s growth and character can be explained by professional self-interests being advanced; however, political and economic forces ultimately stop the “build-out” of social control systems and shape the form that circuits of control take on throughout. So, while political and economic forces set the boundaries of social control, the actions and ideologies of organizations and individual actors help us understand the extent and content of much of this build-out of social control (Cohen, 1985). This is particularly true with community crime control—new alternatives are created to shore up the perceived shortcomings of the previous social control systems. Community alternatives divert people out of the system or serve as “softer” forms of punishment, but then the “wrong” people get caught or the community alternatives actually extend control. Therefore, new alternative systems need to be created to “better serve” these people—be they too “hard” or too “soft” (or too low or high of a risk) for the original alternative (Cohen, 1985). Social control creates more social control, and Cohen calls this the iatrogenic feedback loop phenomenon—illness caused by medical examination or treatment.

This feedback loop phenomenon applies more to the community side of crime control than the carceral because locked institutions are not judged on growth (Cohen, 1985). A good prison is a safe one, not necessarily a large one. But for devolved community crime control there is pressure for growth—in the public and (especially) private sectors of community crime control. How might pay for success alter (in size, but also quality) this feedback loop phenomenon? For one, it might increase the pressure on professionals to create larger systems for reentrants from prison and jail, since more people equals more profit, all other factors being equal. Expanded systems of control have always been in the interest of community crime control professionals—thinking of how they can create new or better ways of classifying and treating reentrants, and have their expertise be recognized as valuable (Cohen, 1985). Pay for success, however, may put a unique spin on this.

Many current recidivism reduction and reentry programs rely on behaviorist models that use some mixture of “carrots” and “sticks” to coax clients into following the organization’s rules and the law (Ward and Maruna, 2007). Under pay for success, the most powerful “stick,” that of jail or prison, would now principally be off limits because this is the “bad” cost that organizations are being charged with reducing; and they are not simply being asked to do this—their viability as an organization now depends on it. Organizations will have to use carrots and find other sticks that are noncarceral.

To illustrate, if a counselor at the PNJ organization suspects a young man is using drugs, the counselor has a vested interest in not turning him over to a police, probation, or parole officer. Still, a corrective action will be needed. After staff members exhaust internal carrots (e.g. extra services, less surveillance, later curfew) and sticks (e.g. verbal reprimands, removal of privileges, early curfew) to correct the young man’s behavior, the PNJ staff member may enroll the participant in the services of another noncarceral social control organization with which it partners: perhaps it will be additional counseling or therapy, perhaps it will be GPS monitoring, or perhaps it will be a make-work type of program. This ratcheting up of assistance and coercion, help and control—and perhaps public cost, if tallied holistically—occurs in the hopes of decreasing the rate of recidivism that investors are wagering on programs to reduce.

What gets complicated here is that these organizations and the investors who fund them have a clear economic incentive to keep people out of the system (at least the “deep end”) but they are part of a larger movement of privatized social control, which has a vested interest in maintaining a wide net of control over people so that organizations can secure profits for investors and ensure a steady flow of clients for the service provider. So an organization needs to have contact (and thus control) over the client to have a chance at profiting, but then to become a successful organization its actions must not facilitate contact with the carceral end of the system (or whatever institution is responsible for the “bad” cost that is trying to be minimized). When used for the purpose of carceral cost savings, pay for success financing will likely encourage organizations to create and strengthen relationships with agencies outside of criminal justice—including social welfare, mental health, and other “alternative” programs—thus further entangling the “penal, welfare, social service, surveillance, governing, economic, and political functions” of government (Gottschalk, 2014: 13).

This we think will expand what Beckett and Murakawa (2012) refer to as the “shadow carceral state” and expand the ways in which the state punishes through seemingly nonpunitive means (Gottschalk, 2014; Hannah-Moffat and Maurutto, 2012; Lynch, 2012). Whether imposed through civil law, administrative law, or noncriminal courts, a range of small policy shifts and civil alternatives will (re)emerge in relative importance in state punishment (Valverde, 2012). These seemingly small or “softer” measures (e.g. fines, registries, drug testing, and host of monitoring technologies) do not necessarily depend on criminal law or actual confinement, but remain punishing in that they draw cognitive boundaries that restrict, exclude, and condition reentrants in their day-to-day lives. We reason that pay for success funding of reentry programming will likely expand the wide array of noncustodial controls at the state’s disposal. As with other reforms of the neoliberal period, this instance of reform may ultimately reinforce “both the carceral state and the sharp right turn in American politics over the long term” (Gottschalk, 2014: 15).

In addition, since custodial costs are the “bad” cost the state is trying to minimize with this incarnation of reform, this may change the way organizations relate to their “clients” and perhaps slow the haste at which violations are brought to the attention of parole officers. Organizations have a financial interest in not turning clients over to the criminal justice system. An inclination to use other noncarceral reprimands may develop, and the result may or may not be a bad thing: while most criminologists would welcome less use of carceral responses for minor parole violations (Petersilia, 2011), many scholars would argue that “alternatives” to incarceration that do not address racialized social and economic inequalities are no alternative at all (Miller, 2014; Story, 2016).

Pay for success may also alter who receives reentry services. Those individuals most likely to be included in a pay for success-funded intervention may be those most amenable to change (e.g. youth) while others will be rejected by the organizations (or investors) as bad risks or as having insufficient future carceral costs attached to them. Maruna et al. (2012: 317) envision a similar phenomenon in the area of early release and the privatization of parole:

Commercial entities will employ financial coercion and a search and seizure authority to ensure that individuals adhere to all the conditions of their release, terminating the contract at will, and/or sending bounty hunters out to return them to prison. Many prisoners will not get this chance, of course, as they will be rejected by the private companies as bad risks.

A broader contradiction arises as well: pay for success financing encourages growth at the organizational level, but it is occurring in the context of broader calls for austerity in state spending (Aviram, 2015). Now, what really matters is what is considered a “cost” and where the line is drawn between “good costs” and “bad costs.” Carceral spending is clearly envisioned as the “bad cost” in need of reduction under these agreements (Deprez, 2014). Other costs associated with reentry (e.g. housing vouchers, drug treatment, mental health services, and education programs) that are at least partially paid by public monies are tougher to categorize as “good” or “bad” and it is not clear how (or if) these will be tallied under these arrangements.

However, what if when trying to determine the success of a recidivism reduction program, rather than only totaling carceral costs savings, your cost–benefit analysis totaled all of the public benefits that a reentrant received and charged community-based organizations with limiting all those costs to prove success? That is, what if the goal was to tamp down all of the social costs a person incurred during reentry, so that the successful reformed citizen was one untethered to any system of quasi-state care or control? In that case, a successful reentrant might look like this: they are not attending a drug treatment center; they are not accessing services at a mental health clinic; they are not in a public employment or housing program; they are not enrolled in an alternative educational course; and, they are not incarcerated.

This might start to sound overly cynical and it might very well be. However, calls for reduced carceral spending are nested within broader calls for austerity in spending on the social (Aviram, 2015; Gottschalk, 2014). At the least, when carceral cost savings—and corresponding profit—become the main driver of policy development and formation other considerations may be crowded out, including professional judgment, theory, research findings, and even ethical concerns. As Mary Bosworth puts it, when

cost savings and lower reoffending rates are presented as ends in themselves—that are somehow separate from thornier matters of human rights, morality, or justice—what is to protect against practices that might be extremely harsh, but reduce overheads or recidivism? (Bosworth, 2010: 196)

Discussion and conclusion

In this article, we reflected on how the logics and practices of community-based programming may change if pay for success arrangements motor recidivism reduction efforts in the US. The limit of this analysis is that we were only able to describe the pay for success movement and theorize what will take place in the area of community-based recidivism reduction. How this all turns out is really an empirical question, and we hope this article motivates researchers to do case studies, qualitative interviews, and ethnographies at pay for success sites. But we raise these questions, in part, to illustrate the historically conditioned politics and sought end games of those who are pushing these initiatives. Linking punishment, and social control generally, to corporate profits has always been a treacherous endeavor (Gottschalk, 2014), as the private investors who devise these social bonds to fund the carceral (and shadow carceral) state are driven to cultivate it.

The reason why private investment companies engage in pay for success deals is meaningful: for profit, or what they likely consider to be virtuous profits: namely, making money while doing quasi-philanthropic work to help the social, but also earning profits under a lower tax bracket status and generating favorable publicity. To be sure, privately funded organizations will find creative ways to achieve goals and outcomes and there are bound to be a multitude of “privatization journeys” as pay for success logics reverberate through different localities and criminal justice agencies with unique institutional histories and work cultures (Robinson et al., 2016; see also Werth, 2016). However, the political and economic interests of those who advocate and fund these initiatives structure the parameters of these journeys. These structural features eventually “stop” the construction of crime producing feedback loops and condition how they function throughout (Cohen, 1985).

Where does the state fall in all of this? We believe that this move toward neo-privatization (i.e. social programs administered by nonstate actors and funded by private investors) should not be read as the state going away. The state steers and legitimates crime control initiatives carried out by private organizations, even when they are funded by private monies (Lea and Hallsworth, 2012). The social impact bond phenomenon ought to be recognized as part of a neoliberal project that transfers wealth to the few and as contributing the installation of security measures that protect the accumulation of capital from the social problems inherent in an unequal and increasingly volatile social world (Wacquant, 2009). The state does not wither away with such privatization; rather, it works through nonstate actors, which it continues to steer.

Private capital, by playing an enhanced role in the provision of security, does not mean that the state loses its sovereign function, even when it appears to be privatizing and rolling back the state. Rather what is occurring is the growth of mechanisms whereby the state delegates aspects of its core functions to others. (Lea and Hallsworth, 2012: 32–33)

Domestically, an important state task is to manage marginal populations for whom there is no place in the labor market (Hallsworth and Lea, 2011; Wacquant, 2009). While some state projects may successfully coerce some marginal people into low-wage work (Soss et al., 2011), nonworkers “recalcitrant to workfare will end up being effectively warehoused out of sight, somewhere along the ‘seamless’ continuum of prison and probation” (Fitzgibbon and Lea, 2014: 26). Private firms are commonly partners on both ends of this continuum already; however, pay for success arrangements amplify this privatization process, as the long-term economic and ideological benefits of securitization may now be realized from a project that is immediately profitable for the same vested interests. This also makes the connection between corporate profits and state legitimated social control less opaque and thus, one might hope, easier to critique and organize against.

As with other aspects of the neoliberal project, inequalities will arise in who will receive supervision and services under pay for success service arrangements (Soss et al., 2011). Pay for success may exclude whole groups of offenders from receiving carceral and reentry services; while the “risk principle” has long called for investing in high-risk categories (Lowenkamp et al., 2006), under pay for success, the best investment will be those groups that are not only at high risk for reincarceration but also most costly to the state (and thus most profitable for private investors). Many groups of reentrants will likely not be invested in because there is too little projected future carceral cost savings to qualify them as a sound investment. Again, juveniles may be the safest investment since they cost a great deal to incarcerate and have the most potential carceral years in front of them. Groups with little projected future carceral cost (though not necessarily the least social, health, or human cost) would appear to be a particularly unwise investment. However, if pay for success arrangements structure other sites for poverty management, low carceral cost groups may become targets for profit-making by healthcare or social welfare entities.

Only with time and empirical inquiry will we know what we got right and where we missed the mark. While much of this article is speculative, we believe that the profound changes that may be on the horizon warrant such speculation. While we do not see pay for success as calling forth some dystopian future, we are not optimistic. In particular, we think pay for success may deepen inequalities in service provisions, call into question all social spending, contribute to the proliferation of securitization projects aimed at marginal populations (Hallsworth and Lea, 2011), and further blur present-day penal boundaries (Hannah-Moffatt and Lynch, 2012). Moreover, even if programs funded through pay for success arrangements successfully lower rates of recidivism, locked facilities need to actually close in order for significant public cost savings to be realized (Gottschalk, 2014). In the end, this iteration of carceral reform may result in net increases in social control for economically marginal social groups while pushing US criminal justice and social policy further away from the noncontingent access to social and material goods guaranteed in more equal, less violent, and less criminalizing countries (Currie, 2016).

In sum, pay for success may enable a prolonged and repositioned form of postcarceral punishment. This article outlined the future role that pay for success may play in “crafting, administering, monitoring and legitimating” community-based recidivism reduction programming (Hannah-Moffat and Maurutto, 2012: 203). Although the noncarceral strategies may appear as a novel way for the state to punish—and this may be true in the exact way they are operationalized—noncarceral punishment has historically been commonplace (Valverde, 2012). Asylums, psychiatric wards, orphanages, and residential schools for aboriginal children were all noncarceral sites for state punishment (Valverde, 2012). We can add countless examples of state punishment from the past that were noncarceral or not based on criminal law, such as those that are philanthropic (e.g. juvenile child saving), medical (e.g. forced sterilization), or place based (e.g. movement restrictions and confinement camps). Our analysis of new and braided forms of assembled punishment practice—which often blur the boundaries of the formal and informal, the criminal and civil, and the legal and quasi-legal—must be situated within the wider project of state punishment (Valverde, 2012).

As we contemplate what is happening or what may be just ahead, our theoretical analysis must be firmly positioned within the wide universe of punishment so often deployed upon those at the social and economic margins. This is not to suggest that nothing is new qualitatively. In fact, just the opposite, as many “reassembled practices of welfare, treatment and punishment” will likely give rise to “new and disparate forms and logics of punishment” (Hannah-Moffat and Maurutto, 2012: 203). Thus, if jail and prison populations decrease and noncarceral punishments increase, rather than advancing empirical claims about the expansion (or contraction) of punishment (especially without running the proverbial numbers), the crucial theoretical task is to chart the vast and shifting cosmos of punishment. We believe that the pay for success phenomenon may fundamentally reorder the frequently shifting (and often familiar) universe of punishment at the state’s disposal.