Every year, JPMorgan Chase and the Global Impact Investing Network report that the top two impact investor complaints are “the lack of … appropriate capital across the risk/return spectrum” and “the lack of … high-quality investment opportunities (fund or direct) with track record.”

Let’s put that to rest, shall we?

This article makes the case that mainstream financial institutions can generate large and enduring deal flow at risk-adjusted, non-concessionary rates of return using an enhanced model of Social Impact Bonds called “Scale Finance.” It takes a close look at the quantitative assumptions and calculations behind a pro forma transaction that would finance the expansion of two certified evidence-based programs — Multisystemic Therapy® (MST) and Functional Family Therapy® (FFT), both of which have been proven to prevent juvenile offending and reduce juvenile recidivism — at their maximum feasible growth rates.

By developing follow-on investments of the same magnitude, a prototypical state that has more than 1,000 juveniles each year who have been committed to some form of custodial “placement” — such as juvenile detention or a group home — could, in time, effectively and permanently end the misguided practice of mass incarceration of juvenile offenders. Scale Finance transactions could accomplish this systemic transformation without up-front cost to the state, pay investors competitive returns, and provide net savings to the state.

Solving pervasive social problems we already know how to fix is a long and arduous journey. Thirty years of rigorous independent research show that MST/FFT work and that they remain effective as long as expansion efforts maintain program fidelity. (See Table 1.) Yet they still reach only 5–10% of the families in need.

Table 1. Two Certified Evidence-Based Programs for Juvenile Offenders

Further, the Essex County Council (UK) MST SIB serving “children in care” shows that MST saves more than it costs and that those savings can be monetized within a reasonable investment horizon. (See Table 2.)

Table 2. Essex County Council (UK) SIB [2]

But could SIBs expand MST and FFT commensurate with unmet population needs and pay for themselves by monetizing future government savings? The Scale Finance investment described below is modeled after the successful Florida Redirection project that served some 10,000 families from 2004–2013. Redirection was not SIB-funded, but it saved the state more than $181 million compared to what it would otherwise have spent on juvenile placements. (See Figure 1.)

Figure 1. Florida Redirection Project Savings

At its peak, Florida Redirection served a little more than 1,000 families per year at an annual service cost of about $10 million. Table 3 presents a pro forma cash flow analysis for a Scale Finance SIB of roughly the same size as Redirection over five years. (To put this in context, juvenile courts committed more than 1,000 youth to placement in seven other states in 2013: California, 4,452; Texas, 2,577; Pennsylvania, 2,337; Ohio, 1,338; New York, 1,236; Michigan, 1,224; and Virginia, 1,014.)

Table 3. Scale Finance Pro Forma Cash Flow

An important feature of Florida Redirection is that it targeted youth who had been adjudicated “delinquent” — that is, they had appeared before a juvenile judge and were found guilty of committing an offense — and they were awaiting “disposition” (i.e., sentencing) to a “residential placement.” In other words, 100% of program participants [cell B4] were not going home after trial. In this way, Redirection focused exclusively on “deep-end” youth who cost the most and whom intensive family therapies like MST and FFT were designed to serve.

The pro forma makes four conservative assumptions:

· Placement Cost. Nationwide, the average annual cost for juvenile placement per youth is nearly $90,000, but this average is reduced by small, primarily southern states that spend little on youth committed to their care. Large, urbanized states with larger placement populations have substantially higher detention costs, such as New York ($352,663), Virginia ($260,019), California ($208,338), Ohio ($202,502), and Michigan ($173,455).[3] The pro forma assumes annual placement cost per youth of $100,000 [B2].

· MST/FFT Cost. The pro forma assumes program cost per family of $10,000 [B3], which exceeds the $7,500 cost of MST, the more expensive of the two interventions.[4]

· Placement Rates. Redirection and the Essex County Council SIB had success rates above 70% for program completers. The pro forma assumes MST/FFT reduce placement rates by only 50% to account for non-completers [B5].

· SIB Overhead Costs. SIBs are private placement transactions, so their transactions costs are generally not public information. A recent Utah PFS analysis estimates “project costs” for “evaluation, project manager, legal and audit expenses” at 12.5% of the services budget.[5] The pro forma assumes SIB overhead costs for legal, data collection, performance management, and evaluation are 25% of MST/FFT costs [B6].

The model also makes three simplifying assumptions that will require more detailed analysis in the future:

1. Timing. It assumes that the impacts of MST/FFT all happen in the year following program completion. That is, youth who receive treatment in Year 1 [B11] either succeed (avoid placement) [C12] or fail (go into placement) [C13] in Year 2, and so on. Hence, the investment proceeds are spent and services are provided in Years 1 through 5 to all 1,000 annual participants [B-F17], with savings (from successful participants) [C-G25] and “residual” placement costs (for unsuccessful participants) [C-G24] accruing in Years 2 through 6. Since MST/FFT treatment take less than six months, actual outcomes will be staggered and distributed across the population.

2. Outcomes. The model assumes that a binary result — a 50% reduction in placements — would satisfy contractual impact targets. It does not include more granular measures such as the number of days in placement.

3. Fixed Costs. The pro forma assumes that savings don’t depend on reductions in large fixed costs for brick-and-mortar facilities. In many states (such as Florida), most placement facilities are small structures like group homes and local detention centers that can be closed whenever commitments decrease even moderately. Other states have large prison-like institutions that can’t be taken offline without significant, long-term reductions in detainees. In those jurisdictions, incremental placement reductions might not produce proximate savings of the size and timing modeled in the pro forma.

Based on these assumptions, 1,000 placements per year (i.e., without MST/FFT) currently cost a prototypical state $100 million per year, or $500 million over five years [rows 14 and 22]. This assumption is reasonable: in 2014, 23 states spent more than $100 million per year on juvenile confinement; 9 other states spent more than $200 million annually; and New York spent more than $350 million.[6]

Providing MST/FFT to those same 1,000 families would (like Florida Redirection) cost $10 million per year, or $50 million over five years [row 17]. Adding in 25% overhead ($2.5 million per year) [row 18] and a 5% intermediary management fee ($625,000 per year) [row 19], would cost about $3.1 million per year, bringing the annual total cost to roughly $13.1 million, or almost $65.6 million over the life of the SIB [row 20]. This would be the SIB principal amount to be raised from investors, otherwise known as the “size” of the SIB.

Assuming conservatively that MST/FFT prevent placements in only half of the 1,000 total cases, then 500 youth per year, or 2,500 over five years, would still cost $100,000 each for placement. If so, “residual” placement costs would be $50 million annually, or $250 million for five years [row 24]. In that event, gross savings would be about $36.9 million per year or nearly $184.4 million over the life of the SIB [row 25], representing 37% of total current state spending. Beyond governmental savings, the SIB would produce large reductions in unnecessary placements and corresponding increases in unseparated families.

The pro forma allocates 5% of funds raised as an annual intermediary management fee [B7], 3% of gross savings as an annual intermediary success fee [B8] and 12.25% of gross savings to investors on the back end [B9]. Thus, total capital costs for the transaction would be $93.7 million [H31], comprising $65.6 million of principal [H27] plus $28.1 million in returns and fees [H29+H30]. Deducting these amounts from gross savings would provide net savings back to the state of $90.6 million [H35].

Thus, instead of spending half a billion dollars [H14] on 5,000 placements over five years [H11], states could cut placements in half [H13] with no up-front funding and keep almost $91 million [H35] in the treasury. Several factors drive these results: the sizable number of youth placed in custodial confinements; the wide cost differential between placements and MST/FFT; and adequate working capital for the high-fidelity implementation needed to achieve documented success rates and savings.

These same factors insulate Scale Finance from concerns that unintended consequences “can occur if service providers are able to choose themselves which beneficiaries are the recipients of the intervention, thereby ‘cherry-picking’ the easy cases and denying services to those most in need.”[7] Highly demanding programs like MST/FFT focus exclusively on stubborn problems with complex causes and devastating effects for tens of thousands of people nationwide. Intensive and comparatively expensive therapies like MST and FFT aren’t designed for low-risk youth, and governmental counterparties won’t pay investment premiums for services families don’t need. “In practical terms, juvenile justice systems will generally get more delinquency reduction benefits from their intervention dollars by focusing their most effective and costly interventions on higher risk juveniles and providing less intensive and costly interventions to the lower risk cases.”[8] For SIBs, the deepest end of need is where the greatest savings are. Scale Finance math doesn’t work without very large cost differentials between prevention and remediation, which “easy cases” simply don’t have.

If the SIB succeeds, investors would commit $65.6 million over five years [H15, H20] and receive back $88.2 million in Year 6 [H39]. The internal rate of return (IRR) for investors would be a healthy 10% per annum [H40]. This would be more than the “financial return of about 7%” that Sir Ronald Cohen believes social investment needs to become “the new venture capital,”[9] but the scope of the undertaking and corresponding risk are greater than the transactions he had in mind. This would be a reasonable reward for raising and managing all of the capital needed to expand proven social programs at their maximum feasible growth rates, and it would be sufficient to attract subsequent rounds of funding to finish the job.

One of the most important features of Scale Finance is that it fairly compensates the intermediary for organizing large, successful transactions, which is what they should be incentivized to do. Standard SIBs are too small with insufficient margins to cover the intermediary’s true costs; indeed, most SIBs require charitable subsidies and pro bono services to make the math work. (For example, Salt Lake County recently stated “philanthropic funds will be used to cover Project Costs — if possible.” The County estimates those costs will be $2 million for two PFS projects with a combined total cost of $11.5 million.[10]) Raising and managing the kind of money needed for systemic change requires a dedicated intermediary whose own funding doesn’t depend on grants and consulting contracts for non-transactional work, as standard SIBs do today.

The usual terms for private equity firms raising similar sums are 2% of the capital raised as an up-front management fee and 20% of the gross returns as “carried interest,” a kind of success fee. By comparison, the Scale Finance pro forma model raises the up-front percentage to 5% annually [B7] to sustain the intermediary organization and a 3% annual success fee on gross savings [B8]. The management fee translates to $625,000 per year [B-H18] (whether or not the SIB is successful) or $3.125 million over the life of the SIB [H19]. If the SIB succeeds, the intermediary success fee would be $1.1 million per year [C-G29], or $5.5 million over five years [H29]. This would be reasonable compensation for raising $65 million, preventing 2,500 placements, and saving the state more than $90 million.

Figure 2 shows why this is a better way to allocate $500 million over five years. Of that amount, 10% would be used to expand MST and FFT to all 5,000 families; only 50% would now go to residual placements; 13% would repay investor principal; 9% would cover SIB overhead, fees and financial returns; and 18% would be returned to the state. And 2,500 youth and their families would be spared custodial confinement, without financial risk to the state.

Figure 2. A Better Way to Allocate $500 Million Over Five Years

The public sector certainly could fund this directly and avoid the financing costs, but nothing has prevented them from doing so over the last thirty years. So we face a choice of (1) maintaining the intolerable status quo, (2) continuing to reduce the problem to a negligible extent or (3) inviting private capital markets to take on the problem at scale. The pro forma analysis shows that Scale Finance offers a small number of potentially game-changing investment opportunities that could plausibly provide risk-adjusted, market-return rates to commercial and institutional asset owners and fund managers with the expectation of large and enduring deal flow.

But, as JPMorgan Chase and the GIIN have shown, those mainstream investors are generally unaware of these profitable opportunities, and they are not capable of understanding them without help from the social investment community.

So the question now becomes: can the foundations, transaction developers and social investors who already understand SIBs and evidence-based programs bring these new opportunities to the attention of their mainstream colleagues? If so, they could help kick-start a new set of Scale Finance transactions that could reengineer public-private partnerships for the new fiscal realities we face.

[1] This article is adapted from “Scale Finance: Steven H. Goldberg. Scale Finance: Industrial-Strength Social Impact Bonds for Mainstream Investors.” Special Report. Center for Community Development Investments, Federal Reserve Bank of San Francisco. April 26, 2017.

Steven H. Goldberg is the founding principal of Caffeinated Capital LLC. Previously, he was the first Managing Director and General Counsel for Social Finance, Inc., where he helped bring Social Impact Bonds to the U.S. In 2009, he published Billions of Drops in Millions of Buckets: Why Philanthropy Doesn’t Advance Social Progress (Wiley), which explores the economics of scaling social innovation.

[2] “Impact Bond Global Database (Essex).” Social Finance Ltd. Accessed May 01, 2017.

[3] Petteruti, A., Schindler, M., & Ziedenberg, J. (2014). Sticker shock: calculating the full price tag for youth incarceration. Washington, DC: Justice Policy Institute.

[4] Aos, S. & Drake, E. (2013). Prison, police, and programs: Evidence-based options that reduce crime and save money (Doc. №13–11–1901). Olympia: Washington State Institute for Public Policy.

[5] “Evidence-Based Interventions for Criminal Justice Reform,” Salt Lake County, Utah. Accessed Dec. 20, 2016.

[6] Petteruti, infra, n. 226.

[7] Gustafsson-Wright, Emily, Sophie Gardiner, and Vidya Putcha. “The Potential and Limitations of Impact Bonds: Lessons from the First Five Years of Experience Worldwide.” Brookings. July 2015. Accessed September 28, 2016.

[8] Lipsey, Mark W., et al. “Improving the effectiveness of juvenile justice programs.” Washington, DC: Georgetown Center for Juvenile Justice Reform (2010).

[9] Cohen, Sir R., and William A. Sahlman. “Social Impact Investing Will Be the New Venture Capital.” HBR Blog Network 17 (2013).

[10] “Evidence-Based Interventions for Criminal Justice Reform,” Salt Lake County, Utah. Accessed Dec. 20, 2016.