Identifying Education Outcomes for Social Impact Bonds for Early Childhood

In the past few years, the social impact bond (SIB) has gained significant momentum around the world. This innovative financing mechanism, called pay for success (PFS) in the United States, harnesses private capital to fund social services. It then engages government to repay those investors only if agreed-upon outcomes are achieved. A variation of this is the development impact bond (DIB); in a DIB, the entity that repays the investor (or “outcome funder”) is a third party, such as an aid agency or a foundation instead of the government. The term “development” refers to the application of a SIB to a developing country context. Globally today,  SIBs and DIBs are being used for nearly 40 programs in various sectors, including for preschool education in Utah and Illinois, and for girls’ education in India.

While still fairly nascent, these innovative funding mechanisms have great potential to drive private capital toward funding social services, to expand programming for those most in need, and to improve the lives of many. At the heart of a SIB contract is a set of agreed-upon metrics on which the repayment is based.

Here at Brookings, in a recent event held jointly with the Institute for Child Success and the Ounce of Prevention Fund, we sought to identify what metrics would best be used for such social impact bond contracts when applied toward early childhood development (ECD). ECD is well known to have an impact on the future health, labor market participation, and criminal behavior of individuals, but in this workshop we focused on outcomes in education.

Our goal was to identify education outcomes that were (1) meaningful, (2) measurable, and (3) would respond in a reasonable timeframe to be applicable to this type of financing strategy. Our conversation helped clarify just exactly what these concepts entail.

  1. Meaning. Meaningful education outcomes should minimally be empirically predictive of the life trajectory of a child. Furthermore, the continuous nature of a child’s development, occurring beyond the scope of a single intervention, suggests that choosing a variety of outcome indicators across a longer timeframe would be beneficial. This could also build incentives for continuity and complementarity in investment in ECD programs throughout the early years.

    While the experts at our workshop agreed that interventions should by no means be limited to the outcomes that are interesting to those repaying the investors, when choosing the metrics for repayment in a social impact bond, ultimately the outcome must be attractive to them. The experts also brought to light some ways to use outcomes metrics to meaningfully target the most disadvantaged groups in a population by varying potential metrics or payment thresholds for those metrics by subgroups in the population, thus incentivizing services to those populations.

    Learning from the Utah and Chicago deals already underway, the experts discussed that, in part, the meaningful nature of the measures was in the composite nature of them. In each of these deals, the special education referral rate is one of the outcome measures chosen. The experts noted that, when administered properly, special education referrals are actually the culmination of a number of evaluations by teachers and other professionals, in addition to being correlated to other later life outcomes. In addition, the Chicago deal was set up with multiple trigger metrics across an eighteen year timeframe. What this means is that sustained support and ongoing evaluation is actually built into the structure of the deal. 

  2. Measurable. In order to make an outcome measurable, it is best to be simple so those structuring the deal can easily have or develop the measurement tools and also install systems to accurately and consistently use them to gather the required metrics. Therefore, the use of administrative data, where available, makes good sense as a source for outcome metrics.

    Nevertheless, it is important to avoid perverse incentives by ensuring the separation of monitoring systems in relation to the payment metrics and those required by the delivery systems. For example, in both PFS deals in the United States, payment metrics are based on regularly collected data across the entire population (i.e., standardized test scores and special education referral rates). In Utah, the payment metrics are based on data not readily available to those delivering the services, such that they are held accountable by the system to deliver high-quality services to the entire population. In both cases, the setup of the PFS relies on regularly collected data without causing any additional burden or requiring additional assessment of children receiving funded services.

    For social impact bonds in developing countries, efforts like the Sustainable Development Goals of the post-2015 agenda could lead to global measurement tools for child development that could be used.

Responding within a reasonable timeframe. It will be critical to gaining the interest of investors and outcome funders that they see the results of their investment within a reasonable timeframe. Private, commercial, and impact investors generally have a short time period in which they are willing to wait for financial returns. Our discussion at Brookings, however, noted that it is important to look at these short-term outcomes not as the end goal for an early childhood intervention, but rather as proxies for the longer-term outcomes that the programs aim for.

For example, in both Utah and Chicago, the PFS payment system begins providing returns to investors based on early elementary school metrics. Avoiding special education courses in third grade (a metric in the case of Utah) and obtaining high third grade reading scores (a metric in Chicago) are not the end goal for children; rather, these metrics provide a meaningful, measurable indicator that can in the short-term serve as a proxy for the overall outcomes we hope that children achieve later on.

That said, experts also discussed the difficulty in using these short-term measures as proxies, given that they do not adequately capture the overall desired effects of high-quality early learning programs.  

Two other parts of the conversation proved interesting as well. First, the narrative in the SIB field around this topic, specifically related to early childhood interventions and the comprehensive outcomes that we have historically used to measure their success, has moved away from identifying outcomes that are strictly monetizable. This means that, although an outcome may not yield budgetary savings as identified by a government agency line item, it may be relevant for broader societal purposes or be thought of as resulting in cost avoidance down the line for interventions that are no longer needed.

Social and development impact bonds, or pay for success financing, may shift the thinking of policymakers toward outcomes that can lead to improved lives. At this point, the long-term policy impact of PFS is unknown, but for it to have a positive impact, the outcomes on which pay for success is focused must be the right ones. Moreover, projects must be implemented carefully, and must be well thought out to avoid pitfalls that can lead to perverse incentives or metrics that drive toward short-term gains without the vision for long-term impact.