Impact bonds and their potential for use in early childhood interventions

Impact bonds have emerged as one of the most talked-about new tools to increase the volume and/or effectiveness of finance for social services. It is estimated that the global impact bond market will triple in size by the end of this decade. This article examines the applicability of impact bonds in the early childhood sector, the associated challenges and possible solutions.

Payment-by-results financing mechanisms, where payments for services are tied to outputs or outcomes, have become increasingly popular in recent years in countries’ domestic financing and in international development finance. These mechanisms are intended to create beneficial incentives, encourage performance management, and provide transparency and accountability. Impact bonds are a form of payment by results where non-state investors provide upfront capital to service providers and are repaid by an outcome funder, contingent on outcome achievement. In a social impact bond (SIB), a government actor is the outcome funder; SIBs are also referred to as pay- for-success (PFS) contracts in the USA and social benefit bonds (SBBs) in Australia. In a development impact bond (DIB), a third party partially or fully supplements government payments for outcomes.

Impact bonds are best suited to situations where there is some, but incomplete, evidence of effectiveness. For completely untested pilot interventions, investors are unlikely to be willing to bear the risk, and a grant may be a more suitable form of finance. Conversely, where the efficacy of service provision is already well established and the likelihood of not achieving outcomes is minimal, there is no need to ask investors to bear the risk of failure – in this case, nationwide provision can be funded without the need for an impact bond structure.

As of April 2016, there were 57 SIBs contracted in high-income countries, two DIBs contracted in middle-income countries, and none in low-income countries. Of the 57 SIBs in high-income countries, nine provide services to children in their early years, across four countries (USA, Canada, United Kingdom and Australia). Two support preschool services, six finance child welfare services related to keeping families together and adoption, and one supports nurse home visiting.

While not all contracts are signed, in March of this year, the Departments of Social Development and Health of the Western Cape province of South Africa committed 25 million Rand (EUR 1.53 million) in outcome funding for three SIBs for maternal and early childhood outcomes (Silicon Cape Initiative, 2016). Outcomes include:

improved antenatal care, prevention of mother to child transmission of HIV, exclusive breastfeeding, a reduction in growth stunting, and improved cognitive, language and motor development.
(Bertha Centre for Social Innovation, 2016)

Another impact bond for ECD is currently being developed in the state of Rajasthan in India -this one would pay private health clinics for reproductive, maternal, and child health outcomes, targeting individuals in the second and third income quintiles. Finally, Grand Challenges Canada, Social Finance, and the MaRS Centre for Impact Investing are working in Cameroon to develop an impact bond to finance Kangaroo Mother Care (KMC) – an intervention known to save and improve the lives of low-birthweight infants.

Even where early childhood programmes promise future benefits to individuals, society and the economy, they often remain unfunded because it is difficult for government agencies to prioritise funding for interventions with delayed benefits. The payment-by-results component of an impact bond could enable government investment in these high-impact interventions by allowing them to pay only if and when outcomes are achieved. Further, under other payment- by-results contracts, service providers lack the necessary upfront capital and capacity for risk absorption. Impact bonds have the potential to address this problem: private investors provide the upfront capital, accepting the risk of
the future benefits not being achieved. Crucially, the involvement of non-state investors in an impact bond can highlight the impressive case for investment in early childhood development (ECD) and encourage increased government commitments of outcome funding. Non-state investors may also increase performance management, or help reorganise a government system of data sharing or provision, beyond what other payment-by-result mechanisms may be able to accomplish. Collaboration across stakeholders – a necessary component of impact bonds – has the potential to align interests and create win-win situations for investors, outcome funders, and programme beneficiaries alike.

By producing evidence of outcome achievement and fostering innovation, experimentation and adaptive learning in service delivery, impact bonds could help identify effective early childhood programmes. As noted above, impact bonds make most sense when there are gaps in knowledge about what works and a desire to learn more, but government is unable or unwilling to take the risk of funding an unproven programme without external support. This is often the case with ECD: despite its high potential impacts, the evidence base from developing countries is still relatively thin. A recent systematic review by the World Bank’s Independent Evaluation Group found only 55 applicable rigorous evaluations, based on only 25 projects (Tanner et al., 2015). Promising interventions which need more robust evidence include micronutrient supplementation, water and sanitation interventions, delivery and ante- and post-natal-related interventions, and disease treatment.

Along with the possibility of high returns, the high participation rate of non- state actors in ECD makes it a promising sector for impact bonds. Globally, in 2012, around 31% of pre-primary students were estimated to be enrolled in private institutions, either for-profit or non-profit, while 12.8% of primary students were enrolled in private institutions in the same year (UNESCO, 2015). The fewer entrenched interests in ECD may allow for more experimentation.

Nonetheless, some challenges are associated with applying impact bonds in the ECD sector. Any payment-by-result financing mechanism requires meaningful outcomes that are measureable within a timeframe that is reasonable to the outcome funder (and investors, in the case of an impact bond). The delay between ECD interventions and later-life results may prove an impediment in some cases.

One solution could be to identify outcomes that are measurable within a reasonable timeframe and serve as proxies for long-term benefits to individuals, society, or the economy. For health and early stimulation programmes, these short-term proxies for long-term benefits could include measures of language development, socio-emotional development, schooling outcomes and child survival (Tanner et al., 2015).

Another challenge could be inability to quantify outcomes or assign attribution of impacts to specific interventions, in circumstances where robust evaluation is not possible or there is potential for a multitude of confounding factors that may influence outcomes. Particularly challenging interventions for quantification of impact may be birth registration or child protection interventions. In these cases, simple outputs either with intrinsic value or an association with longer-term outcomes may be more appropriate.

The lack of appropriate open-source tools to measure early childhood outcomes in low- and middle-income settings may also be a hindrance to applying impact bonds to some outcomes. As evidence increases and more measurement tools become available, however, this will become less of an obstacle. Again, an alternative solution – at least in an intermediate phase – would be to focus on inputs that are known to strongly correlate with development outcomes. In centre-based care (daycare and preschool) interventions, for example, these could include measures of process quality, such as the interaction between teachers and students.

The multi-sectoral nature of early childhood could pose a challenge in selecting outcomes for impact bonds, as outcomes could be linked to multiple ministries. However, this could provide an opportunity to improve coordination between those agencies through a common focus on outcome achievement. Experience from impact bonds in developed countries shows that the coordination of joint data systems resulting from impact bonds has been well worth the effort.

Adequate service provider capacity could be a particular challenge for ECD impact bonds in low- and middle-income countries.  Impact bonds can only be used in countries where legal conditions exist that allow the mechanism to operate.  In particular, it is often challenging or impossible for a government to appropriate funding for contingent outcome payments. Impact bonds should not be implemented where risks of corruption – in procurement, outcome payment design, or evaluation – cannot be mitigated. Nor would it make sense where it is impossible to ensure the outcome funder’s ability to repay investors, as with a government with a poor credit rating. Legislation supporting public–private partnerships and improving the tax status of impact investing may facilitate the development of impact bonds. Finally, impact bonds are likely be much easier to implement in countries that have demonstrated political commitment to the sector – where there is already some interest at the country, regional, or municipal level to expand early childhood services, but insufficient political or constituent support for adequate budget allocation.

The Sustainable Development Goals, with their associated targets linked to measurable outcomes, present an opportunity to demonstrate a commitment to invest in future generations. Leveraging upfront funding, focusing on outcomes through adaptive learning and testing new ways to deliver early childhood interventions more effectively are all means of achieving the ECD-related goals.

Despite the hype around new financing mechanisms, the keys to creating high- quality, locally appropriate programmes remains simple – real-time collection of outcome data, the freedom to fail, and the flexibility to adjust course if necessary. In some circumstances social service provision based on outcomes and adaptive learning may require mechanisms like impact bonds or other payment-by-results mechanisms, while in other circumstances it may not.

With impact bonds still in a nascent stage, it will be crucial for the ECD community – with deep knowledge and expertise in ECD financing, service delivery, programme evaluation and assessment – to be engaged in the discussions about how they can meet the needs of the world’s youngest and most disadvantaged populations.

This article summarises the findings of the authors’ recent publication Using Impact Bonds to Achieve Early Childhood Development Outcomes in Low- and Middle-Income Countries. Full report can be found here.

References can be found in the PDF version of this article.