Bringing Impact Bonds to the Global South

With 44 impact bonds currently in existence and an estimated 100 in development, the outcome-based funding mechanism for governments and development aid organizations has been gaining support across the United States, the UK, and other developed countries. In the coming years, funders will likely deploy hundreds of millions of dollars through these vehicles, potentially impacting government spending by the billions. Meanwhile, only one impact bond has launched in a developing country so far. Why? 

It seems fair to ask whether these instruments—with their complexity, stringent data requirements, and high transaction costs—are appropriate for the developing-world context.

We’ve been working with the South African (SA) government for the past two years on the use of outcome-based procurement, and although there are currently projects in development at the national, provincial and local levels, much remains to be done before the government launches any impact bonds. We believe we’ve uncovered some of reasons why take-up in the Global South has been slow, including some contextual realities that similar geographies will likely face as they attempt to finance services using impact-bond funding:

Prioritizing problems. In developing countries, often the scope and scale of the problems can be overwhelming and thus make prioritizing outcomes difficult. In South Africa, for example, we currently have a 35 percent unemployment rate (more than 8 million people). Job creation is the government’s highest priority, but tackling it involves supporting multiple levels of skill development, the creation of small businesses, and general economic growth. (Conversely, impact bonds may not be suitable for addressing smaller problems. In South Africa, for example, there are proven interventions to improve recycling in urban areas, but many politicians don’t see the value in using such a complex instrument to increase urban recycling rates.)

Multiple risk factors. The majority of potential beneficiaries in low-income communities face a variety of risks, including abject poverty, lack of health and education, and high unemployment. This makes targeting high-risk cohorts and addressing individual outcomes particularly complex. An early childhood development (ECD) impact bond in Utah, for example, aims to increase access to quality development support and improve educational outcomes for 600 pre-schoolers per year who are at risk of repeating a grade. But the challenge of ECD in the developing world goes far beyond classroom instruction. When the South African government designs programs for ECD, it must also contend with the reality that, in many areas of the country, nearly 25 percent of children are stunted (chronic under-nutrition has a lasting effect on neuronal development if not treated before two years old), mother-to-child HIV transmission rates are approximately 3.6 percent, and up to 16 percent of children in some areas suffer from fetal alcohol spectrum disorder. 

Service fragmentation. The majority of service providers in developing countries tend to be small NGOs, with the larger of them turning over about $1 million per year. Because NGOs are small and services are fragmented, deal sizes for impact bonds may be too small to attract private investment. 

Limited government funding. Most NGOs exist in survival mode when it comes to funding; annual revenue depends on erratic public-sector allocations, and the generosity of corporate, private, and international donors. While the reliability of impact bond cash flows could create some medium-term stability that would help stronger organizations grow and create economies of scale, it is difficult for governments to contemplate bearing the full cost of these services on an ongoing basis. HIV/AIDs treatment programs in South Africa, for example, are beginning to face this problem—most of the NGOs in this space have been almost completely subsidized by development aid for the past three decades. Even if government determines exactly how much it costs to fund these programs, it may not have the money in the long-term to cover this growing cost.

Sustaining momentum. Part of government delivering adequate services to the most vulnerable involves working out what it costs, and then generating sufficient political will to shift funding to early intervention or to more-effective services. Impact bonds may address the first requirement, but addressing the second requires a much longer, more complex conversation that will mean sustaining that momentum in the medium- to long-term. Again looking at ECD, even if impact bonds prove that early intervention has significant effects (such as an increase in cognitive ability), politicians won’t see the effect on job creation for another two decades. It will be tricky to avoid shifting away from early education and toward more-immediate issues like unemployment.

Lack of data. Especially in places where social security is inadequate and the public sector is in the process of expanding (in other words, places where access to and effectiveness of services is lacking), governments and organizations generally cannot base outcome costs on cost savings. Unfortunately in developing countries, neither input costs nor outcomes data is readily available to build business cases for early interventions or preventative programs. This has to do with the limited collection (and reliability) of government data, as well as the limited capacity of many NGOs that reflects the paucity of resources they can allocate to evaluation and data collection.

We don’t believe these constraints will stop impact bonds from flourishing in the developing context; however, bonds in these areas will likely be highly contextualized instruments that differ in both structure and process from their predecessors. They could be modified in a number of ways:

• To balance the risk investors and outcomes funders would shoulder in data-poor environments, increasing levels of outcomes payments could be phased into the implementation process. For example, instead of 100 percent outcome payments from the beginning, 40 percent of payments in the first year could be outcome-based and 60 percent fee-for-service, with that proportion increasing each year (or period) of the contract. 

• Impact bonds could be used to tackle problems holistically; for example, they could target increased cognitive and developmental improvements in young children through a combination of nutrition, reduction in maternal alcohol consumption, increased efficacy in parenting skills, and access to quality ECD. Part of addressing the multiple risk factors that target populations encounter may mean paying for a percentage improvement in an entire cohort rather than a defined outcome in an individual. In other words, government could pay for a percentage decrease in stunting in each area, against a baseline collected early in the impact bond’s lifecycle. This would also help deal with the baseline variability that often exists in neighboring communities in developing countries.

• By creating an outcomes fund that finances multiple interventions, investors could spread their money across a range of delivery models, diffusing risk and investing sufficient amounts to make it worth their while. 

• Developing-country government and NGO balance sheets contain substantial amounts of donor capital, so the notion of mixing public and private sector capital is not new. It is, however, relatively novel in the impact bond space generally, and particularly interesting to both parties in terms of the transparency and accountability a bond can create. It is also interesting in middle-income country settings, where traditional donor funding is beginning to decrease as large donors withdraw and look for new, innovative methods through which to distribute their remaining allocation. Finally, the guarantee of outcomes and the promise of private donor funding may generate political will to address difficult issues.

Though the difficulties in designing and implementing impact bonds are immense, we believe impact bonds that are contextualized to meet the specific needs of specific geographies hold great potential for the developing world. We look forward to seeing them develop, and the greater accountability, transparency, and effectiveness they promise for the Global South.

Read more stories by Aunnie Patton & Susan de Witt.