This is the first of two blogs sharing insights from a learning group on innovative finance in education, co-facilitated by the Education Finance Network and the Education Outcomes Fund. The group has brought together experts from 34 organizations in five discussion sessions over a period of nine months for an earnest exchange on opportunities and challenges around using innovative finance mechanisms in education.
Outcomes-Based Financing (OBF), a funding model linking financial resources to the achievement of predefined outcomes, has gained momentum within initiatives aimed at advancing the Sustainable Development Goals (SDGs), including in education, in recent years.
To ensure the success of the OBF model, it's crucial to establish solid metrics and accurately measure and verify outcomes. Selecting the right metrics and verification methods is key to incentivizing the desired results for the target population. Reliable data on programme outcomes and costs also provides valuable evidence for government agencies to identify cost-effective interventions and for service providers to adjust their programmes and optimize their budgets.
Challenges around measuring outcomes
Yet, defining what metrics to use and how to verify them presents challenges, with some being common in OBF programmes across SDGs, and others unique to the education sector. Measuring outcomes is more difficult than measuring inputs or outputs typical in traditional grant-funded programmes, partly because it can be tricky to identify suitable metrics and partly because of the complexity and cost of data collection tools and evaluation methods.
In education, an added challenge emerges in effectively measuring complex outcomes encompassing educational quality, learning attainment, and child development, some of which lack straightforward metrics or reliable proxies. It can also be costly and complex to measure the same learners over time throughout the programme and after the programme ends.
These challenges were discussed by experts in OBF in education in a learning group co-facilitated by the Education Finance Network and the Education Outcomes Fund. Over a series of meetings, group members shared their experiences and insights in OBF in skills for employment, foundational learning, and early childhood care and education (ECCE). In this blog, we share some of the key learnings from the group’s discussions around one particular topic: how to measure outcomes in OBF programmes in education.
Emergence of OBF in education
In 2015, the Educate Girls Development Impact Bond was launched in India, marking the launch of the first impact bond in education in a low- or middle-income country (LMIC). Since then, 15 other education impact bonds have been launched in LMICs, including Social Impact Bonds (SIBs), where at least one of the outcome payers is a domestic government, and Development Impact Bonds (DIBs), where the outcome payer is a donor. Of these 16 impact bonds, seven focused on skills for employment, eight on foundational learning, and one on ECCE, together reaching over 367,000 beneficiaries and raising over USD 22 million in capital.
In the last two years, the launch of two outcomes funds, the Ghana Education Outcomes Programme and the Sierra Leone Education Innovation Challenge, and the LiftEd DIB in India, signals a growing trend of larger-scale OBF programmes, with beneficiary numbers reaching into the tens and hundreds of thousands, and, in the case of LiftEd, millions of children.
Other emerging OBF tools in education include Social Impact Incentives and Impact-Linked Loans, which have been piloted by Impact-Linked Fund for Education since 2021, and Income Share Agreements, which have been introduced in recent years in Sub-Saharan Africa by Chancen International, and in Asia by Waiser.
What results are measured and financially incentivized
In skills for employment OBF programmes, defining which outcomes to measure to determine success and provide financial incentives is usually straightforward, with job placement and job retention the most commonly used metrics. In economies with large informal markets, where OBF programmes may focus on supporting entrepreneurship and micro-enterprises instead of employment in the formal sector, increase in income or household consumption has been considered as an alternative metric.
In foundational learning OBF programmes, defining metrics is often more challenging. This is because foundational learning encompasses multifaceted aspects such as literacy, numeracy, critical thinking, problem-solving, and socio-emotional skills, which are not easily quantifiable. Due to this complexity, almost all foundational learning OBF programmes use metrics related to access (enrollment and retention) and literacy and numeracy assessment scores to measure success. Metrics related to critical thinking, problem-solving, social and emotional learning are notably absent.
In ECCE OBF programmes, defining metrics is even more complicated. ECCE outcomes encompass a wide range of child developmental domains, including cognitive, social-emotional, physical, and language development, which require comprehensive assessment tools and methodologies to measure. Other ECCE outcomes, such as retention and learning in primary education or prevention of special education needs, only materialize many years after the intervention. ECCE OBF programmes in LMICs instead tend to measure system-level outputs (e.g., monitoring and evaluation systems, standards, and curriculum) and expanded access to services, with child development outcomes often taking a back seat.
Measuring and rewarding systems-change outcomes
An additional complexity arises around measuring and rewarding systems-change outcomes, which are often an inherent goal of OBF programmes. For example, in the LiftEd DIB in India, outcome metrics were initially centered around learning outcomes, reflecting the ultimate goal of the programme. However, this DIB also aims to achieve wider and long-lasting change in the education system as a whole. Its interventions are aimed at outcomes that would have a ripple effect beyond the DIB - for example, improving the skills of teachers and district education officers, ensuring adoption of specific pedagogical approaches, improving the quality of classroom observations made by district officers, and so on. This led to the evolution of the outcome metrics to include 'systemic shift indicators,' which are core programme outcomes and targets that form part of the payment mechanism, along with student learning outcomes.
Experiences of measuring and attributing impact
While determining appropriate metrics can be complicated, the process of verifying them and attributing impact can be equally challenging and may present a significant cost and administrative burden for funders and service providers. Therefore, selecting the verification methodology requires careful consideration, balancing the need for rigor against the associated costs and logistical complexities. Concerns have been raised regarding OBF programmes, across sectors, allocating a significant portion of funds to evaluation, potentially at the expense of interventions themselves.
When determining the right balance of cost vs rigor, the learning group discussed several factors to consider. If the purpose of verification is solely to verify outcomes and not attribute them to the intervention, less costly methods such as pre- and post-tests may be considered. However, if the goal is to measure and attribute programme impact or to compare cost-effectiveness of different programme interventions, more rigorous (and more costly) methods may be necessary. Noting that different OBF programmes have found different balance points in this multi-factor equation, the following examples illustrate some of the factors considered in the choice of evaluation methodology.
Focus on generating learning
In the Palestine Finance for Jobs DIB, the first impact bond financed by the World Bank, the programme aimed to test a new approach to funding, designing, and managing skills training, and to demonstrate scalability to tackle larger employment needs. The decision to measure employment outcomes using programme administrative data, alongside a rigorous process evaluation, was guided by the need for a thorough evaluation capable of yielding valuable insights, while avoiding an extensive focus on a randomized control trial (RCT), which might shift attention away from the primary intervention goals.
Attributing impact in large-scale programmes
In the Quality Education India DIB, the quasi-experimental ‘differences-in-differences’ (DID) approach was used, which compared changes in learning outcomes over time between treatment and control schools without randomizing school selection. By examining the differences in outcomes before and after the intervention in both groups, DID helps isolate the effect of the intervention from other factors that may influence outcomes. This can be particularly useful in large-scale programmes where external time-variant factors such as changes in economic conditions or new education policies can influence outcomes and make them more challenging to attribute solely to the intervention.
Tools for measuring change in learning outcomes
When measuring change in learning outcomes in foundational learning OBF programmes, options include using standardized government tests; international tools like the Annual Status of Education Report (ASER) and the Early Grade Reading Assessment (EGRA) and Early Grade Mathematics Assessment (EGMA); or developing bespoke tools tailored to the programme.
In the Quality Education India DIB, because of its focus on generating evidence, a standardized government test was chosen as it would enable comparison across the education system. This would help demonstrate the programme’s effectiveness to the government and thus support efforts to advocate for scale-up.
In the Sierra Leone Education Innovation Challenge (SLEIC), a custom test was developed since reliable national tests were unavailable and none of the international tests were deemed suitable. ASER was excluded due to its focus on fundamental skills, which didn't align with the programme's ambitious objectives. The EGRA and EGMA tests were also unsuitable for two reasons: they are only applicable up to grade 3, while SLEIC targets up to grade 6, and they produce significant floor and ceiling effects (where a large proportion of participants achieve the lowest or highest possible score), limiting their ability to accurately measure progress and differentiate between students' abilities.
The members of the innovative finance in education learning group are united in their commitment to better understand the effectiveness of OBF in education. The challenge of measurement is a common hurdle in programmes using the OBF model, and the choices made around metrics and verification methods in each programme reflect their unique priorities and constraints.
Because there is no one-size-fits-all solution or easily codifiable knowledge to guide these choices, genuine exchanges like those had within this learning group are so valuable. It is our hope that, by sharing insights into these exchanges, we can help make similar choices less difficult and better informed in future OBF programmes, and thereby contribute to better quality education for more children and young people across the world.
We wish to express our gratitude to the individuals who have contributed their valuable insights to the learning group and this blog.
Kommentare