Key Messages

The World Bank has helped build awareness, shared knowledge, and convened global stakeholders around a commitment to quality basic education. Its data and analytics and the World Development Report 2018 have drawn attention to the learning crisis and helped define learning poverty, which has encouraged stakeholder buy-in.

The World Bank is well positioned to help address the learning crisis. Its relationship with governments and its role as the largest external education funder mean it can, with committed clients, help reform key aspects of the education system needed for inclusive learning. Overall, support could be more strategically focused and institutional incentives better aligned to deliver learning outcomes and take advantage of opportunities to reform education systems. Better monitoring—with enhanced disaggregation of data—would provide a feedback loop for adaptive management that is currently missing.

World Bank financing for basic education delivers inputs into education systems and typically tracks outputs, with few operations assessing changes in systems, teaching, and learning. For example, 48 out of 77 projects with project development objectives that address improving learning have learning outcome data, and only 22 out of 188 operations with on-the-job training for teachers tracked the impact of the training on teachers’ practices.

Country analytics focus on specific aspects of basic education systems (that is, what is not working rather than why systems are failing children). This makes it difficult to influence and implement changes related to the intertwined challenges that result in learning poverty.

Reference to marginalized groups and the level of analytic focus on broader equity-related issues increased over the evaluation period in Country Partnership Frameworks and Systematic Country Diagnostics, as well as in Project Appraisal Documents, where targeting of such groups has also increased; however, monitoring and disaggregated reporting of results predominantly focus on gender and not on other groups.

The World Bank responded at the country and global levels to the learning losses arising from the COVID-19 pandemic, with significant additional funding and strengthened partnerships, new working coalitions, and alliances, which are of strategic value. World Bank support was a broad-based emergency response focused on remote learning and on reopening schools.

Tackling Low Levels of Learning Outcomes—A Complex and Costly Development Challenge

Learning losses associated with school shutdowns implemented in response to the COVID-19 pandemic reinforced the international community’s attention to the long-standing issue of the learning crisis phenomenon framed by the World Development Report (WDR) 2018 (World Bank 2018b). In 2019, before the COVID-19 pandemic, learning poverty—the share of children younger than 10 years of age who have not achieved minimum reading proficiency, as adjusted by the proportion of children who are out of school—was at 91 percent in low-income countries, compared with 9 percent in high-income countries (World Bank, UNESCO, et al. 2022). Since the pandemic, the situation has worsened.

Among the children most failed by their education systems are those already disadvantaged by poverty, location, ethnicity, gender, disability, and other factors. The capacity of those children to engage in and benefit from basic education is often stymied by factors that also affect the population at large, such as lack of investment in infrastructure and a shortage of qualified teachers. However, these children also often face a combination of barriers: distance from school, physical access to buildings and related facilities, the availability of adaptive equipment, language barriers, and so on.

Learning for all is a much more difficult and expensive pursuit than access for all. The latter, with an emphasis on more infrastructure and more teachers, is a simpler proposition (although literature reports may be open to clientelism, patronage, and corruption [Grindle 2004; Kingdon et al. 2014]). The former requires greater discipline within the education system and its service delivery, potential loss of power for actors, and greater levels of accountability throughout the system. Improving learning for all involves enhanced teacher quality, alignment of curriculum and textbooks with local contexts and local language and culture, and investment in education equity to ensure equality of opportunity (rather than equal access) for diverse student cohorts. The Global Education Monitoring Report estimates an annual financing gap of $97 billion during 2023–30 in 79 low-income and lower-middle-income countries to achieve the Sustainable Development Goal 4 targets established in 2012, which focus on the reduction of inequality in access to and quality of education (UNESCO 2023). This gap reflects the existing scale of the challenge, which is expected to grow as a result of the projected demographic profile of many of the worst-affected countries.

About This Evaluation

This evaluation assesses the World Bank’s contribution to improving learning outcomes in basic education—defined as primary and lower secondary education—over the 2012–22 decade. It pays particular attention to the extent to which the World Bank has adopted a systems approach to its support for basic education as advocated in Learning for All: Investing in People’s Knowledge and Skills to Promote Development—World Bank Group Education Strategy 2020 and as reinforced since the publication of the WDR 2018 (World Bank 2011, 2018b). It is designed to identify lessons and present recommendations to inform any future education sector strategy. A conceptual framework informed the design of the methodological approaches and data analysis employed. The evaluation draws on portfolio and document analyses, interviews, country case studies, literature, and secondary data analysis.

The starting point for this assessment was to develop and test a conceptual framework that captures the complexity of basic education systems and the system characteristics required to deliver quality education for all. The framework was aligned with the WDR 2018 and was used to assess the extent to which the World Bank response to the learning crisis—at global, regional, and country levels—can be characterized as a systems-based approach (that is, to what extent the World Bank’s approach has taken account of material and intangible elements of basic education systems and the relationships between them, the structuring of basic education systems and subsystems, the systems’ formal and informal functions, and feedback loops that influence behavior within systems). The framework also identifies points at which the World Bank can, and typically does, intervene in basic education systems—strategizing, planning, implementing, and monitoring and evaluation—to provide support through policy dialogue, knowledge, financing, and other means.

The World Bank Has Highlighted the Crisis in Learning

The World Bank’s knowledge and analytic work, partnerships, and global initiatives have contributed to global knowledge and awareness and have encouraged action by country clients to support improvements in system alignment and capacity, teaching, and measurement of learning. With the Commitment to Action, launched at the United Nations Secretary-General’s September 2022 Transforming Education Summit, the World Bank and partners advanced support for foundational learning. Since then, partners of the Global Coalition for Foundational Learning have encouraged more countries to sign the Commitment to Action.1 World Bank vice presidents have played a key role by encouraging ministers of education and finance in client countries to improve learning for all. For example, a high-level meeting in Latin America in March 2023 convened many regional partners and was followed by collaboration with the Inter-American Dialogue, the Inter-American Development Bank, and other partners to raise awareness of the region’s learning crisis (World Bank 2023a). In addition, the Laboratorio Latinoamericano de Evaluación de la Calidad de la Educación at the United Nations Educational, Scientific, and Cultural Organization ensured regular data collection for calculating learning poverty through regional standardized tests (Colombia, Ministry of Education 2023).

The World Bank’s high-quality data and analytics and the WDR 2018 have drawn attention to the learning crisis and addressed improvement in quality and learning. The WDR 2018 renewed attention to the political barriers to progress and the need for more effort in client countries to measure learning and help identify system failures. Valuable global and regional contributions have supported assessing levels of education policy development (Systems Approach for Better Education Results [SABER]) and producing comprehensive regional reports (Facing Forward: Schooling for Learning in Africa [Bashir et al. 2018]). The World Bank provided dialogue and knowledge dissemination for clients who expressed a commitment to implement policies and programs to reform teacher status, evaluation, and remuneration informed by evidence and tailored to the local context (Great Teachers: How to Raise Student Learning in Latin America and the Caribbean [Bruns and Luque 2014]). The evaluation observes that influence with country clients is enhanced where there is support for follow-up and the ability to link global and regional analytics to country programs. However, a supply-demand tension is often found with the more specific, country-level utility of global public goods and global knowledge. Increasing global public goods is difficult, particularly when governments are asked to finance data collection.

The World Bank is the largest source of external financing for the education sector in low- and middle-income countries, although funding for the sector—and for basic education—is a small part of its overall lending. Commitments to the education sector in fiscal year 2022, from preprimary to tertiary levels, were just under 5 percent of total commitments. Primary education, the core of basic education, attracted about 1 percent of World Bank lending, whereas only about 1 percent of all official development assistance goes to education at any level, with less than half of that going to basic education.

The portfolio for this evaluation consists of 236 basic education operations (or portions of them) approved during the evaluation period (fiscal years 2012–22), with a total commitment value of $25 billion. The portfolio, by number of projects, is concentrated in the Africa Region (44 percent), with 16 percent of projects each in the Latin America and the Caribbean and South Asia Regions. Eighty operations in the portfolio (34 percent) address the effects of COVID-19 on basic education systems. The most regularly supported inputs across the entire portfolio are government-level management, in-service teacher training, and school management, which are supported in 88 percent, 80 percent, and 75 percent of projects, respectively. This level of concentration across the portfolio and similar patterns in case studies suggest a lack of nuanced response to the differing contexts and basic education systems within which the World Bank works, without a similar focus on the fundamental causes of context-specific education failure. Our analysis suggests that the types of input likely to be supported by the World Bank remained consistent across the evaluation period and across the primary instrument types—that is, investment project financing and Programs-for-Results. The success of operations is typically measured in outputs: teachers trained, textbooks in classrooms, and schools built. Although these are valid indicators for education access, they are ineffective for measuring changes to systems, teaching, and learning.

Measuring Learning Outcomes Is Critical to Informing Policy Makers

Learning poverty is an easily understood concept that has gained global stakeholder buy-in. The World Bank and multiple partners have supported ambitious targets to motivate global and country stakeholders toward collective action and alignment on a single target and message. Reducing learning poverty was added to the World Bank’s corporate targets at the 2023 Annual Meetings of the World Bank Group and the International Monetary Fund, replacing the previous indicator of human capital (measured as students reached). The increased attention to outcomes is welcome; however, without further attention by the World Bank and partners to address the lack of underlying data (to calculate learning poverty), the indicator will be unable to fill its critical global and corporate monitoring function. In every Region, learning poverty data are lacking for two or more countries. This indicator is represented in 58 out of the 91 countries in the portfolio.

Progress has been made in supporting systems to measure learning from national assessments in 44 countries and from subnational assessments in 6 countries, which will provide data to assess progress on Sustainable Development Goal target 4.1. Data remain particularly scarce at the lower secondary level, with not quite a third of portfolio countries (27 out of 91) reporting data on the proportion of students in lower secondary education achieving at least a minimum proficiency level in reading, compared with half (47 out of 91) of them providing data for end of primary reading. All leading education systems have a learning assessment function in place—a prerequisite to focusing on improvements in learning. Nearly two-thirds of portfolio operations supported learning assessment, learning surveys, capacity building, and dissemination activities. In most cases, the assessments covered grades 6 and lower (92 percent) and evaluated reading (98 percent) and mathematics (84 percent). The SABER Learning Assessment Platform and the Russia Education Aid for Development program have provided numerous tools, reports, and global public goods designed to improve global knowledge. For example, SABER–Student Assessment tools have been applied in about 60 countries, resulting in two regional reports and seven country reports, as well as case studies of lessons learned from learning assessments and learning standards. In addition, the Russia Education Aid for Development program and the SABER Learning Assessment Platform provide technical assistance to a few countries to improve student learning assessments. In recent years, this support has been more strategically focused on measurement in primary grades.

Nevertheless, the World Bank needs to focus more consistently on learning outcomes at both the country and the project levels. Monitoring and evaluation of improved learning outcomes is specified in one-third of project development objectives in the basic education portfolio. The focus on improved learning outcomes in project development objectives is higher in Africa (43 percent of projects). Improving the quality of education is an objective in 62 percent of operations that typically focus on improvements to the learning environment (such as enhanced infrastructure, textbooks, or teachers trained)—often prerequisite conditions to improving learning. Of the 77 projects with project development objectives that improve learning, 48 have outcome indicators. Analysis of Independent Evaluation Group ratings shows that operations with learning indicators receive lower ratings than those without, although this is significant only at the 0.1 level. The lower frequency in measurement of learning compared with outputs of the learning environment is due, in part, to internal incentives that do not encourage country teams or task team leaders to set more ambitious objectives and indicators, which are more challenging to achieve than access-related objectives. The lack of adequate measurement data was also a factor in lower project ratings (Bedasso and Sandefur 2024). The same authors note that task team leaders may account for a significant portion of the variations in type of project activity, suggesting that institutional incentives are needed to ensure that task team leaders support the World Bank’s strategic aim—learning for all—but will not feel pressure to deliver uniformly high ratings.

The measurement of learning outcomes at the country program level is limited. The efficacy of education management information systems remains weak in many client countries despite World Bank operational support. Due attention is not always paid to interactions and interdependencies between the technical (for example, hardware and software) and human components (that is, capacity) within basic education systems that are necessary to support a joined-up, functioning model. The World Bank’s SABER–Education Management Information Systems assessment tool is not “internalized” within the Education Global Practice’s current analytic support for systems strengthening in client countries.

Capacity Development Is Needed across All Levels of Basic Education Systems

The World Bank has put less emphasis on supporting lower levels of basic education systems, such as at the provincial and the district levels. This has negative implications for policy implementation because success depends on actors in the lower levels of the system who are responsible for ensuring fidelity to policy reform and implementation. Inputs at lower levels of basic education systems, where they exist, are often designed to ensure effective World Bank project delivery through, for example, training in financial management and procurement without transfer of capacity to the broader system outside of the project boundaries.

A High-Quality Teaching Career Framework Is an Essential Part of the System

World Bank analysis recognizes that quality teaching is essential to improving learning outcomes and that teaching quality is affected by many factors, including funding, recruitment, monitoring, and motivation. The primary response to the challenges associated with quality teaching is support for on-the-job training (in 188 projects). Such training can be an important input, particularly where there is a cadre of well-qualified teachers and where the training builds on existing knowledge, expertise, and competence. Yet few operations assess the efficacy of training. Only 22 out of 118 operations in the portfolio that supported on-the-job training systematically tracked the impact of that training on teachers’ practices. Typically, the World Bank monitors participation in training. Projects that provide follow-up support to enhance the effectiveness of training do not generally assess the efficacy of the follow-up support provided (assessed by 38 out of 110 operations). Other challenges, such as recruitment, initial training, placement, and retention, were addressed to a lesser extent. There is also a need to expand the evidence base of the efficacy of the World Bank’s Teach and Coach tools to improve teaching practices and student learning. What is needed is consistent examination of whether the interventions and inputs the World Bank finances are having a positive impact on systems, teaching, and learning.

Equity and Inclusion Are a Concern in Many Systems

Despite increased attention to equity and inclusion in documentation, few projects produce disaggregated data related to equity other than for gender, making it difficult to assess the extent to which target groups have been reached or had their needs addressed. World Bank operations planned to support gender-targeted activities, which almost exclusively target girls, account for 67 percent of all projects, with an even higher rate among projects in Africa. Equity issues other than gender are targeted to a lesser extent: 51 percent of projects target children in rural, remote, or nomadic areas; 39 percent support activities related to the educational needs of children with disabilities. Almost all projects addressing gender disparity have indicators with gender disaggregation; only about 30 percent of projects that include other groups in their targeting, such as children with disabilities or those who are rural residents or out of school, have indicators that capture the progress of those target groups.

Global-level advisory services and analytics document inequities in learning for various marginalized groups. Clients would benefit from support in and knowledge about the additional challenges faced by children with disabilities and the learning adaptations they may need for the delivery of education, as this has received modest attention. To maximize its contribution to addressing the learning crisis, the World Bank could provide context-specific evidence and promote equity considerations being fully built into education system planning, implementation, and monitoring.

Key Evaluation Conclusions

The evaluation concludes that where the World Bank has a willing and committed partner, it has been able to better focus on key policy reforms and to lay foundations for a learning-oriented system. Where countries are in tune with, or at least in deliberate, sustained pursuit of certain conditions—teaching and career progression, measurement of learning, financing to achieve equity of learning, and meritocracy in hiring at all levels of the system—World Bank support can be effective. Brazil, Kenya, and Viet Nam all demonstrated strong political and financial commitment in support of learning for all, combined with a strong equity focus (and significant contextual differences). This is evident in clear implementation actions to improve quality, equality, and learning, as well as a commitment to communicate learning data and establish clear goals for learning improvement. These countries have also allocated educational resources to prioritize primary and foundational learning, consistent with leading global education systems.

In such a context, the World Bank has been able to deploy its knowledge, technical assistance, policy dialogue, and financing more effectively in support of reforms that contribute to improvement in learning outcomes. Aspects of the analysis of context and engagement could have been strengthened in each case, but because the World Bank is working with determined and focused partners, it has been able to engage with an enhanced understanding of context and apply its resources to leverage points in support of more effective system reform.

The evaluation concludes that the World Bank is well placed to lead in delivering a more strategic response to the learning crisis and shifting to an outcome orientation. The World Bank typically has well-developed relationships with client governments that can be used to support reform in favor of learning for all. It also has strong research and analytic capabilities, and as the largest provider of development aid to education, it occupies a strong and influential position in relation to other development partners. These comparative advantages can reorient dialogue to emphasize changes in systems to improve learning, which require tracking learning outcomes and improving the quality of teaching. To support an outcome orientation at the country level requires detailed theories of change that define the pathways from enhancements of preservice institutions, teacher recruitment, and teacher monitoring to intermediate outcomes and how those outcomes are to result in improved pedagogical practices in classrooms that increase student learning. This shift will require incentives and signals to staff and Country Management Units that more ambitious objectives and measurements of learning outcomes in primary grades 3 and 6, with disaggregated data, are needed to improve accountability by clients and the institution for improved learning for all.

A stronger contribution would entail the adoption of a contextualized, systems-based approach that gives more attention to political commitment, public funding, and the education system’s capacity to deliver learning for all. Such an approach would recognize the unique political, social, cultural, and economic characteristics of individual basic education systems and facilitate the design and implementation of tailored responses, consistent with the call by the WDR 2018 to identify and address failures in systems for learning.

Country case studies found several weaknesses in the World Bank approach. Documentation from the portfolio and case studies suggests that support to basic education has taken a more uniform, less nuanced approach. For example, documentation rarely emphasized the potential impacts of dynamic interaction among multiple, potentially powerful stakeholders on the achievement of desired outcomes. The case studies also found no assessments of the alignment and capacity of the basic education delivery system, especially for actors in the lower levels of the system on whom fidelity to policy reform and implementation success depends. Analysis undertaken by the World Bank should take account of the level of political will in support of inclusive education reform, the level of financial commitment in support of reform, and the extent of capacity within and across the system. Such assessment may lead the World Bank to prioritize lending for basic education in some countries and to prioritize dialogue and capacity building in others.

Finally, the evaluation concludes that in many countries, World Bank support for basic education lacks intensity and continuity in the face of a particularly challenging problem that requires alignment between many dynamic actors and components in complex systems. Solving the problem requires sequenced engagement leading to incremental reform of systems, measurement of learning, establishment of a teaching career framework, and improvement in learning for all. The relatively limited level of engagement in many instances makes it difficult to provide the traction needed to support the systemic reform necessary to improve learning outcomes. Given the scale and depth of the learning crisis, the limited resources the World Bank and other development partners have to address this challenge need to be strategically deployed.

More comprehensive and scalable approaches to addressing the learning crisis will also require a much greater level of collaboration among development partners. COVID-19 has spurred the emergence of greater collective urgency and innovation among partners, including the World Bank, particularly at the global level. At the country level, however, although partners communicate and cooperate, true collaboration is much more limited and is undermined by the absence of a widely shared understanding of the factors contributing to system failure. A common analysis and understanding could support the co-pursuit of quality education and learning outcomes and the co-pursuit of reform of the teaching career framework and measurement of learning because no leading education system has succeeded without these two critical aspects.

Recommendations

Develop country-specific education engagement plans that include systems-based enhancements to the teaching framework to improve learning outcomes. These plans should be informed by a comprehensive systems analysis of the constraints to implementation of a career framework—teacher recruitment, training, development, motivation, and evaluation—as learning outcomes require capable and motivated teachers. Understanding the underlying issues, such as political will, system capacity, funding, and political economy obstacles and opportunities, will involve eliciting feedback from key stakeholders at all levels and compiling existing and new analysis to guide the development of a medium-term engagement process ideally anchored, where appropriate, within a pillar of the Country Partnership Framework. Sufficient data would be needed to inform adaptive management decisions related to corrective actions and learning during implementation to address the underlying constraints to sustainably improve systems and track intermediate outcomes. Implementation could be measured in Country Partnership Frameworks, supported by analytics and projects, with intermediate outcomes related to the performance of the teaching career framework, rather than just the completion of activities. The success of the recommendation can also be measured in lessons that inform a scaling up of approaches from the World Bank’s engagement.

Collaborate with global and country partners to close the data gaps on learning outcomes (aligned with Sustainable Development Goal target 4.1) and to track progress in ending learning poverty. This would be demonstrated by showing an increase in the number of countries with (i) education projects and Country Partnership Frameworks that include indicators for learning improvements in grades 3 and 6, which may require more ambitious project goals and indicators; (ii) improvement in national educational assessment capabilities and systems for data collection and decision-making; and (iii) participation in cross-national assessments for better data comparability. A focus on those countries that lack quality national assessments and have not been part of international or regional assessments in the last five years is particularly needed.