Schooling for All: Feasible Strategies to Achieve Universal Education


Editor’s Summary: All the Education Money Can Buy

This report debates the case for specific public investments in education in low- and lower-middle-income countries, drawing on evidence of what has worked not just in small-scale experiments but historically and in large-scale national programs. Its messages are intended more for economic policymakers than educators, as they speak to what can be accomplished with fiscal instruments (money) and where trade-offs must be made. CGD does not take institutional positions. Each chapter is authored by a different set of CGD researchers (with some editorial steer), and each commentary is written by external contributors (who were promised space to disagree). This introduction tries to summarize the main arguments across all these contributions, noting points of consensus and ongoing debate.

Money matters

Economists have often shied away from debates about whether developing countries should spend more on education—other than to suggest it doesn’t matter. In 2018, for the first time, the World Bank dedicated its annual World Development Report to the subject of education (World Bank 2018). In the space of 11 chapters and 31 boxes spread across 239 pages, a single box was devoted to education budgets. The headline was that “public spending does not correlate strongly with learning,” and the short discussion concluded that “improvements in learning are unlikely when additional resources are allocated like past funding.”[1]

The core claim of this volume is that spending does matter, for both learning and other outcomes, and a lack of spending is the binding constraint to educational progress in many low- and lower-middle-income countries. We outline a list of shovel-ready investments in education, from abolishing user fees to extending the length of the school day, that have proven technically and politically feasible and have high returns. Of course, the fiscal realities of low- and lower-middle-income countries make some trade-offs unavoidable. The chapters that follow suggest, for instance, that money spent on school meals probably goes further than reducing class sizes. Ironically, it is exactly the kind of spending that the World Bank declared irrelevant—providing more schooling to more kids for more years—that has proven most successful at raising overall learning levels.

Historical and experimental evidence offers a list of educational investments that are technically feasible, politically popular, and effective

In the past 50 years, the developing world has witnessed dramatic improvement in enrollment and learning outcomes. There is a temptation to discount this historical accomplishment as an improvement in “mere” access to schooling, bemoaning the low test-scores of South African eighth-graders (for instance) while ignoring the increased number of kids who make it to eighth grade. In Chapter 1, Lee Crawfurd, Susannah Hares, and I argue that this historical success should inform countries’ forward-looking education policy agenda—especially in places where some of the core policies that did so much good in the 1990s and 2000s, like abolishing user fees and building schools in hard-to-reach areas, have not yet been fully implemented.

The historically proven route to more learning has been more schooling. The massive increase in educational access has also delivered a globally unprecedented increase in learning—with literacy rates in low- and lower-middle-income countries now higher than a half century ago by double-digit percentage points.

Randomized trials and quasi-experimental evaluations show how to do this: subsidize access, take care of kids’ health and nutrition while in school, and don’t expect huge gains from reducing class sizes or buying new books. A 2013 summary of this literature, published in Science by recent Nobel-laureate Michael Kremer, Conner Brannen, and Rachel Glennerster, has held up reasonably well. The abstract is worth quoting in full:

Across many different contexts, randomized evaluations find that school participation is sensitive to costs: Reducing out-of-pocket costs, merit scholarships, and conditional cash transfers all increase schooling. Addressing child health and providing information on how earnings rise with education can increase schooling even more cost-effectively. However, among those in school, test scores are remarkably low and unresponsive to more-of-the-same inputs, such as hiring additional teachers, buying more textbooks, or providing flexible grants. In contrast, pedagogical reforms that match teaching to students’ learning levels are highly cost effective at increasing learning, as are reforms that improve accountability and incentives, such as local hiring of teachers on short-term contracts. Technology could potentially improve pedagogy and accountability. Improving pre- and post-primary education are major future challenges. (Kremer et al. 2013)

Some items on this menu have proven technically and politically difficult for developing countries to scale up in public school systems. In the decade since the end of the trials summarized by Kremer and colleagues, development economists have increasingly recognized the difficulty of translating pilot projects into successful national policies. For instance, pedagogical reforms and accountability initiatives, while successful at trial stage, have often been abandoned or seen their impacts wane as they were taken to scale, encountering implementation failures and political resistance. But not everything falls apart when scaled up.

New public spending will do the most good where money is the binding constraint. Meta-analysis of impact evaluations shows that several categories of policies—which we characterize broadly as less technically demanding, more logistical interventions—show impacts that are more robust in large government-run programs. This includes things like school meals and extending the length of the school day. The core argument of Chapter 1 is that, if countries can spend more on education in the short-to-medium term, these things should be the first priority.

The crucial task of improving pedagogical practices is less about money. In his commentary, Moses Oketch makes the case not to abandon the more difficult reforms, as governments pivot from a focus on schooling to school quality. By construction, that new direction will require new tools. But both Oketch and Rukmini Banerji note that, while spending money has worked well to improve access, it’s less obvious that money can fix the pedagogical failures that Banerji and her colleagues at Pratham India have worked for decades to highlight. In her commentary, Banerji insists that a change in mindset is required across various actors at different levels of the system. Unless people buy into the idea of adjusting the curriculum to the level of the pupil and defining success in terms of learning outcomes, programmatic solutions are doomed to fail.

These are not contradictory agendas: one is a long-term project of research and development or complex system reform; the other is a short-term budget agenda. Most of the chapters that follow focus on this latter fiscal agenda, occasionally neglected by development economists.

Example #1: Free school meals enjoy broad support among education and social protection experts—yet remain limited in scope

Unlike many other policies shown to improve education outcomes, countries have shown they can make school meals work at scale. Many Indian states famously struggle to get teachers to turn up at school—a topic that development economists have fretted over intensely, experimenting with carrots and sticks but having limited success beyond small-scale trials. Yet nationwide, India’s midday meals scheme manages to feed 100 million kids every day, and recent research shows benefits not just in the health and fertility of girls who received meals but also in the nutrition of their own children a generation later (Chakrabarti et al. 2021).

Coverage of school feeding programs has grown quickly in recent years but remains low. In Chapter 2 Biniam Bedasso notes that, as of 2019, only about 1 in 7 children in low-income countries receive meals at school, rising to under 1 in 3 in lower-middle-income countries. This is a policy ripe for expansion.

School meals highlight the limits of traditional comparative cost-effective metrics. The benefits of free school meals are sometimes underappreciated because they are spread across multiple outcomes: better nutrition, higher enrollment, and increased learning for kids who go to school. Combining these various benefits, analysis by the World Food Program suggests benefit-cost ratios between 5:1 and 6:1, where data are available.

The evidence on the benefits of school meals is strongest for nutrition and enrollment, and weakest for learning outcomes. As noted in both commentaries on this chapter (by Farzana Afridi and by Ugo Gentilini and Shwetlena Sabarwal), evidence from randomized controlled trials suggests there are cheaper ways to raise education outcomes. Information campaigns encouraging students to stay in secondary school get more bang for the buck than feeding them. But when viewed through a broader lens, considering not just social protection objectives but also arguments about the social contract between states and citizens and the types of large-scale programs that are politically sustainable, school meals look more and more attractive.

The main obstacle to expanding school meals is simply cost. Universal free school meals aren’t cheap. Low-income countries spend less than 1 percent of their education budgets on meals (compared with about 2 percent in high-income countries). But that reflects their low coverage rates. Raising coverage to just the global median of 21 percent would eat up 5 percent of low-income countries’ education budgets. We return to the question of financing in Chapter 6.

Example #2: The success of free primary education in the 1990s and 2000s provides a template for making free secondary school work

The abolition of user fees has been a key driver of educational progress in the developing world over recent decades. The wave of free primary-education reforms in sub-Saharan Africa and Southeast Asia in the 1990s and 2000s contributed to a massive expansion in education access and a marked global increase in literacy rates. In Chapter 3, Lee Crawfurd and Aisha Ali make the case to build on this success by extending fee abolition to secondary school as well.

Secondary enrollment rates remain low globally. In low-income countries, only about one-third of secondary-school-age children are in school; in lower-middle-income countries, the number is fewer than two-thirds. Hence the need for policy action and the room for potential impact are clear.

There is little doubt that, with straightforward complementary policies in place, fee abolition can boost enrollment. Published econometric studies of the effect of free primary education show that it increased average grade attainment by about one year—of course, with lots of variance across countries. While there’s less direct econometric evidence on the more recent turn to free secondary schooling in lower-middle-income countries, experience from Ghana points to big impacts on enrollment. Crawfurd and Ali note that countries that made free primary education work combined fee abolition with school grants to offset the lost revenue to schools. Another complementary policy that has been crucial to boosting enrollment under free secondary schooling in Ghana, with potential lessons for other countries, is relaxation of exam requirements to enter public high school.

Expanding enrollment has historically not undermined learning levels. While Crawfurd and Ali present a review of the literature to bolster this claim, the commentaries by both Robert Osei and Pauline Rose express some skepticism, or at least caution. Osei notes that the potential for massive enrollment expansion to undermine resources per pupil is real and can’t be dismissed without a big fiscal push.

Free schooling will often be progressive, according to the analysis in Chapter 4, though this is a contested point. Secondary enrollment in low- and lower-middle-income countries remains strongly associated with parental wealth. Crawfurd and Ali argue that the elasticity of new enrollment in response to fee abolition implied by earlier studies on primary schooling suggests that fee abolition will often be progressive—despite the subsidy to middle-class households whose children are already enrolled. Osei cites evidence from Ghana that free secondary has indeed been pro-poor (a slightly different criterion than progressivity). Rose’s commentary remains skeptical.

In fiscal terms, low- and lower-middle-income countries can’t afford the UN’s Sustainable Development Goals for education

It is almost tautological to note that poor countries can’t afford to do everything worth doing in education, as in other sectors. In Chapter 6, Jack Rossiter reviews past efforts to put a price tag on various global programs of improving education outcomes, such as reaching the UN’s Sustainable Development Goal #4 for education, which includes free, universal, high-quality primary and secondary education. Some of these exercises produce eye-popping price tags, with several recommending a trebling of public expenditure on education in low- and lower-middle-income countries by 2030.

We should distinguish normal targets from positive forecasts: current spending trends suggest the ambitious goals laid out by various international bodies are likely out of reach. Rossiter presents simple extrapolations of current spending levels, based on the relationship between GDP growth and education spending by income level, and using IMF growth forecasts to project education spending forward to 2030. Those projections show, for instance, a shortfall of about $20 billion per annum by 2030 for low-income countries relative to the targets for domestic public spending on education set out by Gordon Brown’s 2016 Education Commission report (Education Commission 2016), and a shortfall of about $300 billion per annum in lower-middle-income countries. These shortfalls already factor in large increases in international aid, which may or may not materialize.

Unit costs likely must come down. Rossiter concludes that the only realistic formula for meeting the kinds of ambitious goals laid out by the UN and other entities involves a significant reduction in expenditure per pupil as pupil numbers grow. That is, education must be done more cheaply if countries are going to fulfill the promise of reaching all children.

Of course, any new international action to soften this harsh reality remains welcome. In his commentary, Daouda Sembene acknowledges the inescapable realities laid out in this budget arithmetic, but he advocates greater focus on rooting out corruption and misallocation of education expenditure, and he issues a renewed plea for more international support. On the latter point, Sembene offers a more creative list of potential financing mechanisms, including carbon taxes, financial transaction taxes, debt relief, and allocation of the IMF’s Special Drawing Rights.

For countries forced to contain education expenditure growth, teacher pay and staffing are an obvious place to look

Digging into the costing models underlying the calculations from UNESCO and the Education Commission, the big cost drivers in these projections are the assumptions made about teacher salary levels and pupil-teacher ratios. These models assume that all countries will reach the average of the current top 50 percent of countries in terms of pay and staffing levels by 2030.

Across-the-board increases in teacher salaries or staffing levels are an expensive way to increase education outcomes. Teacher salaries constitute 55 percent of education expenditure in low-income countries and 62 percent in lower-middle-income countries. In Chapter 4, Lee Crawfurd and Alexis Le Nestour make the case that “intensive margin” investments in teachers—that is, raising salaries for existing teachers or reducing class sizes in existing schools with business-as-usual pedagogical approaches—have little impact on learning outcomes.

Higher pay and smaller class sizes make little difference if pedagogy and teacher management are weak. On the basis of their meta-analysis, Crawfurd and Le Nestour argue that the effect of teacher salaries or class sizes on learning outcomes is likely contingent on good pedagogy and teacher management. Even where teacher management systems are functional, as scored by the World Bank, “average effects may be on the order of 0.05 standard deviations per $1000 PPP increase in pay.”

Evidence of limited short-term impacts of salary increases on learning leaves open the question of whether higher salaries will attract more capable teachers over the longer term, as noted by Tessa Bold in her commentary. Interestingly, however, Chapter 4 shows that, in the handful of lower-middle-income countries where data are available, teachers are already recruited from the middle (or higher) of the test-score distribution among university graduates.

Getting teachers to underserved communities may be a higher priority than reducing class sizes in general. While Crawfurd and Le Nestour focus on “intensive margin” investments in new teachers to reduce class sizes, David Evans and Amina Mendez Acosta argue that “extensive margin” investments in teachers—recruiting teachers to staff new or understaffed schools in marginalized communities—are a better use of marginal salary expenditures. Speaking to Malawi’s experience, Esme Kadzamira picks up on the points raised in Evans and Mendez Acosta’s contribution, noting that inequitable allocation of teachers remains a challenge in some rural areas. She also argues that reforms to teachers’ career structure, promotion opportunities, and other nonpecuniary benefits may be important to improving teacher motivation and teaching quality, at lower fiscal cost than unconditional salary increments.

Public-private partnerships have done better at expanding access to underserved populations than improving quality, but questions linger about the sustainability of those cost savings

Across the developing world, the share of children attending private schools has blossomed in recent decades—reaching a fifth of pupils at primary level in Nigeria, a third in Pakistan, and nearly a half in India. Faced with dysfunctional government schooling systems, many policymakers have taken renewed interest in outsourcing public education to the private sector. If you can’t beat ’em, join ’em, as the saying goes.

Public-private partnerships (PPPs) have generated mixed evidence of lifting learning outcomes over traditional government schools. Reviewing the performance of PPPs across various dimensions in Chapter 5, Maryam Akmal, Rita Perakis, and Susannah Hares find weak and inconsistent evidence of learning gains in existing schools. For instance, Liberia’s controversial outsourcing initiative that handed over government schools to private operators delivered only minimal learning gains at enormous per-pupil cost, while reducing enrollment and transitions to secondary school. In his commentary, Jishnu Das piles on. He focuses on five problems bedeviling current attempts to demonstrate gains from PPPs in education and asks whether it’s time to turn back from this policy agenda.

Some PPPs have succeeded in expanding access where public schools are missing. Contrary to many popular conceptions of what outsourcing does, Akmal et al. find important examples of positive impacts on enrollment for marginalized groups. For instance, Uganda’s secondary school PPP or the Sindh Education Fund in Pakistan may not have dramatically improved quality but extended access into underserved communities.

The strongest arguments for education PPPs are financial, not educational. Despite this mixed evidence, the financial imperative for engaging with the private sector remains compelling to many audiences. In his commentary, Moses Ngware notes this urge to leverage private capital for education investments. Viewed this way, education PPPs look more analogous to infrastructure PPPs that, Ngware notes, have a very checkered history in the developing world. He advocates a focus on schemes where money follows the student, to avoid some of the governance pitfalls associated with subsidizing specific firms.

Private schools offer cost savings not because they’re more efficient but because they pay teachers less. That means sustaining the short-term cost savings from PPPs requires a politically difficult, two-tier wage structure for teachers. Teachers doing the same job and paid by the same government will often, quite understandably, militate for equal pay—as seen in India and Kenya.


Sometimes reading reports on topics like education in developing countries can feel like advocacy for “motherhood and apple pie,” as Americans say. Yes, of course we all want these things. Where is the disagreement? For the sake of clarity, it is worth emphasizing what this report is not saying.

This report is not a template for fixing failed schools. Most of the volume is skeptical that new innovations in incentives and accountability for teachers will lead to dramatic improvements in learning outcomes at scale. The agenda of pay-for-performance contracts and outsourcing management of existing public schools remains largely unproven at scale in developing-country contexts. And even the less controversial (and in the opinion of most of the authors here, more promising) agenda around improving pedagogical practices in primary schools is largely offstage here: a topic for ministries of education to pursue internally, while we focus here on budgetary instruments.

This is also not just a plea for more money. It does not rest its case on an appeal for dramatic increases in public expenditure or international aid. Most of the authors gathered here would, I believe, support those things. But we have tried to focus the debate at the level of ambitious but feasible new investments, based on simple extrapolations of spending trends and GDP forecasts in low- and lower-middle-income countries.

Stepping back into motherhood-and-apple-pie mode, if I were trying to sell the messages of this report as broadly as possible, I would note that it is fundamentally a case for building on the amazing educational progress of the last half century in the developing world. It’s a case for letting more kids finish primary and go on to secondary school, extending their instructional time, and providing them with free school meals. Crucially, these are all things that governments from Ghana to India have shown are possible to do nationwide even in imperfect schooling systems, and they are potentially affordable under current budget trends if expenditure growth is contained in other, lower-priority dimensions.


Chakrabarti, Suman, Samuel P. Scott, Harold Alderman, Purnima Menon, and Daniel O. Gilligan. “Intergenerational Nutrition Benefits of India’s National School Feeding Program.” 2021. Nature Communications 12 (4248).

Education Commission. 2016. The Learning Generation: Investing in Education for a Changing World.

Kremer, Michael, Connor Brannen, and Rachel Glennerster. 2013. “The Challenge of Education and Learning in the Developing World.” Science 340 (6130): 297-300.

World Bank. 2018. World Development Report 2018: Learning to Realize Education’s Promise.

[1] Credit where credit is due: more recently during the COVID-19 pandemic, the World Bank and UNESCO have issued joint calls for increased investment in education systems:

Rights & Permissions

You may use and disseminate CGD’s publications under these conditions.