“There are better ways to improve test scores,” “food is expensive,” “most kids would eat anyway,” and other counterarguments contain some truth, but fail to overturn the basic economic logic of free, universal school feeding in poor countries.

Back in February I visited Tassah Public School in rural Bong county, Liberia. It had several nice classroom buildings arrayed in a big open field, and the teachers were all there to meet us. But there weren't very many kids, which was apparently normal. I sat in on a fifth grade English class with just two students. Enrollment had fallen from 360 students down to just 175 this year, and nobody could tell us why. The principal said some children had to walk three hours through the bush to get there. Being sort of a jerk, I persisted. “Wasn't that true last year as well? The school didn't move, did it?” I got the blank stares I deserved.

Finally, one of my colleagues asked the right question: “Do you have a school feeding program?”

“We used to,” the principal said. (I'm paraphrasing a bit here; my field notes are rough.) But the NGO that ran the program had closed it this year. And when the food left, the kids left. “By midday they're hungry and lose interest” he told us, “and now that there's no food, some just don't come.”

The world is full of hungry kids, not learning anything in school

Earlier this year the UN's Food and Agriculture Organization announced that hunger is rising globally for the first time this century. Admittedly, those numbers are shaky, and the uptick is driven somewhat by acute crises in places like Yemen and South Sudan. But there are broader signs that the world is not on an easy march to eradicating hunger. For instance, the data on child malnutrition rates are based on household surveys rather than macro models of the food supply, so they are somewhat more reliable than hunger numbers. They reveal a long-term global decline in malnutrition rates, but also a rising absolute number of malnourished children in sub-Saharan Africa over several years—which represents more than just a blip or a bad harvest.

At the same time, the World Bank's new 2018 World Development Report focuses on the “learning crisis“ in the developing world. Teachers are underprepared and often absent, basic materials are missing, and schools are poorly managed. As a result, children in many countries go to school for years and emerge functionally illiterate.

One shouldn't draw too strong a link here. These twin crises of growing hunger and low learning have distinct causes, even where they overlap geographically. But a problem's solution doesn't need to be a mirror image of its cause, and school feeding may help make progress on both fronts.

As with any public policy, the case for school feeding rests on its effectiveness, affordability, and—a dimension that is particularly acute in fragile states—its feasibility.

Aren't there better ways to improve learning outcomes?

If you survey the research on how to improve learning outcomes in developing countries, you'll find the reviews disagree on what works best, but few put much emphasis on food.

As Dave Evans and Anna Popova at the World Bank note in that link, some reviews declare pedagogical interventions the clear (albeit vaguely defined) winner, while others highlight computers and technology. JPAL, an organization that specializes in RCTs in poor countries, has put a big emphasis on hiring contract teachers to run remedial lessons targeted at students' ability level.

Evidence on the effectiveness of school feeding is more promising than this summary might suggest though.

First, food keeps kids in school. As Harold Alderman and Don Bundy note in the World Bank Research Observer (2011), “numerous studies show that in-school feeding has a positive impact on school enrollment or participation in areas where initial indicators of school participation are low.” The graph below shows results from all the available studies we could find (drawn mostly from earlier reviews here, here , here, here, here, and here).

Second, food facilitates learning. Quoting Alderman and Bundy again, “Improved performance as measured by tests of achievement is often reported for [school feeding programs], although there is a fair amount of variance as to which ages and which skills are most affected.”

Beyond its effectiveness, a potential selling point of school feeding is its feasibility. Abysmal learning levels and chronically absent teachers highlight that many poor countries simply can't deliver basic education to the bulk of their populations. But experiences from India to Mozambique suggest that even weak states—sometimes with the help of aid donors—can successfully deliver meals to millions of kids. Writing about India's massive school feeding program, Abhijeet Singh notes:

“[M]idday meals, which reach about 120 million children on every school day, are probably the most successful of all interventions in education that the Indian state has delivered in the past decade. On any school day, a quarter of teachers are absent from government schools, only 45% of those in school are teaching, but in 87% of schools, a hot meal is served.”

All of this suggests that governments might do more good for more kids if they reallocated money away from some current activities that rank high on the global development agenda in education—including purchasing laptops and even (gasp) raising teacher salaries—in order to feed kids.

Won't a universal school lunch program end up feeding a lot of kids who would eat anyway? Sounds inefficient and wasteful.

A shockingly high share of kids actually won't eat anyway. Back in 2004, Farzana Afridi measured children's nutrition intake over the previous 24 hours in Madhya Pradesh communities served by India's national school meals program, and randomly assigned children to be interviewed so that the recall period would or would not fall on a school day. She found that on average children's nutrient consumption went up on school days by the equivalent of 50 percent to 100 percent of the nutrients in school meals—with no such difference for children in schools where the program wasn't operating. Free food was additional food.

It's not just in India that school meals seem to provide nutrition that kids wouldn't get anyway. In both Kenya and Jamaica (see references in chart), school feeding led to significant weight gain for children and some evidence of increased height. Research in China focused not on school meals per se, but found that randomly assigning nutrition supplements to school children raised hemoglobin levels (confirming a nutritional deficit was present that wouldn’t be met otherwise) and contributed to significant learning gains. (Notably, other research in China shows that nutrition supplements outperformed the government’s policy of one egg per child in school.)

But even if some “undeserving” kids will get free lunch in a universal program, there are good reasons to make any such program universal.

Targeting or “means testing” opens a dangerous can of worms. The state in low-income countries often lacks the capacity to target social programs effectively. In proposing a universal basic income, the Indian Ministry of Finance noted that many current anti-poverty programs are so mis-targeted that they cease to be progressive at all. When the central government lacks the information to target objectively, a common response is to devolve that job to local officials. But giving local politicians the discretion to pick beneficiaries opens them up to corruption and clientelism.

Universality may also be good politics. Political support for anti-poverty and social protection programs often hinges on the participation of the middle class, and simple models of rational voters suggest proposals of targeted anti-poverty programs may lead to lower transfers to the poor than universal proposals.

Is universal school feeding really affordable?

According to the World Food Program, school feeding costs an average of about $56 a year per child in poor countries, rising to about $370 in upper-middle and high-income countries. Poor countries only spend about $82 a year on basic education, so adding meals would be a non-trivial increase in percentage terms in some contexts. Note, however, that some of the biggest programs, like India's, cost just a fraction of this, at about 3 cents per child per day.

Financing universal school meals is a serious obstacle. But it's worth noting that the sticker price may seriously exaggerate the true social cost of these programs. Consider two extreme scenarios.

If households would fill in the gap and kids would eat just as much without school meals, then the social cost of school feeding is only the extra expense of public provision, over and above what households would have spent otherwise. At a dollar a week to feed kids lunch, it's not at all clear that public provision of food costs more than private provision. No study I've seen to date can reliably answer this question, but it seems school feeding could very well be cost saving. The WFP notes that even in the most expensive large school feeding programs, the cost of administration and delivery is about 20 percent of the total, and government food subsidies lower the price of food for non-beneficiaries.

Alternatively, consider the other extreme scenario in which all school feeding is additional, which is closer to what Afridi found in India. In this case the social cost of school feeding is higher—we're spending money that wouldn't have been spent otherwise. So be it. To most audiences, the moral and policy case for school feeding is stronger, not weaker, in this scenario, if the counterfactual is that children go hungry.

State capacity and basic needs trump Econ 101

When I took Econ 101 in Lincoln, Nebraska, we learned that education has positive externalities that justify government intervention, while food staples are commodities that can be efficiently traded on private markets. Lincoln is a town full of high-quality Midwestern public schools, surrounded by private commercial farms producing more corn than America knows what to do with. The world around me looked like my textbook.

Rural Liberia does not look like my Econ 101 textbook. State capacity to deliver basic services and households' urgent need for food are much more important considerations than addressing externalities. Even if food markets functioned perfectly, many households simply lack the means to buy or grow enough food to feed their children. The cash-strapped government does little to help them. Public efforts to provide basic schooling have mostly failed. As of 2015, net primary enrollment was 37 percent and three-quarters of adult Liberian women who had gone to six years of primary school were still illiterate. Meanwhile, 32 percent of their kids are stunted, a sign of severe malnutrition.

Fixing Liberia's schools is an intellectual and political juggernaut. Feeding kids isn't.


Afridi, F. (2010). Child welfare programs and child nutrition: Evidence from a mandated school meal program in India. Syracuse University.

Ahmed, A. U. (2004). Impact of Feeding Children in School: Evidence from Bangladesh. IFPRI.

Alderman, H., Gilligan, D. O., & Lehrer, K. (2012). The Impact of Food for Education Programs on School Participation in Northern Uganda. Economic Development and Cultural Change.

Du, X., Zhu, K., Trube, A., & Hu, X. (2006). Effects of school-milk intervention on growth and bone mineral accretion in Chinese girls aged 10-12 years: accounting for cluster randomisation. British Journal Of Nutrition, 1038-9.

Grillenberger, M., Neumann, C. G., Murphy, P. S., Bwibo, N. O., van't Veer, P., Hauvast, J., & West, C. E. (2003). Food Supplements Have a Positive Impact on Weight Gain and the Addition of Animal Source Foods Increases Lean Body Mass of Kenyan Schoolchildren. The Journal of Nutrition.

Kazianga, H., Walque, D. d., & Alderman, H. (2012). Educational and Child Labor Impacts of Two Food for Education Schemes: Evidence from a Randomized Trial in Rural Burkina Faso. World Bank.

Kleimann-Weiner, M., Luo, R., Zhang, L., Shi, Y., Medina, A., & Rozelle, S. (2012). Eggs versus chewable vitamins: Which intervention can increase nutrition and test scores in rural China? China Economic Review.

Luo, R., Zhang, L., Liu, C., Rozelle, S., Sharbono, B., Yue, A., . . . Martorell, R. (2012). Nutrition and Educational Performance in Rural China’s Elementary Schools: Results of a Randomized Control Trial in Shaanxi Province. Chicago Journals.

McEwan, P. J. (2012). The impact of Chile’s school feeding program on education outcomes. Economics of Education Review.

Powell, C. A., Walker, S. P., Chang, S. M., & Grantham-McGregor, S. M. (1998). Nutrition and education: a randomized trial of the effects of breakfast in rural primary school children. The American Journal of Clinical Nutrition.

Singh, A., Park, A., & Dercon, S. (2015). School Meals as a Safety Net: An Evaluation of the Midday Meal Scheme in India. Economic Development and Cultural Change.

Tan, J.-P., Lane, J., & Lassibille, G. (1999). Student Outcomes in Philippine Elementary Schools: An Evaluation of Four Experiments.

Vermeersch, C., & Kremer, M. (2004). School Meals, Educational Achievement and School Competition: Evidence from a Randomized Evaluation. World Bank.

Whaley, S. E., Sigman, M., Neumann, C., Bwibo, N., Guthrie, D., Weiss, R. E., . . . Murphy, S. P. (2003). The Impact of Dietary Intervention on the Cognitive Development of Kenyan School Children. American Society for Nutritional Sciences.

Text by Justin Sandefur, literature review and graphics by Divyanshi Wadhwa