Measuring Hunger

How many people in the world are going hungry? It’s a complicated question.

© 2011 P. Smith/Panos

In 2010 the UN Food and Agriculture Organization (FAO) announced a stark fact: because of the global food price crisis, the number of people going hungry had spiked to more than 1 billion. The agency launched an online petition for action with the header “1 billion people live in chronic hunger and I’m mad as hell” and hung a banner bearing that message on its Rome headquarters building.

The number sounded right. With food prices going straight up, it seemed likely that hunger among poor people was rising too. But soon another story started to emerge. Gallup polls of people in developing countries did not show a steep rise in the number of people reporting themselves as hungry. Experts questioned FAO’s number, and even the staff at FAO knew they needed to overhaul their data and methods.

In 2012, the agency issued drastically revised hunger assessments. According to its new estimates, there was no spike in hunger during the global food crisis, or at least none that it could measure using the tools available. Instead, it appears that the number of hungry people had been on a slow decline for two decades, reaching about 870 million in 2010.

image of a tape measure wrapped around an apple with an image of the globe on itCounting the number of hungry people is a lot harder than it sounds, and FAO is not the only organization that struggles with it. IFPRI and other institutions have also tried to gauge global hunger. The task raises difficult questions about exactly who is eating what in a wide variety of settings. The challenges vary depending on the scale of the question: Are you most interested in knowing about hunger in a country, a district, a village, a household, or a person? Accurate measurements can help development agencies and political leaders understand trends, direct aid to where it is most needed, design more effective programs, and assess whether interventions to reduce hunger are working—but meeting these different goals can require different approaches to measuring hunger.

Rethinking the Numbers

“If we were only measuring in developed countries, data would be less of a problem, but it’s very partial in most developing countries.”
—Carlo Cafiero, FAO

How did the FAO get from 1 billion to 870 million hungry? To make its annual estimates of people who are undernourished, FAO draws on population data from United Nations agencies. Individual countries supply data on how much food is available there, based on their own calculations of the quantity of major food commodities they produce, import, and export. For many countries, nationally representative surveys supply additional information about the amount of food households acquire. Based on all of these figures, FAO calculates how many people in a country cannot get access to a minimum calorie threshold.

But problems crop up. Up-to-date, accurate data from these various sources are often unavailable. It can, for example, be difficult to get accurate measurements of a country’s grain inventories. People responding to surveys may fail to mention food that is wasted, given away, or bought outside the home. Even population data sometimes turn out to be skewed. In some cases, therefore, FAO must use estimates to fill the gaps and calculate a percentage of undernourishment.

“If we were only measuring in developed countries, data would be less of a problem, but it’s very partial in most developing countries,” says Carlo Cafiero, who leads the FAO Statistics Division team that prepared the measurements of global undernourishment. “We want to make assessments that are consistent across countries and provide insights into the most important questions.”

When its 2010 estimates came into question, FAO launched a two-year consultation process to see how it could improve them. By 2012 FAO had updated its data on food supplies. Deeply revised population assessments were available—Bangladesh’s population was revised downward significantly, for example. The agency got more accurate data on people’s physical stature, which affects people’s calorie requirements. It also estimated the amount of food lost during food retailing. “Recognizing the difference between total food supply and the amount that reaches households is a major new factor,” says Cafiero. Finally, it revised its methods for estimating figures where data were missing.

Nonetheless, says Cafiero, “we recognize that the ability to measure food insecurity is still limited.” FAO’s effort to improve its indicator of food insecurity is ongoing. And even if FAO’s new figures are more accurate, critics say they provide only limited insights into food insecurity. “The revised FAO indicator takes into account households’ access to food and calories but still does not measure access by individual household members, and more important, does not measure any aspects of dietary quality, such as access to diverse diets and micronutrient-rich foods,” says Marie Ruel, director of IFPRI’s Poverty, Health, and Nutrition Division.

Infographic illustrating how the GHI, the FAO prevalence of undernourishment, and the Global Food Security Index measure hunger

Three measures of hunger - FAO prevalence of undernourishment, IFPRI's Global Hunger Index and the duPont/Economist Global Food Security Index (click to enlarge)


Still, the national and global figures from FAO serve a purpose. For example, its national measures of undernourishment are one standard by which countries are judged against Millennium Development Goal 1, which concerns extreme poverty and hunger. “FAO is a global organization that provides overall policy guidance to national governments on general trends, not on specific actions,” says Calestous Juma, a professor of international development at the Harvard Kennedy School of Government.

Beyond Calories

“Calories are just one dimension of food security. People need more than calories for physical and cognitive development and reproduction.”
—Olivier Ecker, IFPRI

Other measures of hunger, such as IFPRI’s Global Hunger Index, come at the problem slightly differently. A precursor to the index was born in the late 1990s, when Doris Wiesmann was a graduate student at the University of Bonn’s Center for Development Research. Joachim von Braun, then a professor at the university and later director general of IFPRI, asked Wiesmann to develop a nutrition index that would reflect not just calorie deficits, but broader aspects and consequences of food insecurity. She constructed an index by combining three indicators: FAO’s measure of undernourishment, the World Health Organization’s data on underweight among children under age five, and UNICEF’s data on mortality among children under age five. Wiesmann’s Nutrition Index represented an attempt to measure food insecurity and its consequences, focusing on the most devastating losses for countries—losses in lives and nutrition and in the future potential of children. Several years later, when Wiesmann joined IFPRI as a postdoctoral fellow, the Nutrition Index became the Global Hunger Index. IFPRI, along with the German NGO Welthungerhilfe and the Irish aid organization Concern Worldwide, now issues the index annually.

Because the Global Hunger Index ranks countries on a 100-point scale, with zero as the best score (no hunger) and 100 as the worst, it’s easy to compare how well countries are performing on addressing hunger. “Indexes of this type can be powerful tools for advocacy when used in international country rankings,” says Wiesmann, now a consultant based in Germany. Indeed, the Global Hunger Index has been widely used to put pressure on policymakers to do more to fight hunger.

Still, like all indexes, it simplifies a complex problem. And although the Global Hunger Index does offer a more multidimensional view of food insecurity, it is only as accurate and up-to-date as the data on which it is based. Given the time lags that affect all of the source data, it is definitely a picture of the recent past, not of current hunger and undernutrition. Conducting surveys takes time and costs money. Nonetheless, according to Klaus von Grebmer, IFPRI research fellow emeritus, the technology exists to get more accurate and up-to-date data: “At a time when data handling has become much easier, the main obstacles to providing this information in a more timely way are lack of political will and lack of funding.”

Diversity on the Plate

A fuller measure of food insecurity, however, would include not just how many calories people are eating, but also how many essential nutrients—such as vitamin A and iron—they are getting. “Calories are just one dimension of food security. People need more than calories for physical and cognitive development and reproduction,” says IFPRI Research Fellow Olivier Ecker. “If you look at nutrition indicators, you have physical indicators that can reflect positive and negative shocks and longer-term developments, though they are often not up-to-date or available in combination with key economic indicators.” Analysts are looking for indicators that provide useful information about specific situations, such as the short-term impacts of economic shocks, and they want more finely grained data that go beyond national averages to give more detail.

Ecker and IFPRI Research Fellow Derek Headey recently reviewed various possible indicators of food security and concluded that measuring the diversity of people’s diets is the most promising way to improve food security measurements. It is relatively easy to measure dietary diversity: surveyors count the different foods or food groups people eat and can weight them according to their nutritional value and the frequency with which they are consumed.

The diversity of people’s diets is highly correlated with certain nutrition indicators, such as people’s intakes of calories and micronutrients, and even—though less strongly—with indicators of chronic malnutrition. And dietary diversity is a useful measure of the impact of economic shocks. “It matches what we know happens when income falls: people reduce the diversity of their diets,” says Ecker. During economic crises families may not reduce their total caloric intake immediately, but they save money by consuming fewer expensive, highly nutritious foods like vegetables, eggs, and meat.

But how many different foods should a person actually eat? “Diets vary quite a bit around the world and are more diverse in some regions than in others, so defining universal thresholds has failed so far,” says Ecker. The best approach, he believes, may be to develop a standard set of food and nutrition security indicators—rather than one index that incorporates several indicators—for various target groups, such as women of reproductive age and preschool children. Among the set of indicators should be dietary diversity.

Aren’t We Already Doing That?

Even if we have better ideas about what kind of information to collect, the problem of how to collect it remains. If we can’t survey everyone in the world about his or her food consumption, a next-best option is to conduct large-scale surveys that would be nationally or regionally representative. Currently, only the Philippines conducts a nationally representative survey of individuals’ food consumption over the past 24 hours every five years. These 24-hour recall surveys are considered one of the best ways of assessing individuals’ food consumption. “They are very costly and require a high level of technical skill and expertise,” says Wiesmann, “but they also provide unique information that is not available from other methods.”

But the World Bank and other agencies do conduct frequent large-scale household surveys of people in developing countries, points out IFPRI Senior Research Fellow Jack Fiedler. In fact, one study showed that in recent years 95 percent of the population in developing countries has been covered by at least one such survey. In many cases these household consumption and expenditure surveys are designed to gather information about poverty, not food and nutrition security—but to measure poverty, they generally collect information on food consumption and expenditures too. The question is how well do they currently do so, and how much they might be improved if they were repurposed with this intention.

Fiedler works for IFPRI’s HarvestPlus program, which supports the breeding of more nutrients into staple food crops. He is interested in knowing not only where people suffer from food and nutrition insecurity, but also which foods people regularly eat that might be suitable vehicles for biofortification. “Food policymakers in many countries have been making a lot of guesses about household consumption patterns and food sources, but household consumption surveys can ask people what they ate, how they got it, and what goods are commercially traded at local markets,” says Fiedler.

Hunger agencies, he says, should look for more opportunities to partner with the agencies that administer household consumption and expenditure surveys and increase their focus on food consumption and nutrition. Household consumption and expenditure surveys are already typically conducted in most countries every three to five years. Questions about dietary diversity and frequency, of the type described by Ecker, can already be addressed using household consumption and expenditure surveys, and other questions could be added to help researchers better estimate the nutritional needs of a household. Are there any pregnant or lactating women? Are there any children under age one? Is anyone suffering from illness? How are foods combined and prepared? In addition, the cost of analyzing households’ nutrient availability from such secondary data sources is about 2 percent of the cost of collecting primary data in a typical 24-hour recall survey. “Affordability is a huge issue,” says Fiedler.

Ecker also points out that using such surveys to reach out to fewer people more frequently—once or even twice a year—would provide valuable information on the effects of seasonal variation and shocks, such as weather events or price changes, on people’s food and nutrition security.

Drilling Down

“Economics has assumed for a long time that households behave as a single unit and that households are all that matters. but individuals are the ones making decisions.”
—Agnes Quisumbing, IFPRI

As researchers make greater use of household surveys to measure hunger, they will need to consider how to determine the food and nutrition security of each individual within the household. “Individuals go hungry and suffer malnutrition, not households or aggregates,” says Christopher Barrett, a professor of economics and agriculture at Cornell University. For example, in recent years researchers have found increasing instances of the “double burden of malnutrition,” in which the same household contains adults who are overnourished (eat too many calories) and children who are undernourished (eat too few calories or nutrients).

“Even when large household surveys collect info on children, this information is typically aggregated up to the household level,” says IFPRI Senior Research Fellow Agnes Quisumbing. “Economics has assumed for a long time that households behave as a single unit and that households are all that matters. But individuals are the ones making decisions.” Although nutritionists collect information on nutritional status at the individual level, the data aren’t always reported in the most useful way. Information on nutritional status reported according to specific age-sex categories would capture food and nutrition security more effectively than looking at average household consumption, Quisumbing argues.

Interviewing individuals costs more than interviewing just the head of the household, the typical practice, but Quisumbing believes the expense is justified. “You learn much more from asking individuals questions, such as why some people in the household are systematically less nourished than others,” she says. However, it may complicate field logistics: interviewers may have to talk to men and women separately, or field teams of male and female interviewers.

Other Yardsticks

To measure a problem on the scale of world hunger, analysts inevitably have to trade off their desire for greater detail against the cost of getting more information. Experts broadly agree that FAO’s national-level assessments only capture part of the problem, and that household or individual data provide clearer information about the scale and distribution of food insecurity.

But some are skeptical about calls for more data. “It’s no small task to collect data every year on every country. There’s no funding for it, and many countries can’t do it for themselves,” says IFPRI’s Ruel. There is also confusion about what different indexes and indicators of food security mean and how they should be used. Ruel and IFPRI Research Fellow Jef Leroy are reviewing how well various indicators measure different aspects of food insecurity. “We want to clarify what indicators have been used, at what level they’re most useful, which ones have been validated, and what they reflect so that we can provide guidance on which ones to use for what,” says Ruel.

In the near term, more frequent household surveys with improved questions on food and nutrition would fill some existing gaps. And FAO is exploring the possibility of collaborating with the global polling firm Gallup to add a module about food insecurity to the Gallup World Poll, which conducts surveys in 160 countries. The poll could ask subjects whether they have gone hungry or had to skip meals over defined periods of time. If done properly, it would provide a fine-grained picture of food insecurity in virtually all countries in the world in a comparable way, and almost in real time. “This would be a real innovation for FAO,” says Carlo Cafiero. “We’ve never had a budget to collect data directly—we do assist national statistical offices, trying to promote common standards, but we have always depended on data collected by others that take time to be processed and are highly heterogeneous in their quality across countries.”

Cornell’s Christopher Barrett has proposed another model: an international network of “sentinel sites” in the developing world that would produce regular household- and individual-level surveys tracking multiple food security indicators and using standardized survey protocols. There’s a template: the US National Science Foundation’s network of 26 sites around the world where researchers can study ecological health and environmental change. “NSF’s network didn’t pay dividends for ten years. But now it’s essential to our understanding of key ecologies,” says Barrett. Similarly, a network of sites focused on food security would be a costly investment but could yield major insights.

In the end, although the alarm about a spike in hunger in 2008–2009 may have been unwarranted, it had a useful outcome. Along with the food price increases and the food riots, the apparent rise in hunger helped refocus the world’s attention on hunger and poverty—problems that remain serious. Better assessment of food and nutrition security could help lead to policies to ensure that all people are well fed and well nourished, although it will be complex and come with costs. “In some cases it might be easier to eliminate hunger than to measure it,” says Calestous Juma, a professor of international development at the Harvard Kennedy School of Government. “But in the end, we would never know what had been achieved.” And perhaps more important, we would not know how we achieve progress.

For more information on this topic:

Each year IFPRI, together with the German NGO Welthungerhilfe and the Irish aid organization Concern Worldwide, issues the Global Hunger Index, which ranks countries on a 100-point scale.

FAO’s annual assessment of global food insecurity appears in The State of Food Insecurity in the World.

The Global Food Security Index was created by the Economist Intelligence Unit and sponsored by Dupont.

Syngenta launched the Rice Bowl Index in 2012 to assess food security across Asia and the Pacific.

The Institute for Development Studies developed the Hunger and Nutrition Commitment Index to measure developing countries’ commitment to reducing hunger and undernutrition.

Save the Children UK and World Vision created the Nutrition Barometer to measure the commitment to nutrition in 36 countries.

IFPRI developed the Africa Food Security Vulnerability Indexes to help determine which countries are most vulnerable to price shocks on global markets.

The risk assessment firm Maplecroft developed a Food Security Risk Index.

In 2010, the Economic Research Service of the US Department of Agriculture published Food Security Assessment, 2010–2020.

In This Issue

Building Bigger Dreams

Can we improve people’s well-being by raising their aspirations?

Into the Spotlight

The grain teff has been consumed as a staple in Ethiopia for centuries but is little known outside the country. Now, researchers are training their attention on this understudied crop.

Untangling the Asian Enigma

Although South Asia has the highest concentration of undernutrition in the world, in the past two decades Bangladesh and Nepal have both achieved striking improvements in the nutrition of their citizens. How did they do it?

Does Money Talk?

FEATURE: Millions of poor people around the world are enrolled in safety net programs that hand out cash or food. What’s the best way to design these transfer programs?