Data in dispute

London - Two years ago, the UN Food and Agriculture Organisation (FAO) launched a petition to fight hunger with the slogan: "1 000 000 000 people live in chronic hunger and I'm mad as hell."

Since then, more than 3.4 million people, including actors, pop stars and footballers, have added their voices to the online campaign calling on governments to make the elimination of hunger their top priority.

But outrage over the "horrifying figure" of 1 billion hungry people around the world, as it was described by former FAO head Jacques Diouf, has turned to embarrassment in some quarters in the light of growing doubts about the accuracy of the number.

Many researchers say the estimate was simply too high.

"The fact that it's 1 billion is a much better story, and that's why it stays in people's minds," said Richard King, a food policy expert with Oxfam.

"It's a great number."

The controversy led the Committee on World Food Security, a top-level UN forum, to urge the FAO to overhaul its calculations using better data and methodology and to call for a set of internationally agreed food security indicators.

The first fruits are due in October, when a new estimate of the number of undernourished people will be published along with revisions for previous years as part of the FAO's annual report on food insecurity.

The figures will incorporate fresher data on world food supplies and more timely and comprehensive household consumption surveys from different countries, said Carlo Cafiero, a senior FAO statistician.

The report will also include supplemental indicators of hunger, such as the share of household budgets spent on food.

"If you only present one number, there is a tendency to over-interpret it and take it as if it were capturing everything, but we want to try and be more explicit in recognising the various dimensions of food insecurity," Cafiero said.

Nutritionists working in the field have long complained that the FAO's hunger estimates focused too narrowly on calorie intake, ignoring the bigger picture - protein, vitamin and mineral deficiencies in diets and the serious health problems they cause.

Methodology

Calculating the number of hungry people around the world at any given moment, let alone predicting how that number is likely to change in the future, is no easy task.

Models for working out how many people don't have enough to eat are not as precise or forward-looking as experts would like, partly due to lags in the release of national-level statistics.

Moreover, shifting economic conditions alter the buying power of the poor day by day, and food harvests - increasingly affected by extreme weather - fluctuate, causing price volatility.

When the FAO came under pressure to say how much hunger was increasing due to skyrocketing food prices and the global financial crisis in 2008, it decided to combine US department of agriculture projections of how economic turmoil would hurt food production, consumption and trade with its own hunger estimates of previous years, and extrapolate from there.

It estimated a "historic high" of 1.02 billion undernourished people, or around one-sixth of humanity, in 2009.

But problems emerged with the assumptions behind the number. Economic conditions did not turn out to be as disastrous as anticipated, and food production and consumption held up better than expected.

In addition, prices didn't rise as much as feared in some developing countries, like India and China, because they used export bans and subsidies to keep them down.

Finally, many people were able to maintain the amount of calories they ate by switching to cheaper foods and cutting spending on other basic needs like education and healthcare, surveys suggest.

"All evidence now is pointing to the fact that the situation was not so desperate in terms of (people's) calorie intake as, at that time, everybody thought it was," FAO's Cafiero said.

In 2010, FAO forecast a drop to 925 million undernourished people and in 2011 it didn't produce a number at all, given the dispute over its methods.

Necessary data

The question is not whether metrics are necessary, but how to collect, interpret and share the data to present a realistic and accurate picture of the food security situation.

Improving the way hunger is calculated could have far-reaching consequences for the way governments and aid agencies respond more effectively to hunger crises, experts say.

Aid groups say information from their work with local communities can contribute to a fuller picture of hunger nationally, regionally and globally, for example.

"We have a responsibility to bring the view from the field... to make sure it's not just a technical exercise, but reflects the reality on the ground," said Alberta Guerra, a Rome-based food policy officer for ActionAid.

In Nairobi's slums, when the cost of food soared in 2008, many poor urban families cut out meat and fish, went without medicine and took their children out of school.

With post-election violence making matters worse, some even stole food, scavenged in garbage dumps, brewed illegal alcohol or turned to prostitution to survive.

But the many aid agencies based in the Kenyan capital, much more used to working in rural hunger crises, didn't have a system to pinpoint when conditions for already poor slum dwellers were becoming an emergency.

"It was very difficult to get funding for urban response, partly because there were no metrics to say we are seeing a critical situation," said Lilly Schofield, research adviser with Concern Worldwide.

The organisation has since begun testing indicators to capture changes in household food security in Kenya's slums, where food has remained expensive.

Turning numbers into policy

Nyauma Nyasani, East Africa nutrition adviser for Action Against Hunger, says frequent, on-the-ground checks are far more effective at anticipating hunger problems than annual nutrition surveys.

For the past year, the aid group has been piloting a food security surveillance system in Kenya's arid northeast, based on household questionnaires conducted every three months.

And in Uganda, after a similar two-year project, it is developing national guidelines to monitor food security with the health ministry.

Funding is an obstacle. Shifting to a more responsive system will require political commitment and long-term financial resources, but rich governments and UN agencies tend to offer money on a short timeframe.

"As long as something like this is donor-driven, the sustainability becomes questionable," Nyasani said.

Ultimately, however, it is not data, but action, that makes a difference.

Saul Guerrero, evaluations adviser with Action Against Hunger, said aid workers detected warning signs months before the onset of last year's severe hunger crisis in the Horn of Africa, where some 13 million people needed food aid because of a regional drought and conflict in Somalia.

"Whoever tells you the data let us down doesn't know what they are talking about," he said. "It was the final bit that didn't work - turning data into policy.

"This is the question no one has the full answer to."