How much is enough?

Hooked on meat

IN HIS 1981 essay, “Poverty and Famines”, Amartya Sen, an Indian economist, argued that the 1943 Bengal famine, in which 3m people died, was not caused by any exceptional fall in the harvest and pointed out that food was still being exported from the state while millions perished. He concluded that the main reason for famines is not a shortage of basic food. Other factors—wages, distribution, even democracy—matter more.

In 1996 the United Nations' Food and Agriculture Organisation (FAO) estimated that the world was producing enough food to provide every man, woman and child with 2,700 calories a day, several hundred more than most adults are thought to need (around 2,100 a day). The Lancet, a medical journal, reckons people need no more than 90 grammes of meat a day. On average they eat more than that now. As Abhijit Banerjee of the Massachusetts Institute of Technology says, “we live in a world that is capable of feeding every person that lives on the planet.”

Indeed, the world produces more than just enough to go round. Allowing for all the food that could be eaten but is turned into biofuels, and the staggering amounts wasted on the way, farmers are already producing much more than is required—more than twice the minimum nutritional needs by some measures. If there is a food problem, it does not look like a technical or biological one.

So why worry about producing more food? Part of the answer is prices. If output falls below demand, prices will tend to rise, even if “excess” calories are being produced. That happened in 2007-08, and is happening again now. Over the past four years prices have been more volatile than they have been for decades. This is bad for farmers (who are left not knowing how and where to invest) and worse for consumers, especially the poor, who risk suddenly being unable to afford basic food.

Another part of the answer is that it is hard to improve distribution and reduce poverty. The world may indeed be growing masses of calories. But the food is not where it needs to be, and biofuel policy is hard to shift (see article). Pushing up supplies may be easier than solving the distribution problems.

But it will still be a daunting task. On one reckoning, in order to keep up with population growth farmers will have to grow more wheat and maize over the next 40 years than was grown in the previous 500. The balance between what is consumed and what farmers produce matters a great deal.

So how do you keep that balance? Start with consumption, the side of the equation that can be forecast with some accuracy. The forecast rise in world's population, from just under 7 billion at the start of 2011 to just over 9 billion in 2050, is the equivalent of two extra Indias. If you include the 1 billion people who are now going hungry, the additional mouths to feed over the next 40 years add up to three extra Indias.

It is not an impossible task. The increase in world population by 2050 will be around 30%, less than in the 40 years to 2010, when it rose by over 80%. Consumption of wheat, rice and maize roughly tracks population growth but at a higher level, so demand for them will add about a billion tonnes to the 2 billion produced in 2005-07. That is much less than during the previous 40 years, when cereal production rose by 250%.

True, the headline numbers somewhat underestimate the problem. Families have been getting smaller for decades, so there are proportionately fewer children in developing countries than there used to be. Many of those countries are benefiting from a “demographic dividend”: an unusually large proportion of young adults in the population, who work hard but eat more than children or older people.

Rise of the carnivores

Moreover, an increasing proportion of the population is living in cities, and dollar for dollar city-dwellers eat more food and especially more processed foods than their country cousins. They also tend to be richer and able to afford pricier food, such as meat. So meat demand will rise strongly. In 2000, 56% of all the calories consumed in developing countries were provided by cereals and 20% by meat, dairy and vegetable oils. By 2050, the FAO thinks, the contribution of cereals will have dropped to 46% and that of meat, dairy and fats will have risen to 29%. To match that soaring demand, meat production will need to increase to 470m tonnes by 2050, almost double its current level. Output of soyabeans (most of which are fed to animals) will more than double, to 515m tonnes.

Overall, the FAO reckons, total demand for food will rise about 70% in the 44 years from 2006 to 2050, more than twice as much as demand for cereals. But that is still less than half as much as the rise in food production in the 44 years from 1962 to 2006. So according to the FAO's Kostas Stamoulis, producing enough food to feed the world in the next four decades should be easier than in the previous four.

Should be, but probably won't be. Increasing food supplies by 70% in the next 40 years may prove harder than it was to raise them by 150% in the previous 40. The main reason: problems with yields.

Yield—tonnes per hectare, bushels per acre or whatever—is the traditional gauge of agriculture's performance. And the growth in yields has been slowing down, from about 3% a year for staple crops in the 1960s to around 1% now.

Yield curb

The earlier period was that of the Green Revolution, an exceptional time. Thomas Lumpkin, the head of CIMMYT, the UN's international wheat and maize research organisation, thinks that farmers in developing countries could often double their harvest by switching to Green Revolution seeds (many of which were developed at CIMMYT by the organisation's most eminent plant breeder, Norman Borlaug). Now, Mr Lumpkin reckons, the best current technologies could perhaps increase yields by 50%—still a lot, but not as spectacular as the earlier improvements. The low-hanging fruit has been plucked and eaten.

The Green Revolution threw resources at plant-breeding, which worked brilliantly. The new seeds enabled grains to absorb more fertiliser and water. But now there is not a lot more water to spare, and fertiliser usage in some places has already passed saturation point (see article), so a new Green Revolution will have to make even more efficient use of existing resources. The next 40 years will also have to deal with the potentially profound damage to farming from climate change, which in some parts of the world could reduce yields by one-third.

And disturbingly, for the first time since the Green Revolution, crop yields are growing more slowly than population (see chart 4). To be more exact: growth in population and demand for food have both slowed down, but crop yields have slowed more. Between 1961 and 1990 wheat yields were rising at nearly 3% a year. During that period the world's population was growing by an average of 1.8% a year. Between 1990 and 2007 population growth slowed down to 1.4%, but the rise in annual wheat yields slackened to 0.5%. The growth in rice yields between the two periods halved. Yields of mankind's two most important crops are now almost flat.

Some argue that these figures are not as worrying as they seem. After all, population growth is slowing and yields of some crops, notably maize, are still rising at a steady pace. And the growth in yields may have slowed not because agricultural technology has hit a wall but because farmers are cutting inputs for environmental reasons, or because they are focusing on quality more than quantity. Indeed, in America farm output is rising but the use of fertilisers and other inputs has been cut back. Breeders have lately been working on wheat for extra protein, not just yield.

If this is the correct explanation, farmers' overall productivity is still increasing, since they are using fewer inputs to get the same output. And that is what some researchers find. Keith Fuglie of the United States Department of Agriculture reckons that total factor productivity in world agriculture—a measure which includes capital, labour and other inputs—is still rising at a healthy 1.4% a year. This reflects a more efficient use of resources. And if farmers are choosing to reduce yields now, they could also push them back up again later.

Other researchers, however, think global productivity is indeed slowing down, especially outside China. According to a study using different definitions from Mr Fuglie's, growth in land productivity fell by over one-third between 1961-90 and 1990-2005, and growth in labour productivity fell by two-thirds. And as Mr Fuglie says, even if productivity is rising, it needs to rise more, from an annual gain of 1.4% to 1.75%, he thinks—a big leap. And though farmers might choose to increase yields later, their choice would depend partly on food prices rising more than prices of inputs such as fertilisers, which they may not (in 2007-08 fertiliser prices rose much more than food prices). So even if productivity is increasing—and that is not clear—on its own it is not enough.

And what if the slower rise in yields reflects something more fundamental, the approach of some sort of biological limit in plants? The worry is not that yields are flattening out in farmers' fields, where agronomic practices or the weather or any number of things may be responsible. It is that there may be a problem in breeders' fields where the potential of plants is tested. This possibility is controversial and many breeders reject it. But the idea should not be dismissed out of hand.

A fancier kind of maize

The Green Revolution had little to do with making plants bigger: rather, it produced higher yields by persuading more plants to grow in the same space and by getting them to put less effort into growing stalks and leaves and more into seedpods, the part people eat. The nagging fear is that both trends may be reaching a limit.

The number of maize plants in a hectare has risen from roughly 40,000 to 90,000 in the past half-century. There must come a point where plants can no longer be sardined any closer together, stalk crushed against stalk. Similarly, at some point it may no longer be possible to persuade plants to put ever more energy into seeds. In animals biological limits are already clear: turkeys are so bloated that they can no longer walk; chickens grow so quickly that they suffer stress fractures. Racehorse speeds have levelled out.

The evidence on whether plants are reaching similar limits is mixed. On the one hand, maize yields in rich countries are continuing to rise, thanks to huge investments by seed companies. A recent Australian study found cereal yields in India, Britain, and Australia are increasing by about 1% a year. Britain's government forecast in 2009 that wheat yields there would rise from 7.7 tonnes per hectare to 11.4 tonnes in 2025 and 13 tonnes in 2050.

On the other hand, trials at CIMMYT's principal wheat-breeding station at Obregon in Mexico indicate a slowdown in potential yield growth to only 0.4% a year. It is also notable that the latest research into plants is focused not on redistributing growth towards seeds but on making the whole plant bigger.

Yields may still be growing, but more slowly, and even that slower growth can no longer be taken for granted. The question is, how much do they need to grow to keep abreast of the growth in the world's population of about 1.2% a year?