How Much Is Enough?
IT IS 7am at Kabiyet Dairies in the emerald hills of western Kenya. The dairy is five miles down an almost impassable track, and you would think milk would turn to butter long before it arrives. Yet the place is heaving with farmers waiting for their produce to be tested, carrying it in pails on trucks, on the backs of motorbikes or on their heads. The dairy opened only 18 months ago and may seem basic, yet it has just struck a deal to sell milk to an international processing plant in Nairobi. Farmers get 26 shillings a litre, more than twice what they were paid before the dairy opened its doors.
Laban Talam, a 30-year-old villager, has a smile on his face. He farms just under a hectare on a hillside overlooking the dairy. Two years ago he was scratching a living, supplementing his earnings from one cow, a native longhorn, with odd jobs outside farming. Now he has five cows, three of them Holsteins who give twice as much milk as the native breed. He rents extra land from his neighbour, has rebuilt his house, grows pineapples for export and has installed a biomass pump. His children go to a private school.
Kabiyet Dairies is only one agricultural success story among many. Brazil, by investing heavily in research, has turned itself into the first tropical farm giant, joining the ranks of the temperate-food superpowers such as America, Europe and Canada. It did so in a single generation, thanks mainly to big commercial farms. Vietnam, through policy changes (especially freeing up small-scale private agriculture), turned itself from one of the world’s largest importers of rice into the world’s second-largest exporter.
So it is possible to grow more food, more efficiently, on both a regional and a national scale. But can it be done on a global scale, which is what is needed to feed 9 billion people? If so, how?
Because of the constraints described in the previous section, there will not be big gains in food production from taking in new land, using more irrigation or putting more fertiliser on existing fields. Cutting waste could make a difference, but there are limits. The main gains will have to come in three ways: from narrowing the gap between the worst and best producers; from spreading the so-called “livestock revolution”; and—above all—from taking advantage of new plant technologies.
The huge gap between the best and worst producers in roughly comparable farming areas shows the scope for improvements. Both eastern and western Europe are good for growing wheat. Yet west European farmers achieve yields of up to 9 tonnes per hectare, whereas east European ones get just 2-4 tonnes. The discrepancy is much wider than differences in incomes or soil quality might suggest.
Or take the example of maize seed. According to Pioneer, a big seed company, central Ghana has some of the best maize land in the world, yet only 3% of the country’s seed is the hybrid kind that can take full advantage of it. In contrast, Brazilian land is less good, but 90% of its seeds are hybrid ones. The country is now the world’s third-largest exporter. If Ghana bought more hybrid seeds, it could presumably achieve something closer to Brazilian yields.
Why don’t the laggards catch up? A good place to look for an answer is Africa, the part of the world that has most conspicuously failed to feed itself over the past 50 years. Five years ago, says Joe deVries, head of crop research at the Alliance for a Green Revolution in Africa (AGRA), the big problems in Africa were prices and investment. Farmers were getting too little for their produce and no one was doing any research into “African crops” such as sorghum and cassava. Now prices are higher—a benefit to producers, at least—and the “African crop” problem is being solved. New semi-dwarf sorghum has three times the previous yield, and genetic research has shown how to control cassava’s great scourge, viral disease.
The problem now is to get those improved seeds to the farmers. Around Kabiyet, Western Seed Company, a small outfit that develops its own varieties of maize for smallholders, doubled production in 2010 and still sold out two months early. It is one of 45 seed companies set up with AGRA’s backing, and Mr deVries reckons they will need 100 to meet prospective demand. At present only 10% of Kenya’s farmers are using new seeds, but Mr deVries hopes that by 2015 the figure will have risen to half.
When India began its Green Revolution in the 1960s, it had 388km of paved roads per 1,000 sq km of land, and only about a quarter of its farmland was irrigated. Ethiopia now has just 39km of roads per 1,000 sq km, and less than 4% of its land is irrigated. So the remaining problems in Africa are vast. Moreover, says Don Larson of the World Bank, farming in that continent is intrinsically harder to change than in east Asia because it is more varied. In east Asia, if you invent an improved rice variety, every farmer for hundreds of miles around can use it because the land and climate are much the same. In Africa, soil and climatic conditions are much more diverse and farmers a few hundred yards apart may need different seeds.
But better technology is removing some of the barriers. Since 2008 African food production per person has been rising for the first time in decades. Rwanda and Malawi have begun to export food (admittedly in Malawi’s case thanks to massive and unaffordable fertiliser subsidies). For now Africa is still a net food importer, but a recent Harvard study for the presidents of East African countries argued that it could feed itself in a generation. Even if that proves optimistic, Africa could surely increase food production by more than 1.5% a year. “We didn’t know how the Green Revolution would come to Africa,” says Mr deVries. “Now we do.”
The second main source of growth will consist of spreading a tried and tested success: the “livestock revolution”. This consists of switching from traditional, open-air methods of animal husbandry, in which chickens and pigs scratch and root around the farm, eating insects, scraps and all sorts of organic waste, to closed “battery” systems, in which animals are confined to cages and have their diet, health and movement rigorously controlled. This entails huge losses in animal welfare, and European consumers are reacting against the system. But there are also gains in productivity and sometimes even in welfare, by reducing losses from diseases and predators that in traditional systems can be distressingly high.
Improving livestock farming is important because of meat’s growing share in the world’s diet. Meat consumption in China more than doubled in 1980-2005, to 50kg a year per person. Between now and 2050, meat’s share of calories will rise from 7% to 9%, says the FAO; the share of dairy produce and eggs will rise more.
Livestock matters for many reasons. It provides financial security in poor countries, where herds are often a family’s savings. It can affect people’s health: new infectious diseases are appearing at the rate of three or four a year, and three-quarters of them can be traced to animals, domestic and wild. Avian flu is just one example. Livestock also plays a part in global warming. Much of the methane in the atmosphere—one of the worst greenhouse gases—comes from cattle belching.
Since the 1980s livestock production has far outstripped that of cereals. World meat output more than doubled between 1980 and 2007. Production of eggs rose from 27m tonnes to 68m over the same period. Some countries have done better still. India has the world’s largest dairy herd. Its milk production trebled, to 103m tonnes, over a period when global milk output increased by half. Brazil increased its production of chickens fivefold in 1987-2007 to become the world’s largest exporter. Most spectacularly, China raised it output of both eggs and milk tenfold.
For sheer efficiency, there is little question that battery systems do a better job than traditional methods. A free-range hen scratching around might lay one or two eggs a week. Feeding her costs nothing, giving a net gain of 50-100 eggs a year. A battery chicken will lay six eggs a week. She might cost the equivalent of 150 eggs to feed, producing an annual net gain of 150 eggs. And selective breeding has made her more economic to keep. Battery chickens used to need 4kg of feed for 1kg of eggs; now they need only 2kg.
Moreover, it is almost impossible to scale up a farmyard operation: there are only so many insects to eat, and so many hens one family can look after. And to breed the most productive hens which convert their feed most efficiently into eggs and are most resistant to disease, you need large flocks.
So there are two reasons for thinking that the livestock revolution will continue. One is that some countries still lag behind. An example, surprisingly, is Brazil, which has just one head of cattle per hectare—an unusually low number even for a country with so much land. Roberto Giannetti da Fonseca, of the São Paulo industry federation, says Brazil should be able at least to double that number—which could mean either doubling beef production or using half the area to produce the same amount.
Carlos Sere of the International Livestock Research Institute thinks traditional systems could borrow some of the methods of closed battery-farm systems—notably better feeding (giving a small amount of animal feed makes a big difference to the weight of range-land cattle) and the introduction of new breeds for better yields (as Kabiyet did by switching from longhorn to Holstein cattle).
The second reason for expecting further gains is that recent genetic analysis could improve breeding dramatically. About a third of the livestock revolution has come about through selecting and breeding the best animals. Another third comes from improved feeding and the remainder from better disease control. In the 1940s and 1950s breeding relied on the careful recording of every animal in the herd or flock; in the 1970s on artificial insemination by the best sires; and in the 1980s on embryo transfers from the best females into ordinary breeding animals.
New genetic analysis now promises to bring in another stage, says the FAO’s Henning Steinfeld. It allows breeders to select traits more precisely and thus speeds up breeding by reducing generational intervals: if you know which genetic traits an animal has, there is no need to wait several generations to see how things turn out.
This will not happen everywhere. Europeans and—to some extent—Americans are increasingly influenced by welfare concerns. They jib at confining animals. The European Union has banned certain kinds of cages, and California is following suit. But, so far, people in emerging markets, where demand for meat and animal products is growing fast, are less concerned about such things, so the next stage of the livestock revolution will mainly be concentrated there.
GM and after
This will make some difference but the change likely to generate the biggest yield gains in the food business—perhaps 1.5-2% a year—is the development of “marker-assisted breeding”—in other words, genetic marking and selection in plants, which includes genetically modifying them but also involves a range of other techniques. This is the third and most important source of growth.
“Until recently we knew little about how plants function, how they perceive heat and cold, how they flower, and so on,” says Caroline Dean of Britain’s John Innes Centre. That is changing, thanks to greater understanding of plant genetics as well as to a dramatic fall in the costs of gathering genetic information.
Ms Dean has worked out, for example, how plants “remember” the length of time winter has been going on and do not therefore mistake a mild spell in January for spring. The answer, it seems, is by turning off a gene after a certain period of cold weather. This process is finely adjusted, so in Sweden the plant switches off the gene later than one of the same species in southern England.
At the same time the price tag on gathering genetic data is now much lower than it used to be. Gary Atlin, a maize breeder at CIMMYT, reckons that whereas a couple of years ago the cost of identifying a single gene in a single plant was $2, now it is about 15 cents. That is still too high for many breeding operations, but work now being done jointly by CIMMYT and Cornell University should cut this to $30 per million genes. “This is Moore’s law for plants,” says Mr Atlin, referring to the rough rule that computing power doubles every two years for the same price. The cost of genetic identification will soon stop being a serious constraint.
The public debate on plant genetics focuses almost entirely on the pros and cons (mostly cons) of genetic modification—putting a gene from one species into another. A gene from a soil bacterium, Bacillus thuringiensis, for example, when spliced into maize, makes the plant resistant to herbicides; this enables farmers to plant maize, spray the crop with a weedkiller and end up with a field of nothing but maize. In Europe it is illegal to plant such maize. The biggest advantage of genetic selection, however, is probably not that it makes it possible to grow transgenic crops (“Frankenfoods”), but that it allows faster and more precise breeding.
Imagine the genetic material of plants as a vast library, with billions of books. This library has no catalogue, and none of the books has an index or table of contents. It is still possible to discover what is in the library by reading every volume. That is roughly what plant breeders have done in the past, painstakingly planting hundreds of varieties of a single species and discovering traits by breeding numerous generations from them.
Genetic marking is the equivalent of giving every book a title, table of contents and index—and with much greater speed and accuracy than any librarian could manage. Monsanto has a “corn chipper” which takes a small amount of genetic material and generates a DNA profile of hundreds of maize seeds simultaneously in seconds. It leaves the seed alive, so breeders, having mined the computer data from this and every other seed in Monsanto’s vast library, can go back to a seed they like and breed from it. It is possible literally to find one plant in a billion.
Such gains are likely to snowball. In 1997 Monsanto introduced a variety of corn resistant to various pests. It fully controlled four of 15 common “above-ground” corn pests like corn borers, cutworms and stinkbugs, and partially suppressed three more. In 2004 the company introduced a successor that controlled nine of the 15 above-ground pests and seven of the eight that lived in the soil. The 2010 version controlled nine above-ground pests and seven in the soil, and suppressed three more.
At the moment the genetic evolution is just beginning. The genomes of most important crops have been sequenced only fairly recently, and that of wheat is only partly done. There are only a handful of genetically modified crops. Commercial firms have concentrated most of their efforts on only one or two traits controlled by individual genes, such as disease resistance. But the future, argues Giles Oldroyd of the John Innes Centre, lies in traits controlled by multiple genes and genetic “pathways”, that is, interactions between groups of genes.
The most important of these is yield. Over the next 40 years yields need to rise by around 1.5% a year to feed mankind adequately. Maize, which has had by far the most genetic research, is the only crop whose yield is growing by more than that. If genetic selection can be extended to wheat, rice and soyabeans, that should go a long way to feeding the world by 2050.