The Economist gets some well-deserved derision these days, but it still delivers lots of interesting data, illustrated by graphs that are usually well designed and informative. Via Kenny Easwaran I found this table (published by EconomistDailyChart, but I haven’t yet located the chart) of annual meat consumption per person by country. The data set has plenty of anomalous features, but looks accurate enough for my purposes.
I’ve previously argued that we can feed the world if we make the right choices. . More precisely, our current food system produces more per person than is needed for adequate nutrition, and can continue do so in future if the right policy choices are made. The key problem is distribution, not production.
But the meat consumption data leads me to a more surprising conclusion. Using current technology and with no additional diversion of food grain, the world could produce enough meet to give everyone an intake comparable to that of the average person in the Netherlands[fn1].
Here are the numbers we need to start with from the data table. Current average annual world meat consumption is 9.5 kg beef, 15kg pork and 12.5kg chicken for a total of 37kg per person per year. Netherlands average is 70 kg.
Each kg of grain-fed beef requires about 8kg of grain, compared to 2kg for chicken, and the trade-off similar when cattle are pastured on land that could be used for grain. So, 5kg of beef could be replaced by 20 kg of chicken.
The other main user of grain (apart from human consumption) is ethanol production which now takes something like 140 million tonnes a year. Fed to chickens that would produce around 70 million tonnes or 10kg per person per year.
That would give an average of 62kg per person per year, not far below the Dutch average. To fill the remaining gap, I’ll call on the usual suspects, reductions in inefficiency and waste.
The reduction in methane emissions from cattle would almost certainly outweigh any adverse impact from reduced ethanol production (numbers on both of these effects vary so wildly that I’m not going to attempt a calculation for now).
How feasible is all this? The use of food grain for biofuels is discredited as a policy, and even the US Congress has withdrawn some support. The shift towards chicken makes economic sense, and would be accelerated if carbon pricing were applied to agriculture, which might well happen in the next couple of decades. So, world meat production could increase steadily over the next few decades, well ahead of population growth.
That still leaves the crucial problem of distribution. People in some rich countries, notably the US and Australia consume much more than the Netherlands, and that the billion or so poorest people in the world can’t afford enough grain to eat, let alone meat. Until this changes, increasing average meat production isn’t going to solve the problem. 
There’s no real answer to this within the current world order, except to wait for poor people to become richer, as they have done in much of South-East Asia and are now doing, in large numbers, in China and India.
But a large part of my reason for doing exercises like this one is to consider the feasibility of a better world, even if it might be considered utopian at present. The ability of the world to feed itself, and to do so with a diet that should satisfy any reasonable person, is an important precondition. Until recently it has not been met – the total food output of the world has been barely adequate in normal times, and quite inadequate in famine years. But now, as I’ve argued it’s entirely possible.
fn1. I’ve picked the Dutch because they are supposed to be the tallest people in the world, which implies an adequate diet.
fn2. Even in a world where everyone had enough, substantial differences would persist. For example according to the data in the table, meat consumption (I’m not sure if they have a good handle on fish) in Japan is very low by developed country standards, and obviously this reflects preferences and national policies, rather than poverty.