How Much Land to Feed a Person?

I've just been reading this post on Naked Capitalism, which raised the question of how much land you need to feed a person. In the past I've seen numbers in the range of 1 - 2 acres per person, but do you really need that much?

I think working out the answer is a bit complicated, but I thought I'd start with something simple. What yield would a 1 acre apple orchard give in the UK? Apple trees on MM106 are supposed to yield 30 - 60 kg a year on average when fully grown, and be about 4m wide. That means that one tree covers 12.57 m2.

Now, an orchard doesn't have a closed canopy, so lets say that the canopy area is 50% of the ground area. That means that in an acre you can fit 159 trees, with a yield between 4.8 and 9.5 metric tonnes. That is 13 - 26 kg of apples a day, and each kg contains about 10 apples. Of course, you can't eat the apples evenly through the year without some method of preservation, but let's pretend that you can.

Next, an apple contains about 50 calories. 13 kg of apples is therefore 13 * 10 * 50 = 6500 calories. The average intake for a man should be around 2500, so this is far more than required. If we only care about energy, and ignore all the other nutrients people need, then our hypothetical man in fact needs at most 0.4 acres, and perhaps as little as 0.2 acres.

Of course, this is a massive simplification, for a few reasons. These are:

1. No-one would try to survive by eating 5 kg of apples a day
2. This assumes that the yield can be consumed evenly throughout the year
3. It assumes that there is no spoilage in storage
4. It assumes that there is no need to worry about variations in yield due to weather etc.
5. It assumes that all that matters is energy, and not other nutrients
6. It assumes that there is no need for additional land to grow fertility / mulch materials

But allowing for some pessimism, land to grow a bit of food for chickens (meat/eggs) etc., 1 acre per person for a good diet does seem like it might be about the right answer for the UK.

Another way to approach the problem is to divide the total amount of agricultural land in the world by the number of people (it doesn't work for individual countries due to food imports, but luckily the world as a whole doesn't import any food) . This gives an answer of 1.8 acres per person, based on numbers from Wikipedia. But again, this answer is imperfect for any particular place because of wide variations in climate across the globe, and because the world's land isn't used to produce a healthy, sustainable diet in the most efficient way possible. This is obvious since on the one hand we have overconsumption of luxury foods like meat in much of the world, and hunger in much of the rest. In addition, the current agricultural system relies on massive injections of mined / artificial nutrients, so a sustainable system could actually require more than 1.8 acres per person at a global level.

So what's the right number? A minimum is probably at least 0.5 acres per person. A realistic number with a bit of meat/eggs is probably more than 1 acre per person. How much more I'm not sure.

Rational behaviour is maximising ... what?

Utility maximisation

In economics, rational actors are assumed to maximise their utility. However, the fact that world is uncertain means that you can't always know what would maximise your utility, so this is instead amended to expected utility. This means: a rational actor makes the decision that on average produces the highest utility. Many people then take a step further and move from utility to money: rational individuals make decisions that maximise their expected worth.

This assumption also feeds through into a lot of financial modelling and optimisation. For example, valuations using methods such as dynamic programming typically assume that decision making maximises expected utility and nothing else.

But... does this make any sense? To see why it might not, consider what you would do if a billionaire offered you a choice. If you accept, he will flip fair coin. If it comes up heads, he will give you a million pounds, and if it comes up tails you will give him all your assets. Now let's consider the expected value here. If your assets have value A, and you take the bet, then your expected financial worth after the bet is:

0.5 x (A + 1,000,000) + 0.5 x 0 = 0.5 x A + 500,000

Obviously, if you don't take the bet then your assets will continue to be worth A. Therefore, the expected improvement in your financial position if you take the bet is:

(0.5 x A + 500,000) - A = 500,000 - 0.5 x A

So if your assets are worth less than 1,000,000 then you should take the bet if you are maximising your expected worth. What's more, the less your assets are worth, the stronger the expected benefit of the bet, and therefore the more "rational" it would be to take it as an maximiser of the expected.

But do people behave this way? And should they? This bet would mean, for a middle income home owner, that they have a 50% chance of becoming a millionaire, and a 50% chance of becoming a homeless pauper. I think a significant proportion of people would not risk becoming homeless even for the chance of becoming a millionaire. It has been demonstrated experimentally that people are strongly averse to losing what they have, even when there are significant rewards to taking the risk.

As to whether they should... well, despite the fact that many economists equate rational with maximising the mean, it does not seem irrational at all to decide not to risk homelessness. In fact, it seems pretty rational to protect your access to necessities before trying to get the nice-to-haves. But this kind of maximising the minimum, rather than the mean, behaviour is too often ignored.

Agricultural modernisation
 
The same phenomenon crops up in traditional agriculture. New high yield varieties of major crops have slowly spread through the world, but in traditional communities there was a lot of resistance to their use. Why? Don't those peasants want to improve their lives? Modern agricultural experts tended, I think, to see peasants in poor countries as backwards, since the new high yield varieties would make them more money.

But it is seeing yield from a commercial perspective that makes you miss why standardisation on high yield varieties was resisted. From the point of view of the subsistence peasant, the product of agriculture is not just food to sell, but survival into the next year. For this reason, an extremely high value was place on reliability, and the ability to reliably produce enough to survive. The natural consequence of this was:

1. to grow crops that could tolerate tough conditions and still produce something
2. to grow diverse crops so that if one failed there would be another source of food

In other words, the backwards peasant was making the perfectly rational decision to maximise the minimum yield instead of to maximise the average yield. Peasants and subsistence farmers are conservative because their system, while it might leave them hungry, at least minimises the risk of widespread starvation. Any change is risk, and risk is what they are trying to minimise.

The book below, which I would like to read from cover to cover one day, discusses some of these issues in more detail:

A book partially about how peasants maximise the minimum

Extreme levels of Roundup in Soybeans

How much of this ends up in your food?

Following on from my previous post on genetic engineering for pesticide or herbicide production/resistance, I found this article via Naked Capitalism. The article is about farmers using ever higher doses of Roundup (a herbicide) to fight herbicide-resistant weeds, and about the ingredients of the herbicide accumulating in crops. To quote from the article:

Seven out of the 10 GM-soy samples we tested, however, surpassed this "extreme level" (of glyphosate + AMPA), indicating a trend towards higher residue levels. The increasing use of glyphosate on US Roundup Ready soybeans has been documented (Benbrook 2012). The explanation for this increase is the appearance of glyphosate-tolerant weeds (Shaner et al. 2012) to which farmers are responding with increased doses and more applications.

 Regulatory agencies have responded to increasing usage not be intervening or even investigating its potential health effects, but by raising the legal limits to align with the current practice of farmers. God forbid that the precautionary principle is applied to public safety if it costs money.

The authors have studied the impact of Roundup on a kind of water flea, and their findings were as follows:

Our own recent study in the model organism Daphnia magna demonstrated that chronic exposure to glyphosate and a commercial formulation of Roundup resulted in negative effects on several life-history traits, in particular reproductive aberrations like reduced fecundity and increased abortion rate, at environmental concentrations of 0.45-1.35 mg/liter (active ingredient), i.e. below accepted environmental tolerance limits set in the US (0.7 mg/liter) (Cuhra et al. 2013). A reduced body size of juveniles was even observed at an exposure to Roundup at 0.05 mg/liter.

Water fleas are very different to human beings, but it does suggest that there might be cause for concern. And if there are the same insidious effects in humans, noticing it will take a while - if something makes people drop dead immediately then it's pretty obvious, but if a chemical has a more subtle effect like reducing human fertility and/or increasing the rate of miscarriage then proving it can take years or decades. Think how many chemicals were widely used in the past but are now banned as unsafe, simply because their danger was more subtle than immediate death.

Of course, many plants we eat do contain chemicals which are hazardous in large doses, so in a sense this is nothing new. But for most of human history, diets were varied enough that exposure to individual toxic chemicals was typically low. However, industrialised agriculture has meant that people now mostly eat a small number of highly productive crops, raised in a similar way. This means that exposure to these major crops is massive, and exposure to any toxins in them is also large. The same kind of market concentration applies to herbicides, so if glyphosate (the active component of Roundup) does turn out to be dangerous to human health, so will a very diverse range of crops that it's been used on or with. A range that is almost certain to increase as the Monsantos of the world engineer resistance to it into more and more crops.

If I had a choice, I would prefer to avoid foods containing high levels of herbicides or pesticides. But I don't have a choice, really - the level of such chemicals is not on the packet, so how can I really know? The only way I can control my own exposure is to grow all my own food, which is not really possible even in a large urban garden. So I'm therefore forced to trust regulators to protect me - the same regulators who have been taken over by those they're supposed to be regulating.

I suppose the only bright side is that so far GM foods have been resisted more successfully in Europe than the US, no thanks to my own government in Westminister. If GM food ends up being herbicide resistant crops, rather than, say, higher yielding crops, then I hope that GM continues to lose the battle for Europe.

Genetic engineering fails to overcome the flaws of monoculture

I saw an interesting story on the Independent today, claiming that pests are developing resistance to the Bt proteins genetically engineered into some crops. Here is the link:

Worm evolves to eat corn that was genetically engineered to kill it

Although the Independent has gone downhill over the last few years and is rapidly descending into red-top rag territory, I don't see any reason to doubt the story, since evolved resistance is what you'd expect to happen over time.

The reason I think this story is interesting is that it highlights the downsides of misapplying our understanding of genetics. I am not necessarily against genetic engineering for some purposes - fixing genetic diseases in people has my complete support, and introducing genes for some traits into food would be OK too. I have no problem with anyone trying to genetically engineer bigger tastier potatoes for example.

An example of corn monoculture

The problem is when genetic engineering is used to promote and enable the "exterminate everything else" monoculture approach of modern farming. This is doomed to fail because ultimately the pests don't really have anywhere else to go. If you give a plant better disease resistance in a varied ecosystem, the pests will move onto other plants, but if you only grow one plant across wide areas, then the pests must adapt or die. And there are good reasons to expect that adaptation should generally be possible, such as the following:

1. The plant itself must tolerate any toxic chemicals it contains or that are applied, so the chemical cannot be inimical to basic life processes without killing the desired crop

2. Some part of the plant will be consumed by humans, and therefore must also be non-toxic to humans.

3. Plants have been engaged in chemical warfare against pests for as long as both have existed, and yet pests still exist

4. Many plants depend on a supporting web of other life, from pollinators to soil micro-organisms, and in this case any chemical weapon must be selective. Pests may be able to copy non-affected organisms.

The attempt to wipe out all the pests while at the same time retaining a non-harmful, edible plant appears to me to be a war that we simply can't win. The best that can be attained is the fluctuating stalemate which has held throughout history, with one side gaining a temporary advantage for a while which is then neutralised by changes in predator populations or changes in the plants themselves.

Genetic engineering plants to either contain or tolerate poisons is just a risky way to gain a temporary advantage, since there is a high risk of these poisons having adverse effects either on other parts of the ecosystem (e.g. pest predators, beneficial soil microbes, ...) or even on people themselves. How many chemicals were used for decades in the belief that they were safe, only for scientists to later show that they increased the risk of cancer or some other health problem?

A much better approach to pest control is:

1. diversify away from monoculture - give the pests more targets so they either can't specialise, or if they do they cannot wipe out your entire crop

2. Encourage predator populations to maintain a limited but constant stock of pests instead of insisting on complete extermination. This means providing predator and beneficial species with habitat and uncultivated spaces.

3. Use limited and targeted chemicals only when necessary, accepting that they do harm as well as good

But this won't happen as long as the model is industrial monoculture agriculture, where the goal is to minimise the human labour in farming, and to de-skill and automate the process as much as possible.