Figure 1
Summary: This page deals with food at 'the aggregate world level'. We take issue with the optimists, especially Tim Dyson [], Donald Mitchell [], and Nikos Alexandratos []. We think that they fail to take proper account of the recent 30 year trend in falling increases in grain yields. They assume (a) that these yield increases are not going to decrease any further, an (b) that they are going to get larger. This optimism is, we think, both unjustified and dangerous.Lady M What is your 'agenda'?* What is that of your opponents?Those who wish can skip to the critical Figure 8 at the end of this page. Meanwhile Lady M has a few questions to ask.
MHK. There is no such thing as a 'naked fact'. We all look at 'the facts' from a particular point of view.
Our major concern is with local demographic entrapment, which we see as the greatest problem in public health. If we can show that the world as a whole is demographically trapped, this is a powerful argument for lifting the Hardinian taboo and promoting disentrapment locally - there will be botha local and a global case for a 1-child world. In this respect local and global disentrapment reinforce one another.
The concern of most social scientists, is not to rock the political boat.
Lady M. What do you mean by 'Global demographic entrapment ???
MHK. It is a strict parallel with local entrapment. If we humans exceed the carrying capacity of our earth, we cannot migrate to Mars. Anyway, you don't want us! Globally, we have no hope of any exports, or therefore of any imports - except photons of light energy from the sun, upon which we depend absolutely. So if we exceed the carrying capacity of our earth, or we are projected to do so, we are indeed demographically trapped.
Lady M. Are you going to exceed earth's carrying capacity?
MHK. Pessimists, like myself think we are. Optimists think that 'science will fix it', and there is no need to worry.Lady M. Then what is the evidence?
MHK. We eat many foods, but much the most important one is grain of various kinds. This provides about 60% of our calories, either directly, or by eating animals which have been fed on grain. FAO have data on grain production going back to about 1950. The data are not perfect, but they are much the best food data we have. Also other foods behave in much the same way - or worse. Our fish catch, for example, plateaued out some time ago. []
The data for global grain are worrying. The 'grain-land' in the world is approximately constant. The area built on or made into roads or golf courses, etc. (usually the best land) is about balanced by the forests cut down (usually less good land). Since our population is going up, there is less and less grain-land for us each year.
So increasing numbers of people have to be fed by increasing yields - more kilos of grain from each hectare each year. Here is a figure from Dyson in which he nicely balances the scales so that yields in kilos per head go up equally as grain-land area per head goes down.
Figure 1
MHK. Grain production rose faster than people until about 1985. This was partly because there was still more land to plough in the 1950s and 1960s, and because grain yields grew steadily, and increasingly fast - until the middle 1960s. Then from the middle 1960s, although yields are still increasing, they have been increasing more and more slowly.
There then came a point, about 1985, when yields were increasing less fast than population. Since then grain per head has been falling. Each year there is less and less grain for each of us - on average.
Here is Tim Dyson's graphs to show world grain production per head in recent years:Figure 2. (After Tim Dyson)
The key question's are:
3. How much are we all going to share
the grain that the earth produces?
Let's think first about yields. In 1996 Tim Dyson published a splendid book: Population and Food, global trends and future prospects.[] It is guardedly optimistic - what he calls 'tempered hope'. The colours are mine:
Underpinning this general progress there has ben a steady rise in food crop yields. At the aggregate world level there is no reason to be particularly alarmed by recent trends (though, clearly, things could be better if yield increments were greater). Importantly, we have seen that yields tend to increase 'arithmetically' rather than 'geometrically'. In most regions recent increments in average cereal yields per hectare compare favourably with previous experience. It equally important to appreciate that raising yields is an immensely complicated and multifaceted process. It is folly to put one's hopes in wonder breakthroughs; it is folly to conceptualize future food production prospects largely in such terms. At the aggregate level, even the effects of 'green revolutions' (an over simple expression) become rather subsumed in yield trends.The 'aggregate world level' is our concern here - inadequate yields at the local level too often indicate local demographic entrapment, as discussed elsewhere in this website.
Tim Dyson takes comfort in arithmetic increase, rather than geometric increase. 'Arithmetically' means that yields rise by a particular number of kilos per hectare per year. 'Geometrically' means that they rise as a given percentage of last year's harvest - which continually increases as one year succeeds another.
Say yields increase start at 50 kg a hectare, and increase by 50 kg a hectare. This is an arithmetic increase of an additional 50 kg each year. The first year they will increase 100% - a geometric increase of 100%. After 10 years yields will be 550 kg. A further increment of 50 kg will only be a percentage or geometric increase of about 9% (50/550 x 100=9).
Here is an graph of Donald Mitchell's to show the arithmetic nature of yield increases. [Figure 3, to follow] In fact global grain yields have been rising by about 40 kg a hectare for many years and continue to do so.
Which kind of increase best matches population increase? Unfortunately, as Malthus [] argued in his most famous passage:Population when unchecked, increases in a geometrical ratio. Subsistence [food] increases only in an arithmetical ratio. A slight acquaintance with numbers will show the immensity of the first power in comparison with the second.
By that law of nature which makes food necessary to the life of man, the effects of these two unequal powers must be kept equal.
This implies a strong and constantly operating check on population from the difficulty of subsistence. This difficulty must fall somewhere and must necessarily be severely felt by a large portion of mankind.Unfortunately population increase at the aggregate global level is far from being a steady arithmetic progression. Here is data from Tim Dyson to show that the annual population increase, which is now nearly 100 million a year, is far from simply arithmetic. It has recently getting more each year. Fortunately it is about to become less.
Figure 4. (Data after Tim Dyson)Figure 5. (After Tim Dyson)
If population increase is geometric, grain yields have to be considered geometrically also, so that grain and population can be matched with one another. Percentage increases in population have to me matched with percentage increases in yields. Do they match or don't they?I therefore read Tim dyson's book with great care. Where, if at all, had he failed to take proper account of reality? I was particularly interested by this figure:
Note that the rate of growth of of yield is always positive (yields greater each year), and grain-land is always negative - an ever smaller area for each of us. From yields per hectare and grain-land per head, it is easy to calculate grain per head - the middle line. On first glance the pattern seems to be random. Two things however stand out.Here are these two features inserted on the graph:1. The steep dip in all three around 1960. This is the Chinese famine, when there was massive starvation, mostly the result of the political mismanagement of Chairman Mao's great leap forward. The Chinese famine was a unique event - so far. We will therefore ignore it and consider only the trends since 1965.
2. The almost linear decline in grain per head after 1965, which eventually becomes negative in 1985. Why? Are yield declines responsible, or is it grain land per head? Although there is some grain-land effect, it seems mostly to be the result of slackening yield increases.
Figure 6. (After Tim Dyson)
Here are 5-year moving averages, and the regression line (trend line) which summarizes them, during the 32 years from 1965 to 1997.Even 5-year moving averages fail to take out the wide swings in yield increases. These are mostly due to good harvests and bad ones in the major grain producing areas, especially North America. The regression line is however disturbingly downwards - from 3.02% in 1965 to 0.87% in 1987 . As would be expected it crosses the line for population increase in 1985 - the point at which per capita grain started to fall in Figure 1.
Population growth is however remarkably uniform, decreasing in almost a straight line on this scale, from 2% in 1965 to 1.48% at the present time. Figure 7. (FAO data) 
1. Disaster ! The rate of increase of global grain yields, not only continues to decline, but gets steeper. This is not impossible.
2. 'Business as usual'. The present downward tend continues at its present rate. Since nature abhors straight lines for too long, this is unlikely.
3. The downward trend continues, but becomes less steep. We think that this is the most likely outcome.
4. The downward trend is. This is what various experts project:
Alexandratos [] projects yields to rise 1.5% annually between 1991 and 2010 [note 1]
Dyson projects a rise of 1.27% between 1990 and 2020. [note 2]
Mitchell projects a rise of 1.4% between 1990 and 2010. [note 3]
Figure 8 shows that s The chances of these projections being reached do not seem good.
Lady M. I think you read much too much into a widely wobbly graph. Look at that downward dip in 1985!
Figure 8. (Data from various sources)
UN projections
Lady M. Your UN Population Division have made three projections, its high, medium and low variants. How much have grain yields to grow if the people in these three variants are to be fed? I see the medium fertility variant is 9.4 billion in 2050. The high one is 11.1 billion and the low one is 7.7 billion.MHK. This is what needs to happen to grain yields if the people in these projections are to be fed.
Figure 9 Lady M. Let me see if I have got it right. Is what yields are expected to do in respect of population essentially this ? Figure 10 Lady M. What about 'green revolutions'?
MHK. We should not be too confident about them. The 'green revolution' in India contributed to the yields of the 1970s and but effect cannot be distinguished in the yield trend of the world as a whole.
Lady M. I seem to remember that calculation have been made as to how many people your world can support?
MHK. Indeed they have, here are some of them in a figure from Joel Cohen using data up to 1990. It shows what people thought in the past was the maximum population. In the 17th, 18th and 19th centuries there were several estimates of around 10 billion. More recently estimates have ranged widely, from as low as a billion to as much as a trillion!
The pink constant fertility path is what would happen if the 1990 fertility had continued. The green instant replacement projection shows what would happen if the average woman had 2 children from 1990 onwards (actually 2.06 children). Note that the population continues to rise for a time before levelling off. The UN low variant is in brown.
Estimates of this kind vary so greatly as to be ridiculous. What worries me is not the number of people the world could support, given all sorts of unlikely conditions,
MHK. I am about to ask him. I looked through his book with great care to find where he has been over-optimistic, and then wrote our BMJ paper 'Tighten your belts for CAIRO II'. Tim thought it was the best paper that I had ever written! Imagine it - congratulating a critic with such a criticism. What magnanimity of spirit! However, he seems to have had doubts, as we quoted above "...(though, clearly, things could be better if yield increments were greater)....: Indeed they could be!
There are said to be many 'respectable pessimists' around, but I have yet to find anyone who has looked at grain yields in this way.
Lady M. I look on major uncertainties in global grain yields as a strong additional indication for a 1-child world.
Lady M. Figures 8 and 9 show very elegantly how much yields need to be to meet various population projections. But how much grain do they assume that the average person is going to eat?
MHK. They assume that levels of per capita grain consumption are going to remain the same as when the projection was made, usually in 1990 - say about 330 kg a head. If per capita consumption were to be lower, yield increases need to be proportionally less. Say consumption was to fall by a quarter (a lot!) to, say 250 kg, then Alexandratos's projection of a 1.5% yield increase being needed, could fall to 1.125%.
Lady M. So a lot is going to depend on how much you people share.
MHK. Indeed it is !!! This is why we need to eat a photon-efficient diet.
Lady M. How do you think Figures 8 and 9 will end?
MHK. They will lend when the world ends! In a more practical sense, when world population levels off and is assumed to become constant, say in 2150, the rate of increase of population growth will become zero. The rate of increase of yield growth could become zero too. This is of course quite hypothetical, Even so, it is useful to see it as a theoretical endpoint.
Lady M. I have been reading The Word Food Outlook by Mitchell Ingco and Duncan. They are very optimistic. They say: "Crop yields continue to grow faster than population, world cereal yields grew more rapidly during the 1980s than during the 1960s or the 1970s" (page 1) MHK. I fail to see how they arrived at this. FAO data [3] shows that yields increased more rapidly in the 1960s (3.0%) than in the 1970s (2.13%), the 1980s (1.78%) or the 1990s (0.97%). What is worrying is the progressive fall in the rate of increase of grain yields since the 2960s....
Note 1 An annual increase in yields of 1.5% is derived from Table 2.1 page 45. This projects a rise of total production from 1756 mt in 199/92 to 2335 mt in 2010.
Note 2. Dyson's Table 4.6 which projects a rise in yields from 2.711 tonnes ha in 1990 to 3.989 in 2020 requires an annual rise of 1.27 %.Note 3. Mitchell (page 137) expects consumption (for which read production) to grow by "...only 1.4%..." to 2010.
1. Cohen Joel, How Many People can the Earth Support? New York, WW Norton 19962. Dyson Tim, Population and food, global trends and future prospects. London Routledge, 1966.
3. FAO. Faostat grain data from the FAO website, April 1998.
4. Malthus, Thomas Robert , An Essay on the Principle of Population, Penguin Books.
5. Mitchell, Donald O, Ingco Melinda D, Duncan Ronald C. The World Food Outlook. Cambridge University Press. I997.
