12 July 2012

A Closer Look at Gobal Food Supply

According to the Food and Agricultural Organization, "On average, a person needs about 1800 kcal per day as a minimum energy intake." (A kcal - kilocalorie - is a measure of food energy, also known as the Calorie).  For comparison, the US government recommends 2,500 per day, on average (a recommendation that is certainly exceeded by most people, but I digress).

The unit of kcal/person/day provides a useful basis for evaluating total food supply as compared to population. The graph at the top of this post shows such an evaluation, based on data downloaded from FAOSTAT (thanks RTC!).

The data show that from 1961 to 2007, when the dataset begins and ends, global food supply in kcal/person/day has steadily and consistently increased such that it has been for many decades comfortably above the level deemed necessary to meet individual nutritional needs.

In fact, if food supply distribution were perfectly efficient (which of course it is not) the world could feed an additional 1-3 billion people with the food produced in 2007 (depending on your view of nutritional requirements). This can be hard to reconcile with the fact that in 2007 the UN found about 1 billion people globally to be "undernourished." So there is considerable "headroom" for progress even without increasing global food supply, and UN data show progress in recent years.

It is such simple math that leads the OECD and FAO to conclude:
Food production has not only kept pace with population growth, it has outstripped it. The world now produces more food than ever, and even countries that were once practically synonymous with famine have achieved self-sufficiency in staple foods... hunger is a problem of poverty, not scarcity.
And also:
[W]hen you look at the facts, there is no “agricultural” reason for hunger today. Global food production has increased more quickly than population over the past half century, and the EU and USA even had to bring in policies to get rid of “mountains” and “lakes” of food and drink.

If people are hungry, it’s because they can’t afford to buy food, not that there is no food to buy. There are many reasons for this. Politics, policies and poverty all intertwine, and as Nobel-prize winning economist Amartya Sen said “There is no such thing as an apolitical food problem.”
Let's dig a little deeper into the numbers.
The graph above shows the annual rate of change in kcal/person/day for the 20 years ending in 2007. The red line shows the linear trend in kcal/person/day, and shows that the annual rate of growth has just about doubled over that time period. The data illustrate that food supply has been growing faster than population, and this trend has been accelerating (which probably owes to a slowdown in population growth rates in addition to effects from agricultural intensification).

What does this data mean from the standpoint of discussing agricultural policies? I can think of several things.

1) It can be misleading to talk of a global "food supply" problem. Certainly, sustained improvement in agricultural productivity will continue to be important, but at present does not appear to be a limiting factor in meeting global nutrition goals. Talk of the need for a "second green revolution" not only fails to accurately reflect the so-called "first green revolution" (more on this to come) but also distracts from the fact that supply is presently not a limiting factor in meeting global nutritional goals.

2) Over many decades the global agricultural system has shown itself to be very robust to a range of shocks, including the widespread pattern of climate anomalies of the early 1970s, financial crises, and rapidly changing prices of inputs (notably energy). However, there could be unforeseen major shocks yet to come, including disease (e.g., wheat rust) and rapid climate changes (e.g., from a massive volcanic eruption). Policy should focus on the robustness of agricultural systems to such shocks.

3) The issue of global food is ultimately as much (if not more) a problem of distribution, poverty and governments as it is an issue of technological innovation. From where I sit it seems that far more attention it paid to the latter than the former. Every call for a "second green revolution" should be met with a reply focused on the social and political factors that affect meeting nutritional goals.

Finally, in my explorations of issues of food (the subject of several papers in the works and a chapter in my coming-into-focus new book), it seems that a focus on yield, productivity and total production (all important, of course) can obscure the larger focus on the ultimate goals of food policy -- feeding people. This post argues that an important but underutilized (in policy discussions, that is, perhaps not among specialists) metric of kcal/person/day can help to keep our attention on that larger focus.


  1. Roger, good post.
    Just a small clarification on food needs:
    1. The estimates of food energy needs per person are based on (a) the age-sex structure of the population, b) an assumption about the acceptable body weight for height for each age and sex; and (c) an assumption about the amount of physical activity per person.
    2. The average energy expenditure of a person to keep bodily weight at the midpoint of the range of "acceptable" weights for height while performing a "moderately active" level of physical activity varies from about 2050 in least developed countries to about 2400 or 2500 in developed countries (because people are taller in developed countries, in part due to nutrition, and mostly because there are fewer children and more adults: children are smaller and consume less energy). The US recommendation of 2500 kcal per person/day derives from this kind of definition.
    3. Bodily weight (for each height) may vary within a certain range without endangering health. There are standard curves in this regard (the ones in international use are compiled and provided by the World Health Organization).
    4. FAO estimates the minimum energy needs based on keeping body weight at the lower bound of the acceptable range, and performing "light" (rather than "moderately active") physical activity. This yields an average minimum requirement of about 1800 kcal (varying from about 1650 to about 1950 across developing countries, and up to 2000-2100 in developed countries). An energy intake below this minimum makes you "undernourished" by FAO definition. FAO publishes an annual report on The State of Food Insecurity (SOFI) with estimates of the prevalence of undernourishment across countries and regions. The level is practically zero in rich countries. Levels below 5% are regarded by FAO as statistically non significant due to inherent imprecision of the norms and estimates.
    5. Even if people cover all their food energy needs, they may still have a deficit of specific nutrients, mostly vitamins and minerals contained especially in milk, fruit and vegetables, and in some cases also protein (supplied mainly by animal products and by the combination of cereals and legumes).

    We discuss extensively these matters in our recent book (from a father-son collaboration that may sound familiar to you):
    H.Maletta and E. Maletta, 2011. Climate change, agriculture and food security in Latin America. Brentwood (Essex): Multi-Science Publishing.

  2. -1-Hector M.

    Thanks! Very useful ... I will (evetually) initiate a discussion on the social construction of the calorie based on the excellent work of Nick Cullather (Indiana). I will also track down your book!

  3. The calorie, like the mile or the meter, is certainly a "social construction", because it is only a unit of measurement. Nowadays the international standard units of energy (joules and kilojoules) are the recommended ones, though the calorie is still around. What is not a social construction is the amount of energy required to stay alive and perform physical activity, no matter the unit used. This amount is a physical amount, determined by thermodynamic and biological factors, and varies by known reasons such as height, weight, and age, and may be affected by some other factors such as the waste of energy due to infections through fever or diarrhoea, and also the special needs of pregnancy and lactation in women, and bodily growth in children and adolescents (all these factors are included in the estimates referred in your post and in my previous comment). Per capita needs are of course an average for all the population of a country, and admit inter-individual and inter-temporal variation for various reasons.

  4. Poverty is a problem caused by government. And without poverty, there is not distribution problem.

    If we could rid the world of the nasty kleptocracies, problems solved. If we don't get rid of them, it is impossible to be rid of poverty and associated hunger.

  5. I believe looking at agriculture in terms of total factor productivity including both storage and distribution would provide a useful point of view.

    In areas ranging from microprocessors to solar panels, looking at the entire value chain allows for estimates of and plans for incremental innovation that combine across individual sectors to yield logarythmic improvements.

    The solar power sector is particularly useful. Long hoped-for gains in module efficiency and the arrival in the marketplace of many new playes caused first, polysilicon and later, module prices to plummet. Rather than helping the solar industry it almost killed it. Where further innovation is desperately needed in solar power is in permitting, installation costs and balance of systems,especially inverters.

    Looking across the value chain of agriculture in a similar fashion, the Green Revolution needs to expand horizontally to include storage and shipping. It remains incomplete.

    Grab the next logistics chief who retires from Walmart and put him on the case.

  6. I don't disagree that distribution is a big problem, but I wonder how much of the distribution problem is related to the consumption problem; does the average American (whatever that may mean) consume more food (is 'afford to' a better thought?) than the average Haitian? If I can, and do buy more, won't the distribution channels be skewed to serve my demand?

  7. Has there been an analysis of the increased carbon production in the atmosphere to the increased crop yield?

  8. Where is food production measured in the chain of farm to plate?

  9. Has there been an analysis of the increased carbon production in the atmosphere to the increased crop yield?

    Increased crop yields will reduce carbon dioxide levels. It's what plants absorb to grow, after all.

    Hey, but don't let that worry you. If what you need is some reason to keep people hungry, then fretting about trivial levels in CO2 emissions due to crop growing will make you lots of friends among the wanna-be Malthusians.

  10. A post on distribution problems (tax, trade barriers, subsidies, protectionist policies, ...) would be welcomed.

  11. Notice that food production has gone up both in absolute terms during this period as well as on a per capital basis.

    During the 1960s and 70s there was a great deal of starvation in the world due to inadequate good production. In fact, a number of books were published (i.e., "The Population Bomb", etc) proclaiming "the race to feed the humanity" was "over." Now, as Roger points out, there is adequate food and widespread starvation has disappeared from headlines.

    What changed?

    Two things: The Green Revolution and warmer temperatures. Google the former.

    With regard to temperatures, a plot is available here: http://woodfortrees.org/plot/hadcrut3gl/from:1900

    Temperatures fell from 1944 to 1977. The shorter growing seasons and accompanying weather patterns could not sustain the population's need for food. With the rise in temperatures from 1977 to 1998 we have adequate calories.

    Of course, we have a much larger world population now. If temperatures were to cool to 1977's level, it would be catastrophic to humanity.

    I often tell the pro-global warming crowd that they need to be careful what they wish for.

  12. Roger, who if anyone has done work on the change over time in the amount of food wasted vs produced. I assume that agricultural production, via its interaction with consumption (transport, storage, marketing, refinement), has always generated waste or losses of the available, total yield. My question is: how useful or telling is the spread between total production and consumption when we know that distribution can never be completely efficient (though certainly could see great progress). Best, G

  13. -12-www.gregor.us

    The efficiency of distribution is a key variable that we should be paying attention to, in addition to other metrics of productivity.

    So the spread between production and consumption is IMO incredibly important.


  14. I think both Matt Ridley (Rational Optimist) and
    Bjorn Lomborg (TSE) had sections on this.

  15. The efficiency of distribution is a key variable

    In most cases this is actually due to the rulers of the countries involved, either official or unofficial.

    Distribution in Africa is a problem because at least some of the people there want it to be a problem.

    I think that if the governmental issues, including poor decisions regarding supporting subsistence farming and central planning of the economy, are removed then distribution issues will sort themselves out pretty quickly.

  16. If we don't really need to grow any more food to feed everybody, then by the same token we don't need more economic growth to reduce poverty. Just spread the GDP more evenly.

    But of course that is too simplistic. We need to grow more food because the people eating so much now (meat in particular) aren't inclined to eat less.