This LRB article by Jeremy Harding articulates an increasing fear that food markets will not operate smoothly over the next decade or two. He gives some major reasons (only partially reproduced here) to be pessimistic:
The first is the nature and extent of population growth: we are six billion now and by 2030 we’ll be eight billion…
The second is ‘the nutrition transition’: generations that once lived on grains, pulses and legumes have been replaced by more prosperous people with a taste for meat and dairy. Crops like maize which once fed many of us directly now feed fewer of us indirectly, via a costly diversion from which they emerge in the value-added form of meat. Global production of food – all food – will have to increase by 50 per cent over the next 20 years to cater for two billion extra people and cope with the rising demand for meat.
The third factor is energy: the industrial production of food is sure to become more expensive as fuel costs rise. It takes 160 litres of oil to produce a tonne of maize in the US; natural gas accounts for at least three-quarters of the cost of making nitrogen fertiliser; freight, too, depends on fuel.
Land is the fourth. The amount of the world’s land given over to agriculture continues to grow (in the UK, roughly 70 per cent of land is agricultural), but in per capita terms it’s shrinking. As with oil, it’s possible to envisage ‘peak food’ (the point of maximum production, followed by decline), ‘peak phosphorus’, i.e. the high point in the use of phosphate fertiliser (one estimate puts it at 2035), and, as the FAO suggests in its diplomatic way, ‘peak land’: the point at which the total area of the world’s most productive land begins to diminish (soil exhaustion, climate change) and marginal land comes up for reassessment.
[Fifth] Alternative fuels are reducing the amount of land available for growing food.
The Julian Simon-savvy crowd that reads MR might not be so impressed, but I wouldn't write off these worries so quickly. On the list, #1 and #2 do not impress me per se, but they do require that market mechanisms of adjustment be allowed to operate. Note that agriculture and land markets are highly regulated around the world and that you don't have to read this as a story of market failure. As for #3, most energy is mispriced today. Keeping it cheap means growing pressure on that externality, while taxing it means a solid whack to a lot of food markets. #5 stems from bad government policies. Another problem, mentioned later by Harding, is that very often water for agriculture is subsidized and unsustainably so. Keeping water cheap means growing pressure on that externality, while removing the subsidies (which I favor) means a solid whack to a lot of food markets, at least in the short run. The world as a whole is consuming its capital of aquifers and the like and engaging in short-term thinking by refusing to let the price of water rise as it ought to. Internalizing all the relevant externalities, and increasing sustainable long-run production, would in fact mean big "tax" hikes on growing food today.
There is also a critical scale at which fertilizer run-off and erosion externalities start to matter at a level beyond which we are accustomed to seeing.
I believe these factors mean a stronger case for agricultural free trade, rather than "localism," yet at the same time removing the subsidies for sprawl. Yet so far the people worried most about these issues are often the ones with the least economically informed answers. It would be a mistake to, say, mock Paul Ehrlich's earlier doom-saying predictions and ignore these problems altogether.