In March 2011, a senior Brookings Institution official wrote that “the crux of the food price challenge is about price volatility, rather than high prices per se” and that “[i]t is the rapid and unpredictable changes in food prices that wreak havoc on markets, politics, and social stability.”

In a recent research paper, however, I use monthly data on food prices and news reports of social unrest worldwide to tease out the causal relationship between food prices and social unrest. The results indicate that it is rising food prices that cause social unrest and that increases in food price volatility are actually associated with decreases in the number of food riots. It thus would be a critical mistake to work toward price stabilization instead of curbing rising food prices.

Consumers, Producers, and Social Unrest

If you have taken a principles of economics class, you know that everything else equal, an increase in the price of a good means that you can afford less of that good. So if you value a given good, the consequence of an increase in the price of that good is that you are worse off.

If you live in the United States, your household dedicates about 13 percent of its budget to food. But if instead you lived in a developing country, that figure would be well over 50 percent.

It should thus come as no surprise that increases in the price of food are especially bad for the poor in developing countries.

This is true for urban households, who are almost all net consumers of food, but also for many rural households who, for a variety of reasons – ranging from failures of the credit, input, or land markets to adverse meteorological conditions – fail to produce enough to feed themselves and are also net consumers of food.

There are many more net consumers than there are net producers of food worldwide. For those net consumers of food, rising food prices can have disastrous consequences. Worse, the greater the share of its budget a household dedicates to food, the more disastrous the consequence of a rise in food prices for that household (Deaton, 1989).

Last March, Annia Ciezadlo, a writer whose memoir Day of Honey explores the social importance of food in the Middle East, wrote in an article for Foreign Affairs:

Change is sweeping through the Middle East today, but one thing remains the same: the region once known as the Fertile Crescent is now the world’s most dependent on imported grain. Of the top 20 wheat importers for 2010, almost half are Middle Eastern countries. The list reads like a playbook of toppled and teetering regimes: Egypt, Algeria, Iraq, Morocco, Yemen, Saudi Arabia, Libya, Tunisia.

For decades, many of these regimes relied on food subsidies to ensure stability (…). But over the past several years, grain prices reached record levels, and these appeasement policies lost their luster. In Tunisia, pro-democracy demonstrations began in late December 2010 with protesters brandishing baguettes. In just a few months, a wave of uprisings rippled across the region, toppling Tunisian President Zine el-Abidine Ben Ali and Egypt’s longtime ruler, Hosni Mubarak.

Indeed, after attaining a peak during the summer of 2008, food prices started rising rapidly again in the second half of 2010 to hit an all-time high in March of 2011.

Likewise, as illustrated in Figure 1 above, the 2008 and 2011 spikes in food prices (denoted by the red line) coincided with spikes in the number of food riots reported in the news (denoted by the blue line).

Correlation Is Not Causation

But as I constantly remind the students in my development seminar, correlation is not causation, and a key component of critical thinking is the ability to question correlations presented as causal claims. In other words, social unrest may lead to high food prices just as much as the opposite is true.

In my paper, using a statistical technique called instrumental variables estimation, I attempt to identify one side of this relationship by first conditioning food prices on the number of natural disasters worldwide (i.e., droughts, episodes of extreme temperature, floods, insect infestations, storms, volcanic eruptions, and wildfires). Not only do natural disasters constitute shocks to the supply of and demand for food, they presumably affect social unrest only through food prices. In principle, this statistical apparatus allows teasing out the potential causal relationship flowing from food prices to social unrest from the correlation between the two.

Intuitively, this is possible because conditioning food prices on natural disasters (which are themselves uncorrelated with social unrest) allows eliminating the variation in food prices that is purely due to variations in social unrest.

The results are fairly robust: to be sure, rising food prices result in more instances of social unrest, and this remains true whether:

• The price of food is considered broadly or the scope is narrowed to only the price of cereals, which constitute the bulk of an average developing-country diet;
• The food crises of 2008 and 2010-2011 are controlled for; or
• Alternative definitions of what constitutes a natural disaster are considered.

Surprisingly, however, food price volatility – unexpected departures from the food price level, holding the food price level constant, which includes both rises and falls in the price of food – is actually associated with fewer instances of social unrest.

This is likely because unlike food producers, food consumers tend to slightly benefit from food price volatility.

Food producers make production decisions on the basis of expected prices, long before uncertainty over food prices is resolved (Sandmo, 1971); volatility therefore is something producers would like to avoid, and so they typically favor price stabilization policies.

Food consumers, however, make consumption decisions knowing exactly what food prices are at that moment, and so an increase in the uncertainty surrounding food prices means they might get to enjoy relative price discounts between food commodities (Turnovsky et al., 1980). This underappreciated theoretical point has recently found empirical support.

Note, however, how I avoid causal language in the case of food price volatility. That’s because my data do not allow establishing whether there is a causal relationship between food price volatility and social unrest, and the most that can be said is that the two are correlated.

Policy Implications

What does this mean for policy? First, for domestic policy makers who want to prevent social unrest, it is crucial to ensure that food is affordable.

In many cases, this means not doing away with food subsidies for urban consumers (Bates, 1981). This is especially true in places where people already have other reasons to be upset, such as countries with high rates of unemployment. Here, both Tunisia and Egypt at the end of 2010 come to mind.

Second, if international policy makers want to prevent social unrest, it is better to work toward preventing sharp increases in food prices rather than preventing increases in food price volatility, as Chris Barrett and I argued in Foreign Affairs last summer.