Mark Notaras has a great review of the UN's World Food Program on the links between food insecurity and conflict:
Brinkman and Hendrix point out that the relationship between food insecurity and violent conflict is not one way. High food prices are both exacerbated by and exacerbate the chance of civil unrest. In other words, food security is a pre-condition for political stability and political stability is a precondition for food security. The question of what came first is perhaps not as relevant as attempting to understanding how to escape what Collier and his colleagues refer to as the "conflict trap".
The paper identifies some general policies that even fragile states can take in times of high food prices, in order to prevent conflict breaking out. These include reducing import tariffs and taxes and increasing subsidies to lower prices, as well as releasing food reserves to increase supplies. In fact, in the 2007-2008 round of food price spikes, 77 of 84 developing countries surveyed did implement such policies in order to stabilise prices.
The challenge for policy-makers is to ensure that feeding people in the short term does not exacerbate the food insecurity-conflict link in the long term. Subsidies, for example, seem sensible because they reduce food prices for all people, including the most vulnerable. Yet, they are also inherently regressive since those who spend a much smaller percentage of their income on food receive the same financial benefit in absolute terms. If applied without care, subsidies can increase income inequality and thus augment the potential for conflict down the track.
Worth a read - in the end, the conclusions aren't that new, but it is always interesting to see what the thinking is.