The food crisis of 2007 – 08 and the urban riots that ensued in some 40 developing countries placed the question of food price instability at the heart of policy debates. Since the 1980s, the prevailing idea has been that the best solution is managing risk without 'affecting prices' through private instruments (such as crop insurance, futures markets) in conjunction with the provision of safety nets for vulnerable populations. Nevertheless, this strategy did not prove effective: private risk-management instruments did not come to fruition, and safety nets did not succeed in preventing the deteriorating nutritional situation of vulnerable house- holds. This paper shows that the arguments against price stabilisation (the informational role of prices and the 'natural insurance' of producers) do not hold when the different causes of price instability are taken into account. The author proposes a typology of these causes. The paper closes by proposing relevant combinations of instruments for each cause of instability and discussing ways to implement them.