Famine early warning systems, many of which were established in the 1980s, have improved capacity to predict when famine is likely to occur, but this has not been matched by preventative response. What is needed is a systematic analysis of what happens to early warning information once it enters the decision-making process. There are four categories of reasons for which early warning information is not fully used; these concern the nature of the information, the institutional context and institutional links to decision-makers, the broader political climate, and the logistical obstacles to an optimal response. The international relief system responds to famine once it is underway, but is badly equipped to respond to early warning; further, relations between donors and governments are often the most important determinants of response. The case for an early response is made on the basis that assistance must be developmental in nature and arrive before lives have been seriously disrupted by the necessity to sell off productive assets. The rising incidence and severity of short-term shocks make countries in the Horn of Africa and the Sahel particularly susceptible to famine, and nearly half the population faces chronic food insecurity. Mixed results of structural adjustment have left the poor more vulnerable to internal and external shocks, and macro-economic decline has increased inequalities between social strata, and people in famine-prone areas have little direct input into decision-making procedures relating to famine. It is important for early warning systems to detect pockets of acute food stress. -b/m