DRY SEASONS AND FOOD SECURITY: A MARKOV-SWITCHING VAR ANALYSIS ON NIGER RIVER AND FOOD PRICES IN NIAMEY

摘要

In Niger, structural and seasonal problems hamper the scope for sustainable management of land and water resources, and cause recurrent food insecurity. Each year in many areas of the country, residents suffer from increased malnutrition and morbidity during the 'hunger season' preceding harvests. This period is typically characterised by water dearth in the Niger basin and jumps in food prices, with seasonal co-movements between these prices and river flow rates. For the analysis of seasonal food insecurity in a drought-prone environment, this study proposes a two-regime Markov-switching vector autoregression (MS-VAR), with transition probabilities varying according to the time spent in a state. The model is applied to monthly flow rates of the Niger River and seasonal changes of retail food prices in the city of Niamey. These changes turn out to co-vary closely with concomitant price shocks affecting food markets in Bamako, but to be weakly related to food price variability in Cotonou (the closest major transit seaport for Niamey). Based on model results, over the last two decades the risk of seasonal food crisis has tended to spread over longer periods of the year, along with a gradual shifting forward of its timing. Transition probabilities are found to quickly switch from one regime to the other.