Capturing Weather-related Shocks In Smallholder Agriculture

Abstract:

Rainfed agriculture accounts for 93% of total cultivated area in SSA (FAO, 2002) and smallholder is the predominant form of farm organization (33 million with less than 2 ha, representing 80% of all farms). At the same time, 75% of extreme poor live in rural areas and 90% rely on agriculture. A poor harvest can have significant impact on livelihood and welfare of a household with limited coping mechanisms and little savings.

This study aims to shed light on how small farmers are affected by growing conditions in a given season and how well these impacts can be predicted using objective measures derived from time-series of remote-sensing data. We investigate the covariate vs idiosyncratic nature of shocks experienced at the household level, using both community-level and household-level observations. We also test the relative efficiency of different metrics in predicting shocks, acknowledging that both spatial resolution and temporal frequency are likely to be limiting factors in the objective measures.

To support this analysis we draw on surveys from the Living Standards Measurement Study Integrated Surveys on Agriculture (LSMS-ISA) project. The LSMS-ISA surveys are georeferenced, multi-topic panel surveys with a strong focus on agriculture, conducted in 7 African countries. We combine the LSMS-ISA data on production and agricultural shocks with objective measures (rainfall deficit/surplus, phenological variables, water requirements satisfaction index) derived from various data sources (CHIRPS, MODIS Land Cover Dynamics, NDVI3g). Objective measures will be considered in both absolute terms as well as deviation from mean in the survey season.