Overdependence on rain-fed agriculture by rural households in Malawi puts them at risk of crop losses tied to extreme weather events, like floods and droughts. In recent years, Malawi has experienced an increase in frequency, intensity and spread of these events.
In January 2015, severe flooding in Malawi affected over a million families, displacing 230,000 and killing over 100 people. A team of researchers from the World Bank and LEAD Analytics, Inc., in collaboration with the National Statistical Office (NSO), subsequently leveraged the existing longitudinal household survey infrastructure in Malawi to study the micro-level impacts of exposure to flooding on agricultural and food consumption outcomes. The team, consisting of Nancy McCarthy, Talip Kilic, Alejandro De la Fuente, and Josh Brubaker, also investigated the role that risk management strategies (e.g. sustainable land management) and risk coping measures (e.g. access to social safety nets) play in mitigating the adverse effects of flooding.
In an IFPRI-Malawi seminar on November 16, 2018 Talip Kilic, a Senior Economist at the World Bank and a member of the Living Standards Measurement Study (LSMS) team, gave a presentation on their paper, which was published earlier this year in Economics of Disasters and Climate Change and can be accessed here. Their study used the Flood Impact Assessment Survey (FIAS), which was implemented by the NSO and followed a subsample of households that had been interviewed by the Integrated Household Panel Survey (IHPS) in 2013 (pre-flooding period) in Southern and Central region districts.
Looking at the data on flowering season rainfall, Talip showed that all (georeferenced) household locations had received rainfall above the historical average during the 2014/15 season, with an average of 55 percent in 2014/15 compared to an average of 14 percent during the 2012/13 season. Since flooding is caused not only by excessive rainfall and is mediated by geology and topography, he explained the construction of a flood index that was used, alongside the absolute percent deviation of flowering season rainfall from the long-term (1983-2013) mean, to capture the full extent of household exposure to flooding.
Their flood index was constructed using principal components analysis, and is composed of three geospatial variables (defined at the georeferenced household locations): (1) the mean flood intensity measure – from the Global Flood Monitoring System maintained by the University of Maryland, (2) household elevation, and (3) household (Euclidean) distance from the nearest river. The tercile of this index were in turn used to define three flood exposure categories in 2015: low, medium and severe.
The descriptive findings showed that maize yield and value of crop production per capita were lower across all the flood exposure categories in 2015 versus 2013, and that the percent reductions in maize yield were greater among households that faced moderate and severe flooding.
In terms of food consumption, Talip focused on three household-level outcomes: (1) food expenditure per capita, (2) caloric intake per capita, and (3) food consumption score. He showed that mean food expenditures were significantly lower in 2015 compared to 2013 among households that faced low and moderate flooding - though the reductions were significantly muted with respect to the dynamics in agricultural outcomes.
Despite the reduction in food expenditures, caloric intake in fact increased in 2015 compared to 2013 across all the flood exposure categories, mainly due to a shift in household consumption from high- to low-quality maize flour, in part facilitated by a decline in maize flour unit values.
On the other hand, the food consumption score (which captures dietary diversity as well as consumption frequency and relative importance of main food groups) was lower, on average, in 2015 compared to 2013, across the broad and particularly among households located in areas with moderate and severe flooding.
Talip subsequently presented the findings from the econometric analysis of flood impacts on food consumption outcomes. The results showed that deviation from expected rainfall had significant negative impacts on all consumption outcomes, as households were not able to self-insure and maintain consumption when rainfall differs from expectations, even for non-extreme rainfall events.
While exposure to flooding, over and above deviation from expected rainfall, was shown to lead to a significant shift towards less diverse food consumption among households located in areas that experienced moderate and severe floods, it did not necessarily generate additional negative impacts on per capita measures of food expenditures and caloric intake – in line with the descriptive findings. Overall, the analysis depicted a reduction in quality, rather than quantity of food consumption, with dynamics that were considerably muted compared to the changes in maize yields and value of production per capita
Further, while risk management strategies such as sustainable land management, and risk coping measures, such as access to off-farm employment income, were generally ineffective in alleviating the negative effects of flooding, the analysis showed that having access to food assistance increased food consumption scores for households located in severe flood areas vis-à-vis those located in low flood areas. On the other hand, access to school feeding led to improvements in all three of our consumption measures, but only for those residing in moderate flood areas. Similarly, access to MASAF led to higher food expenditures and calories consumed, but only for those in moderate flood areas.
Talip concluded with a description of the ongoing work to evaluate the impacts of single and multi-year flood and drought shocks using the IHPS 2013, FIAS 2015, and IHPS 2016 data. The preliminary findings showed significant reduction in maize yields in 2016 compared to 2013 across low, moderate and extreme rainfall categories that capture different levels of absolute percent deviation of flowering season rainfall from the long-term mean. The reduction was indeed the highest, and stood at 41 percent among households that faced the most extreme rainfall conditions during the 2015/16 season. The early results regarding the impacts on food consumption outcomes, however, mirrored the muted dynamics picked up in the analysis of the flood impacts, lending support to the consumption smoothing hypothesis. The future work will provide a more robust econometric basis for these early findings and will continue to look at the role of risk management and coping strategies in shielding households from extreme weather events. In particular, the team will aim to gain a better understanding of the effectiveness of humanitarian cash and food aid in response to the recent climate shocks.
The seminar presentation may be viewed below.
