Ethiopia’s Chronic Vulnerability Index

Summary of published research¹

The concept of vulnerability has become an important part of food security analyses since the 1980s. It is seen as having two sides; exposure to external hazards, and an inability to cope with those shocks, attributed to social, political, and economic factors. Numerous attempts have been made to construct models to determine levels of vulnerability among populations. A paper, recently published in the journal Disasters, analyses one such attempt, the Chronic Vulnerability Index (CVI) - developed to measure levels of vulnerability to food insecurity in Ethiopia.

The CVI consists of an indicator-based series of maps of more and less vulnerable woredas or districts in Ethiopia. The example of the CVI reveals many of the difficulties associated with producing even a basic model of vulnerability that can be used in development or emergency response planning - it is intended for both. These problems emerge when vulnerability is assumed to be a linear, additive phenomenon with discrete causes and effects. An information product such as the CVI is ultimately of limited utility, given not only the data constraints in a country such as Ethiopia, but also the intricacy of interactions between hazards and the human systems that produce and complicate them. This renders the task of determining causation difficult.

The strength of the CVI is that it draws together data that previously were held in separate organisations. It also attempts national coverage, while many existing vulnerability studies are more limited in scale. The CVI also provides an overview of what data either are missing from its pooled sources or are inadequate at the woreda level. However, one of the problems the CVI faces is institutional. While the Vulnerability and Analysis Mapping (VAM) unit of the World Food Programme (WFP) currently takes responsibility for revising the CVI, its institutional location currently sits somewhere between the Disaster Prevention and Preparedness Agency (DPPC) and the Food Security Coordination Bureau (FSCB).

The authors suggest a number of means to help the CVI better address need. One possible modification is to improve the CVI as a targeting tool that defines more specifically the hazards to which people are considered vulnerable and map their tendency to occur in Ethiopia. This would produce a form of 'hazard atlas'. Another role would be to use the CVI to produce woreda profiles, which would be more descriptive and informative than numeric rankings. Many of the current problems with the CVI lie in the construction and mapping of the composite index. The profiles could be compared with what agencies find in assessment missions.

This may not solve the problem of modelling vulnerability, but anything more conceptually complex than this requires tools that can effectively account for the interactions between the hazards and shocks and the human systems that produce and interact with them. One possible solution to this problem would be to find ways to incorporate more sophisticated indicators. Many of these are based on household-level surveys, which may present a problem for obtaining data on a national scale in the immediate future, but this may give organisations a goal to strive for. Looking beyond Ethiopia is instructive, e.g. the Food Security Assessment Unit for Somalia (FSAU).

The FSAU classification is mapped on a national scale along with immediate and underlying causes, the number of people affected and an indication of the phase trend, whether it is stable, worsening, improving or mixed. The FSAU is based on a wide range of sources, including agricultural production surveys, household surveys, market analysis, nutritional assessments, participatory research methods, reviews of secondary data, and satellite imagery and Geographic Information Systems (GIS). The CVI is restricted to secondary data analysis, satellite imagery and GIS.

In sum, the CVI indicates some of the difficulties of operationalising the concept of vulnerability. The method by which the index is constructed erases the socioeconomic relationships that are, or should be, the hallmark of vulnerability analysis. It also erases any understanding of the dynamic causal relationships between the hazards and coping factors it incorporates and the conditions of the population. Perhaps the main methodological question raised by the CVI is whether or not data collected for other purposes can be used effectively to assess vulnerability. The answer is not yet definitive, as the CVI is still being improved, and the product needs to be evaluated once it is put into use. But it would seem that a data construct such as the CVI can, at best, provide an indication of where more detailed assessments are needed, rather than an indication of the causes, levels and locations of chronic vulnerability.

¹Burg. J (2008). Measuring populations' vulnerabilities for famine and food security interventions: the case of Ethiopia's Chronic Vulnerability Index. Disaster, 2008, vol 32 (4), pp 609-630