Abstract/Summary

This thesis developed and validate a multidimensional food security indicator called the Vulnerability to Food Insecurity Index (VFII). Currently, there is no standard indicator of vulnerability analysis in food security research and this thesis responds to this challenge. The first research objective was to demonstrate how to develop this indicator and establish its validity through comparison with other traditional food security indicators such as per capita calorie consumption (PCC), food consumption score (FCS) and coping strategy index (CPI). The second objective was to systematically evaluate the effect of some assumptions on the robustness of the VFII. The aim was to examine how data type, weighting scheme, normalization method and excluding/including of variables, affect the output of the index using sensitivity and uncertainty analysis. The third objective was to verify the result of the VFII with real-life experience and to understand why households are vulnerable to food insecurity using qualitative insight. The research applied both quantitative and qualitative method. The study used the World Bank LSMS panel dataset for households in South-South Nigeria to design the index, while fieldwork was used to verify the results of the index. In designing the VFII four steps were used. The first developed a conceptual framework for vulnerability to food insecurity, which helped to select indicators for the index. Structurally, Vulnerability to Food Insecurity Index is a multidimensional index of the probability of covariate shock occurring (exposure), the accumulative experience of food insecurity (sensitivity) and coping ability of households (adaptive capacity). The second step applies equal weight to each component of the index based on the evidence from the sensitivity and uncertainty analysis. In the third step, variables were normalised using the min-max normalization method. In the last step, a linear aggregation method was applied to generate the score of the index. For the uncertainty and sensitivity analysis, the one-at-a-time and global sensitivity approach were applied to examine the robustness of the index. Using the one-at-a-time approach, the research explored how the VFII output responds to different weighting schemes, normalisation method and inclusion/exclusion of variables. For the global approach, Monte Carlo simulation and Sobol first-order index and total-effect index were used to explore the uncertainty and sensitivity of VFII. In the qualitative phase, the results of the index from the quantitative phase were verified in the field using qualitative methods. Food vulnerability maps for households in South-South Nigeria were used to purposively select Akwa Ibom State for the verification exercise. Key informant interviews