Copula Joint Modelling of Food Insecurity Indicators with Application to Food Insecurity Prevalence (Fip), Household Dietary Diversity Score (Hdds) and Months of Inadequate Household Food Provisioning (Mihfp)

Laina Mbongo, University of Namibia
Lillian Pazvakawambwa, University of Namibia
Lawrence Kazembe, University of Namibia

Food insecurity is expressed using various indicators to measure availability, access, utilization and stability. Some of the indicators used are household food insecurity prevalence (HFIP), household dietary diversity score (HDDS) and months of inadequate household food provisioning (MIHFP). These measures are often assumed to be independent, since they capture different spectrums of food insecurity. However, these are correlated to each other, and their dependence has rarely been analyzed. This study used generalized joint regression models through copulas to estimate the relationship between food security outcomes/indicators and exposure variables. The results indicated that both the Frank copula and bivariate normal copula fitted the data better of establishing the relationship between HFIP and HDDS (AIC=2287.296), and between HFIP and MIHFP (AIC=2072.708) respectively. The chapter thus concluded that copula approaches provide an advantage of analyzing jointly two outcomes in order to test for significant relationships between high-level hierarchical effects (e.g., random effects).