Implications of crop yield distributions for multiperil crop insurance rating in Ghana: a lasso model application

Abstract

Purpose – The objective of this paper is to examine crop yield predictions and their implications on MPCI in Ghana. Farmers in developing countries struggle with their ability to deal with agricultural risks. Providing aid for farmers and their households remains instrumental in combatting poverty in Africa. Several studies have shown that correctly understanding and implementing risk management strategies will help in the poverty alleviation agenda. Design/methodology/approach – This study examines the importance of crop yield distributions in Ghana and its implication on multiperil crop insurance (MPCI) rating using the Lasso regression model. A Bonferroni test was employed to test the independence of crop yields across the regions while the Kruskal-Wallis H test was conducted to examine statistical differences in mean yields of crops across the ten regions. The Bayesian information criteria and k-fold cross-validation methods are used to select an appropriate Lasso regression model for the prediction of crop yields. The study focuses on the variability of the threshold yields across regions based on the chosen model. Findings – It is revealed that threshold yields differ significantly across the regions in the country. This implies that the payment of claims will not be evenly distributed across the regions, and hence regional disparities need to be considered when pricing MPCI products. In other words, policymakers may choose to assign respective weights across regions based on their threshold yields. Research limitations/implications – The primary limitation is the unavailability of regional climate data which could have helped in a better explanation of the variation across the regions. Originality/value – This is the first study to examine the implications of regional crop yield variations on multiperil crop insurance rating in Ghana.

Description

Research Article

Keywords

Crop insurance, lasso regression, MPCI, Risks

Citation