Insurance companies in Mozambique: a two-stage DEA and neural networks on efficiency and capacity slacks

Abstract

This article analyses the efficiency of Mozambique insurance companies using a DEA model. Two DEA models are used and a bootstrap approach adopted. Furthermore, the efficiency scores are predicted based on neural networks. The results reveal that Mozambique insurance companies’ output-increasing potentials are severely constrained, particularly in terms of the ceded reinsurance increasing potentials. Policy implications are derived.

Keywords:

• insurance
• Mozambique
• DEA
• bootstrap
• neural networks

JEL Classification:

• L89
• M21

Acknowledgement

The authors would like to thank Henrique Ewbank, doctoral student, for helping in performing the neural network analysis.