Efficiency in Non-Life Insurance Market in Argentina and its sensitivity to the inputs/aoutputs selection

GUSTAVO FERRO; ZILLA SINUANY-STERN

Quebec

Conferencia; 21st Conference of the Internacional Federation of Operational Research Societies IFORS 2017; 2017

IFORS

In this study we measure the relative efficiency of the non-life insurance companies in Argentina and its evolution during 2009-2014 via DEA (Data Envelopment Analysis) ? the Variable Returns to Scale (VRS) version. The sensitivity of the efficiency to the inputs/outputs specification is tested in regard to two well-known approaches in the insurance efficiency measurement: Model A by Cummins et al. (1996), and Model B by Luhnen (2009). In Model A the inputs are Employees, Physical Capital, and Financial Capital and the outputs are Losses Incurred, and Invested Assets. In Model B, the inputs are Operative Costs, Production Costs, Financial Capital, and Liabilities, while the outputs are Claims Paid and Invested assets. Overall data from 83 companies was used over the 6 years. We found that the selection of inputs/outputs is important. In general, as the number of inputs/outputs increases, the efficiencies and number of efficient units also increase, as we found for model B which has one more input. Over time, in both model the average efficiencies of the insurance companies decreased from 2009 till 2012, and increased afterwards. However, Model A average efficiency is maximal at 2009, while Model B at 2014. In model A none of the companies was efficient every year, and 58 were not efficient in any of the years; while in Model B, 3 companies were efficient every year and 58 companies were not efficient in any year. In 2014 we found 14 efficient companies by both models, and 45 non-efficient companies by both models. Using Chi-square test we found significant dependence between the two models (P-value