Modelling Claims of Nigeria Oil & Gas Insurance Class of Business with Extreme Value Theory

Abstract

Oil and gas sector has been faced with huge losses in Nigeria as a result; insurance companies pay high premium due to these loses and Extreme Value Theory (EVT) is found to be suitable in modeling the extreme losses. Claims resulting from Nigeria Oil, and Gas insurance class of business for five insurance companies were modeled with EVT to estimate Value-at-Risk (VaR), where VaR measures the minimal anticipated loss over a period with a given probability and under exceptional market conditions. The mean excess plot was obtained which helps in determining the threshold value to be chosen and gives the shape of the distribution in the tail. Q-Q plot being linear and curved tail plots reveals that parametric model fits the data well. VaR based on EVT-Generalized Parameter (GPD) was carried using the chosen thresholds at 5% confidence interval. Results obtained were compared with VaR based on Historical and Gaussian method; it was established that Extreme VaR is most suitable to calculate VaR as against the Historical and Gaussian method.