The insurance industry has recently witnessed evidence of models both under and over estimating loss estimates. While it is still too early to review the modeling firms’ performance in the recent earthquake in Japan, the models overestimated the actual incurred loss in last year’s Chile earthquake.
This is in contrast to the US hurricane events* of the past decade where the catastrophe models have consistently underestimated the actual incurred losses, in some cases by a factor of two or greater. This illustrates the need for insurers to understand as much about uncertainty and the vendors’ assumptions before deciding upon capital reserves and reinsurance levels.
In addition, today’s Solvency II environment means users of model output need to be able to demonstrate a robust understanding of the inner workings of the ‘black box’ models.
Aon Benfield’s dedicated team of model evaluation experts works with insurers to assess the reliability of loss estimations and identify the strongest model for each territory and peril. The team combines natural hazard expertise and previous experience at the vendor firms to deliver the following evaluation process:?1. Initial overview of the new model loss estimates with a high level review of the model changes?2. In-depth analysis of the key components: the hazard, vulnerability and financial loss ?3. Benchmarking individual insurers against the industry and advising on reasons behind any deviations?4. Knowledge sharing with clients to assist with their discussions with regulators and rating agencies.
Paul Miller, head of International Catastrophe Management at Aon Benfield Analytics, said: “Model evaluation is a crucial part of ensuring a company is adequately capitalized to meet its catastrophe exposures. Through our understanding of the robustness of each catastrophe model and assessing its uncertainty, insurers are armed with the necessary information to make appropriate adjustments to the models’ output.”
Ben Fox of the Model Evaluation team at Aon Benfield Analytics added: “There has long been appetite for more transparent catastrophe models as components remain hidden from the end users. The proposed Solvency II regulation has further driven the need for more transparency as insurers are required to explain why they have chosen a particular model. Until ‘open’ models – such as those from Impact Forecasting – are widely available, evaluation is the only way to drill down into the inner workings of these products. Without this insight, it is impossible to critically evaluate their relative strengths and weaknesses.”