Tornado risk analysis in the United States

Purpose - Perils of tornado and hail cause large amounts of loss every year. Based on the data provided by Property Claims Services, since 1949, tornado, hail and straight‐line‐wind losses account for more than 40 percent of total natural losses in the USA. Given the high frequency of tornado and damaging hail in the continental USA, quantifying these risks will be an important advancement in pricing them for insurance/reinsurance purposes. In the absence of a realistic physical model, which would look at these perils on a cluster/outbreak basis, it is not possible to underwrite these risks effectively. The purpose of this paper is to focus on the tornado risk. Design/methodology/approach - A tornado wind‐field model is developed based on the model used by Wen and Ang. The model is calibrated to the specifications given in the Fujita intensity scale. To estimate the tornado hazard, a historical database is generated and de‐trended using the information provided by Storm Prediction Center along with the dataset given by Grazulis. This new historical database together with a reinsurance timeframe criterion in mind was used to define outbreaks. These outbreaks are used in a Monte‐Carlo simulation process to generate a large number of outbreaks representing 35,000 years of simulated data. This event‐set is used to estimate spatial frequency contours and loss analyses. Findings - The results focus on the spatial frequency of occurrence of tornadoes in the USA. The losses are tallied using multiple occurrences of tornado and/or hail per outbreak. The distribution of loss, both on per occurrence and on aggregate basis, are discussed. Originality/value - This paper is believed to be the first one to use a tornado wind‐field model, outbreak model, and vulnerability models, which estimate both spatial distribution of hazard and location‐based distribution of losses. Estimation of losses due to hail is also provided.