Dynamic modelling in loss frequency and severity estimated: Evidence from the agricultural rice loss due to typhoons in Taiwan

The paper first adopt the BDS test to show that the BDS statistics of the time series of typhoons is a chaotic behaviour while the associated rice damage is random. The authors' investigations show that the time series of typhoons and rice damages are described by nonlinear. The result of the assessment shows that the model based on the AR(1)-GARCH(1,1) model is the best performing model in describing the rice loss severity due to typhoons and may have a chaotic behaviour if the variation of parameters is large enough. The best forecasting models in loss frequency with chaotic and severity predictions with random walk are superior to the best forecasting models in the current traditional or official estimated. This paper find that the fitting insurance price decision process by the loss cost charged in our method is different from the actuarial premium approaches of comparing evaluating effectiveness under the non-crop insurance program.