Underwriting Risk Prediction Based on Moran Process and Fourier Fitting

Abstract

Extreme-weather events are becoming a crisis for property owners and insurers, which has drawn widespread attention. For insurance companies, predicting regional losses from Extreme-weather events and studying the participation decisions of insurers can inform insurance strategies. This paper primarily focuses on the impact of extreme-weather events on insurance behavior, based on ten years of extreme-weather data of Michigan, we use Fourier fitting to predict the frequency of extreme-weather events in the subsequent year. Next, to measure underwriting risk, this paper considers economic losses and pay-out ratios, etc. to apply Biotic community Strategy-updated Moran process to simulate the social leaning and reciprocal mechanism in catastrophe insurance purchase decision-making combined with the prediction results and consumers with different levels of risk-aversion and expected utility perception. Based on the above analysis, the URA model is developed, and the validity of the model is demonstrated using Michigan as an example. The evolutionary dynamic analysis based on Moran process expands the research perspective of the insufficient demand for Extreme-weather insurance, and can be a reference for insurance company's insurance strategy.