Comparison Of The Typical Random Mortality Models And Its Application In Commercial Pension Insurance

With the development of social and the improvement of health conditions, themortality rate of China has decreased steadily; the human life expectancy has alsobeen increasing significantly, these have brought increasing mortality risk to the lifeinsurance industry of China. Therefore, how to accurately predict the future mortalityrate is particularly important. This paper makes focus on seven popular randommortality models, iterative method is used to fit these models with experience data ofthe death population by age and sex from1996to2009and ARIMA model is used topredict the mortality rate. Through the comparative analysis of the fitting results andforecasting effects of these existing random mortality models, the most suitablerandom mortality model can be chosed, and then the selected model can be used tomeasure the impact of decreasing mortality rate to China’s life insurance companies.By theoretical and empirical researches, this paper gets several conclusions.Firstly, a good fit to historical data does not guarantee sensible forecasts. Therefore,when selecting mortality model, people should consider both fitting results andforecasting effects. Secondly, depending on the sex, the best random mortality modelmay be different. The fitting results and forecasting effects of one of the extensions ofCBD models is ideal for men, but the APC model is more excellent for women.Finally, due to the improvement of mortality, the mortality risk assumed by insurancecompanies is growing. Life annuity products, for example, based on the calculationresults of the mortality before and after improvement, the male policyholders’ annualpremium is undervalued by about4.2%. Unexpected mortality risk has a significantimpact on the payment of the present value of individual annuity products, so there ismortality risk in our commercial insurance companies.