Improve Bootstrap Method With Normal Distribution And Use For Claims Reserving

Among the methods to estimate outstanding claims reserve, the most widely used methods are deterministic model estimations. Although these methods are simple, easy to understand and operate. There are some defects; Such as the esti-mation accuracy is not high and it can only get the average value of the outstanding claims reserve, can not measure the uncertainty.As China’s insurance industry matures, the reserve volatility studies have been necessary。 The dynamic stochastic model is applied to the reserve estimation by using the theory of linear generation probability and mathematical statistics. The Bootstrap method is a typical example. But the calculation of the bootstrap lim-its the range of the self-help sample which is only from the original sample. So it will make the bootstrap distribution departure to the real distribution. For the shortcomings of bootstrap method,Liujian and Wuyi proposed improved bootstrap method.In this paper, an improved method based on bootstrap, combined with the normal distribution,improve bootstrap method further, and proposed normality im-proved Bootstrap method. Combined with the chain ladder method, used in the extraction reserves, and achieved good results.