我国车险未决赔款准备金评估方法研究

发布时间：2018-02-25 22:03

本文关键词： 车险未决赔款准备金链梯法 B-F模型卡尔曼滤波模型　出处：《南京大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着中国保险业已经进入了一个深化的改革、全面开放加快发展的新阶段,中国正在成为新兴的保险大国。保险业参与到更多的经济和社会服务领域,承担了越来越重的社会责任。在这一时期,中国的外国分支机构的数量是不断增加的,这就增加了中国保险企业的压力。在这种压力下,能够增加竞争力的实力主要就体现在投资策略上。而投资比率很大程度上由准备金的提取决定。因此,为了评估的利润,我们必须考虑到未决赔款准备金的储量。总之,考虑用科学的方法来估计未决赔款准备金,具有非常重要的实践意义。目前,我们在实务中最广泛使用的评估方法就是链梯法了。在准备金评估模型中,这是最基本的方法。虽然链梯法计算方法非常简单和便捷,但是在遇到历史数据异常和出现意外波动很大的情况下,对准备金估算的结果就会不够准确。所以,本文考虑将动态随机的思想引入到评估模型中,选取了卡尔曼滤波模型来作为未决赔款准备金的基本评估模型,并且,把多期准备金进展年的数据做成时间序列,代入ARMA模型作为对卡尔曼滤波中的状态方程。本文结合了精算评估实务,对于非寿险公司短尾业务选取了车险数据作为评估对象,研究了赔款次数和赔款额度的不同评估方法,分别对已决和未决数据使用不同的模型进行评估,并且对比后得出结论。本文研究发现,在数据充分的情况下,两种模型得出的评估都可以接受。但在进展年较短的情况下,卡尔曼滤波模型的评估结果不是那么令人满意。进展时间增加后,卡尔曼滤波模型的效果比较明显。如果在有经验的精算师的主观判断下,也可以使用B-F法对结果进行加权后得出最终的评估结果。
[Abstract]:As China's insurance industry has entered a new stage of deepening reform and opening to the outside world and accelerating its development, China is becoming a new insurance country. The insurance industry is participating in more economic and social service fields. Taking on more and more social responsibility. During this period, the number of foreign branches in China has been increasing, which has increased the pressure on Chinese insurance companies. The strength to increase competitiveness is mainly reflected in the investment strategy. And the investment ratio is largely determined by the withdrawal of reserves. Therefore, in order to evaluate the profits, we must consider the reserves of the outstanding claims reserve. It is of great practical significance to consider the scientific method to estimate the reserve of outstanding claims. At present, the most widely used evaluation method in practice is the chain ladder method. In the reserve assessment model, this is the most basic method. Although the chain ladder method is very simple and convenient, However, when the historical data are abnormal and unexpected fluctuations occur, the results of reserve estimation will not be accurate. Therefore, this paper proposes to introduce the idea of dynamic randomness into the evaluation model. The Kalman filter model is selected as the basic evaluation model of the pending claim reserve, and the data of the multi-period reserve progress year is made into a time series, and the ARMA model is used as the state equation in the Kalman filter. In this paper, combined with actuarial evaluation practice, the vehicle insurance data is selected as the evaluation object for the short tail business of non-life insurance companies, and the different evaluation methods of indemnity number and compensation amount are studied. Different models are used to evaluate the determined and pending data, and the results are compared. In this paper, it is found that the evaluation of both models is acceptable when the data is sufficient. However, in the case of shorter years of progress, the results of the Kalman filter model are not so satisfactory. The effect of the Kalman filter model is obvious. If the subjective judgment of the experienced actuary, the B-F method can be used to weight the results to get the final evaluation results.
【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F842.63

【相似文献】