【作者】 邵学清；

【导师】 袁卫；

【作者基本信息】 中国人民大学 ， 应用数理统计， 2004， 博士

【摘要】 I.选题意义BMS即奖惩系统(Bonus-Malus System),在我国机动车辆保险中将它称作无赔款优待系统,它是机动车辆保险中普遍采用的这样一种体系:在续保时,对有索赔记录的投保人在保费上给予一定的惩罚;而对无索赔记录的投保人给予一定的奖励。保险实践证明,在机动车辆保险中引入BMS非常必要,同时意义重大。1.必要性不同投保人的风险水平是不同的,表现为在相同的投保时间内,不同投保人的出险次数不尽相同,我们称之为风险的非同质性。对不同风险水平的投保人,保险公司收取的保费也是不同的。在保费确定时,保险公司的主要依据是驾驶员的年龄、性别、职业、居住地区以及车辆的种类、用途等,这些变量被称作先验变量。保险公司之所以采用这些先验变量,是为了使分组以后的投保人在各组中具有风险同质性。然而,即使我们选用了这样一些变量分组,各组中的投保人仍然会存在一定的风险非同质性。其主要原因为:第一、一些影响驾驶风险或出险次数的因素是不可度量的。例如驾驶员的驾驶熟练程度,驾驶时注意力的集中程度,在紧张情况下做出反应的速度以及判断的准确性等;第二、由于分类变量有限,可度量的因素也不可能全部包含在这些变量中;第三、由于道德风险的存在,对投保人的分组很难肯定是正确的。如有的国家的机动车保险中要求投保人告知他们的年行驶里程数,在这种情况下,保险公司很难采取什么措施避免投保人的故意低估。如何克服这些不足,使保费体系能够比较准确地反映投保人的风险水平呢?机动车辆保险中的BMS保费体系正是用来达到这一目的的,而且它是一个行之有效的方法。它是在变量分组的基础上,进一步通过每个投保人的索赔记录来对其保费水平进行调整。这在实际中是容易理解的,因为每个投保人的索赔情况正是他们不同的风险水平所决定的,这样,在影响风险水平的诸多因素难以被考虑的情况下,索赔记录无疑是对这些因素的一个较好的综合。2.重大意义加拿大机动车辆保险业于1950年最早采用BMS,以后在北美、欧洲等地的国家普遍采用,而我国机动车辆保险中无赔款优待即是一种BMS保费系统。BMS在实际应用中被证明具有以下的作用:第一、它使每个投保人缴更多还原

【Abstract】 I. SignificanceThe BMS ( Bonus - Malus System ) is namely prize and punishment system, in our motor vehicle insurance it is called to be no-claim bonus system. It is like this a kind of widely adopted system in the motor vehicle insurance: in the next insurance year, the policyholders who have some claim records will be punished by the premium from the insurance company, while those who have no claim record will be rewarded. The insurance practice has proved that, in the motor vehicle insurance, the BMS is very necessary, and the meaning is important.1. NecessityThe risk levels of the different policyholders are different, that is to say, in the same term of insurance, the numbers of claims of different policyholders are not always equal, and we call it heterogeneity of the risk. To the policyholders with different levels of risk, the premiums the insurance company takes in are also different. When establishing the insurance fee, insurance company mainly considers policyholder’s age, sex, occupation, habitation, the use and category of the vehicle...etc, and these variables are called priori variables. The reason why priori variables are adopted is to make the policyholders who are divided in the same group has homogeneity of their risks.But, even we chose such variables to divide the policyholders; there is still some heterogeneity of their risks in each group. The main reasons are: the first, some characters influencing the risk of driving and the number of accidents that may be happen are not measurable, such as the well- trained degree of the drivers, the concentration degree of the attention during the driving hour, the speed of the reaction and the accuracy of the judgment in the nervous circumstance; The second, because the classification variables are limited, the measurable characters of the risk are impossible to be contained in these variables; The third, because of the existence of the moral risk, it very difficult to affirm that the classifying of the policyholders is right. For example, in some countries the policyholders are requested to tell their mileage更多还原