基于损失分布法的健康保险欺诈风险度量研究

发布时间：2018-01-25 03:46

本文关键词： 保险欺诈损失分布法蒙特卡洛模拟度量　出处：《青岛大学》2014年硕士论文　论文类型：学位论文

【摘要】：保险业是金融行业的三大支柱之一,保险行业的健康发展关系着一个国家的经济命脉、社会安定和人民生活。但自保险业诞生以来,保险欺诈便如影随形,而且近年来保险欺诈案件不断增多,严重扰乱了保险市场秩序,侵害了诚实投保人的合法权益,威胁着保险行业的健康发展。基于此,本研究对保险业较为严重的健康保险欺诈问题进行了研究。本研究利用2006年至2013年的新农合欺诈损失数据,以一年为一个周期,借鉴损失分布法原理,给定置信区间,以VaR方法为基础,对欺诈损失频率和欺诈损失强度分别建模,拟合出它们的最优分布函数,然后根据这两个分布函数拟合总损失的最优分布函数,并采用蒙特卡洛模拟方法进行拟合,从而计算计提欺诈风险损失准备金。实证结果表明：新农合欺诈损失频率的最优拟合为正态分布,欺诈损失金额的最优拟合为韦伯分布。在此基础上,使用Matlab软件进行总损失分布的蒙特卡洛模拟10000次,得出10000个欺诈损失金额的模拟结果。根据VaR模型定义,在一年内新农合欺诈损失金额有99%的可能性不会超过1749.46万元,即在99%的置信水平下,新农合基金每年需要计提1749.46万元的欺诈风险损失准备金。鉴于每年健康保险欺诈损失金额巨大,本研究还针对健康保险欺诈问题的预防提出了几点防范措施。
[Abstract]:The insurance industry is one of the three pillars of the financial industry, the healthy development of the insurance industry is related to a country's economic lifeline, social stability and people's livelihood. And in recent years, the increasing number of insurance fraud cases seriously disturbed the order of the insurance market, violated the legitimate rights and interests of honest policy holders, and threatened the healthy development of the insurance industry. This paper studies the serious health insurance fraud in the insurance industry. In this study, the data of fraud loss of NCMS from 2006 to 2013 are used for a period of one year, the principle of loss distribution method is used for reference, the confidence interval is given, and the VaR method is used as the basis. The frequency of fraud loss and the intensity of fraud loss are modeled, and their optimal distribution functions are fitted, and then the optimal distribution functions of total losses are fitted according to the two distribution functions. Monte Carlo simulation method is used to calculate the reserve for fraud risk loss. The empirical results show that the optimal fitting of the loss frequency of the new rural cooperative fraud is normal distribution. The optimal fitting of the amount of fraud loss is Weber distribution. On this basis, the Monte Carlo simulation of total loss distribution using Matlab software is carried out for 10000 times. According to the definition of VaR model, the probability of new rural cooperative fraud loss within one year is less than seventeen million four hundred and ninety-four thousand and six hundred yuan. Namely, under the confidence level of 99%, the new rural cooperative fund needs to raise seventeen million four hundred and ninety-four thousand and six hundred yuan of fraud risk loss reserve every year. In view of the huge amount of annual health insurance fraud loss. This study also puts forward some preventive measures for the prevention of health insurance fraud.
【学位授予单位】：青岛大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F842.684;F323.89;R197.1

【参考文献】