Research On Health Insurance Fraud Risk Measurement Based On Loss Distribution Method

Insurance is one of the three pillars of the financial industry, and the healthy development of insurance industry is of vital importance to the economic lifeline, social stability and people’s living standards of a country. But since the birth of insurance industry, insurance fraud will go hand in hand. In recent years, the insurance fraud cases is increasing, which disrupts the order of the insurance industry infringes the legitimate rights and interests of the bona fide insured, threatening the healthy development of the insurance industry.The paper made a research on the health insurance frauds appearing in high frequency in insurance industry.The paper made a model for the fraud loss frequency and intensity respectively and fit out their optimal distribution function with a year as a cycle in a given confidence interval, based on the loss distribution method and VaR method, by using the2006to2013NCMS fraud loss data. Then it fit out the optimal distribution of the total loss of function based on these two distribution functions and calculate the accrued fraud risk boss reserves by using Monte Carlo simulation method. The empirical results show that the best fit for NCMS fraud loss frequencies is the Normal distribution and the best fit for loss amount is the Weibull distribution. On this basis, the paper calculated the total losses for10000times by using Matlab, and got10000fraud loss amounts. According to the definition of the VaR model, NCMS fraud losses amount likely won’t exceed17.4946million yuan with99%chance, namely the NCMS fund need to withdraw17.4946million yuan for risk loss reserves every year under the confidence level of99%. Given the annual health insurance fraud loss amount is huge, this study also puts forward some preventive measures for the health insurance.