Research On The Model Of Car Insurance Rate Determination Under The Condition That The Number Of Claims And The Intensity Of Claims Are Dependent

At this stage,the rapid development of our country's auto industry has promoted the development of our country's auto insurance market.Auto insurance accounts for nearly 70%of the market share in property insurance and is the most important link in our country's property insurance.However,the competition of various auto insurance businesses lacks differentiation,and there is serious homogeneity among products.Insurers are launching price wars in order to compete for the market,resulting in a loss in the auto insurance industry.In order to change the status quo of auto insurance,our country has gradually introduced commercial fare reform clauses and gradually transferred pricing power to insurance companies.However,the previous research on auto insurance pricing by insurance companies is often based on the independent assumption of the number of claims and the intensity of the claims,but the dependence of the number of claims and the intensity of the claims will lead to unreasonable determination of the auto insurance rate.In order to make auto insurance pricing more accurate,provide suggestions for insurance companies' auto insurance pricing,and change the loss situation of the auto insurance industry,this article considers the auto insurance rate determination model under the assumption of dependency.Based on previous studies on the determination of auto insurance rates,this paper uses data from an insurance company in Shandong Province to conduct research on the determination of auto insurance rates.Previous studies are often carried out under the assumption that the number of claims and the intensity of claims are independent,ignoring the interaction between the number of claims and the intensity of claims.However,this article analyzes the influencing factors of the claim intensity and tests the correlation,and finds that the number of claims has a negative impact on the claim intensity.Therefore,this paper studies the method of determining the auto insurance rate under the assumption that the number of claims and the intensity of the claim are dependent,and compares the fitting effect of the model established with the same data under the assumption that the number of claims and the intensity of the claim are independent.In this paper,the zero-expansion Poisson-inverse Gauss model and Tweedie model are established under the condition that the number of claims and the amount of claims are independent,and the two-step model,generalized linear mixed model and Copula model are established under the assumption that the number of claims and the amount of claims are dependent.The fitting effect of the comprehensive model and the prediction effect of pure premiums.Although the model established under independent assumptions can fit the claim data well,its predicted value is higher than the actual value,and the determination of pure premiums is not accurate enough.The fitting and predictive abilities of the three models built under dependent assumptions are different.Simply modeling the number of claims as an independent variable can only simply indicate the adjustment effect of the number of claims on the amount of claims,unable to express the complex function adjustment relationship between the number of claims and the amount of claims.Adding the number of claims as a random effect to the established generalized linear mixed model and Copula can make the complex correlation between the number of claims and the intensity of the claims work in the model,which better reflects the negative effect of the number of claims on the amount of claims.Finally,it can be obtained through comparative analysis that the Copula model has the best fitting effect and prediction effect.Although the generalized linear mixed model predicts pure premiums slightly inferior to the Copula model,it is easy to popularize and apply,and its algorithm can be further refined.This article recommends that insurance companies fully consider the impact of the interaction between the number of claims and the intensity of the claims on the accuracy of the model when pricing auto insurance,and establish a Copula model or a generalized linear mixed model to reversely adjust the predicted value to fit the number of claims and the amount of claims.Dependency relationship.Secondly,since this article found that vehicle weight,plate type,and number of seats have a significant impact in any model,the current stage of car insurance rate determination should fully consider the impact of vehicle weight,plate type and number of seats.Pay attention to the interaction between the insignificant related factors in the auto insurance influencing factors.