The Application Of MCMC In Enterprise Property Insurance Pricing

In property insurance companies, risk premuim method is an important method to price enterprise property insurance. Since experience rating is a very important part of risk premuim method, both insurers and governments focus on it. Bühlmann-Straub credibility model method has been widely used in experience rating, which is very simple, and many insurers prefer it. Yet Bühlmann-Straub credibility model method also has many disadvantages. For computing the premium in the future, when the payment data is insufficient , the estimated experience premium doesn't reflect the stochastic characteristic and is not dependable. As a result, the Bühlmann-Straub credibility model method provides the value of point estimation, people can not get the distribution of the estimation interval. In order to solve the problem, I'm trying to modify the Bühlmann-Straub credibility model method in this thesis.With the development of statistic and computer technology, stochastic simulation has been used in risk premuim rating. In this thesis, we construct the DAG credibility model to get the payments patterns in the future years on the basis of traditional Bühlmann-Straub credibility model by using MCMC algorithms and with the implemented use of WinBUGS. Comparing with risk premium method based on traditional Bühlmann-Straub credibility model method, the stochastic simulation method can objectively simulate the average claims, risk exposures, credibility factor and credibility premium in the future years as well as improve the precision of the numeration, which provides more information about insurer'price making for property insurance companies.