Research On The Rating Of Automobile Insurance Based On Generalized Linear Models

With the boost of automobile industry, Chinese automobile insurance industry has developed fast. Nevertheless Chinese auto insurance industry has shadowed by three problems i.e. low efficiency, repetitive insurance products and twisted price system, which hinder auto insurance industry's further and sustainable development. Accordingly it's equally important to explore how to establish fair and stable rate making system for the exploration can enhance the rate market's opening-up and it can help achieve the auto insurance's scientific operation and promote Chinese auto industry's health growth.At present, generalized linear model is one of the most commonly used methods in premium ratemaking. The basic theory of traditional premium ratemaking is to find a suitable model which makes the minimum difference between the observed cost of loss and net premium, but the weakness of this model is the limited variable, which is not practical when there are lots of related pricing variables. What's more, the iteration of this method is quite hard and the result of it is not accurate. In addition, the traditional linear model makes a wide analysis on significant level and correlation. But the assumption of normality, linearity, constant variable, makes the model unpractical in lots of researches. The GLM makes a huge progress in theory, structure of model and diagnosis of the model that makes it acceptable for rating.Automobile factor, the driver factor and circumference of driving make a whole risk system to categorize the multiple systems, a feasible way is to get the corresponding risks rating by some methods of classifying risks. The reform of Auto insurance rates was carried out in 2003, many experts believe that the auto insurance's rating must consider adding the variable factors about drivers and areas. The paper based on generalized linear mode focus on automobile insurance in China, which factors about drivers and areas will significant impact the claims. Through analysis on Sample data we can conclude that factors of drivers and areas impact significantly the claims. The paper will divide all drivers into several groups according different rating factors, and determine those groups' premium. Rating automobile insurance must consider loss frequency, loss severity, additional costs and other financial factors. Take a case for example, the paper respectively discuss about the generalized linear model modeling of loss frequency and loss severity, from the statistical analysis to model selection and diagnosis; at the same time, the paper makes sure that human and area factors are important in rating of automobile insurance.