Fire Public Liability Insurance Pricing Based On Non-parametric Methods

In recent years, the fire accidents take on frequently, and have brought great loss to the society. The fire accidents of public places gave birth to the urgent needs of people to fire public liability insurance. But now, the narrow fire public liability insurance coverage, less-developed pricing technology and so on, make the fire public liability insurance difficult to develop. This paper explores the fire public liability insurance pricing problems in perspective of the fire public liability insurance, provides a guideline for the fire public liability insurance scientific and rational pricing.Before the study of fire public liability insurance pricing, this paper firstly clears the fire public liability insurance coverage, on the basis of summing up the experience of overseas development, analyzes the development status and problems of our fire public liability insurance, provides a train of thought and points out the direction for our fire public liability insurance. In order to ensure the scientific and advanced of the fire public liability insurance pricing theory, the paper systematically combines relatively mature and relatively advanced liability insurance pricing theories, by comparing the theories, the final choice is the actuarial pricing theory which is well developed and has less assumptions as the fire public liability insurance pricing methods, then presents the entire pricing process of the fire public liability insurance.When using an actuarial approach to the fire public liability insurance pricing, the method of fitting a loss distribution is an important guarantee for the fitting precision of the loss distribution. So this paper introduces the non-parametric method for the loss distribution. Compared with the parametric method, the method does not presuppose the loss data follow a specific distribution, based on the sample data’s own characteristics, makes the loss distribution more reasonable and realistic to ensure the accuracy of the loss distribution and premium. Fitting the loss distribution in the use of non-parametric kernel density estimation method, we obtain the expected mean loss of the fire losses, and then get the pure premium.On this basis, taking the particularity of fire public liability insurance into account, combined with factors such as business area, buildings, structures, fire safety facilities and other factors on the influence of the total premium, this paper uses the discriminatory rating of fire public liability insurance, which make the estimated results more scientific and reasonable, and is conducive to long-term stability of the insurance company’s operations and the sound development of fire public liability insurance.To verify the feasibility of the pricing method, this paper makes an empirical research by Chinese wholesale and retail industry fire accident loss data.Innovative explore in the following areas:(1)When fitting the distribution of the fire public liability insurance losses, introduces the non-parametric kernel density estimation method, when fitting the data, there are no prior assumptions that the sample data follow a particular distribution, but from the characteristics of the sample data, makes the fitting results more accurate and realistic, overcome the defects of the actuarial pricing method using parameter fitting loss distribution.(2)Improving the fire loss distribution payment standard. In dealing with the empirical data, this paper thins the casualties compensation standard, ensures the accuracy of the loss compensation.(3)Examining the applicability of non-parametric methods in the fire public liability insurance, and expanding the idea of a fire public liability insurance pricing. Using the wholesale and retail industry as the research object, this paper provides the fire public liability insurance process.