Study Of The Predicting Models Of Solvency In Property Insurance Companies

The insurance industry has been developing quickly, since the insurance business was recovered in1980s. The total assets of insurance reached4.9trillion in2010, and china has been one of the most important countries in the word. The stable development of insurance is significant to our country’s economic development. The adequacy of the solvency of insurance companies not only affects insurance companies’sustainable operation, but also affects the stable development of China’s insurance industry and financial markets.Currently our insurance supervision system is consisted of solvency, market conduct, company governance, which the solvency supervision living in the heart of the regulatory system. However, there were still7insurers were shorting with solvency. How to play the central role of solvency regulation in risk prevention is an important task for China’s insurance regulator. It is significant to establish a sensitive solvency warning system for improving our countries solvency supervision system.The principle of principal component analysis is dimension reduction. It reduces many indexes into a few integrated indexes. BP Neutral Networks can imitate, simplify and abstract biological brain nervous system.It has the advantages of adaptability, fault tolerance and self-organization, which traditional statistical methods can’t compare.This paper sum domestic and foreign papers which study the influencing factors of insurance solvency or solvency prediction. After comparing several typical Influence Factors of Solvency and solvency projections econometric models, we chose principal components analysis to study the factors of solvency and BP Neutral Networks to predict solvency.In this article we use principal components analysis reduce13determinants that affect solvency to6principal components. Taking if Solvency adequacy ratio reach 100%as standard, we use BP neural network prediction to forecast the solvency of the sample companies in the coming year and the next two years. The results prove that BP neural can identify the insurance companies which are short of solvency at the rate of90percent. Finally, according to the empirical results, a number of advices to improve solvency regulation are given.