Study About The Predicting Models Of Solvency For Property Insurance Company In China By The Artificial Neural Network Model

The rapid development of Chinese insurance industry makes the insurance penetration has been increasing in the social and economic life, in the meantime it also makes the insurance industry do more responsibility for economic security. The current shortage of solvency is a major problem in the development of the insurance industry, and twelve Chinese insurers were insolvent in 2008.The solvency problem of Insurance company is related to the sustainability of operations and interest of the insured, what's more, it will influence the safety and steadiness of the insurance market operation. Insurance companies should try to achieve the capital adequacy and to establish the mechanism of capital management according to the company development strategy and business planning management, in order to improve the insurance competitiveness imperative. We need to establish predicting models of solvency to improve the solvency regulatory system for Chinese insurance industry. BP neural network model is a kind of artificial neural networks, which mimic biological functions of brain information processing system. It can achieve an arbitrary and optimized non-linear mapping between the input and output. It has a adaptability that traditional statistical methods can not match, and it also have fault tolerance and self-organization. It has certain advantages in the insurance solvency forecasting.This paper reviews the domestic and international insurance solvency prediction research literature, and explain the concepts of insurance solvency. And compared to the other typical predicting models of solvency, the BP neural network model is more applicable. It selects the property insurance company data in China from 2004 to 2007 with complete financial statements, and makes solvency adequacy ratio of 100% as a solvency standard. And it selects the eleven indicators that influence insurance company solvency and use BP neural network model to make solvency prediction of the sample companies in the coming year (T+1)and the next two years (T+2).It proves that BP neural network model of the property insurance companies has a good predictive result compared two methods, especially the second prediction method is good at finding a company with insufficient solvency. Finally, it gives several recommendations to improve the use BP neural network prediction in solvency based on empirical research findings.