A Systematic Study on Financial Crisis Early Warning Model
|School||Dongbei University of Finance|
|Keywords||financial crisis early warning model indicator system|
Financial crisis which happened in 2008 brought the world a great loss. The financial crisis spread to China soon after it happened because China is very open these years. Therefore, if we can predict financial crisis precisely, the government can take measures to reduce the economic loss.The article is based on the financial crisis warning methods and finally work out what kind of methods is suitable for China through the research. The research is focus on various types of financial crisis early warning model and the method of selection index system. In-depth investigation and analysis on each kind of specific methods can be found in the article. The article also discusses the characteristic of each method with the purpose to get a Chinese method which is suitable for China.Firstly, I learnt the financial crisis early warning system which advanced by foreign scholars. It mainly includes FR probability model, STV cross-section regression model and KLR signal analysis model. "The triple estimation" of FR probability model greatly reduces the accuracy of the model. Besides, it adopts abroad sample data from many countries, which may be not suitable for the characteristics of our country. In addition, the number of the data used in FR probability model is very little, therefore it is not satisfied "the large numbers law". STV cross-section regression model only considered three factors which are the real exchange rate, loan growth and international reserve/M2. However, the number of the factors is not enough. What’s more, it adopts linear regression which lacks the theory supporting. Besides, it is hard to find the similar countries. From the above two models, we can speculate that we cannot put all nationals together to study the early warning models. KLR signal analysis model does not consider political factor. This model is not a structural economic model, so the factors why and how to affect the likelihood of a crisis is not achieved. We can combine KLR signal analysis model with FR-logistic probability model based the factor analysis to overcome the defects. Secondly, the overseas scholars’financial crisis early warning models and applying them in China are introduced. Some of them modify economic indexes and definition of monetary crisis. We can say that they are exploring the suitable models for China.Furthermore, I introduce the domestic financial crisis early warning models which include the financial crisis pre-warning system structure, multiple time-scale warning model based, P-S financial crisis warning method and PHM warning model. All of these models are immature.Finally, By studying the existing research results at home and abroad, we know that KLR signal analysis model system is the most perfect system in choosing indexes. FR probability model can forecast a financial crisis accurately. And I get a financial crisis early warning model suitable for china, that is combining KLR signal analysis model system and FR-logistic probability model based the factor analysis together.From now on, we should resolve the following problems:Firstly, we should consider the contagion of financial crisis. Secondly, it is better to unify the qualitative analysis and quantitative analysis together. Thirdly, we should modify models according to the characteristics of China.