Research On China’s Financial Systemic Crisis Warning Based On Dynamic Logit Model

Taking all the financial crises from1970s into consideration, we can find that thedestructiveness of financial crisis was becoming stronger and stronger, and the rangeof its influence was becoming larger and larger. Especially the2007-2009international financial crisis, which was triggered by the U.S. subprime mortgagecrisis and quickly swept through the whole world, caused huge impact to globalfinancial markets, exposed a series of serious weakness of the global financial system.After this global crisis, systemic financial crisis early warning and macro-prudentialregulation received extensive attention of governments and financial regulators, andsoon becoming a hot spot of research. China joined World Trade Organization (WTO)in late2001, and then gradually opened the financial market. With the deepening ofglobal integration, it is difficult for any country to escape in the volatile internationalenvironment. Now, China’s economic indicators are running good, but make acomprehensive survey on the economic situation at home and abroad, China’seconomy still have a lot of unstable factors. Therefore, to build a financial crisis earlywarning system, which conform to the actual situation of our country, has animportant and profound realistic significance.First, this paper introduced external shock pressure index to reflect the impact ofchanges in the global economy on China, which was based on currency crisis pressureindex, the banking crisis and asset bubble crisis pressure index, and then synthetized asystemic financial crisis pressure index based on the above four pressure index. Whenconstructing the systemic financial crisis early warning index system of China, we chose21representative alternative economic indicators from the aspects ofmacroeconomic, financial system, asset prices and foreign impacts. Then take all thealternatives early warning indicators and the systemic risk pressure exponentconstructed by this article into Granger causality test pairwise, ultimately selectedfour indicators which have significant influence on systemic financial crisis pressureindex, which are as follows: change rate of stock index, growth rate of industrialadded value, actual deposit interest-rate spread at home and abroad, and the ratio offoreign direct investment GDP.Second, we constructed a dynamic Logit model with lagging of binary variables,and by comparing with the traditional static Logit model, found that this dynamicLogit model’s fitting effect is better than that of static Logit model.Therefore, this paper simulated the systemic financial crisis of China in2001-2012based on the dynamic Logit model. Simulation results show that theoverall economy in our country during2001-2012running stability, and thesimulation curve has several larger fluctuation in June2002, in March2005, January2009-February and December2010, especially in2009, the model shows that theprobability of systemic financial crisis in China reached more than80%; Finally, thesimulation curve has several certain amplitude fluctuations in2012, which shows thatChina’s current financial system may still in a state of instability, the prediction of2013crisis warning conditions of China is necessary.And then, we predicted the condition of China in2013by using the dynamicLogit model, with forecasted the four indicators by ARIMA model. Finally, theprediction results showed that China’s probability of systemic financial crisis in2013is extremely low, almost to zero. This result conforms to the fact that theinternational economic situation and the present situation of the domestic economy.The last part is the conclusion of this article. The conclusion simply summarizedas follows: First, the external impact factors have a great influence on the outbreak of asystemic financial crisis in China. Among the four indicators which have significantinfluence on systemic financial crisis, there have two foreign impact indicators.Second, the method of Granger causality test is more in line with the actualsituation of China which have no systemic financial crisis.Third, the dynamic Logit model which containing the lagged binary variable notonly considered the effect of macroeconomic variables, but also considered theimpact of financial crisis on the sustainability of time dimension. By comparing withthe traditional static Logit model, found that the dynamic Logit model is better thanthe static Logit model.Fourth, the prediction on the possibility of a systemic financial crisis in2013ofChina based on dynamic Logit model is very small, almost to zero.Under the current situation of increasingly close global economic ties, the resultsof this article have very important practical significance for our country to build anearly systemic financial crisis warning system, and effectively preventing systemicfinancial crisis.