Research On International Contagion And Warning System Of Financial Crisis

In the past 20 years three financial crises have shown transnational contagion,with the economic globalization and financial liberalization continue to deepen,the international contagion of financial crisis has shown the trend of normalization.,and the mechanism of transmission has also shown a complex and diverse state.The depth and breadth of the impact of the financial crisis are also expanding,not only in economic issues,but also in political issues and livelihood issues.While the global economy has begun to recover,the rise of conservative forces and protectionism has added to the downside risks,in the meantime China's economy has entered the "new normal" phase,debt problems and imbalances in the financial system highlight the contradictions between increased economic fragility and reduced policy swing space.At present,China is in the critical period of economic restructuring and financial restructuring,the construction of “One Belt And One Road” and the process of RMB internationalization are advancing,under the double pressure of the increasing external environment uncertainty and the deepen internal contradictions,we should prevent external financial risk from the impact on our country,on the other hand,we also need to be vigilant about the internal risks that could trigger a financial crisis.Based on this background,we will focus on the two core issues in the financial crisis which are "international contagion of financial crisis" and "financial crisis warning",in order to provide a theoretical and reliable reference of external financial risk prevention and internal risk warning basis.The main works of this paper are as follows:In Chapter 2,we estimate the asymmetric Granger-Geweke causality of different stock market by using data of forty countries and regions' stock index from 2004 to 2015 to construct global stock networks based on the framework of constructing complex network-Analysis of network structure-simulation of financial contagion.Meanwhile,we quantify the contagion to global and China's stock market through the perspective of network structure variation,utilize epidemic model to simulate the process of financial contagion and compare with the result of DCC-GARCH model.The results indicate that the structure of global stock networks breaks when shocked by financial crisis.The subprime crisis' s infectiousness is stronger than European debt crisis to Global stock market but weaker than European debt crisis to China.Neither S&P 500 nor ASE infect SSEC which is infected indirectly.The impact of external shocks has the trend of aggrandizement though the internationalization level of Chinese stock market is still lower.There exists diffusion threshold and collapse threshold for contagion on Subprime crisis structure.However,Greece as the single infection source can't lead to the collapse of European crisis network.Besides,Countries can adjust macroeconomic policy for coping with financial crisis in latent period,but the longer latent period is,the stronger infectiousness is.The test results of complex network analysis on the contagion of financial crisis are verified in the DCC-GARCH model,and its conclusions are robust.In Chapter 3,we use the Space panel model to test the contagion mechanism of financial crisis empirically.We have found that: the risk level of each country's financial system increased markedly during the financial crisis,and the contagion phenomenon of financial crisis is obvious.The subprime crisis caused global contagion through three mechanisms of monsoonal effect,spillover effect and pure contagion,and the contagion of financial crisis has obvious economic-institutional spatial aggregation;the contagion mechanism of financial crisis to emerging economies and developed countries is greatly different.In particular,the liquidity of the US financial market is the only transmission channel to the developed countries,the role of economic growth is more obvious to the developed countries and the emerging market economies are influenced by the price channels;for the trade spillover effect,the financial crisis contagion to the developed countries is through export channels,Emerging markets are contagious through the competitive devaluations of currencies.The financial spillover effect of emerging market economy is more obvious,the contagion of financial crisis in developed countries has shown economic institutional space dependence,pure contagion phenomenon is more remarkable.In Chapter 4,we use the 2007 international trade data to construct a global macroeconomic network which consists of 148 countries and regions.Then we use the S.I.R model to simulate a situation which is when a country suffers from different levels of the financial crisis what is the number of countries it can affect and make a difference to these countries' total economic output,we have found that there exist differences between the national infectious which show regional characteristics,many small economies also have the potential to cause a serious crisis;using dynamic clustering method,we classify the infectious of 148 countries,the countries in category 1 to category 4 are the potential source of infection of the international financial crisis,by calculating the Spearman correlation coefficient between infectious and a set of data(GDP,imports,network topology indicator),we have found that countries with higher GDP and import but with low clustering coefficient tend to perform negative overflow when facing the low level crisis,however,countries in the center of the network and with more trading partners have the potential to cause a serious crisis during a higher level of crisis,using data in 2009,the main conclusions remain robust.In Chapter 5,we use the Logit model and machine learning method(the classification tree and classification tree combination mode)to build a warning system for the financial crisis and compare the warning effects of different models.We have found that in Logit model,in the case of selected financial crisis warning index,the warning effect of currency crisis has the best performance,followed by the warning of the sovereign debt crisis and the last is that for a systemic bank crisis,there are many differences between the marginal contribution of probability even for the same warning index during different types of financial crisis;For classification tree and combination classification tree model,in the case of early warning indicators having been selected,random forest model perform better in warning effect than that of Bagging model,binary classification tree model is even not good as Logit model;In the situation of including all the indicators,the warning performance of the random forest model,the Bagging model and the binary classification tree model are all improved,but the binary classification tree model is still not better than the Logit model.In addition,for the three crises,the warning effect of 3 years in advance is the best,followed by 2 years in advance,and 1 year in advance.There are differences between the leading indicators during different types of the crisis,the real effective exchange rate,M2 growth,public debt/GDP,trade openness,domestic credit/GDP,foreign debt/export and foreign exchange reserves/GDP are common indicators causing systemic banking crises,currency crises,and sovereign debt crisis.In Chapter 6,we have chosen related indicators of the banking sector,the stock market,bond market and the foreign exchange market,then we use factor analysis method to construct high frequency financial pressure index of China,the United States,Germany and Japan.Furthermore,we use the financial stress events and zone distributions to judge the effectiveness of financial stress index.Finally,we use the TVP-VAR model to test China financial stress index's macro-economic effect empirically.We have found that the financial stress index in this chapter can reflect the risk status of the financial system and identify the financial stress zone characteristics effectively,which we can tell this index has a good applicability and use it as an effective tool to set up the early warning of domestic and foreign financial risk in real-time;the China financial stress index has obvious macroeconomic effects,the rise of financial risk will restrain economic growth,raise interest rates level and reduce the inflation level.This index also has a warning effect on the trend of macroeconomic.