Research On RMB Exchange Rate Based On MS-GARCH Models

The RMB has become one of the IMF's official reserve currency,from the beginning of October 1,2016.With the RMB on the international currency stage,increasing the risk of foreign exchange,effective prevention and control of foreign exchange risk have become a priority.The MS-GARCH(Markov-switching GARCH)models have the characteristics of revealing the influence of the unobservable variables in the fluctuation of the financial sequence on the fluctuation of the sequence.Therefore,we use the MS-GARCH models to study the RMB exchange rate risk.This thesisi is divided into five chapters.In chapter 1,We mainly introduce the background,significance of this study and then introduce the content and innovation for the research.In chapter 2,The ARCH family models?GARCH family models?the basic theory of Markov switching models and the factors that affect the exchange rate volatility are introduced.In chapter 3,The moment,the kurtosis and the skewness of the MS-GARCH(1,1)model are calculated.The results show that the difference of the means of the state is a important factor for the Markov switching models produce the skewness.The variances of the two states is not enough to produce skewness.The mean value of the RMB exchange rate is very small,it is suitable for the GARCH models with Markov switching.At the same time,VaR and Kupice are introduced.In chapter 4,Choosing the exchange rate data from July 2005 to June 2015 as samples.Firstly,the exchange rate data are processed by logarithm,and the stationarity test and correlation test are carried out.Secondly GARCH(1,1)model and EGARCH(1,1)model are structure to simulate the rate data based on normal distribution,t distribution,GED(Generalized error distribution),at the same time,we calculate the VaR value and use Kupice to estimate the validity of the model,the optimal model is GARCH-G.Finally,the GARCH-G model and the MS-GARCH model are compared to analysis the rate data.The empirical analysis shows that:(1)The rate of return of RMB exchange rate is non normal stationary sequence,thick tail,aggregation and asymmetrition.(2)The thesis prediction effect and failure frequency test data show that the EGARCH-G model is better than the other GARCH family model.(3)The MS-GARCH model can be a more comprehensive exchange rate fluctuation in reaction,and can advance the identification of major events,comparing with GARCH-G model can better reveal the impact of financial events of China's foreign exchange market.In chapter 5,the conclusions of this paper are aummarized.