Study On Forecasting The Exchange Rate Of RMB Based On A Series Of MIDAS Models

As the rate of one currency exchanges to another currency,exchange rate can not only influence the balance of a country's domestic economic development,but also have important influence on the attainment of a balanced development of the international financial market.With the deepening of globalization,China has become more and more influential in the world economy.As becoming a member of Special Drawing Rights(SDR)basket,RMB is becoming more and more appreciated by the international market.Exchange rate of RMB is an important reference for the international financial institutions to measure the value of RMB,as well as an important reference for the monetary policy of the People's Bank of China.Therefore,the real effective exchange rate(REER)of RMB,which can be regard as a leading indicatorhas of the value of RMB,has attracted the attention of the whole world.This paper intends to introduce the MIDAS,using the data of exchange rate of RMB against dollar,euro and yen as high-frequency variables,to forecast low frequency variable of RMB real effective exchange rate.On one hand,this paper uses univariate Mixed Data Sampling(MIDAS)model and multivariable MIDAS model to forecast RMB real effective exchange rate,and at the same time compare the prediction effect with the benchmark model.On the other hand,considering the characteristics of the nonlinear of exchange rate data,this paper adds Markov regime switching to MIDAS model and establish MS-MIDAS model to forecast RMB real effective exchange rate.In order to simplify the model,this paper uses the weighted method of combination MIDAS model to change multivariable model to combination model,establishing combination MS-MIDAS model to forecast RMB real effective exchange rate.The results show that the simulation effects of the univariate MIDAS model and multivariate MIDAS model are better than benchmark model.Adding the autoregressive item to MIDAS model,the simulation effects get better than before.The combination MIDAS model is superior to the single variable MIDAS model and greatly simplifies the modeling degree.In series of MS-MIDAS models,MSIH(2)-MIDAS has obtained more effective prediction effect.The real effective exchange rate of RMB can be divided into two regional systems.According to the results of the forecasting model,it is expected that there will be a small appreciation space in the first period of the future.However,the actual effective exchange rate of RMB will fluctuate in the second and third period of the future.