With financial liberalization and economic globalization, international trade and capital transactions result in larger exchange rate fluctuations than in the past. Besides, it can¡¦t be ignored that the change of exchange rate influences the economics and real exchange rate which be regarded as the indicator of external competitiveness becomes more important than before, so my paper aims to know not only whether there is stochastic segmented trend in their fluctuation but also the factors which are closely related to regime switches. As we all know that it is significant to forecast the volatility of exchange rate in the global society. A number of previous studies discussed the relationship between exchange rate and fundamentals under the monetary models, however many people found that these models are handicapped in out-of sample forecasting. Therefore, I compare the forecasting performance of the real interest differential monetary (RID) model of Frankel (1979) with the models which I built in the paper. By using the market share of the top ten currencies in 2010 which is published by Bank for International Settlements (BIS) and the fundamentals. The empirical results indicate that fundamentals do not only matter for real exchange rate, but are also related to the switches between the regimes. Besides, the real exchange rates are highly persistent in each regime and the effect of fundamentals is different in different countries. At last, my result suggest that the models which I built in the paper provide better forecast in the yen, pound sterling and New Zealand dollar than the RID model.

This study mainly concentrates on the exchange rate problems between Taiwan and China, so cross-strait economic and the evolution of exchange rate regulation regime would be the first work so as to provide some policy suggestions about the cross-strait trading settlement .The empirical work has two parts. The first part is examining the Currency Substitution(CS) of Taiwan, then uses Cointergration and Vector Estimate Correction Model for the short and long run condition. Currency Substitution would cause the volatility of exchange rate, and the next procedure is using the Markov Regime Switch Model to analyze the exchange rate of two countries from 03 January,1994 to 30 April,2012. The main purpose of this study is examining whether the two markets have a significant regime switch or not, then the empirical result finds that both markets have regime switch .Considering the difference of exchange rate regime in two countries, the decision of the cross rate becomes more prudent because China authority may underestimate the exchange to disturb the export of Taiwan.