Structure Time series model Analyses and Application in Taiwan Dollar per U.S. Dollar real Exchange Rate
| Autor: | Yi-Che Kuo, 邱乙哲 |
|---|---|
| Rok vydání: | 2004 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 92 Using structural time series model studies the NT/US real exchange rate of the unobservable component, the trends, seasonal variables, and cycle variables. The advantage of structural time series is that it can produce the estimator, filter, confidence interval. Furthermore, it also forecasts the time series data and has the good property of the statistics. The studied object of the structural time series is NT/US real exchange rate from the period of Jan in 1995 to Dec in 2003. In the research, we discover that the Taiwan dollar per U.S. dollar real exchange rate is affected by the trend variables, not by the seasonal variables and cycle variables. Furthermore, there is an intervention on the residual plot of the trend variable in April 2000. According to the purchasing power parity and uncovered interest parity, we use two explanatory variables, which are the different of the interest of two countries and the ratio of price index of the two countries, to fit the model. Finally, we can find out the best forecast model that composed of the trend variable, intervention variable, and explain variables that these are the index of the U.S. and the price index of the Taiwan. Finally, we get the conclusion. The first conclusion is that the purchasing power parity and uncovered interest parity don’t well explain the the Taiwan dollar per U.S. dollar real exchange rate perfectly but the price index is still considered the important variable. The second conclusion the structural time series model compared with the random walk model, the random walk model has better the forecasting ability. The third structural time series model has the stable predicted interval. Finally, the trend variable increases 7.87% per year, and the NT/US real exchange rate tend to depreciate on the first half in 2004. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|