英文摘要 |
This paper examines the real exchange rates of the New Taiwan dollar, Korean won and Singapore dollar with a view to determining whether there is stochastic segmented trend in their fluctuation. An econometric model is used to identify transitions in state in the real exchange rates of the three currencies resulting from changes in market fundamentals, and to present the alternation of upward and downward swings. The empirical results indicate that, in terms of in-sample fitness capacity, the performance of the two-regime-drift random walk model with Markov switching mechanism is superior to that of the traditional single-regime random walk with drift model. The New Taiwan dollar presented five regime changes with long depreciation swings and short appreciation swings. The Korean won and Singapore dollar presented four and nine regime changes, respectively, with long large-depreciation swings and short small-depreciation swings. However, in terms of out-of-sample forecasting performance, the traditional single-regime random walk with drift model is superior to the two-regime-drift random walk model with Markov switching mechanism. |