| dc.coverage.temporal | 計畫年度：90 起迄日期：20010801～20030731 | en_US |
| dc.creator (作者) | 郭炳伸 | zh_TW |
| dc.date (日期) | 2001 | en_US |
| dc.date.accessioned | 18-Apr-2007 16:36:11 (UTC+8) | en_US |
| dc.date.accessioned | 8-Sep-2008 15:48:36 (UTC+8) | - |
| dc.date.available | 18-Apr-2007 16:36:11 (UTC+8) | en_US |
| dc.date.available | 8-Sep-2008 15:48:36 (UTC+8) | - |
| dc.date.issued (上傳時間) | 18-Apr-2007 16:36:11 (UTC+8) | en_US |
| dc.identifier (Other Identifiers) | 902415H004013.pdf | en_US |
| dc.identifier.uri (URI) | http://tair.lib.ntu.edu.tw:8000/123456789/3794 | en_US |
| dc.identifier.uri (URI) | https://nccur.lib.nccu.edu.tw/handle/140.119/3794 | - |
| dc.description (描述) | 核定金額：800100元 | en_US |
| dc.description.abstract (摘要) | We set out to compare the out-of-sample forecasting performance of a linear AR model and a nonlinear TAR model. The idea was to see if the intuitively appealing suggestion of nonlinear adjustment of real exchange rates toward purchasing power parity, due to market frictions, could be utilized in forecasting. Indeed, we found the tendency of the US/DEM real exchange rate to converge faster when the previous real exchange rate was farther away from the estimated purchasing power parity level. However, this tendency did not appear to be strong enough for the TAR specification to lead to consistently better out-of-sample forecasting performance. One reason may be that the data sample is not long enough to reliably estimate the non-linear specification which requires more parameters to be estimated than the more parsimonious linear specification. The results presented in this paper are very preliminary. The next step we want to pursue is to check the robustness of the results to the assumption on the inner regime process. Specifically, we want to a restrict the inner regime to follow a unit root process. This is reasonable given that the estimated inner regime does not appear mean reverting. Also, we would like to tackle the forecast uncertainty further. Recently, Kitamura (1999) has paid attention to the uncertainty involved in the coefficient estimates and the resulting risk involved in the forecasts. He suggests using bootstrap smoothing to reduce the variance of the forecasts. The intuition is that since the forecast is sensitive to the realization of training period samples, its variance is reduced by first perturbing the training observations using the bootstrap and then averaging over the perturbed series. Following this idea we intend to combine bootstrapping with the MC forecasting hoping to reduce the variations in the TAR forecasts and thus potentially improve their overall accuracy. Also, we want to check for the robustness of the results for the assumptions about the error process. Doing the estimations for other currencies would help in gaining some understanding of the generality of the results as well. If our current results are found to hold with these robustness checks, we would conclude that the theoretically and intuitively attractive idea behind the nonlinear threshold adjustment may not be empirically significant enough to help us in understanding the future behavior of real exchange rates. | - |
| dc.format | applicaiton/pdf | en_US |
| dc.format.extent | bytes | en_US |
| dc.format.extent | 140683 bytes | en_US |
| dc.format.extent | 140683 bytes | - |
| dc.format.extent | 20409 bytes | - |
| dc.format.mimetype | application/pdf | en_US |
| dc.format.mimetype | application/pdf | en_US |
| dc.format.mimetype | application/pdf | - |
| dc.format.mimetype | text/plain | - |
| dc.language | zh-TW | en_US |
| dc.language.iso | zh-TW | en_US |
| dc.publisher (Publisher) | 臺北市：國立政治大學國際貿易學系 | en_US |
| dc.rights (Rights) | 行政院國家科學委員會 | en_US |
| dc.subject (關鍵詞) | 實質匯率;預測;門檻迴歸模型;自動迴歸模型 | - |
| dc.subject (關鍵詞) | Real exchange rate;Forecasting;Threshold autoregressive model;Autoregressive mode | - |
| dc.title (題名) | 實質匯率預測：TAR與AR之比較 | zh_TW |
| dc.title.alternative (其他題名) | Forcasting Real Exchange Rate Out-of-sample: TAR vs. AR | - |
| dc.type (資料類型) | report | en |
