| dc.contributor | 國貿系 | |
| dc.creator (作者) | Tsong, Ching-Chuan;Waung, Yie-Yuh | |
| dc.creator (作者) | 欉清全;汪義育 | zh_TW |
| dc.date (日期) | 2000-09 | |
| dc.date.accessioned | 6-Oct-2015 16:34:04 (UTC+8) | - |
| dc.date.available | 6-Oct-2015 16:34:04 (UTC+8) | - |
| dc.date.issued (上傳時間) | 6-Oct-2015 16:34:04 (UTC+8) | - |
| dc.identifier.uri (URI) | http://nccur.lib.nccu.edu.tw/handle/140.119/78880 | - |
| dc.description.abstract (摘要) | 當干擾項為移動平均且參數接近負1時, 許多單根檢定量將產生嚴重型一誤差偏誤。 本文探討 Phillips-Perron 檢定量與其類似型式檢定量在此參數設定下的性質。 比較工具變數估計式與簡單迴歸估計式所形成之 Phillips-Perron 型式檢定量, 我們發現兩者的漸近分配分別以 $O(\\sqrt{T})$ 與 $O(T)$ 的速度發散; 前者型一誤差表現相對優於後者, 而且增加樣本數並無法改善兩者型一誤差扭曲。 文中接著提出工具變數估計之 「修正」 型式檢定量, 並導出其漸近分配。 由漸近分配可推論: 大樣本下, 且干擾項為移動平均, 「修正」型式檢定量具有穩健的型一誤差表現。 文中的模擬結果顯示: 即使移動平均干擾項參數接近負1, 且在一般實證常用的樣本數下, 利用 「修正」 型式檢定量進行靴帶反覆抽樣法, 其型一誤差相當接近事先給定的顯著水準。 | |
| dc.description.abstract (摘要) | The article investigates the asymptotic properties of Phillips-Perron (PP) unit root tests and some of their modified variants in the presence of moving average errors. We show that, in the local-to-unity context, the PP tests based on the OLS or IV estimators cannot converge without re-normalization when the moving average coefficient is close to unity. This implies a size distortion that cannot be resolved by an increase in sample size. In particular, tests with the OLS estimator diverge at a faster rate than those with the IV estimator. However, the modified PP test based on the IV estimator is not subject to the size problem because its limit distribution converges. Our simulations reveal that the empirical size of the bootstrap counterpart is close to the nominal level in small samples. | |
| dc.format.extent | 120 bytes | - |
| dc.format.mimetype | text/html | - |
| dc.relation (關聯) | 經濟論文叢刊, 28(3), 377-400 | |
| dc.subject (關鍵詞) | 單根檢定量;型一誤差扭曲;靴帶反覆抽樣法 | |
| dc.title (題名) | 移動平均干擾項下的單根檢定量 | zh_TW |
| dc.title.alternative (其他題名) | Improved Inference for Unit Root in the Presence of Moving Average Errors | |
| dc.type (資料類型) | article | en |
