The distribution of the Yule-Walker estimator for autoregressive time series with a unit root
スポンサーリンク
概要
- 論文の詳細を見る
This paper investigates the exact finite sample and limiting distributions of the normalized Yule-Walker(YW) estimator of α in the model yt = αyt-1 + ut where α=1 and the ut are independently identically distributed N(O, σ?). For the practical purpose of testing a unit root in time series regression, the tabulated values of the critical points for various sample size and levels of significance are obtained by numerical integration. By a numerical power comparison it is shown that the YW estimator can be used as an alternative test for detecting a unit root however none of the normalized OLS estimator, the t-statistic, the normalized Durbin-Watson statistic and the normalized YW estimator is uniformly superior.
- 香川大学経済研究所の論文
著者
関連論文
- E-3 非定常SURモデルの推定と検定の漸近理論(日本統計学会第67回大会記録 : 計量経済学方法論)
- Nonsense Regressions in Econometrics, I(1) with drift vs. Trend Stationary(Dedicated to Professor Hideo Watanabe in Appreciation of His Many Years of Service)
- The Schur Complements and the Le Chatelier-Braun Principle
- On the deficiency of tests for structural change in a non-cointegration regression model
- 非定常SURモデルの推定と検定の漸近理論 : 改訂版
- 非定常SURモデルの推定と検定の漸近理論 (秦隆昌教授記念号)
- 複数の構造変化点の不偏推定について
- The distribution of the Yule-Walker estimator for autoregressive time series with a unit root
- The limiting distribution of the Durbin-Watson statistic for testing a unit root in economic time series
- F test for unit roots in a VAR model