題 目:Robust detection of spurious regressions involving processes moderately deviated from a unit root
主講嘉賓:塗雲東研究員 北京大學
講座時間:2017年11月8日(周三) 15:00--17:00
講座地點:學術會堂606
嘉賓簡介:
塗雲東,北京大學光華管理學院商務統計與經濟計量系和北京大學統計科學中心聯席助理教授,研究員。2004年獲武漢大學數學與統計學院信息與計算科學專業學士學位,2006年獲武漢大學經濟與管理學院數量經濟學專業碩士學位,2012年獲美國加州大學河濱分校經濟學博士學位,同年6月加入北大光華。曾獲世界計量經濟學會(Econometric Society),加州計量經濟學會議等學術組織提供的青年學者研究資助以及Phi Beta Kappa International Scholarship Award。學術論文發表在Journal of Econometrics, Econometric Reviews, Journal of Business and Economic Statistics,Statistica Sinica等國際一流專業雜志。同時擔任以下學術期刊匿名評審:Annals of Statistics, Econometric Reviews, Empirical Economics, Journal of Business and Economic Statistics, Journal of Econometrics, Studies in Nonlinear Dynamics and Econometrics, Journal of Quantitative Economics。理論研究領域涵蓋非參數/半參數計量經濟模型,模型選擇和模型平均,網絡數據建模,金融計量,信息計量經濟學,模型設定檢驗等;應用研究包含宏觀經濟預測,價格指數建模,網絡數據分析,股票市場預測,生産率建模等。
内容摘要:
This paper studies spurious regressions involving processes moderately deviated from a unit root, and establishes the limiting distributions for the least squares estimator, the associated $t$-statistic, the coefficient of determination $R^2$ and the Durbin-Watson statistic. We find that these limiting distributions depend on nuisance parameters, which makes spurious effects detection impossible using the conventional $t$-statistic in practice. As a cure, we propose a robust $t$-test based on the balanced regression model, where the lagged regressor and the lagged dependent variable are augmented to the original regression. The induced $t$-statistic via such an augmentation is shown to be asymptotically standard normal and free of nuisance parameters. Furthermore, the limiting properties of other statistics are show to be similar to those in the classical regression theory. Such spurious detective method is very easy to implement in practice. Finally, the finite sample properties of the robust detective method are demonstrated through Monte Carlo studies and empirical examples.