-
首页
-
- 首页
- WORKSHOP预告
计量、金融与大数据分析workshop：Nickell Meets Stambaugh: A Tale of Two Biases
发布日期：2024-03-29 12:00 来源：
主讲人：史震涛（香港中文大学）
主持老师：（北大经济学院）王熙
参与老师： （北大经院）王一鸣、王法、刘蕴霆
（北大国发院）黄卓、张俊妮、孙振庭
（北大新结构）胡博
时间：2024年3月29日（周五） 10:00-11:30
地点（线下）： 北京大学经济学院606会议室
报告摘要：
In panel predictive regression, the Nickell bias is mixed with the Stambaugh bias when regressors are persistent, imposing challenges against valid inference. This paper proposes a new estimator, which we call IV-X-differencing (IVXD) to restore the standard inferential procedure based on t-statistic. This estimator is constructed based on a panel data extension of the IVX technique in time series. The baseline panel IVX is asymptotically biased when the cross-sectional dimension n and the time dimension T are both large. To make bias correction feasible, we plug in the X-differencing estimator into an analytic formula. The bias-corrected estimator delivers a unified inferential procedure for the predictive regression that covers the stationary, mildly integrated, and the local-to-unity cases. In contrast, unified inference collapses if either of the two components is estimated by the familiar within-group estimator.
主讲人简介：
史震涛，香港中文大学经济系副教授。他的研究聚焦于机器学习算法在计量经济学场景中的渐进理论，涵盖截面数据、时间序列和面板等多种数据类型。他的研究成果已经发表在包括Econometrica, Review of Economics and Statistics, International Economic Review, Journal of Econometrics 等国际一流期刊上。
国家发展研究院官方微信
Copyright© 1994-2012 北京大学 国家发展研究院 版权所有, 京ICP备05065075号-1
保留所有权利，不经允许请勿挪用
-
首页
-
计量、金融与大数据分析workshop：Nickell Meets Stambaugh: A Tale of Two Biases
发布日期：2024-03-29 12:00 来源：
主讲人：史震涛（香港中文大学）
主持老师：（北大经济学院）王熙
参与老师： （北大经院）王一鸣、王法、刘蕴霆
（北大国发院）黄卓、张俊妮、孙振庭
（北大新结构）胡博
时间：2024年3月29日（周五） 10:00-11:30
地点（线下）： 北京大学经济学院606会议室
报告摘要：
In panel predictive regression, the Nickell bias is mixed with the Stambaugh bias when regressors are persistent, imposing challenges against valid inference. This paper proposes a new estimator, which we call IV-X-differencing (IVXD) to restore the standard inferential procedure based on t-statistic. This estimator is constructed based on a panel data extension of the IVX technique in time series. The baseline panel IVX is asymptotically biased when the cross-sectional dimension n and the time dimension T are both large. To make bias correction feasible, we plug in the X-differencing estimator into an analytic formula. The bias-corrected estimator delivers a unified inferential procedure for the predictive regression that covers the stationary, mildly integrated, and the local-to-unity cases. In contrast, unified inference collapses if either of the two components is estimated by the familiar within-group estimator.
主讲人简介：
史震涛，香港中文大学经济系副教授。他的研究聚焦于机器学习算法在计量经济学场景中的渐进理论，涵盖截面数据、时间序列和面板等多种数据类型。他的研究成果已经发表在包括Econometrica, Review of Economics and Statistics, International Economic Review, Journal of Econometrics 等国际一流期刊上。
访问学者申请
EMAIL
北京大学
ENGLISH
