[数量经济学研讨会]Limit Theory of Restricted Maximum Likelihood Ratio Test in Predictive Regression
发文时间:2013-12-27

数量经济学研讨会

【201308】

报告题目:Limit Theory of Restricted Maximum Likelihood Ratio Test in Predictive Regression

报告人:陈烨

报告时间:2013年12月31日下午12:30-13:30

报告地点:明德主楼729

内容简介:

Chen et al. (2013) provide the limit distribution of quasi restricted maximum likelihood ratio test (QRLRT), based on the weighted least squares approximation of the restricted likelihood (WLSRL). Their result shows that the limit distribution of QRLRT is well approximated by λ?(1) when the predictor follows a stationary, local to unity, mildly unit root or unit root process. In this paper, we first provide the limit distribution of restricted maximum likelihood ratio test (RLRT) based on its Taylor series expansion. This expansion formula also helps to develop the limit distribution of RLRT in more general cases, such as multivariate predictors. Further, we explore the role of intercept term in predictive regression for both estimation and inference. It is worthy noticing that the RLRT by Chen and Deo (2009) fails if the true data generating process of the predictor has an intercept. We provide an explanation for this point based on the localizing drift specification (Phillips et al. 2013).

报告人简介:

陈烨,新加坡管理大学经济学博士候选人。她的主要研究方向为Financial Econometrics, Econometric Theory 和 Forecasting.

中国人民大学经济学院数量经济教研室

中国人民大学运筹学与数量经济研究所

2013年12月

为了加强与国内外高水平数量经济学学者的交流和合作,更好地促进数量经济学与院内其他学科之间的合作交流,数量经济学教研室将举办数量经济学研讨会。研讨会侧重于理论计量经济学、应用计量经济学及数理经济学的最新研究成果,欢迎各位学者参加或报告。如果您有研究成果想要报告,请与数量经济学教研室章永辉博士联系(yonghui.zhang@hotmail.com)。