Finite-sample Corrected Inference for Two-step GMM in Time Series

Jungbin Hwang and Gonzalo Valdés
Additional contact information
Gonzalo Valdés: Universidad de Tarapacá

No 2020-02, Working papers from University of Connecticut, Department of Economics

Abstract: This paper develops a nite-sample corrected and robust inference for e¢ cient two-step generalized method of moments (GMM). One of the main challenges in e¢ cient GMM is that we do not observe the moment process and have to use the estimated moment process to construct a GMM weighting matrix. We use a non-parametric long run variance estimator as the optimal GMM weighting matrix. To capture the estimation uncertainty embodied in the weight matrix, we extend the nite-sample corrected formula of Windmeijer (2005) to a heteroskedasticity autocorrelated robust (HAR) inference in time series setting. Using xed-smoothing asymptotics, we show that our new test statistics lead to standard asymptotic F or t critical values and improve the nite sample performance of existing HAR robust GMM tests.

Keywords: Generalized Method of Moments; Heteroskedasticity Autocorrelated Robust; Finite-sample Correction; Fixed-smoothing Asymptotics; t and F tests. (search for similar items in EconPapers)
JEL-codes: C12 C13 C32 (search for similar items in EconPapers)
Pages: 40 pages
Date: 2020-01
New Economics Papers: this item is included in nep-ecm and nep-ets
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://media.economics.uconn.edu/working/2020-02.pdf Full text (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:uct:uconnp:2020-02

Access Statistics for this paper

More papers in Working papers from University of Connecticut, Department of Economics University of Connecticut 365 Fairfield Way, Unit 1063 Storrs, CT 06269-1063. Contact information at EDIRC.
Bibliographic data for series maintained by Mark McConnel ().