A one covariate at a time, multiple testing approach to variable selection in high-dimensional linear regression models: A replication in a narrow sense

Héctor M. Núñez, Jesús Otero

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Chudik, Kapetanios, & Pesaran (Econometrica 2018, 86, 1479-1512) propose a one covariate at a time, multiple testing (OCMT) approach to variable selection in high-dimensional linear regression models as an alternative approach to penalised regression. We offer a narrow replication of their key OCMT results based on the Stata software instead of the original MATLAB routines. Using the new user-written Stata commands baing and ocmt, we find results that match closely those reported by these authors in their Monte Carlo simulations. In addition, we replicate exactly their findings in the empirical illustration, which relate to top five variables with highest inclusion frequencies based on the OCMT selection method.

Original languageEnglish (US)
Pages (from-to)833-841
Number of pages9
JournalJournal of Applied Econometrics
Volume36
Issue number6
DOIs
StatePublished - Sep 1 2021

All Science Journal Classification (ASJC) codes

  • Social Sciences (miscellaneous)
  • Economics and Econometrics

Cite this