HAC Corrections for Strongly Autocorrelated Time Series

Applied work routinely relies on heteroscedasticity and autocorrelation consistent (HAC) standard errors when conducting inference in a time series setting. As is well known, however, these corrections perform poorly in small samples under pronounced autocorrelations. In this article, I first provid...

Full description

Saved in:
Bibliographic Details
Published inJournal of business & economic statistics Vol. 32; no. 3; pp. 311 - 322
Main Author Müller, Ulrich K.
Format Journal Article
LanguageEnglish
Published Alexandria Taylor & Francis 03.07.2014
American Statistical Association
Taylor & Francis Ltd
Subjects
Approximation
Autocorrelation
Consistent estimators
Correlation analysis
Critical values
Density estimation
Estimators
Failure analysis
Inference
Local-to-unity
Long-run variance
Normal distribution
Spectral energy distribution
Statistical inference
Studies
Time series
Weighted averages
Online AccessGet full text

Cover

More Information
Summary:Applied work routinely relies on heteroscedasticity and autocorrelation consistent (HAC) standard errors when conducting inference in a time series setting. As is well known, however, these corrections perform poorly in small samples under pronounced autocorrelations. In this article, I first provide a review of popular methods to clarify the reasons for this failure. I then derive inference that remains valid under a specific form of strong dependence. In particular, I assume that the long-run properties can be approximated by a stationary Gaussian AR(1) model, with coefficient arbitrarily close to one. In this setting, I derive tests that come close to maximizing a weighted average power criterion. Small sample simulations show these tests to perform well, also in a regression context.
ISSN:0735-0015
1537-2707
DOI:10.1080/07350015.2014.931238