Bias-reduced log-periodogram and whittle estimation of the long-memory parameter without variance inflation

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

The bias-reduced log-periodogram estimator d̂ LP(r), r ≥ 1 of Andrews and Guggenberger (2003, Econometrica 71, 675-712) for the long-memory parameter d in a stationary long-memory time series reduces the asymptotic bias of the original log-periodogram estimator d GPH = d̂ LP(0) of Geweke and Porter-Hudak (1983) by an order of magnitude but inflates the asymptotic variance by a multiplicative constant c r, for example, C 1 = 2.25 and c 2 = 3.52. In this paper, we introduce a new, computationally attractive estimator d̂ WLP(r) by taking a weighted average of d̂ LP(0) estimators over different bandwidths. We show that, for each fixed r ≥ 0, the new estimator can be designed to have the same asymptotic bias properties as d̂ LP(r) but its asymptotic variance is changed by a constant c* r that can be chosen to be as small as desired, in particular smaller than c r. The same idea is also applied to the local-polynomial Whittle estimator d̂ LW(r) in Andrews and Sun (2004, Econometrica 72, 569-614) leading to the weighted estimator d̂ WLW(r). We establish the asymptotic bias, variance, and mean-squared error of the weighted estimators and show their asymptotic normality. Furthermore, we introduce a data-dependent adaptive procedure for selecting r and the bandwidth m and show that up to a logarithmic factor, the resulting adaptive weighted estimator achieves the optimal rate of convergence. A Monte Carlo study shows that the adaptive weighted estimator compares very favorably to several other adaptive estimators.

Original languageEnglish (US)
Pages (from-to)863-912
Number of pages50
JournalEconometric Theory
Volume22
Issue number5
DOIs
StatePublished - Oct 1 2006

Fingerprint

inflation
trend
normality
time series
Long memory
Estimator
Inflation

All Science Journal Classification (ASJC) codes

  • Social Sciences (miscellaneous)
  • Economics and Econometrics

Cite this

@article{42844fb93d424a6f8fa2926df75574b1,
title = "Bias-reduced log-periodogram and whittle estimation of the long-memory parameter without variance inflation",
abstract = "The bias-reduced log-periodogram estimator d̂ LP(r), r ≥ 1 of Andrews and Guggenberger (2003, Econometrica 71, 675-712) for the long-memory parameter d in a stationary long-memory time series reduces the asymptotic bias of the original log-periodogram estimator d GPH = d̂ LP(0) of Geweke and Porter-Hudak (1983) by an order of magnitude but inflates the asymptotic variance by a multiplicative constant c r, for example, C 1 = 2.25 and c 2 = 3.52. In this paper, we introduce a new, computationally attractive estimator d̂ WLP(r) by taking a weighted average of d̂ LP(0) estimators over different bandwidths. We show that, for each fixed r ≥ 0, the new estimator can be designed to have the same asymptotic bias properties as d̂ LP(r) but its asymptotic variance is changed by a constant c* r that can be chosen to be as small as desired, in particular smaller than c r. The same idea is also applied to the local-polynomial Whittle estimator d̂ LW(r) in Andrews and Sun (2004, Econometrica 72, 569-614) leading to the weighted estimator d̂ WLW(r). We establish the asymptotic bias, variance, and mean-squared error of the weighted estimators and show their asymptotic normality. Furthermore, we introduce a data-dependent adaptive procedure for selecting r and the bandwidth m and show that up to a logarithmic factor, the resulting adaptive weighted estimator achieves the optimal rate of convergence. A Monte Carlo study shows that the adaptive weighted estimator compares very favorably to several other adaptive estimators.",
author = "Patrik Guggenberger and Yixiao Sun",
year = "2006",
month = "10",
day = "1",
doi = "10.1017/S0266466606060403",
language = "English (US)",
volume = "22",
pages = "863--912",
journal = "Econometric Theory",
issn = "0266-4666",
publisher = "Cambridge University Press",
number = "5",

}

Bias-reduced log-periodogram and whittle estimation of the long-memory parameter without variance inflation. / Guggenberger, Patrik; Sun, Yixiao.

In: Econometric Theory, Vol. 22, No. 5, 01.10.2006, p. 863-912.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Bias-reduced log-periodogram and whittle estimation of the long-memory parameter without variance inflation

AU - Guggenberger, Patrik

AU - Sun, Yixiao

PY - 2006/10/1

Y1 - 2006/10/1

N2 - The bias-reduced log-periodogram estimator d̂ LP(r), r ≥ 1 of Andrews and Guggenberger (2003, Econometrica 71, 675-712) for the long-memory parameter d in a stationary long-memory time series reduces the asymptotic bias of the original log-periodogram estimator d GPH = d̂ LP(0) of Geweke and Porter-Hudak (1983) by an order of magnitude but inflates the asymptotic variance by a multiplicative constant c r, for example, C 1 = 2.25 and c 2 = 3.52. In this paper, we introduce a new, computationally attractive estimator d̂ WLP(r) by taking a weighted average of d̂ LP(0) estimators over different bandwidths. We show that, for each fixed r ≥ 0, the new estimator can be designed to have the same asymptotic bias properties as d̂ LP(r) but its asymptotic variance is changed by a constant c* r that can be chosen to be as small as desired, in particular smaller than c r. The same idea is also applied to the local-polynomial Whittle estimator d̂ LW(r) in Andrews and Sun (2004, Econometrica 72, 569-614) leading to the weighted estimator d̂ WLW(r). We establish the asymptotic bias, variance, and mean-squared error of the weighted estimators and show their asymptotic normality. Furthermore, we introduce a data-dependent adaptive procedure for selecting r and the bandwidth m and show that up to a logarithmic factor, the resulting adaptive weighted estimator achieves the optimal rate of convergence. A Monte Carlo study shows that the adaptive weighted estimator compares very favorably to several other adaptive estimators.

AB - The bias-reduced log-periodogram estimator d̂ LP(r), r ≥ 1 of Andrews and Guggenberger (2003, Econometrica 71, 675-712) for the long-memory parameter d in a stationary long-memory time series reduces the asymptotic bias of the original log-periodogram estimator d GPH = d̂ LP(0) of Geweke and Porter-Hudak (1983) by an order of magnitude but inflates the asymptotic variance by a multiplicative constant c r, for example, C 1 = 2.25 and c 2 = 3.52. In this paper, we introduce a new, computationally attractive estimator d̂ WLP(r) by taking a weighted average of d̂ LP(0) estimators over different bandwidths. We show that, for each fixed r ≥ 0, the new estimator can be designed to have the same asymptotic bias properties as d̂ LP(r) but its asymptotic variance is changed by a constant c* r that can be chosen to be as small as desired, in particular smaller than c r. The same idea is also applied to the local-polynomial Whittle estimator d̂ LW(r) in Andrews and Sun (2004, Econometrica 72, 569-614) leading to the weighted estimator d̂ WLW(r). We establish the asymptotic bias, variance, and mean-squared error of the weighted estimators and show their asymptotic normality. Furthermore, we introduce a data-dependent adaptive procedure for selecting r and the bandwidth m and show that up to a logarithmic factor, the resulting adaptive weighted estimator achieves the optimal rate of convergence. A Monte Carlo study shows that the adaptive weighted estimator compares very favorably to several other adaptive estimators.

UR - http://www.scopus.com/inward/record.url?scp=33749352042&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33749352042&partnerID=8YFLogxK

U2 - 10.1017/S0266466606060403

DO - 10.1017/S0266466606060403

M3 - Article

AN - SCOPUS:33749352042

VL - 22

SP - 863

EP - 912

JO - Econometric Theory

JF - Econometric Theory

SN - 0266-4666

IS - 5

ER -