Estimation bias and bias correction in reduced rank autoregressions

Heino Bohn Nielsen

doi:10.1080/07474938.2017.1308065

Estimation bias and bias correction in reduced rank autoregressions

Heino Bohn Nielsen^*

^*Corresponding author for this work

1 Citation (Scopus)

Abstract

This paper characterizes the finite-sample bias of the maximum likelihood estimator (MLE) in a reduced rank vector autoregression and suggests two simulation-based bias corrections. One is a simple bootstrap implementation that approximates the bias at the MLE. The other is an iterative root-finding algorithm implemented using stochastic approximation methods. Both algorithms are shown to be improvements over the MLE, measured in terms of mean square error and mean absolute deviation. An illustration to US macroeconomic time series is given.

Original language	English
Journal	Econometric Reviews
Volume	38
Issue number	3
Pages (from-to)	332-349
Number of pages	18
ISSN	0747-4938
DOIs	https://doi.org/10.1080/07474938.2017.1308065
Publication status	Published - 16 Mar 2019

Keywords

Bias correction
bootstrap
cointegration
estimation bias
stochastic approximation
vector autoregression
C32
C13

Access to Document

10.1080/07474938.2017.1308065

Cite this

@article{56616c5283804dbabffd3e43d3c3d95c,

title = "Estimation bias and bias correction in reduced rank autoregressions",

abstract = "This paper characterizes the finite-sample bias of the maximum likelihood estimator (MLE) in a reduced rank vector autoregression and suggests two simulation-based bias corrections. One is a simple bootstrap implementation that approximates the bias at the MLE. The other is an iterative root-finding algorithm implemented using stochastic approximation methods. Both algorithms are shown to be improvements over the MLE, measured in terms of mean square error and mean absolute deviation. An illustration to US macroeconomic time series is given.",

keywords = "Bias correction, bootstrap, cointegration, estimation bias, stochastic approximation, vector autoregression, C32, C13",

author = "Nielsen, {Heino Bohn}",

year = "2019",

month = mar,

day = "16",

doi = "10.1080/07474938.2017.1308065",

language = "English",

volume = "38",

pages = " 332--349",

journal = "Econometric Reviews",

issn = "0747-4938",

publisher = "Taylor & Francis",

number = "3",

}

TY - JOUR

T1 - Estimation bias and bias correction in reduced rank autoregressions

AU - Nielsen, Heino Bohn

PY - 2019/3/16

Y1 - 2019/3/16

N2 - This paper characterizes the finite-sample bias of the maximum likelihood estimator (MLE) in a reduced rank vector autoregression and suggests two simulation-based bias corrections. One is a simple bootstrap implementation that approximates the bias at the MLE. The other is an iterative root-finding algorithm implemented using stochastic approximation methods. Both algorithms are shown to be improvements over the MLE, measured in terms of mean square error and mean absolute deviation. An illustration to US macroeconomic time series is given.

AB - This paper characterizes the finite-sample bias of the maximum likelihood estimator (MLE) in a reduced rank vector autoregression and suggests two simulation-based bias corrections. One is a simple bootstrap implementation that approximates the bias at the MLE. The other is an iterative root-finding algorithm implemented using stochastic approximation methods. Both algorithms are shown to be improvements over the MLE, measured in terms of mean square error and mean absolute deviation. An illustration to US macroeconomic time series is given.

KW - Bias correction

KW - bootstrap

KW - cointegration

KW - estimation bias

KW - stochastic approximation

KW - vector autoregression

KW - C32

KW - C13

U2 - 10.1080/07474938.2017.1308065

DO - 10.1080/07474938.2017.1308065

M3 - Journal article

AN - SCOPUS:85019263841

SN - 0747-4938

VL - 38

SP - 332

EP - 349

JO - Econometric Reviews

JF - Econometric Reviews

IS - 3

ER -

Estimation bias and bias correction in reduced rank autoregressions

Abstract

Keywords

Access to Document

Fingerprint

Cite this