On determination of cointegration ranks

Qiaoling Li, Jiazhu Pan, Qiwei Yao

Research output: Contribution to journalArticle

Abstract

We propose a new method to determine the cointegration rank in the error correction model (ECM). The cointegration rank, together with the lag order, is determined by a penalized goodness-of-fit measure. We show that the estimated cointegration vectors are consistent with a convergence rate T, and our estimation for the cointegration rank is consistent. Our approach is more robust than the conventional likelihood based methods, as we do not impose any assumption on the form of the error distribution in the model. Furthermore we allow the serial dependence in the error sequence. The proposed methodology is illustrated with both simulated and real data examples. The advantage of the new method is particularly pronounced in the simulation with non-Gaussian and/or serially dependent errors.
LanguageEnglish
Pages45-56
Number of pages12
JournalStatistics and Its Interface
Volume2
Issue number1
Publication statusPublished - 2009

Fingerprint

Cointegration
Error correction
Serial Dependence
Error Correction Model
Goodness of fit
Convergence Rate
Likelihood
Methodology
Dependent
Simulation
Model

Keywords

  • cointegration
  • error correction models
  • penalized goodness-of-fit criteria

Cite this

Li, Qiaoling ; Pan, Jiazhu ; Yao, Qiwei. / On determination of cointegration ranks. In: Statistics and Its Interface. 2009 ; Vol. 2, No. 1. pp. 45-56.
@article{c18235429fd14224972bd7459d80bed0,
title = "On determination of cointegration ranks",
abstract = "We propose a new method to determine the cointegration rank in the error correction model (ECM). The cointegration rank, together with the lag order, is determined by a penalized goodness-of-fit measure. We show that the estimated cointegration vectors are consistent with a convergence rate T, and our estimation for the cointegration rank is consistent. Our approach is more robust than the conventional likelihood based methods, as we do not impose any assumption on the form of the error distribution in the model. Furthermore we allow the serial dependence in the error sequence. The proposed methodology is illustrated with both simulated and real data examples. The advantage of the new method is particularly pronounced in the simulation with non-Gaussian and/or serially dependent errors.",
keywords = "cointegration, error correction models, penalized goodness-of-fit criteria",
author = "Qiaoling Li and Jiazhu Pan and Qiwei Yao",
year = "2009",
language = "English",
volume = "2",
pages = "45--56",
journal = "Statistics and Its Interface",
issn = "1938-7989",
number = "1",

}

Li, Q, Pan, J & Yao, Q 2009, 'On determination of cointegration ranks' Statistics and Its Interface, vol. 2, no. 1, pp. 45-56.

On determination of cointegration ranks. / Li, Qiaoling; Pan, Jiazhu; Yao, Qiwei.

In: Statistics and Its Interface, Vol. 2, No. 1, 2009, p. 45-56.

Research output: Contribution to journalArticle

TY - JOUR

T1 - On determination of cointegration ranks

AU - Li, Qiaoling

AU - Pan, Jiazhu

AU - Yao, Qiwei

PY - 2009

Y1 - 2009

N2 - We propose a new method to determine the cointegration rank in the error correction model (ECM). The cointegration rank, together with the lag order, is determined by a penalized goodness-of-fit measure. We show that the estimated cointegration vectors are consistent with a convergence rate T, and our estimation for the cointegration rank is consistent. Our approach is more robust than the conventional likelihood based methods, as we do not impose any assumption on the form of the error distribution in the model. Furthermore we allow the serial dependence in the error sequence. The proposed methodology is illustrated with both simulated and real data examples. The advantage of the new method is particularly pronounced in the simulation with non-Gaussian and/or serially dependent errors.

AB - We propose a new method to determine the cointegration rank in the error correction model (ECM). The cointegration rank, together with the lag order, is determined by a penalized goodness-of-fit measure. We show that the estimated cointegration vectors are consistent with a convergence rate T, and our estimation for the cointegration rank is consistent. Our approach is more robust than the conventional likelihood based methods, as we do not impose any assumption on the form of the error distribution in the model. Furthermore we allow the serial dependence in the error sequence. The proposed methodology is illustrated with both simulated and real data examples. The advantage of the new method is particularly pronounced in the simulation with non-Gaussian and/or serially dependent errors.

KW - cointegration

KW - error correction models

KW - penalized goodness-of-fit criteria

UR - http://eprints.lse.ac.uk/24106/1/On_determination_of_cointegration_ranks_(LSERO).pdf

UR - http://dx.doi.org/10.4310/SII.2009.v2.n1.a5

M3 - Article

VL - 2

SP - 45

EP - 56

JO - Statistics and Its Interface

T2 - Statistics and Its Interface

JF - Statistics and Its Interface

SN - 1938-7989

IS - 1

ER -