ScholarLens Back to search
1996•Applications of MathematicsOpen access

A comparison of cointegration tests

Petr Mariel

Open full text 5 citations

Abstract

In this paper some of the cointegration tests applied to a single equation are compared. Many of the existent cointegration tests are simply extensions of the unit root tests applied to the residuals of the cointegrating regression and the habitual $H_{0}$ is no cointegration. However, some non residual-based tests and some tests of the opposite null hypothesis have recently appeared in literature. Monte Carlo simulations have been used for the power comparison of the nine selected tests ($ADF$, $\hat{Z}_{\alpha }$, $\hat{Z}_{t}$, $DHS$, $J1$, $H1$, $H2$, $C$, $LBI$) using several types of data generating processes.

Open-access reader

About this research paper

What this paper is about

In this paper some of the cointegration tests applied to a single equation are compared. Many of the existent cointegration tests are simply extensions of the unit root tests applied to the residuals of the cointegrating regression and the habitual $H_{0}$ is no cointegration. However, some non residual-based tests and some tests of the opposite null hypothesis have recently appeared in literature. Monte Carlo simulations have been used for the power comparison of the nine selected tests ($ADF$, $\hat{Z}_{\alpha }$, $\hat{Z}_{t}$, $DHS$, $J1$, $H1$, $H2$, $C$, $LBI$) using several types of data generating processes.

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

In this paper some of the cointegration tests applied to a single equation are compared. Many of the existent cointegration tests are simply extensions of the unit root tests applied to the residuals of the cointegrating regression and the habitual $H_{0}$ is no cointegration. However, some non residual-based tests and some tests of the opposite null hypothesis have recently appeared in literature. Monte Carlo simulations have been used for the power comparison of the nine selected tests ($ADF$, $\hat{Z}_{\alpha }$, $\hat{Z}_{t}$, $DHS$, $J1$, $H1$, $H2$, $C$, $LBI$) using several types of data generating processes.

Key concepts: Cointegration, Mathematics, Unit root, Monte Carlo method, Null hypothesis, Econometrics, Residual, Null (SQL)

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
A comparison of cointegration tests — Research Paper | ScholarLens