Bootstraping time series regressions with integrated process

DOI10.1111/1467-9892.00235MaRDI QIDQ2744935zbMATH OpenOpenAlexFDO

Authors Zhijie Xiao, Hong-Yi Li

Publication date 9 October 2001

Published in Journal of Time Series Analysis (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1111/1467-9892.00235

zbMATH Keywords

bootstrapintegrated processesregressionunit root tests

Mathematics Subject Classification ID

Nonparametric statistical resampling methods (62G09) Time series, auto-correlation, regression, etc. in statistics (GARCH) (62M10)

Review text

The authors consider the bivariate time series regression model \(y_t=\beta x_t+u_t\), where \(x_t\) is an I(1) process \(x_t=x_t+v_t\), \(u_t\) and \(v_t\) are AR processes. A bootstrap procedure is constructed for the estimation of the distribution of the convenient least squares estimator \(\hat\beta\) for \(\beta\). It is shown that the bootstrap distribution approximates the asymptotic distribution of \(\hat\beta\). Hypothesis testing based on this technique is considered. Results of simulations are presented.

Recommendations

• On bootstrapping regressions with unit root processes
• Bootstrapping cointegrating regressions
• Bootstrapping time series models
• Bootstrapping unit root tests for integrated processes
• A simple bootstrap test for time series regression models

Cited in

(5)

• On bootstrapping regressions with unit root processes
• Behavior in small samples of some tests of non-nested hypotheses in nonstationary regressions and their bootstrap versions
• Multiple Time Series Regression with Integrated Processes
• New bootstrap inference for spurious regression problems
• Block bootstrap theory for multivariate integrated and cointegrated processes

This page was built for publication: Bootstraping time series regressions with integrated process

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2744935)