Bootstrap Tests: How Many Bootstraps?

Russell Davidson and James MacKinnon ()

No 951, Working Papers from Queen's University, Department of Economics

Abstract: This paper discusses how to choose the number of bootstrap samples when performing bootstrap tests. There are two important issues that arise when the number of bootstraps is finite. One is bias in the estimation of bootstrap $P$ values or critical values, and the second is loss of power. We discuss an easy way to avoid bias and thus obtain exact tests if the underlying test statistic is pivotal. We also propose a simple pretest procedure for choosing the number of bootstrap samples so as to avoid power loss, and we illustrate its performance using sampling experiments.

Keywords: Bootstrap testing; Bootstrap samples (search for similar items in EconPapers)
JEL-codes: C12 C15 (search for similar items in EconPapers)
Date: Written
View citations in EconPapers

Published in Econometric Reviews, 19, 2000

Downloads: (external link)
http://www.econ.queensu.ca/working_papers/papers/qed_wp_951.pdf First version 1997 (application/pdf)

Related works:
Working Paper: Bootstrap Tests: How Many Bootstraps? (2001)
Journal Article: Bootstrap tests: how many bootstraps? (2000)
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: http://EconPapers.repec.org/RePEc:qed:wpaper:951

Access Statistics for this paper

More papers in Working Papers from Queen's University, Department of Economics
Contact information at EDIRC.
Series data maintained by Mark Babcock ().