To Pool or Not to Pool: A Closer Look at the Use of Sub-Regressions in Antitrust Class Certification

Journal of Competition Law & Economics, vol 13, Issue 4

Posted: 13 Mar 2017 Last revised: 20 Apr 2019

See all articles by Ai Deng

Ai Deng

NERA Economic Consulting; Johns Hopkins University

Date Written: July 22, 2017

Abstract

Did the anticompetitive conduct in question impact all or nearly all class members? This is a question central to a court’s class certification decision. To answer this question, a methodology — known as sub-regressions — is being increasingly employed, particularly by defendants’ expert witnesses. A key step of a sub-regression type analysis is to partition the data into various subgroups and then to examine data poolability. In this article, I focus on three areas of interest pertaining to sub-regressions. First, I review the related law and economics literature related to this methodology. I then analyze courts’ recent class certification decisions in cases where parties introduced sub-regression analysis. I then discuss several methodological challenges, many of which have not been previously acknowledged, and present potential ways to address these challenges. I emphasize that a disciplined and rigorous implementation is crucial for the reliability of sub-regressions.

Keywords: Class certification, Common impact, Data mining, Poolability

JEL Classification: K00, K21

Suggested Citation

Deng, Ai, To Pool or Not to Pool: A Closer Look at the Use of Sub-Regressions in Antitrust Class Certification (July 22, 2017). Journal of Competition Law & Economics, vol 13, Issue 4. Available at SSRN: https://ssrn.com/abstract=2931455 or http://dx.doi.org/10.2139/ssrn.2931455

Ai Deng (Contact Author)

NERA Economic Consulting ( email )

1255 23rd Street, NW, Suite 600
Washington, DC 20037
United States

Johns Hopkins University ( email )

1717 Massachusetts Ave NW
Washington, DC DC 20036
United States

Register to save articles to
your library

Register

Paper statistics

Abstract Views
473
PlumX Metrics