Senior Vice President, London
This article focuses on the assessment of statistical evidence in the first cartel damages claim to reach final judgment in the English courts. To assess the quality of the court’s scrutiny, we consider key concepts that affect the probative value of statistical evidence: statistical significance, positive predictive value and statistical power. Using this framework, we conclude that the judge’s approach amounted to a generally sound examination, although there could have been more focus on the concept of statistical power. In support of our views, we refer to recent academic literature which raises serious concerns about the credibility of social science research due to a lack of attention to statistical power. We also note that such concerns likely cause a gross inflation of expected effects reported by the meta-study used in the European Commission’s antitrust damages guidelines. We conclude that there is a need for more overt scrutiny of statistical power of models used in cartel damages cases, in particular for the upfront assessments of methodologies at the certification stage in class action procedures. Ultimately, however, in our view the BritNed judgment demonstrates the English court’s keen ability to deal with complex statistical evidence, and although rejected on this occasion for case-specific reasons, we expect such evidence to continue to feature heavily in cartel damages cases before the English courts.