When economic expert reports refer to statistically significant or insignificant results, they make a qualitative statement about the degree of statistical uncertainty in the estimates. This statistical uncertainty influences the evidential weight of the estimates and should be taken into account in the legal assessment of the facts.
In a new article published in WuW – Wirtschaft und Wettbewerb (Competition Law & Economics), Johannes Dittrich, Sebastian Panthöfer and Lars Wiethaus delve into the concept of statistical significance and explain essential concepts such as standard errors, hypothesis testing, p-values, and confidence intervals in simple terms.
When applied correctly, these concepts help to determine how reliable an estimate of an overcharge or any other parameter relevant to competition law is. However, the article also highlights the limitations of statistical significance, noting that it cannot distinguish between correlation and causation. Therefore, the evaluation of antitrust infringements requires an overall assessment that combines findings from regression analyses – such as effect sizes and statistical significance – with insights from a qualitative analysis of the facts.
You can read the full article in German here. Please contact the authors if you would like an English translation.
