Number and Coverage as Structural Margins for Identifying Cartel Markets in EU Manufacturing
Year: 2026
Type: preprint
Abstract: Competition authorities must choose ex ante where to look for cartels. Market concentration, the usual guide, combines two features acting through different mechanisms: the number of strategically significant firms needed to coordinate, and the share of the market those firms cover. This paper separates the two and asks whether they support a transparent enforcement screen and where that screen stops being informative. The outcome is an established infringement decision rather than the opening of proceedings: of 62 European Commission manufacturing cases, 45 ended in a decision and 23 are analyzed, organized as 630 country-by-industry units across 27 countries and 24 industries. Most cartels span the whole European Economic Area and appear in every member state, so a country holdout leaks the same cartel into training, and the valid design holds out entire industries. Under it the score attains 0.68 ROC and 0.53 precision-recall against a base rate of 0.35; reviewing the top fifth of ranked markets captures a third of infringements at 1.66 times the random rate. It ranks industries but not markets within them, high coverage is not decisive, and no threshold offers a safe harbor. The result is a one-way screen that sharpens targeting within clear limits.
Source: Research Square
Author: Shidong Li
Institution: North University of China
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7225873447
