Optimal Enforcement and Decision Structures for Competition Policy

The first aim of this project was to extend in a number of important directions a 2009 paper by the investigators that gave a rigorous evaluation of the conditions under which Effects-Based procedures – now very widely adopted by most competition authorities over past 20 years – are superior to Per Se Procedures. The second aim was to engage closely with two competition authorities – the Office of Fair Trading (UK) and the Autorite de la Concurrence (France) – to explain our emerging research findings and identify issues of concern to them on which our research could shed light. Both were undertaking major reviews of their penalty policies and guidelines.

Some key findings:

(i) Appeals procedures reduce deterrence which is good if actions are on average benign but bad if actions are on average harmful. They make effects-based procedures more attractive in former case and less attractive in latter.

(ii) Contrary to received legal scholarship: (a) effect-based procedures need not entail legal uncertainty; (b) if penalties are set appropriately welfare may be higher with some degree of legal uncertainty than with none; (c) effects-based procedures unambiguously welfare dominate per se; (d) appropriate penalties may be higher the greater the degree of legal uncertainty. Contrary to Economic scholarship appropriate penalty may be zero if legal uncertainty is very great.

(iii) Existing theory of optimal penalties ignores possibility that competition authorities sometimes stop on-going actions. This can lead to penalties being 25% lower than existing theory suggests.

(iv) This supports OFT’s proposal to raise penalties to 30% of revenue.

(v) Imposing penalties on revenue creates a distortion that raises prices. This can be greater the tougher the penalty regime.