A CEPR workshop on ‘Recent Developments in the Design and Implementation of Competition Policy’ took place in Florence at the European University Institute on 29/30 November 1996. The workshop was supported by the Robert Schuman Centre of the European University Institute and by a grant from the European Commission under its Human Capital and Mobility Programme. The workshop, which was organized by Louis Phlips (European University Institute) and Damien Neven (Université de Lausanne and CEPR), gave economists and lawyers an opportunity to exchange views on the requirements for efficient implementation of competition policy.

Luc Peeperkorn (European Commission) presented ‘Competition Policy Implications from Game Theory: An Evaluation of the Commission’s Policy on Information Exchange’. He argued that, even though non-cooperative game theory leads to rather strong implications for competition policy, it is still of limited value in answering practical policy questions. In particular, existing models rely on unrealistic assumptions and do not answer questions concerning the influence of significant factors – like the number of firms – on the likelihood of collusion. Moreover, experiments indicate that cooperation is much more likely to occur than theory would predict. Consequently, any implications for competition policy drawn from non-cooperative game theory have to be considered as the minimum action that a competition authority should undertake. Massimo Motta (Universitat Pompeu Fabra and CEPR) agreed with the general point that game theory still has many limitations in answering practical competition policy questions. Some of the criticisms made by Peeperkorn, however, no longer hold thanks to the most recent developments in industrial organisation theory. In particular, collusion can still arise in game theory even when one assumes that the number of firms in the industry is large and when the game is repeated a finite number of times.

Andrzej Baniak (University of Liverpool) presented ‘Antitrust Enforcement with Asymmetric Information and Cournot Strategies’, written with Louis Phlips. Baniak argued that one of the principal problems faced by anti-trust authorities is lack of appropriate data, including firm costs, consumer demand, prices and market shares. The focus of the paper was on situations where costs are known to firms but not to the authority, which has only subjective beliefs about them. Hence, the authority is not able to distinguish a priori between a competitive price that is high on account of high marginal costs and a collusive price. Moreover, firms’ costs and the occurrence of collusion may be discovered only by costly legal action. The authors had modelled the optimal anti-trust policy under such circumstances, based on previous work by Besanko and Spulber, but with the assumption of a somewhat different objective on the part of the competition authorities. Their interpretation of Articles 85 and 86 of the Treaty of Rome was that the objective was not to get price down to marginal costs, but to create a unique and competitive market among the member states. Therefore, they had defined ‘normal’ competitive equilibria as Cournot-Nash equilibria, and had allowed for the corresponding positive profits in the authority’s objective function.

The main result was that, while the authority induces firms to stay in the Cournot-Nash equilibrium under full information, under asymmetric information, collusion by a low-cost industry is always allowed, no matter how high the fines and how low the enforcement costs. Paul Grout (University of Bristol) noted that the costs faced by an authority should be distinguished in two categories: the investigation costs (to determine the real value of the firms’ production costs) and the prosecution costs (the trial).

Penelope Papandropoulos (Université Libre de Bruxelles) presented ‘The EU Procedure Towards Agreements’, written in collaboration with Damien Neven and Paul Seabright (Churchill College, Cambridge, and CEPR). The paper analysed the European Commission’s decision-making procedures for implementation of competition rules, using information from a survey of firms involved in competition cases since 1989. The authors claimed that the Commission’s decisions and its decision-making processes were influenced not only by the characteristics of the cases under investigation, but also by the behaviour of the firms involved. Two types of distortions were possible: first, the Commission may clear an agreement that is anti-competitive. Second, it may condemn an agreement that has overall beneficial effects on competition.

A simple econometric model of lobbying activities was used to analyse the determinants of lobbying and its influences on the Commission’s decisions. There were four main findings. First, the more difficult a case (i.e. the greater the suspicion of infringement), the higher the probabilities that firms will lobby at a high level and bring external support for their case. Second, transport firms appear most likely to lobby at a high level. Third, lobbying does not appear to produce the desired effect on the decision-making process. This is probably because firms are more likely to lobby when they are rightly suspected of infringement. Fourth, by contrast, external support appears more likely to induce favourable decisions in difficult cases.

In ‘Designing Competition Policy Towards Essential Facilities’, Kai-Uwe Kuhn (Institut d’Anàlisi Econòmica, CSIC, Barcelona, and CEPR) analysed some of the latest ‘refusal to supply’ cases, involving access to an infrastructural input, such as a port or an airline facility. The paper was written with Cristina Caffarra (Lexecon Ltd, London) and Daniel Maldoom (University College, Oxford). In all the cases, some infrastructural element was deemed operationally essential by the European Commission, in that access could be seen as a complementary input necessary for serving a downstream market. The concern of competition policy is that monopoly power in the ownership of assets producing such an input may be used to leverage market power in a related market from which potential competitors are excluded.

The authors suggested three fundamental elements for a precise definition of an essential facility. First, the input is essential for production if, without access, entry at the minimum efficient scale would not be possible. Second, users of the input must be at least potential competitors in some output market. Third, the input must exhibit some public good characteristic, like excess capacity. This definition allows, in principle, a test of whether an input should be treated as essential. Failure to satisfy these criteria should give a dominant provider of an input protection from competition policy intervention. The definition provides the basis for development of a procedure for enforcement of competition policy towards essential facilities. Eric van Damme (CentER, Tilburg University) pointed out that identifying the three aspects could be extremely difficult in practice. Moreover, a formal model – which the authors propose to develop in the future – is necessary to give a scientific basis to the argument.

Paul Seabright (Churchill College, Cambridge, and CEPR) presented ‘European Policy Towards Vertical Restraints: A Proposal and an Assessment’. Seabright argued that, in the early days of competition policy, vertical restraints were considered anti-competitive as such. Subsequently, it had been demonstrated by economic theorists that these constraints could, in some cases, have beneficial effects on productive efficiency. Given that assessment of the precise impact of a given vertical restraint on competition will often be a delicate matter, the benefits of discretion in implementing competition policy in this field are important. But discretion also has its costs, in that it makes it harder for firms to predict decisions by the authorities, and increases the risk that firms may be breaking the law when they have been trying in good faith to abide by it. Seabright argued that practical policy towards vertical restraints therefore should seek to balance theoretical rigour with simplicity and predictability. In other words, it should attach significant weight to diminishing the extent of discretion exercised by the competition authorities. This could be done by making use of a sequence of rules of thumb. In particular, Seabright argued that vertical restraints should be presumed not to be anti-competitive in three cases. These are: (1) if the parties to the contract do not operate in the same product market; (2) if inter-brand competition is not so weak that, if the parties to the contract were vertically integrated, they would together possess substantial market power in the combined market for their final product; and (3) if consumers, existing competitors or potential entrants are not likely to be damaged by the contract.

Hans-Theo Normann (Humboldt Universität) noted that German anti-trust law actually prohibits only resale price maintenance and other restraints on resale conditions. Abuses of vertical links and tying agreements in general are controlled simply by a rule of thumb. Consequently, the legal environment in Germany is very close to that suggested by Seabright, even though the rule of thumb is not based on economic reasoning.

John Fingleton (Trinity College, Dublin, and CEPR) presented ‘Competition Policy in Ireland’. Fingleton argued that, until 1991, Irish competition policy was particularly weak and ineffective, being a mixture of legislation concerning restrictive practices and a law controlling monopolies and mergers. The intention of the system was to control anti-competitive behaviour by operating on the principle of fairness, rather than of economic efficiency. Thus, it often resulted in price controls that protected competitors. Although EC law had direct effects in Ireland, enforcement again was weak for a variety of reasons. The 1991 Competition Act marked a complete overhaul of competition policy. It introduced Articles 85 and 86 of the Treaty of Rome into domestic Irish legislation, and established the Competition Authority as the administrative enforcer of the new law. A 1996 amendment had given the Authority new powers and responsibilities and had introduced criminal sanctions for breaches of competition law.

From an analysis of outcomes to date, Fingleton concluded that substantial progress had been made in Irish enforcement standards. None the less, enforcement remained weak, even non-existent, in markets where competition is weakest, and policy decisions were not immune from political influence. William Bishop (Lexecon, London) argued that in Ireland, as in the EU more generally, more staff – particularly economists – are needed to improve enforcement of competition law.

The evolution of competition policy in Ireland was also the subject of the paper ‘Enforcement of Competition Policy’, presented by John Meade (Arthur Cox Solicitors). Dealing with some of the practical enforcement issues, Meade argued that competition law is difficult to apply. Competition issues tend to be complex and sensitive, and practical difficulties of time and costs arise in obtaining evidence. Drawing on Irish experience under both EC and domestic competition law, he suggested that, for companies wishing to rely upon competition law, an adequately empowered regulatory authority is to be preferred to a system based upon private enforcement through litigation in the courts. Mark Williams (Exeter College and RPI, Oxford) pointed to four major factors that determine the effectiveness of competition policy: the reliability of decisions, which depends on the ‘quality’ of decision-making; the speed of decisions; the existence of fines; and the availability of remedies. Enforcement in the United Kingdom is based not on fines, but only on remedies; in the EU, exactly the opposite applies.

Massimo Motta and Michele Polo (Università Bocconi, Milano) presented ‘Concentration and Public Policy in the Broadcasting Industry’. They argued that scarcity of delivery frequencies on the radio spectrum has always led to high concentration in the broadcasting industry. This absolute barrier to entry justified public regulation of the industry together with – in most countries – direct intervention through public television provision. Recently, however, cable and satellite modes have offered new technological opportunities for overcoming the spectrum constraint, leading to more heterogeneous structures with private groups competing on equal terms with public television.

Using an endogenous sunk-cost model to predict medium-term market-concentration tendencies, Motta and Polo found that competition depends on the attractiveness of programme schedules, as well as on broadcast varieties. More popular programmes permit increased advertising and direct subscription revenues, but they also imply higher fixed costs; the market is therefore dominated by the few high-revenue television companies able to finance costly programme schedules. While further entry might eventually be possible in small market niches through single-scheme TV, the top segment would remain concentrated. Although persistent concentration calls for some intervention, close regulation is no longer suitable in today’s dynamic and diversified environment. Instead, competition policy should become the leading form of intervention. Giuliano Amato (Autorità Garante della Concorrenza e del Mercato, Roma) agreed with the authors’ economic reasoning, but pointed to two special characteristics of broadcasting that had to be protected, namely freedom of expression and pluralism. These requirements, together with the technological barriers to entry, explain why this industry has always needed regulation.