The design of competition policy and the regulation of public utilities have recently attracted considerable attention in both Europe and the United States. These topics formed the theme of a CEPR joint workshop with the European Centre for Advanced Research in Economics (ECARE), Université Libre de Bruxelles, on `The Political Economy of Competition Policy', on 26/27 November. The workshop was organized by Mathias Dewatripont, Professor of Economics at the Université Libre de Bruxelles and Co-Director of ECARE, Damien Neven, Professor of Economics at the Université de Lausanne, and Paul Seabright, Fellow and Director of Studies in Economics at Churchill College, Cambridge, all Research Fellows in CEPR's Applied Microeconomics programme. It formed part of the Centre's research programme on `Market Structure, Industrial Organization and Competition Policy', supported by a SPES grant from the Commission of the European Communities.

Mark Armstrong (Gonville and Caius College, Cambridge, and CEPR) opened the workshop with `Liberalisation and Regulatory Capture', which examined the possible effects of the risk of regulatory capture on public policy towards liberalization in an industry with economies of scale. The government is imperfectly informed about industry conditions and appoints a regulator with specialist knowledge of the industry. The regulator must decide how many firms should be allowed to operate in the industry and risks capture by industry and consumer pressure groups with competing interests. He found that an increased likelihood of capture reduces both the maximum attainable welfare level and the degree of discretion the government will grant the regulator over industrial policy. Finally, he argued that his framework of analysis is readily applicable to sentencing procedures.

David Salant (GTE Laboratories Inc) noted that this model accords well with the recent shift towards price capping (rather than rate-of-return regulation), which has changed the level of discretion enjoyed by regulators. Mathias Dewatripont pointed out that the issue was not just how many but also which firms should be allowed to enter the market. He also suggested that the government could organize auctions to reduce informational asymmetry.

In his paper, `Dualling Regulators: Overlapping Jurisdictions and the Regulation of Investment by Public Utilities', David Salant modelled regulation as a dynamic game between a firm, which operates in many jurisdictions with a unique technology (or network), and their regulators, who are each concerned only with the welfare of their own constituents. He compared three situations: when the firm's and regulators' finite horizons coincide; when their terms overlap; and when multiple regulators have been merged into one. He found that independent regulation could dominate the outcome with a unified regulator.

Mathias Dewatripont recalled that multiple equilibria exist in repeated games models and pointed out that the outcome depends not on the number of regulators but rather on the existence of staggered terms which provide incentives to attain the `good' equilibrium. He noted that the development of `corporate cultures' within firms may cause them to behave like infinitely-lived individuals which, in the context of Salant's paper, increases the likelihood of reaching the cooperative outcome.

Paul Seabright provided an alternative to the Tiebout model of government in `Accountability and Decentralisation in Government: An Incomplete Contracts Model'. While the Tiebout framework of local public goods militates in favour of policy differentiation by region but fails to provide any insights on the allocation of power, Seabright's model compared different allocations of powers as alternative means of motivating politicians to act in citizens' interests. While centralization brings benefits in the form of policy coordination, it also entails costs as public accountability is diminished. He derived conditions under which a given region will be better off under central or regional government.

Erik Berglöf (ECARE and CEPR) warned against economists' tendency to oversimplify the political process. He also questioned whether local government is really less prone to regulatory capture. Several participants suggested introducing equity considerations into Seabright's analysis, while Kotaro Suzumura (Harvard University) emphasized the intrinsic value of decentralization.

In `Common Agency with Horizontally Differentiated Principals: Regulating a Multinational Firm to Extract Rent from Abroad', Claudio Mezzetti (University of North Carolina) examined non-linear taxation schemes. A multinational operates in two countries and produces differentiated products for export on the international market. Each country (principal) wants to maximize its tax revenue and offers a non-linear tax scheme to the multinational. The countries seek to extract rents from foreign consumers by influencing the firm's production decision, but they can only offer incomplete contracts since they observe only the levels of its domestic output. Mezzetti showed that regulatory cooperation between two countries makes them better off in terms of tax revenues, while a multinational firm will always prefer a decentralized regime.

Maria Maher (Department of Applied Economics, Cambridge) maintained that this model was very complex and relied too heavily on restrictive assumptions, such as the inversely proportional variation of cost conditions in the two countries. Kai-Uwe Kühn (Universitat Autònoma de Barcelona and CEPR) pointed out that the results depend on the particular shape of the multinational's profit function.

In `Multiprincipals Charter as a Safeguard Against Opportunism in Organizations', David Martimort (Université des Sciences Sociales, Toulouse) showed that this organizational form could limit the extent of regulatory opportunism, since the ex ante allocation of control of the agent's performance among multiple principals reduces the level of rents committed in a renegotiation-proof equilibrium. This gain must be set against the inefficiency stemming from principals' non-cooperative behaviour in the ex post game. Martimort discussed the trade-off between rent extraction and efficiency, whose outcome depends on the agent's production function: when its activities are complementary, a multiprincipal arrangement is at least weakly dominating. Finally, he examined the speed of information revelation within the organization to show that a multiprincipal arrangement may increase it, depending on the discount rate.

Khalid Sekkat (ECARE) noted that this framework could be applied to the functioning of government, where the executive is often controlled by two chambers. Japanese keiretsus are also characterized by multiple principals, while competition may serve as an alternative to introducing a second principal for a regulated industry. Paul Seabright suggested that the model could be applied to the GATT negotiations, since the US president could face multiple principals (Congress) if agreement were not reached before the fast-track procedure expired.

In his joint paper with Daniel Spulber, `The Capital Structure of a Regulated Firm', Yossef Spiegel (Bellcore) examined the strategic role of debt for a public utility. In a model of regulation as a sequential game which determines the equilibrium level of investment, the financial structure of the firm, and the price set by the regulator, the latter seeks to reduce the likelihood of bankruptcy by setting a higher price when the firm has issued debt. Debt therefore raises the regulated price and reduces the scope for regulatory opportunism. Underinvestment nevertheless persists in equilibrium because the regulator does not commit to the efficient level of prices.

Sudipto Bhattacharya (Université Catholique de Louvain) showed that the model could be simplified without loss of generality. David Martimort questioned the empirical basis of Spiegel's results, arguing that the high level of debt issued by public utilities results from the nature of the investments they have to undertake.

Kai-Uwe Kühn then presented his joint paper with Xavier Vives, `Excess Entry, Vertical Integration, and Welfare', which provided a systematic analysis of vertical mergers by a monopoly input supplier into a monopolistically competitive downstream industry. While the welfare effects of vertical mergers are difficult to assess in the existing theoretical literature, which entails the comparison of two totally distinct allocations, Kühn instead analysed small moves in the direction of the vertically integrated outcome. He reported that using the SpenceDixitStiglitz `love-of-variety' specification to examine the trade-off between higher output levels and fewer available varieties indicated that vertical integration leads to welfare improvements for the most plausible class of preferences (those in which love-of-variety is increasing).

Patrick Rey (Ecole Nationale de la Statistique et de l'Administration Economique, Paris, and CEPR) pointed out that the welfare assessment of vertical mergers relied entirely on the love-of-variety parameter, which is difficult to interpret in this context. He added that these theoretical insights should be translated into guidelines for competition policy design and implementation.

The workshop ended with a presentation by Barry Nalebuff (Yale University) on `Designing an Auction for PCS Licences' For the first time, licences for PCSs (personal communication systems) in the US are to be awarded in the near future by auction rather than lottery. Nalebuff described the complex system of allocating different radio frequencies for PCSs at the local, state and national levels. The criteria used to evaluate the auction are, among others, speed of execution and maximization of public revenue. In his proposal, state airwaves' space would be sold sequentially, starting with peripheral ones on the first day (Hawaii and Alaska), immediately followed by larger agglomerations (e.g. New York and Los Angeles).