As originally envisaged the research continues to be classified into roughly three areas:

Task 1: The relationship between Regulation and Competition
Task 2: Analysis of Investment in New Networks
Task 3: The duality of Networks and Geography and of Networks and Globalization

In terms of the overall structure there are two developments that are becoming increasingly apparent as the network matures. First, the number of research projects that cut across the various themes is increasing. As documented below, not all projects fit nicely into one of the above three tasks and so we have decided not to structure the fourth report by tasks.

A second development (and not unrelated to the first) is that the interactions between the various network nodes across Europe is now very high. As can be seen from the number of projects between researchers located in different countries, the network has now produced a very coherent and interactive group of researchers in the field of network industries.

Progress has been made in these areas and a number of activities have been organised during the reporting period where the results have been presented and discussed (see also below). What follows is a brief overview of the main findings as well as the ongoing research.

WZB

The group at the WZB has been working on issues surrounding regulation and competition policy. In particular, research has addressed how the endogeneity of policy decisions affects the impact of policy on the economy as well. In other words, how the incorporation of political economy concerns alter our understanding of markets. An example of this is the paper by Duso and Röller (WZB). This paper argues that the study of policy incidence in industrial organization needs to take the endogeneity of government into account. The point is made by investigating whether political considerations are important in terms of understanding the causes and effects of deregulation using data provided by the OECD. In particular, the paper addresses two interrelated questions: (i) do political and institutional factors matter in a systematic way in terms of the decision to deregulate, and (ii) if so, what does this imply in terms of the policy incidence of deregulation. The results indicate that political considerations do matter. Most importantly, by introducing political and institutional variables into the empirical analysis of policy incidence, it is found that policy conclusions are substantially different from an analysis that treats political factors exogenously. They concluded that the evidence is suggestive of the claim that a full understanding of the effect of government intervention in the marketplace implies a closer integration of political economy with industrial organization.

A second paper using this approach was produced by Neven (GIIS) and Röller. The paper analysed how the choice of a standard (either consumer surplus or welfare) that an antitrust agency is held accountable to is affected by the political economy environment that it may be operating in. They considered a framework in which the antitrust agency can be influenced by third parties (at a cost in terms of real resources) and in which the agency is imperfectly monitored. A welfare comparison between the two standards revealed that neither standard dominates. The main findings were that the consumer surplus is attractive relative to a welfare standard, when lobbying is efficient, when accountability is low, where mergers are large and when a marginal increase in merger size is highly profitable.

A final paper in this area investigated the potential for conflict between antitrust agencies (an example would be the recent GE/Honeywell case). In a first paper Neven and Röller analysed the scope for conflict between national merger control agencies which simultaneously assert jurisdictions. They considered a positive model of merger control in which market definition and the analysis of dominance are both explicitly specified. They found that conflict in international merger control is less likely to occur when economic integration is high. Hence, ''globalisation'' should alleviate rather than exacerbate conflict. In addition, conflicts are more likely to arise between countries of similar market size and for moderate competition policy rules.

Co-operation between firms, such as RJVs has been analysed by Lars-Hendrik Röller (WZB) and Konrad Stahl (Universität Mannheim), who have finished a paper on positive and normative aspects of research networks (RJV’s), when firms who produce differentiated products pursue cost reducing R&D. This work is a generalisation of Kamien, Mueller and Zang’s 1992 American Economic Review classic. In particular, they are in the process of showing that KMZ’s claim of the universal welfare optimality of RJV’s that has become the basis for important industrial policy considerations no longer holds when firms produce differentiated products. They also showed benefits from co-operative research vary with the degree of product differentiation, and thereby with the productivity versus the competition effects of co-operative research and of R&D spillovers respectively.

Paul Heidhues (WZB) and Nicolas Melissas (IAE) continued investigating a simple dynamic global game. Typically, in static games with positive network externalities there exist multiple Nash equilibria with self-fulfilling beliefs, because if a player believes that other players will be active then it is profitable for him to become active. If he believes, however, that other players refrain from becoming active then the best response is to remain passive. The recently developed theory of global games overcomes this multiplicity problem through relaxing the unrealistic assumption that players know the payoffs of the game they are playing perfectly. Within a static framework, this enables researchers to predict behaviour in situations with positive network externalities. An open question - addressed in this research project – is whether global game techniques can be extended to dynamic environments. First results indicate that if network effects are not period-specific, a unique prediction can be found. If, however, the network externalities are partially period-specific multiple equilibria with self-fulfilling beliefs may re-emerge.

There are also a number of other projects that involve the analysis of merger control. Christian Wey (WZB/CEPR) and Roman Inderst (LSE/CEPR) have continued their research projects on horizontal mergers (for an overview see the 3rd Annual Report that covers the period 01/03/2000 – 28/02/2001). The project that deals with the welfare effects of recent concentration trends in the European retail sector shows that retailer mergers across independent consumer markets may increase welfare. Retailer concentration make marginal cost reductions relatively more profitable for suppliers because they have to bear more of the them. Therefore, even if retail merger reduce suppliers’ overall profits they may increase the suppliers’ incentives to tradeoff lower marginal costs with higher fixed or operating costs. Besides other things, these results have been published in a CEPR discussion paper (no. 2981) entitled "Bargaining, Mergers, and Technology Choice in Bilaterally Oligopolistic Industries”.

Wey and Inderst published their research “The Takeover Incentives in Oligopoly” as a CEPR discussion paper (No. 3163), in which they model a takeover as an auction where the target company commits to an optimal reserve price. In the course of completing this project an important insight concerning the relative efficiency of Bertrand and Cournot competition has been derived. As is well-known, Bertrand competition involves lower prices for a given number of firms than does Cournot competition; and therefore increases welfare. If goods are substitutes the analysis shows that Bertrand competition leads, however, to a higher propensity for takeover when compared with Cournot competition. This should typically produce a countervailing effect on welfare, which may more than outweigh lower Bertrand-prices for a given number of firms.

Rainer Nitsche (WZB and Humboldt University) investigated issues of merger control in network industries with low sunk costs (e.g. local transport, postal services). One important characteristic of these industries is that competition is localised due to the multimarket nature of operations. In his work “Incremental Mergers and Competition Policy in Multimarket Industries” he argues that competition authorities should be more restrictive than at present in controlling incremental mergers and acquisitions in multi-market industries. Mega-mergers are not the only source of concentration that damages consumers' interests. In particular, in multi-market industries, the number of players is reduced gradually via a sequence of acquisitions. Yet merger control is lenient on mergers that either have a small impact on market share or do not significantly affect competition in any given product market. Using a merger race model, this paper argues that large firms have an incentive to buy small firms until there are none left to buy, even where there are no synergies.

ECARES

Patrick Legros (ECARES) and Konrad Stahl (Mannheim) continued with their work on ‘Global versus Local Competition’, in which they analyze the impact of increased outside opportunities brought to consumers by access to a global market on local market performance under monopoly vs. oligopoly. If consumers have to choose once where to shop we show that under all forms of organising the local market, increased competition from the global market will crowd out variety in the local one. The effect of increased global competition on prices is much less clear. While it yields a price reduction under monopoly, prices may increase under oligopoly. They therefore checked the robustness of these results in various extensions and drew conclusions on competition and industrial policies.

Legros, Dewatripont (ECARES/CEPR) and Mathews (Pennsylvania) developed a dynamic moral hazard model in which effort in the first period is observable but not verifiable. Contracts can be renegotiated before output is realised but after effort is chosen by the entrepreneur. The contribution of their work to this field of study is to show that under reasonable conditions on contracts (monotonic payoffs in particular), the optimal initial contracts are debt contracts. This model is well adapted to contracts signed between entrepreneurs and financiers in the new economy or to venture capital.

Legros and Newman (UCL/CEPR) showed that the possibility of interference in court proceedings, or more generally jamming other agents' messages, has significant consequences for the form of optimal contracts and the flexibility of decisions that can be made inside firms. Their approach offers a new view of authority, basing it on the ability of parties to have their say in court. Interference gives authority a role in worlds where it is traditionally absent in contract theory, like simple employment relationships without relation specific investments. They also studied frictionless matching in large economies with and without market imperfections, providing sufficient conditions for monotone matching that are weaker than those previously known. Necessary conditions, which depend on a key analytical object they call the surplus function are also offered. Changes in the surplus yield valuable information about the comparative statics of matching patterns across environments. They applied their framework to some examples adapted from the literature, accounting for and extending several comparative-static and welfare results and also explored the dependence of the matching pattern on the type distribution.

Legros and Newman also looked at the progress in the application of matching models to environments in which the utility between matching partners is not fully transferable and has been hindered by a lack of results analogous to those that are known for transferable utility that help to characterize equilibrium matching patterns. They presented sufficient conditions for matching to be monotone that are simple to express and easy to verify and illustrated their application with some examples that are of independent interest: the formation of financial co-operatives to share risk and the formation of principal-agent relationships.

Legros and Boccard (Universitat de Girona), provided a simple framework for analysing how competition affects the choice of audit structures in an oligopolistic insurance industry. When the degree of competition increases, fraud increases but the response of the industry in terms of investment in audit quality follows a U-shaped pattern. Following increases in competition, the investment in audit quality will decrease if the industry is initially in a low competition regime while it will increase when the industry is in a high competition regime. Moreover, they showed that firms will benefit from forming a joint audit agency only when the degree of competition is intermediate; in this case, cooperation might improve total welfare.

INSEAD

Previous literature has concentrated on examining the impact of a single regulatory constraint on the final product. However, an industry such as electricity requires two network inputs from naturally monopolistic industries i.e. gas transmission and electricity transmission. Matthew Bennett (Young Researcher, INSEAD) modelled the impact of such dual regulation schemes. Firstly, they found that the result of previous literature that regulation causes increasing prices for higher cost consumers may be reversed. Secondly, dual input regulation creates distortions if regulators do not explicitly co-operate. Where regulators place some weight on their respective sector’s consumer surplus, these differing agendas lead to competition in regulatory strictness resulting in a significant sub optimal welfare outcome relative to a joint regulatory body. Lastly, he found that the levels of the two regulatory constraints have a significant impact on new generator’s location decisions and hence the mix of transmission inputs any new generator decides to utilise.

Models of spatially differentiated markets become analytically intractable before becoming rich enough to offer conclusive insights. For example, existing models are unable to describe the trade-off between diversity and quality without making distortionary assumptions. In his paper, Randal Heeb (INSEAD) offered an alternative methodology in which complex equilibria are simulated computationally, and analytic results are derived from closed form approximations from the simulations. The technique is demonstrated for a differentiated product duopoly. A unique interior equilibrium exists, whereas conventional models predict maximum differentiation in one dimension and minimum differentiation in the other. The standard result is adversely affected by the standard assumption of market coverage. Heeb’s new methodology allows the use of more reasonable assumptions, and obtains several new results.

Universität Mannheim

The Mannheim group has worked on a number of themes related to the interaction between firms in complex industries. This has resulted in papers on the telecom industry, on the banking industry, and on co-operation between firms (research joint ventures, mergers) and their effects on an industry.

As to the telecom industry, Matthias Blonski (Mannheim) finished a paper on the demand and the market structure in telecommunications markets. Based on a theoretical model, he argues that price differentiation may lead network service suppliers to retain (localised) monopoly power even under deregulation. This work is based on the author’s earlier work on the existence and characterisation of anonymous games with binary actions.

Malte Cherdron (Mannheim) has finished a paper on negotiated access charges for network interconnection. Network providers can use such charges to collude in the final market when consumers’ calling patterns are sufficiently biased, and when consumers cannot easily switch networks. High access charges lead to high prices for off-net calls. If consumers call users within ‘their’ network more frequently than others, this leads them to perceive the networks as being differentiated even though, technically speaking, they are not. By agreeing on high interconnection fees, the network providers can thus turn themselves into local monopolists in their respective sub-markets rather than competing with their undifferentiated product in the full market.

On the banking industry, Thomas Roende (Young Researcher, Universität Mannheim) has completed a paper called ‘Regulating Access to International Large Value Payment Systems’ (joint with Cornelia Holthausen). As a result of European Monetary Union, there have recently been established several European networks which allow banks and other financial institutions to do cross-border transfers (so-called large value payment systems). The design and regulation of these payment systems are crucial for financial stability because of the large volumes transferred but have so far received little attention. The paper studies regulation of access to different types of cross-border payment systems when banking supervision, as in Europe, is kept at the national level. The authors build a model in which the communication between the national regulators is endogenous. It is shown that the national regulators’ incentives are not perfectly aligned when deciding upon access. As a result, systemic risk is excessive under public regulation. However, the authors show that leaving access regulation in the hands of the private sector can only be optimal if the private banks have superior information about the risk of their foreign counterparts in the payment system.

In a related work (Holthausen and Roende, 2002), the authors analyse co-operation between home and host country in the supervision of an international bank. According to the principle of ‘home country supervision’, it is the supervisor in the bank’s home country that has the consolidated oversight of the bank with all its foreign branches. The host country supervisor participates in the supervision of the local branch to make use of its local expertise. The authors consider the question whether to close the bank or to leave it open. The home country supervisor has the formal authority to close the bank. However, before making this decision, it wishes to consult the host country supervisor about the financial state of the foreign branch A political economy approach is taken to regulation and it is assumed that supervisors maximize the welfare of their own country. The communication between the supervisors is modeled as a 'cheap talk' game. It is shown that: (1) first best closure regulation cannot be implemented; (2) the more aligned the interests are, the higher is welfare, but the lower are the profits of the bank; (3) the bank can allocate its investments strategically across countries to escape closure.

Cherdron and Stahl have completed two papers on merger decisions in complex industrial networks, such as that involving the upstream firms in the automotive industries. Building on Segal’s model of merger decisions in a co-operative game, they develop a general framework, and give specific examples of how merger decisions between two firms change the structure of bargaining power within an industrial network. They give conditions for two central results to emerge: First, that a merger may benefit uninvolved firms more than the merging firms themselves; and second, that one merger may induce a merger wave.

Lisbon

Tommaso Valletti (CEPR), Steffen Hoernig (Young Researcher, Lisbon) and Pedro Pita Barros (Lisbon) looked at how universal service objectives are pervasive in telecommunications, and have gained new relevance after the introduction of competition in many markets. Despite their policy relevance, until now little work has been done allowing for a thorough discussion of instruments designed to achieve universal service objectives under competition. Their intention was to fill this gap, and disaggregate the problem into interacting forms of regulatory intervention such as uniform pricing and coverage constraints. It is shown that these are not competitively neutral and may have far-reaching strategic effects. Under uniform pricing, equilibrium coverage of both incumbent and entrant may be lower than without regulation. These effects depend on which measures are imposed at the same time, thus no single measure can be evaluated in isolation. They also pointed out that different groups of consumers are affected in different ways, making welfare comparisons difficult.

Hoernig showed that under decreasing returns to scale there are continua of non-zero profit-mixed Bertrand equilibria where firms randomise between any finite number of pure equilibrium prices. In addition, he worked on defining a notion of stability of equilibrium in an infinitely repeated step-by-step R&D race. The unique symmetric equilibrium is shown to be unstable, and stable asymmetric equilibria arise, if product market competition is intense, firms are patient, imitation is difficult and innovations are large. Some predictions based on symmetric equilibria, e.g. that less patient firms always invest less in research, or that more intensive competition leads to higher economic growth, are reversed for ''realistic'' values of the underlying parameters.

Hoernig carried out further work on analysing the effects on entry and welfare of universal service obligations, such as uniform pricing, price caps and unbundling, on allocations in markets that were newly opened to competition, e.g. telecommunications. He found that if these obligations are imposed not only on incumbents but also on entrants, entry may not result in competition if installed capacity is low, or may lead to the neglect of high-cost areas if installed capacity is high. These results hold no matter whether a price cap is in place or not, and do not depend on whether entry is capacity-based or facilitated through unbundling. If on the other hand uniform pricing is only imposed on the incumbent, welfare rises after entry with unbundling. In the presence of a low price cap, welfare without unbundling may actually decrease after entry due to the duplication of capacity.

Pita Barros and Martinez-Giralt (IAE, Barcelona) found that in several instances, third-party payers negotiate prices of health care services with providers. They showed that a third-party payer may prefer to deal with a professional association than with the sub-set constituted by the more-efficient providers, and then apply the same price to all providers. The reason for this is the increase in the bargaining position of providers. The more efficient providers are also the ones with higher profits in the event of negotiation failure. This allows them to extract a higher surplus from the third-party payer.

There is currently little knowledge on the consequences of diffusion of cellular technology on the incumbent fixed-link telephony service. Pita Barros and Cadima (Lisbon) addressed this issue by estimation of diffusion curves for both technologies, allowing for potential cross-effects, using data from a small European economy. Their main findings were a negative effect of the mobile phone diffusion on the fixed-link telephony penetration rate. The effect is, roughly, a ten percent decrease in the fixed-link penetration rate (in comparison with the absence of mobile phones). No effect on the reverse direction seems to exist. Mobile phone market growth seems to be determined essentially by technological advances.

Recent years have witnessed a renewed interest in the efficiency defence in merger evaluations. Antitrust authorities adopted a softer position regarding this argument and put through approval of merger operations. Curiously, most discussions implicitly assume welfare to be evaluated at the expected value of cost savings. This introduces a bias in merger relative evaluation to the use of an expected welfare measurement. Cabral (Lisbon) and Pita Barros showed that no general presumption can be held about the sign of the bias. Depending on the particular details of each industry, the bias can be either towards approval of expected-welfare decreasing mergers or towards rejection of expected-welfare increasing mergers. A detailed case-by-case analysis is required, adding further complexities to the proper use of efficiency arguments for merger approval.

People vote although their marginal gain from voting is zero. Amaro de Matos (Lisbon) and Pita Barros contributed to the resolution of this paradox by presenting a model for equilibrium configuration of voting attitudes. Each individual is seen as an element of a social network, within which pairs of individuals express ideas and attitudes, exerting mutual influence. They modelled the role of such networks in propagating the mutual influence across pairs of individuals and showed that it suffices that a small set of individuals have a strong feeling about voting to generate a significant turnout in elections.

The European Commission issued a proposal for reform of some aspects of Communitary competition policy. In particular, proposes a shift from ex-ante control (notification system) to an ex-post control regime of agreements (or decisions) conflicting with Article 81 (Ex-Article 85). Pita Barros highlighted the expected effects of this shift on the type of agreements that firms will implement. The type of agreements has been taken as exogenous in most analysis of the reform proposal. He contended that significant economic effects may result from recognizing the endogeneity of agreements but predicted that the proposed reform will result, in general, in firms implementing less restrictive agreements.

Instituto d’Analisi Economica (IAE), Barcelona

Vasiliki Skreta (Young Researcher, IAE) considered the hypothetical situation in which a risk neutral seller owns a single object and faces a risk neutral buyer whose valuation is private information and supposed that there are two periods and that the seller wants to design a mechanism to sell the object. However if the buyer does not obtain the object in the first period, the seller cannot commit to not try to sell the object at the second period using a different mechanism. Skreta characterised the revenue maximizing allocation mechanism in this environment, and showed that posting a price in each period is optimal. In order to do so, she developed a methodology that can be used in order to derive optimal incentive schemes in asymmetric information environments without commitment. Bargaining theory acknowledges that individuals cannot commit to stop negotiating at a point where no agreement is reached, but has only examined possible outcomes of negotiations when the offers are take-it-or-leave-it. She showed that this is in fact the best thing to do among all the possible procedures. This is the first complete characterisation of the optimal mechanism under non-commitment in an environment where the agent’s type is drawn from a continuum.

Skreta extended this work to incorporate the situation in which the seller faces many buyers. In a two-period model, and assuming ''independent private values and risk neutral buyers,'' she derived the set of allocation rules implemented by a Perfect Bayesian Equilibrium, (PBE), of the game and assumed that the buyers' distribution of valuations satisfies the monotone hazard rate assumption and finds the optimal allocation rule among all PBE-implementable ones. The seller can implement this allocation rule by employing in each period a mechanism that assigns the object to the buyer with the highest virtual valuation if it is above a cutoff value. In the symmetric case, this rule can be implemented by a sequence of second or first price auctions with a reservation price in each period. The reservation price decreases over time.

In addition to her work above, Skreta showed how revenue generated from agreed access fees can alleviate the problem of providing high quality telecommunication services at low prices even to consumers located in high-cost areas. She demonstrated that a rural network may generate revenue from selling access, high enough to cover its losses from the retail sector. The possibility of cross-subsidization via access fees has so far been ignored by policy makers. It is generally believed that cross-subsidization between profitable and less profitable areas, common in the regulation era, is incompatible with deregulation. It turns out that cross-market subsidization can take place without outside intervention, through access fees that networks pay to each other. The lessons apply to other industries that have the network structure, for instance electricity, postal services and transportation.

Nicolas Melissas (Young Researcher, IAE) compared two different auctions in the oil-drilling industry. The first one, used by the American government, is a first-price sealed bid auction in which each player’s bid becomes public information prior to the first drilling date. As firms infer each other’s signal by inverting the bidding function the ensuing war of attrition (to decide who will drill first) is characterised by symmetric information. The second one is identical to the first one except that each player’s bid is not divulged to the other players. He showed that the second auction can be more efficient than the first one.

Melissas also worked on a paper introducing cheap talk in a dynamic investment model with information externalities. He first showed how the credibility of cheap talk statements is adversely affected when players can learn by observing other players' actions. He then proceeded to show how informational cascades help in restoring this credibility. Subsidising investments favours truthful revelation of private information. The more precise the sender's information, the higher his incentives to truthfully reveal his private information.

Roberto Burguet (IAE) (with Yeon-Koo Che, University of Wisconsin) looked at competitive procurements which usually entail assessment of dimensions other than price, that we term quality. An agent has to assess these other dimensions therefore giving rise to the possibility of manipulation in exchange for bribes. They analysed a model of simultaneous competition in multidimensional bids and bribes. The higher briber obtains a favorable assessment of quality. Unless the manipulation ability is small with respect to the efficiency differential, the equilibrium result is inefficiency in the sense that inefficient firms win the contract with positive probability. On top of this inefficiency cost and the cost of the bribe, the price competition is softened so that the buyer supports an inflated price.

Burguet (with Perry, Rutgers University) also looked at the case of an auctioneer who takes bribes for one of the bidders in order to allow for a revision of the bid in case of unsuccessful bidding. The bribe takes the form of a percentage of the surplus. They showed that this type of bribery need not increase the price paid by the buyer if the percentage bribe is high. If firms invest in cost reduction before competing, bribery decreases the total incentives to invest, but may increase that of the favoured firm. If ex ante competition for favouritism takes place, a more efficient, bigger firm wins that competition.

London Business School

The LBS organised a Conference in May 2001 on Corporate Control and Industry Structure in Global Communications. Most of the papers have been collected in a forthcoming issue of Telecommunications Policy edited by FrancescTrillas (Young Researcher, LBS) and Leonard Waverman (LBS). The papers in the issue analyse the problems arising from the interaction between corporate finance and industrial organisation in telecommunications.

This special issue will include a paper by Trillas on "Mergers and Acquisitions in European Telecommunications". This paper is a study of the twelve largest acquisitions in European telecommunications up to 2000. It shows that bidding telecommunications firms did not create value for their shareholders in their acquisitions activities, on average. A case study of the Spanish firm Telefonica shows that firms can actually destroy value for shareholders, mostly due to political constraints.

A revised version of Trillas' paper on "The Enersis takeover" has been accepted for publications in Utilities Policy. In this paper, the author shows how takeovers by electricity firms may destroy value for shareholders, due to the political mobilization of stakeholders preventing the bidding firm from capturing any gains.

Francesc Trillas has continued working on the Political Economy of Regulation. He is involved with Paul Levine (Surrey University/LBS) and Jon Stern (LBS) in a research project on independence and governance of regulatory agencies in telecommunications. The two first outputs of this research are:

- Regulation with Rationed Information or Delegation: Solutions to the Underinvestment Problem? (joint work with Levine). Regulation Initiative Discussion Paper.

-Regulation in Telecommunications: Lessons from Central Banks, published in the Business Strategy Review.