70 be that the incumbent, for some reason, has access to a more efficient production technology or lower input prices than potential entrants, for instance, due to control over natural resources. When the requirements for long-term cost-advantages are not fulfilled, then entrants have access to the same efficient production technologies as the incumbent. Short-term cost-advantages of the incumbent, which Bain includes among the barriers to entry, will disappear over time and can therefore, according to Stigler, not deter entrants from entering a market in which above competitive profits persist. Product differentiation activities, for instance, can be undertaken by either firm. In fact, product differentiation is a central expression of competition in markets. The costs of differentiating one’s service or product are the same for both incumbents and newcomers. The fact that the expenditures for product differentiation and advertising may occur at different points in time for the incumbent and the new entrant is negligible when both have access to efficient capital markets.59 Even economies of scale and scope in the relevant region of market demand cannot be considered barriers to entry in the long run. When both the incumbent and the entrant have access to the same production technologies, then the entrant can potentially replace the incumbent and produce the same level of output at the same average cost of production (Knieps, 2005b: 82). In the extreme, even a natural monopolist has to fear being replaced by a potential entrant, when the entrant can expect to produce on the same long-run cost function as the incumbent.60 4.4 A reference model for justifying sector-specific regulation in network industries Given the typical market characteristics of network industries, it is easily recognizable that the use of market shares as a criterion to localize market power will promote errors of the first order. In these industries especially, it is important to use a model of monopolistic competition as a reference model to judge the effectiveness of competition in the market. This section introduces the theory of monopolistic bottlenecks as a reference model, which was designed explicitly to localize stable market power in network industries (Knieps, 2005b: 82). It is a theory grounded in network economics and therefore able to differentiate between typical characteristics of network industries and evidence of market power. Based on this theory it is possible to identify stable market power ex-ante and therefore to justify ex-ante sector- 59 This argument is backed up also by empirical studies which do not support the hypothesis that advertising expenditures are a barrier to entry (Schmalensee, 1989: 981). 60 Empirical studies on the correlation between concentration and profitability have produced mixed results and in sum do not support the hypothesis of a positive concentrationprofitability relationship (Schmalensee, 1989: 973-977). 70 be that the incumbent, for some reason, has access to a more efficient production technology or lower input prices than potential entrants, for instance, due to control over natural resources. When the requirements for long-term cost-advantages are not fulfilled, then entrants have access to the same efficient production technologies as the incumbent. Short-term cost-advantages of the incumbent, which Bain includes among the barriers to entry, will disappear over time and can therefore, according to Stigler, not deter entrants from entering a market in which above competitive profits persist. Product differentiation activities, for instance, can be undertaken by either firm. In fact, product differentiation is a central expression of competition in markets. The costs of differentiating one’s service or product are the same for both incumbents and newcomers. The fact that the expenditures for product differentiation and advertising may occur at different points in time for the incumbent and the new entrant is negligible when both have access to efficient capital arkets.59 Even economies of scale and scope in the relevant region of market de and cannot be considered barriers to entry in the long run. When both the incumbent and the entrant have access to the same production technologies, then the entrant can potentially replace the incumbent and produce the same level of output at the same average cost of production (Knieps, 2005b: 82). In the extreme, even a natural monopolist has to fear being replaced by a potential entrant, when the entrant can expect to produce on the same long-run cost function as the incumbent.60 4.4 A reference model for justifying sector-specific regulation in network industries Given the typical market characteristics of network industries, it is easily recognizable that the use of market shares as a criterion to localize market power will promote errors of the first order. In these industries especially, it is important to use a model of monopolistic competition as a reference model to judge the effectiveness of competition in the market. This section introduces the theory of monopolistic bottlenecks as a reference model, which was designed explicitly to localize stable market power in network industries (Knieps, 2005b: 82). It is a theory grounded in network economics and therefore able to differentiate between typical characteristics of network industries and evidence of market power. Based on this theory it is possible to identify stable market power ex-ante and therefore to justify ex-ante sector- 59 This argume t is backed up lso y empirical studies which do not support the hypothesis that advertising expenditures are a barrier to entry (Schmalensee, 1989: 981). 60 Empirical studies on the correlation between concentra ion and profitability have produced mixed results and in sum do not support the hypothesis of a positive concentrationprofitability relationship (Schmalensee, 1989: 973-977). 71 specific regulation. The disaggregated regulatory approach is introduced as a regulatory approach which applies the theory of monopolistic bottlenecks to network industries. 4.4.1 The theory of monopolistic bottlenecks The theory of monopolistic bottlenecks (Knieps, 1997) is based on a strict application of Stigler’s definition of entry barriers in the context of network industries. It states that only network areas which combine the characteristics of a sustainable natural monopoly with irreversible investments have entry barriers that lend stable market power to incumbent firms. The formal definition of these network areas, socalled monopolistic bottlenecks, is given as (Knieps, 2006: 53): • The incumbent controls a facility which is necessary for reaching consumers and no second or third such facility exists such that there is no active substitute for the facility in the market. This is the case when, due to economies of scale and economies of scope, it is less costly to have only one supplier of this facility in the market (natural monopoly situation). • The facility cannot be duplicated by a potential entrant in an economically feasible way. Hence, there is no potential substitute for the facility. This is the case when the costs of the facility are irreversible. Both conditions need to be fulfilled if an incumbent is to have stable market power. A natural monopoly alone is not sufficient to substantiate network-specific market power because it is not established that potential competition will not effectively discipline the incumbent.61 Further, irreversible investments alone are not sufficient to deter market entry as long as several active firms can operate in a market. Only when an incumbent natural monopolist has made irreversible investments does this give him an asymmetric cost-advantage over potential competitors in the sense of Stigler’s definition of entry barriers. The irreversible investments are no longer decision-relevant to the incumbent, but are part of the calculation made by entrants contemplating market entry. The credible threat that the incumbent will lower its prices to the level of variable costs prevents potential entrants from entering the market. Potential competition is therefore not effective when large sunk costs are associated 61 That natural monopoly alone does not lend market power to an incumbent is the central proposition of the contestable markets theory by Baumol et. al (1982). This theory can be seen as a precursor of the bottleneck theory (Knieps, 2006: 55). The aim of the contestable markets theory is to expand the traditional reference scenario of perfect competition to an alternative and more general reference scenario, compatible also with natural monopolies (Mantzavinos, 1994: 56). It derives the precise market conditions which guarantee market entry into markets with above competitive profits. The bottleneck theory focuses on the opposite viewpoint. It seeks to define market characteristics that lend stable market power to incumbent firms. It derives those market conditions that discourage market entry, even when above competitive profits are being made. 72 with market entry. Network industries are particularly vulnerable to this type of market power because large infrastructure investments create scale advantages and are often to a large degree sunk.62 The preconditions which need to be fulfilled in order for a network element to be classified as a monopolistic bottleneck are very close to the definition of what is called an “essential facility” in U.S. antitrust laws. A facility is deemed essential when (Hausman and Sidak, 2000: 467): • it is controlled by a monopolist, • competitors are not able to practically or reasonably duplicate the facility, • the monopolist has denied competitors the use of the facility while • it is practically possible to make the facility available to competitors.63 In fact, the difference in these two concepts is not in their definition of what constitutes an essential facility or a monopolistic bottleneck but rather that an essential facility is a concept used in the context of general competition policy, which takes effect only ex-post and is applied on a case-by-case basis. A monopolistic bottleneck, on the other hand, is a concept based on economic theory and developed to identify ex-ante a class of cases which are in need of sector-specific regulation (Knieps 2006: 54). 4.4.2 The disaggregated regulatory approach The disaggregated regulatory approach (Knieps, 1997 and 2006) uses the theory of monopolistic bottlenecks to identify network-specific market power as a justification for sector-specific regulation. Since the theory of monopolistic bottlenecks uses a narrow definition of entry barriers, it is a suitable basis to limit regulatory intervention. According to the disaggregated regulatory approach, the task of regulation is to design a system for guaranteeing access to monopolistic bottlenecks (Knieps and Zenhäusern, 2008: 129).64 This regulatory framework recognizes the costs of regulation and calls for minimally invasive intervention into market processes. It argues that network industries consist of vertically and horizontally integrated markets of which many are (potentially) competitive and only some lend network-specific mar- 62 Von Weizsäcker points out that when economies of scale are present, very often the incumbent firm will own plant and equipment dedicated to the particular industry, such that the theoretical and empirical work on economies of scale as a barrier to entry may have misjudged the actual cause of the entry barrier (v. Weizsäcker, 1980: 401). 63 The federal courts first applied the essential facilities doctrine in 1983. In the case MCI Communications Corp. vs. American Telephone & Telegraph Co. the Seventh Circuit defined a four-part test by which the plaintiff needs to show “(1) control of the essential facility by a monopolist; (2) a competitor’s inability practically or reasonably to duplicate the essential facility; (3) the denial of the use of the facility to a competitor; and (4) the feasibility of providing the facility.” (Hausman and Sidak, 2000: 467). 64 The design of access regulation within the disaggregated regulatory framework will be introduced in more detail in Chapter 9.