135 power. In this case a vertically integrated monopolist could avoid the doublemarginalization problem of successive monopolies and increase its profits compared to the non-integrated case (Kaserman and Mayo, 1995: 305). These conclusions drawn from the theory of vertical integration hold in the absence of market-power regulation. They are not applicable when regulation hinders the upstream monopolist from extracting monopolistic rents on the upstream market. In this case a monopolist can have an incentive to vertically integrate also into a competitive downstream market. As long as the downstream market is not price regulated an integrated monopolist can hope to foreclose the downstream market and extract above competitive returns on this level. Whenever a monopolistic bottleneck is identified for regulation, it is therefore important to consider also the incentives of a regulated monopolist to foreclose a related, unregulated downstream market. Foreclosure in this context does not only refer to practices by which access on the wholesale level is completely denied to competitors or offered only at discriminatory prices. Foreclosure can also refer to the practice of discriminating competitors by non-price methods relating especially to the quality of the service offered (Laffont and Tirole, 2000: 161). The threat of foreclosure adds additional demands to the regulation of bottleneck network elements. In addition to applying price regulation as an instrument to guide the efficient allocation of resources in the market, in which monopolistic bottleneck elements are found, policy makers must also monitor the quality of service of the wholesale products to guarantee equal chances for competitors in the downstream market. 8.2 The regulatory framework Since market-power leveraging can extend inefficiencies from monopolistic bottleneck network areas to related competitive markets, leveraging is often used as a justification for extending the regulatory basis into competitive market segments. Laffont and Tirole (2000: 163), for instance, argue: “…the deregulation of competitive segments is costly, since it substantially increases the monitoring requirements.” They favor a global price-cap regulation which includes both the intermediate (access) goods as well as the competitive retail goods of the bottleneck owner (Laffont and Tirole, 2000: 170ff.).118 A global approach to sector-specific regulation, however, neglects the fact that regulation in itself is always costly. The present analysis takes the stance that a minimization of the regulatory basis should be aspired to by avoiding the regulation of market segments that are competitive. When market power is effectively regulated, then this will hinder the bottleneck owner from leveraging market power into competitive markets. 118 Laffont and Tirole, however, also caution that a global price cap could be used by the incumbent to price squeeze competitors out of the market by meeting the price-cap requirements through a substantial decrease in the final goods prices combined with an increase in the intermediate goods prices (Laffont and Tirole, 2000: 174). 136 The disaggregated approach to sector-specific regulation introduced in chapter 4 is a regulatory approach that pursues the objective of minimal regulation. It uses the theory of monopolistic bottlenecks to identify the minimal regulatory basis and formulates normative rules for the proper regulation of monopolistic bottlenecks. The present analysis uses this approach as the regulatory framework within which the regulation recommendations for the market for Internet service provision are derived. In what follows the principles of regulation laid down by the disaggregated regulatory approach will be presented in more detail. The symmetry principle The underlying principle for all sector-specific regulation within the disaggregated regulatory framework is that regulation must be symmetric. By this is meant that all market participants should be treated equally, be they incumbents or new entrants (Knieps, 1997: 326). In the absence of active or potential competition, incumbents must offer access to monopolistic bottlenecks on a non-discriminatory basis, both with respect to price and non-price characteristics. All technologies should be treated equally, which implies that if monopolistic bottlenecks exist in the infrastructures of new entrants, for instance in upgraded cable-TV networks, then entrants should face the same open-access regulation as is applied to the infrastructure of incumbents. Furthermore, all market participants should face the same obligations with respect to, for instance, financing fixed costs of network infrastructure, financing universal service obligations, etc. Finally, incumbents and entrants should face comparable price signals in their decisions to invest and to innovate. Accounting separation The disaggregated regulatory framework proposes accounting separation as one means to enforce the symmetry principle (Knieps, 2006: 67). By requiring the owner of the monopolistic bottleneck to keep separate accounts, the regulator can monitor the regulated and non-regulated divisions of the regulated operator as separately as possible and can more effectively hinder the regulated firm from cross-subsidizing or from treating competitors differently from internal divisions. The information required of the bottleneck owner should also include details on the service quality vis-à-vis own divisions and external divisions such as time to deliver and other performance characteristics, such that quality discrimination can be detected. Price-cap regulation The disaggregated framework endorses price regulation of bottleneck network elements. Optimally, the regulation of access prices to monopolistic bottlenecks should perform three objectives. Firstly, prices should provide signals for allocative efficiency. In competitive markets, marginal-cost prices steer resource allocation into efficient usage. In a regulated market, regulated prices should be close to incremental costs. Secondly, prices should encourage productive efficiency by sending the right signals for the decision on whether joint usage of a bottleneck or bypass investments by new entrants are economically efficient (Armstrong, 2002: sec. 2.4.1). 137 Prices that reflect the incremental costs of additional usage of the bottleneck resource would meet this requirement. Lastly, access charges should send signals for dynamic efficiency and cover the total (efficient) costs of the monopolistic bottleneck, not merely its incremental cost. Only then does the regulated firm earn a rate of return that provides it with an incentive to maintain the network and invest into innovations. These objectives for optimal price regulation are incompatible when the regulatory instrument consists of only one access charge. Any mark-up on incremental costs that is required to cover the large fixed and common costs of network infrastructures will lead to inefficient usage of the bottleneck network elements. When only a linear access price is available to pursue the conflicting policy objectives of access-price regulation, then regulatory theory proposes a Ramsey-pricing structure as the second-best alternative to incremental cost pricing.119 Ramsey-prices are obtained by maximizing social welfare under the constraint that a given revenue must be raised. This pricing approach yields prices that include mark-ups on marginal costs, which are inversely proportional to the elasticity of demand (Baumol and Bradford, 1970: 269f.). The lower the elasticity of demand for a given product, the higher can be the mark-up on marginal costs. By this pricing rule, the loss in social welfare which results from the need to cover total costs of production is minimized as quantity reactions to a price increase are lowest for those portions of demand that are inelastic. A regulatory authority cannot “set” the welfare-maximizing Ramsey-prices because it lacks the necessary information on the demand elasticities of consumers as well as the production costs of the regulated products. However, the regulatory authority can create an environment in which it is in the interest of the regulated firm to set prices that approximate the positive welfare effects of Ramsey prices (Laffont and Tirole, 2000: sec. 3.4.1). As will be shown below, price-cap regulation provides incentives for setting prices in inverse proportion to the elasticity of demand. Price-cap regulation fixes the price level, but not the price structure, for a given basket of regulated goods.120 119 The term Ramsey-pricing originates from work by Frank Ramsey on the question of how a government can raise a given tax revenue with a minimum of utility loss (Laffont and Tirole, 2000: sec. 2.1.1). 120 Price-cap regulation was developed by Prof. Littlechild for the British Department of Industry in the early 1980s during the privatization process of British Telecom. Littlechild criticized the prevalent rate-of-return regulation for two main reasons. In rate-of-return regulation the regulated firm provides the regulator with information on its operating costs and its cost of capital. The regulator, upon auditing these cost accounts, allows a fair rate of return on the invested capital. This provides the regulated firm with incentives to inefficiently increase its capital stock (Averch and Johnson, 1962). Furthermore, rate-of-return regulation is often applied to the regulated firm as a whole. Littlechild was looking for a regulatory instrument that is applied more specifically to monopoly areas only. 138 The key features of this price control are that, for a pre-specified period of four to five years, the company can make any changes it wishes to prices, provided that the average price of a specified basket of its goods and services does not increase faster than RPI-X, where RPI is the Retail Price Index (i.e. the rate of inflation) and X is a number specified by the government. At the end of the specified period, the level of X is reset by the regulator, and the process is repeated (Beesley and Littlechild, 1989: 455). Price-cap regulation gives the regulated firm the freedom to allocate common costs among the products in a product basket in any way it chooses. It increases the incentives for productive efficiency by the bottleneck owner because for the regulated time period the operator gets to keep any gains in productive efficiency.121 Since the regulated firm’s revenues are maximized by setting the prices inversely proportional to the demand elasticities of the products, the resulting prices have similar efficiency properties as Ramsey-prices. A further advantage of price-cap regulation as compared to cost-based regulation is its lower information requirement. The price of the product basket is calculated as the average price of the products in the relevant product basket, determined with weights that reflect the products’ previous period’s turnover shares. To determine the RPI, the regulator needs to select an appropriate price index. Most difficult is the setting of the right level of X. Here, however, the regulator also can take more liberties than in cost-based regulation. Phasing-out of regulation The last principle of disaggregated regulation is that the dynamics of market evolvement and technological development need to be taken into account. All regulation must be reviewed regularly in order to evaluate whether demand changes or supply changes from innovations or investments have altered the extent of the monopolistic bottleneck network area. If the bottleneck no longer exists or the extent of the bottleneck has shrunk, then the application of regulatory measures must end or be modified (Knieps, 1997: 331ff.). 8.3 Regulation recommendations The first step in applying the disaggregated regulatory approach is to perform a thorough analysis of the extent of the monopolistic bottleneck network area to which regulation is to be applied. Gabelmann (2003, 161ff.) approaches the question of how to define the exact extent of a monopolistic bottleneck by asking what type of competition can be expected in the end-user market and what would be the minimal invasive regulation into the incumbent’s property rights to facilitate this competition. She differentiates especially whether competition can be expected to bring forth price differentiation only or technology variation in the end-user market as 121 In cost-plus regulation the efficiency gains are theoretically “regulated away” as soon as the firm lowers its production costs.