U Privatesector
P B L I C P O L I C Y F O R T H E The World Bank Group January 1997 Note No. 104
Competition in Network Industries Where and How to Introduce It
If privatizing network industries is to bring lasting public benefits, governments should also attempt to introduce competition. Otherwise, the regulation required in areas with monopoly segments may become so intrusive that it undermines the reform. The scope for such competition is growing with increasing deregulation and technological innovation. And as technology continues to improve, the use of smart marketscomputer-assisted auction systems to clear competitive but complex marketsis likely to become feasible for an ever-expanding group of products and countries. This Note outlines the opportunities for introducing network competitioncompetition for the market, competition over existing networks, and competition among networks. It briefly considers in each case whether regulation will still be necessary and whether it will become easier or more complicated. It looks at how these opportunities could be applied in different networks. And it concludes with some basic guidelines for introducing competition. price regulation. Only rebidding promises an escape from the old-style regulated monopoly. But if there are significant sunk investments involved, assets will need to be transferred at the end of the franchise, leading to complex exercises in asset valuation. Still, by placing time limits on the franchise and requiring some form of competitive rebidding, governments can ensure regular challenges to the incumbent, and the incentive to maintain reputation will reduce the temptation to slacken efforts during the duration of the franchise.
Michael Klein and Philip Gray
Competition over existing networks
There are three types of competition over existing networks, described here as open access, pooling, and time-tabling arrangements. Which type is suitable for a particular network industry depends on the technical characteristics of the goods and of the networks over which these goods are provided.
Open access
Competition for the market
One way of bringing competitive forces to bear on natural monopoly segments of an industry is to delineate a monopoly franchise and auction it off to the bidder offering the lowest price to consumers.1 But monopoly franchises, especially long-term ones, still involve regulationindeed, some commentators argue that this form of competition is simply a way of facilitating regulation. Prices and related terms of the franchise (often known as a concession) have to be adjusted in response to events. These adjustments can be complex, but there are basically two options for making them: rebidding the franchise periodically or using traditional Open access regimes are found in many gas pipeline systems in Europe and the United States. Parts of telecommunications networks are also under open access regimes, including long-distance satellite communications, parts of the major carriers long-distance networks, and the local loop in systems where there is competition in long-distance services. In essence, open access occurs when allowing competition in one segment of an industry requires ensuring access to the remaining natural monopoly bottlenecks, provided that there is available capacity. If the owners of a gas pipeline have no interest in supply, for example, it will always pay them to allow access to additional
Private Sector Development Department Finance, Private Sector, and Infrastructure Network
Competition in Network IndustriesWhere and How to Introduce It
gas suppliers. When capacity constraints are binding, rationing of access (interconnection) to the bottleneck will be needed. This interconnection can be achieved efficiently without regulation, but the owners of the bottleneck facility may receive monopoly profits. If these profits are politically unacceptable, the access prices charged by the owners of the bottleneck facility and the prices for any other services they provide to final customers of the network industry will need to be regulated. If the incumbent owns some of the competing supply facilitiesfor example, power plants, gas fields, or long-distance telephone transmission facilitiesit may try to raise prices for network access to preclude competitors in the nonmonopoly segments of the network. To prevent such predatory behavior by owners of monopoly segments, regulators may impose access obligations and matching pricing principles, such as forcing the incumbent to pay a price for transport equal to the price it charges its competitors. If such limits cause the owner of the bottleneck facility to try to exploit market power in the competitive segments, there may be a case for imposing limits on vertical integration and separating ownership from other parts of the system.
Pooling
ing used in electricity systems throughout the world and in gas systems in the United Kingdom and the United States.
Time-tabling
In power or natural gas supply systems, the source of the electrons or molecules a customer receives does not matter because the product is sufficiently homogeneous. But for airlines, railways, or telecommunications, where freight, passengers, or callers need to reach a particular customer or point in the network, the requirements for network optimization are more complex than simply that total inflows match total outflows. If rights to use railway tracks, for example, were defined and allocated to multiple parties, secondary trading should yield the optimal set of paths through the networkthe set that maximizes welfare given producers and consumers valuations of the service. The optimal set of paths forms the optimal delivery schedule or timetable (delivery of person or good x to point y at time z). The issue is whether an optimal timetable can be generated through decentralized bargaining. Because the value of each right to use a segment of track at a particular time depends on what happens with all adjacent segments (all segments are indirectly adjacent to all others), a single, optimizing smart market may be needed. Sweden and the United Kingdom are investigating whether such smart markets can be established for railways. Another potential application is airport slots. For now, experiments with such smart markets have been limited to computer simulations.
The open access approach attempts to allow competition over the network by selling rights to network capacity to competing firms on a nondiscriminatory basis. But it may be difficult to define, adjust, and enforce such rights in a way that allows effective competition. For example, in a power system, the capacity used or unused at any moment in any part of the system is a function of all physical flows throughout the system, not of bargaining or individual transport decisions, so it may not be practical to define capacity or access rights. An alternative is to use a central dispatch system that optimizes system flows, instantaneously matching supply and demand. This ensures open access in the sense that winning bidders will always and by definition be dispatched. Such smart pooling systems are be-
Competition among networks
The discussion has suggested that the hard core of natural monopoly is the smart market, whether for dispatch or for optimizing timetabling. But in some cases, competition among multiple networks may be desirable. For longdistance telecommunications networks, petroleum product distribution systems competing with natural gas systems, or railways competing with trucks, for example, the theoretical
benefits of complete and integrated scheduling are probably less important than the practical benefits of allowing competition among networks. Competition is most useful where the central planning problem is hardest because of great uncertainty or complexity. There are thus dynamic or informational benefits from incomplete scheduling, which allows competition on the basis of some duplication. This duplication is necessary to try out new things and to check monopolistic behavior. Centralized scheduling is likely to be inevitable only when temporary congestion is very costly such as systemwide electricity blackouts.
where goods and services are not homogeneous and where starting and end points of network flows matter, though possible are solutions being debated in the context of the Swedish and U.K. railway reforms. The currently preferred option for such networks is some form of open access or common carriage system with regulation of interconnection. This approach is particularly appropriate for telecommunications, but it is also used in such networks as railroads, airports, and natural gas pipelines. One way to introduce unregulated competition is to rely on competition between networks, or intermodal or substitute competition. Railways, for example, often face competition from trucks. And competition from the petroleum products market can discipline pricing behavior in the natural gas market, as in Finland, Germany (for large users), and Hong Kong. An international comparison of regulatory regimes shows that countries are most likely to leave the rail and natural gas sectors unregulated, relying on substitute competition to provide pricing discipline. In telecommunications, line-based networks are increasingly exposed to competition from wireless services and in many cases even from new line-based networks being established as the cost of such infrastructure falls. Further technical progress may obviate the need for regulation. Countries with limited regulatory capability can already rely on competition from wireless services to provide basic consumer protection. The toughest regulatory challenges remain in electricity, water, airports, and roads. As previously indicated, in electricity, the solution may lie in competitive power pools. In water, competitive forces may be conceptually similar to power pools, but their effective introduction is a fair way off. In the roads sector, operations may be revolutionized as electronic traffic management in conjunction with congestion pricing may soon become more widespread as a result of tests in such countries as Italy, Norway, Singapore, and the United States. In California, for example, a private toll lane has been financed by charging congestion tolls that are inversely proportional to the traffic on the free
Industry guidelines
How competition is introduced and how effectively and easily it is implemented will vary from one network industry to another, depending on the physical characteristics. Introducing competition is generally easiest in industry segments where sunk costs are unimportant, such as for many transport vehiclesships, planes, trucks, and taxis. The policy solution here is free entry without economic regulation. Where economies of scale due to scheduling are important, such as in urban bus transport or solid waste collection services, awarding monopoly franchises competitively may be efficient. As long as sunk costs are not important, repeated franchise bidding can provide a good level of competition without a need for extensive regulation. There has been positive experience with competition in all these transport industry segments. Where sunk costs are important, introducing competition becomes more complex. For electricity and natural gas systems, which produce and transport fairly homogeneous products, the best solution appears to be smart competitive pools wherever a sufficiently large market can be created to sustain workable competition. This argues, of course, for fostering international trade in energy services wherever possible. Although still at an experimental stage, competitive pools have shown clear promise. Smart markets have yet to provide practical solutions for introducing competition in networks
Competition in Network IndustriesWhere and How to Introduce It
alternative. Auctions of airport landing rights still await the arrival of appropriate smart markets, which would also be instrumental in efficiently managing road networks and decentralizing investment decisions in these networks. The key to finding new solutions for introducing competition is technical progress. In telecommunications, technical change offers hope for workable competition among networks. In other industries, such as transport and energy, new solutions will depend on advances in telemetry and telecommunications combined with computer-based smart markets.
ture firms or issuing concessions to build new facilities. Monopoly rights do lower the cost of capital and make financing easier. But they do so by shifting risk to customers, not by reducing overall risk. In the last century, investments not protected by entry barriers were nevertheless funded. And today, new investments in competitive segments of network industries also are being financed, such as power plants in competitive markets in Argentina, Chile, and the United Kingdom. The technological changes and new thinking discussed in this Note suggest the following guidelines for policymakers introducing competition: The more complex the network and the lower the sunk cost, the more value there is likely to be in introducing competition from other networks. Where technical change is rapid, defining the bounds of natural monopoly will be m...