Chapter 4. Competition between and within networks
Face - Contents - Bottom - Previous/Next "Telecompetition" Chapter 4. Competition between and within networksTelecommunication is a network industry as electricity, gas, rail and postal services. Network industries have a range of common features setting it apart from ordinary industries. The prime distinguishing feature is the dependence on a network, for example cables, tracks and pipelines, as a conduit of the services produced in the industry. Networks have been very expensive to build with high fixed costs, but low marginal costs: Adding a new user to an existing network has not been very costly. For this reason, efficient utilisation of society's resources has historically been closely associated with monopoly provision of network services. A second distinguishing feature is the existence of network externalities. Network externalities imply that private and social gains from using a network grow as the number of customers connected to the network increases. It is clearly more beneficial for a person to be connected to a telecom network the more friends and family members can be contacted using the network. A non-trivial consequence of the presence of network externalities is that neither a competitive market, nor a monopolist may by them selves implement the best market outcome assuring an efficient utilisation of society's resources. Alone for this reason, the role of regulation is much more outspoken in industries with network externalities than in industries without. In many network industries there is only a single monopoly network. This is for example the case in electricity, gas and rail. These networks are characterized by significant economies of scale making it socially undesirable to duplicate the networks. In these network industries competition has to take place within the same network. The key competitive concern is to assure access to the network for all companies, actually and potentially, in the market. In most cases, governments have chosen to unbundle the network. Unbundling implies that ownership52 is transferred to a system operator independent of any commercial interest within the industry and rules are designed to assure non-discriminatory and transparent access to the network for all market participants. However, telecommunications is different. In all Nordic countries, the incumbent still owns and controls the Public Switched Telephone Network (PSTN) even though the incumbent at the same time competes fiercely with other companies in telecom markets that depend on access to the very same network. The dual role of the incumbent raises significant competitive concerns and accentuates the need for regulatory intervention to moderate the obvious conflicts of interest. The lack of unbundling between network ownership and provision of telecom services may be justified, if there is strong technical interdependence between the services provided on the network and the network itself. This may be the case if the introduction of new technologies requires significant investments that are sunk in case of failure and are associated with considerable uncertainty. If the ownership of the network is split between two entities, overall risk will be amplified due to the danger of coordination failure and it may limit the incentive to invest in the first place. This argument is supported in a recent report for the European Commission (Ovum, 2003) and even more forcefully by OECD (2003). In telecommunications, the case for unbundling of the entire network seems to be much less convincing than in other network industries. However, at the same time technological advancement has markedly reduced the costs of transmitting and switching a telephone call and has made the existence of competing networks and competition between networks much more realistic than in other network industries: In the mobile market, several competing networks are active in all Nordic countries; In the broadband market, network access - in particular via cable-TV - has already created competitive pressure on network access through PSTN. Furthermore, a range of promising network technologies is on the verge of becoming technologically and commercially viable on a grand scale: Optical Fibre, Fixed Wireless Access (FWA) and Power Line Communication (PLC), the latter exploiting the electricity network as the carrier of telecom signals. It seems that competition between and within networks has the potential to be significant and complementary means of promoting overall competition and to achieve efficient utilisation of society's resources in the telecom sector. The challenge will be to give the proper priority to the two types of competition and to make sure that policies designed to foster one type of competition does not undermine the other type of competition. Overall, competition between networks seems to be superior to competition within networks for the following two reasons: First, if competition takes place between alternative networks, competition is putting a competitive pressure on all parts of the value chain and generates maximum freedom to create and compete in differentiated products. In contrast, when competition takes place within the same network, some part of the value chain is exempted from the competitive pressure. In the telecom sector, competition within the (PSTN) network can, in principle, take three different forms: Bit stream access, Shared line and full unbundling of the local loop, cf. Figure 4.1. Bit stream access allows a service provider to resell the same retail products as the network owner without requiring significant specific investments. Shared line allows him to develop and sell a limited range of own broadband products but requires some specific investments. Full unbundling of the local loop implies that he rents the full local loop and can develop a broad range of narrow- and broadband services, but with significant investment costs. In contrast, setting up an alternative network to create competition between networks implies very significant specific investments, but the freedom to operate commercially is also unbounded. The above examples demonstrate the basic trade-off between entry and scope of competition. The less specific investments needed, the smaller is the risk, the easier is immediate entry, but also the more limited is the scope of competition. On the one hand, promoting access with low investment requirements leads to rapid entry but gives rise only to limited competition. On the other hand, promoting access with high investment requirements makes entry difficult, but may give rise to full competition. Only in the case of investments in alternative networks the entire value chain exposed to competition, except call termination to which we will return in the next chapter. Figure 4.1: Competition between and within networks
Second, if competition takes place within the same network access to the network must be mandated and strongly regulated (or alternatively the network must be unbundled). This is, in particular, important when the network owner competes on the same downstream market as the service providers who are granted access to the network. The network owner has a strong incentive to obstruct the service providers by raising access prices or otherwise making access problematic, costly and unreliable. In contrast, service providers who own their network owner do not depend on (one-way) access to resources controlled by competitors, although they may have a reciprocal interest in (two-way) access to each others networks. Consequently, more competition between networks not only extends the competitive pressure in several dimensions, it also reduces the need for regulatory intervention. It follows that any policy aiming at an efficient utilisation of society's resources as a prime concern should have to create competition between networks. However, competition between networks is not always superior to competition within networks. Competition between networks requires entry and if entry is difficult and carries high risk, competition within networks may be the only feasible option. Competition between and within networks in the Nordic countriesWe now give a brief overview of the status of and prospects for competition between and within networks in the Nordic countries. We distinguish between the three main telecom markets53: Fixed telephony, mobile telephony and broadband Internet services54. Fixed telephony Fixed voice is by far the largest and most mature telecom market. The fixed penetration rate, that is, the number of fixed voice subscriptions per 100 inhabitants, is above 65 in most Nordic countries and the call volume is at least 10 times as large as the call volume for mobile voice. However, total revenue for fixed voice and fixed broadband is not more than four times as large as the revenue for mobile voice. In all Nordic countries, the key network on the fixed voice market is the PSTN network owned by the incumbent operator(s). True competition between networks is non-existing as all service providers rely on PSTN-access, cf. Figure 4.2. In contrast, the number of service providers competing within the PSTN network is quite significant, in Norway, Denmark and Sweden having a market share between 30 and 50 percent. Figure 4.2: Fixed voice, competition between and within networks, markets shares for networks and competitors, 2003
Note: Competition within networks is measured on the basis of total outgoing calls in minutes. No market share data available for competition within networks in Finland. Source: Telestyrelsen, Denmark; Post- og Teletilsynet, Norway; Teleavisen AS, Norway; Kilpailuvirasto, Finland; Post- och Telestyrelsen, Sverige. However, measured by the share of domestic outgoing calls55 the market share of the incumbent operator(s) is still very significant, in all Nordic countries above 60 percent. Measured by the share of subscriber lines, competition is less intense as the incumbent's market share is close to 100 percent in all countries except Denmark with about 86 percent56. The prospects for attracting investments in new networks on the fixed voice market seem bleak. The number of fixed network subscribers has started to decline slowly and the call volume even faster. In the last three years the number of subscription lines in Sweden has declined by 6 percent while the call volume has slumped by almost 12 percent. The call volume stagnates as Internet traffic moves from ISDN and modems on the PSTN to high speed connections on the broadband market and as mobile voice gradually takes over volume from fixed voice. IP-telephony, that is, telephony via the Internet seems to be a promising alternative to fixed voice telephony. IP-telephony is a telephony service operated via Internet that requires access to a broadband network. If IP-Telephony succeeds, it is very likely that fixed voice becomes a product bundled with Internet access and Cable TV and provided on a broadband network. As a consequence, the fixed call volume may decline drastically. New investment in fixed voice is costly, especially in the local loop connecting directly to consumers, and investments are sunk cost in the sense that once investment has taken place, it has no alternative value if the investment fails. Large, sunk costs significantly increases the risk of the investor and deters investment in new networks. To sum up, the significance of the fixed voice market is already declining. It is not likely that investments in new infrastructure will take place in the market and there does not seem to be a strong role for promoting more competition between networks. However, we expect a significant part of the call volume to become integrated on the broadband market where the prospects for competition between networks are much more promising. Fixed broadband In all Nordic countries there are already several broadband access networks in place covering significant parts of the population. In addition to the PSTN network, cable TV networks cover 50-70 percent of the population, most significantly in Denmark, least in Norway. In Sweden, networks using the optical fibre technology also have some significance. Thus, compared to fixed voice there seems to be some room for competition between networks. Overall, the number of service providers is rather high, but lower than on the fixed voice market. One reason may be that independent service providers have (mandated) access to the PSTN network, but not to cable networks. However, the competitive pressure on the fixed broadband market is significantly muted due to the incumbent's joint ownership of the PSTN network and the main cable operator in all Nordic countries, except Sweden. Joint ownership of two potentially competing networks reduces the incentive to compete on prices and on creating new innovative services due to the risk of cannibalizing market shares of a sister company. This effect is larger the larger the combined market share of the jointly owned networks is. In Denmark, the incumbent, TDC, owns the PSTN network and the largest cable TV operator giving rise to a combined markets share at more than 70 percent measured in the number of subscription lines, cf. Figure 4.3. Hence, the combined TDC market share is larger on the expanding broadband market than on the declining fixed voice market. Figure 4.3: Broadband, competition between and within networks, markets shares for networks and competitors, 2003,
Note: Market shares within networks are measured on the basis of total outgoing calls in minutes. Market shares in Finland are for the Helsinki area where Elisa Corporation and TeliaSonera Finland are the two dominating market participants. The same story holds for Norway and Finland and for Sweden until the merger of Telia and Sonera in 2003. As part of the merger approval procedure TeliaSonera accepted to divest their cable TV operations and they now have separate ownership. However, it is still too early to gauge the impact on the market. The prospects for further investments in alternative networks seem promising on the broadband market. The market is expanding and margins are much larger than on the fixed voice market. Furthermore, a whole range of new technologies is gradually becoming technologically and commercially viable. It is expected that the promotion of new technologies as optical fibre, fixed wireless access and power line communication will have a significant, but still uncertain, potential for strengthening competition. In sum, there seem to be considerable opportunities for expanding competition between networks on the fixed broadband market. However, strict separation of ownership of competing networks, now and in the future, must be enforced. Nordic competition authorities should give high priority to the separation of horizontal ownership in all Nordic countries and to the prevention of ownership concentration across the new and old technologies. In addition, it is important to assure that consumers have real substitution possibilities between alternative networks and service providers. If switching costs are high between networks, due to different standards, need for new converters, switching fees and, possibly, administrative harassment, consumers will be locked in, in small pockets of monopoly power in reality nullifying the benefits of competition between networks. Mobile voice In all Nordic countries, there are at least two alternative networks on the market giving rise to considerable competition between networks. Furthermore, the number of competing networks is going to increase with the introduction of the third generation of mobile technology, but the speed of penetration is hard to predict (and depends on the choice of regulation). The incumbent owner of the PSTN networks owns the leading mobile network in all Nordic countries, but its market role is more modest than on the other two markets. In all Nordic countries, the incumbent has market share of between 40 and 60 percent in all Nordic countries, cf. Figure 4.4. However, in some countries the incumbent has started to buy up independent service providers with sizeable market shares, thus increasing the market share controlled by the incumbent. Figure 4.4: Mobile voice, competition between and within networks, markets shares for networks and competitors, 2003,
Note: Market shares within networks are measured on the basis of total outgoing calls in minutes. Market shares in Finland are for the Helsinki area where Elisa Corporation and TeliaSonera Finland are the two dominating market participants. At this stage, mobile and fixed voice services seem to be complementary services rather than substitutes. For this specific reason, we do not find that joint ownership of fixed and mobile networks needs to be a significant impediment to competition between networks in either of the two markets. However, Nordic competition authorities should be aware of the incumbent's potential for discriminatory treatment of competitors in the market for bundled products covering both mobile and fixed services and should follow the market closely. The new network technologiesWe now briefly discuss the key properties of new network technologies that may play a more significant role for the provision of telecommunication services in the future: Cable-TV, optical fibre, power line communications (PLC) and fixed wireless access (FWA). Cable-TV However, the cable network technology is mature and the technological challenges are modest, cf. Box 4.1. Currently, cable networks can handle up to 200 Mbps, but the capacity is expected to increase as TV transmission is digitalized. Box 4.1: Cable network
Source: Ovum, Barriers to competition in the supply of electronic communications networks and services, November 2003 and Marilyn Kemper Littman, Building Broadband Networks, 2002, CRC Press Cable network already has a significant coverage and customer base in all Nordic countries, more than 60 percent of the population in Finland, Denmark and Sweden and slightly lower in Norway. On the Nordic market for broadband access, cable networks have market shares in the range between 20 and 30 percent. However, the market share has been declining recently as access through digital subscriber lines (DSL) on the PSTN network has exploded. The costs of gaining access to a cable network are modest and comparable to the costs of broadband access on the PSTN network. Likewise, the costs of switching between cable and PSTN networks seem to be modest. Independent service provision on cable TV networks is technically more challenging than service provision on the PSTN networks, as consumers do not have a dedicated line each, but must share a common capacity. The key problem with the cable networks is the fact that the incumbent owner(s) of the PSTN network also owns the largest cable network in all Nordic countries, except Sweden. In Denmark, TDC owns TDC Kabel TV with 60 percent market share; In Norway, Telenor owns Canal Digital Kabel TV with 55 percent market share; in Finland, the (many) incumbent operators own all providers of cable TV, except in Helsinki. In contrast, Telia had to divest its cable network in Sweden when they merged with Finnish Sonera to create TeliaSonera Sweden. We find that the positive impact on competition between networks will be significantly muted as long as the incumbent owners of the PSTN networks continue to have large ownership shares in cable network. Optical fibre networks Optical fibre networks are a promising alternative to broadband access and have already gained some coverage, especially in Sweden. Optical fibre networks use mature and wellknown technologies, and the capacity is massive and measured in Giga-, not Megabytes, cf. Box 4.2. Box 4.2: Optical fibre
Source: Ovum, Barriers to competition in the supply of electronic communications networks and services, November 2003 and Marilyn Kemper Littman, Building Broadband Networks, 2002, CRC Press Information about the number of subscriptions on optical fibre networks is hard to find and should not be directly compared to each other. Nevertheless, there is no doubt that fibre access is much more developed in Sweden than in the other Nordic countries. There are about 140,000 fibre users in Sweden, in Denmark and Finland about 30,000 users and In Norway about 14,000 users. These figures correspond to market shares below 5 percent. The number of fibre users in Iceland is not known. Optical fibre networks suffer from significant installation costs that have to be justified by equivalent traffic volumes. A recent study57 by OVUM calculated the average cost per subscription line of different network technologies for broadband access and found that the costs of access on optical fibre networks are eight times larger than the similar costs of DSL- access on the PSTN-network. This level will probably restrict the use of optical fibre networks to installations with very high traffic volumes, for example companies or housing associations. All incumbent operators in the Nordic countries are active in the development of optical fibre networks, but we are not aware of their ownership share of optical fibre networks. However, the same conclusion holds as in the previous section. In order to create or maintain competition between networks, the incumbent owner of the PSTN network should not be allowed to acquire significant ownership shares in optical fibre networks. Power line communications (PLC) However, the technological feasibility is still in doubt. Noise and interference in the un-shielded electricity wires is still significant and it is not clear whether these interference problems will be solved in the near future, cf. Box 4.3. The commercial viability can also be questioned, but the on-going liberalisation of electricity sector distribution companies in all Nordic countries render it probable that the commercial viability of the Power Line Communication will be tested. PLC networks have been tested for a number of years in different countries. However, the actual take-up of the technology is still very low. Norway has about 300 users; Finland about 600 users and Iceland about 700. In Denmark and Sweden public figures are not available. To our knowledge none of the incumbent operator(s) are currently engaged in testing or operating PLC networks. As in the previous sections, it is essential that the ownership of a future PLC network remains separate from the ownership of the PSTN network. Box 4.3: Power Line Communication
Source: Ovum, Barriers to competition in the supply of electronic communications networks and services, November 2003 and Marilyn Kemper Littman, Building Broadband Networks, 2002, CRC Press Fixed wireless access The term "Fixed" refers to the fact that users can only receive signals from stationary positions, FWA networks typically have an omni-directional antenna located at the service provider and a directional antenna at the end-users' premises pointing towards the omni-directional antenna cf. Box 4.4. The typical distance between the two antennas is 5km. All the Nordic countries have issued and distributed licenses for FWA deployment, but the roll- up is still very limited. Denmark has about 2,000 users, Sweden about 3,500, and Norway about 7,500 users. We have no figures for Finland. FWA networks are quite expensive to implement and the commercial viability of grand-scale FWA-networks has been questioned. A recent study by OVUM found that the costs of access on FWA networks are likely to be four to eight times larger than the costs of providing DSL- access on the PSTN network. This level will probably restrict the use of FWA-networks to installations with high traffic volumes. Box 4.4: Fixed Wireless Access
Source: Ovum, Barriers to competition in the supply of electronic communications networks and services, November 2003 and Marilyn Kemper Littman, Building Broadband Networks, 2002, CRC Press Promotion of competition between and within networksWe now turn to an examination of how to maintain and stimulate competition between and within networks in the telecom industry. Our starting point is that competition between and within networks is complementary and that both types of competition have an important role on the telecommunications market and both should be encouraged. However, we tend, generally, to rate competition between networks higher than competition within networks, as the former but not the latter promote a competitive pressure on the entire value chain and require less regulatory intervention, thus lowering the administrative burden on both regulator and industry. This is an important consideration, because under some circumstances promotion of competition within networks may bring forward more competition in the short run, but may also – by lowering incentives for investment - hold back competition between networks reducing competition in the long run. Thus, there may be a trade-off between gains from competition in the short and in the long run. It should be noted that competition between networks is harder to bring about than competition within networks due to higher investment needs and risks. Clearly, competition between networks is only superior to competition within networks if competition between networks is feasible. We now consider the case, where alternative technological network solutions already exist and discuss under which circumstances competition will prevail. First, from a competition perspective it is imperative to limit horizontal integration, mergers or cooperative agreements involving two or several networks, especially if the merger involves a dominant player in the market. A merger between owners of two potentially competing network technologies, in particular with large market shares, may significantly reduce the incentive to let the networks compete because the success of one network, and higher market shares, may only be achieved at the expense of lower market shares for the other network. Vertical integration, mergers down the value chain, may be a cause of worry or not. It depends on the circumstances. If competition is alive on all levels of the value chain, vertical integration may bring about significant benefits by coordinating operations and investments across the value chain. However, if some of the markets are concentrated, vertical integration may be a tool to leverage market power from one level of the value chain to another. In this latter case, vertical integration should cause worry at competition authorities. See also Chapter 6 For this reason, we find that there is a strong case for divesture of companies controlling two or more alternative network technologies, in particular if the company has a dominant position on the market. We also find a strong case for promoting policies that limit the possibility for companies controlling one network technology to gain control over significant shares of alternative, potentially, competing network technologies. One example of such policy was the deliberate decision by the Danish government not to cede a FWA-license to the incumbent operator of the PSTN network. Second, the mere existence of alternative networks with separate ownership is, however, not sufficient to guarantee that consumers have real choice between networks and that competition prevails. It also requires that the costs of switching between two alternative networks, for example cable network and PSTN network are such that a certain number of consumers views switching supplier of broadband access as a realistic and desirable option when faced with a jump in access prices. By costs of switching we mean all cost incurred by the consumer irrespective of being financial or administrative. They can, among others, include switching fees, the cost of new converters or modems, installation costs and waiting time. Accordingly, it is important to assure a minimum amount of technical compatibility between access to alternative networks and transparent, low-cost rules for switching. If not, lock-in will destroy competition once customers have chosen their network. Third, regulation of access to the incumbent's network – competition within the fixed network – must be regulated such that it does not destroy the incentives for investing and rolling out alternative networks. The problem is that if service providers are given access to the incumbents fixed network (or any other network) on very favourable terms, it becomes commercially more attractive to be an independent service provider and less attractive to be an investor in an alternative network. It makes service provision within the existing network more attractive than the provision of new networks. It stimulates entry and competition within network today. It drives down prices and margins and it deters investment in new networks and reduces competition between networks tomorrow. In contrast, if conditions for network access become very unattractive, it makes new network provision more profitable than service provision within the existing network, deters competition within network today, but also promotes investment in new networks and increases competition between networks tomorrow. Thus, the choice of access regulation to the current network carries significant implications, not only for the state of competition within networks today, but also for the state of competition between networks tomorrow. Optimal access regulation is, basically, an exercise in the efficient monopoly pricing on which an extensive literature exists, at least from a static perspective. The dynamic perspective (that is, access pricing over time) is much less understood (Valletti, 2002). We shall develop some basic principles here, but also refer to Appendix 1: One way access. Efficient price regulation requires three steps:
The economic literature gives a rather precise description of how the efficient access price is determined as the sum of three elements:
The actual size of the three elements depends on a number of firm and market specific conditions. In principle, therefore, the calculation of efficient prices requires detailed knowledge of the circumstances concerning products, firms and markets. Many calculation methods exist. There is a fundamental difference between methods that takes as their starting point the costs of the firms in question and those that start from actual prices. Among cost-based methods, a distinction is made between methods using actual costs and methods that aim to establish minimum required costs. And, finally, the cost base may be backward-looking (historic costs) or forward-looking (projected costs). Cost-based methods (Cost-plus methods) lead to prices that in principle cover the incumbent's costs. If calculations are based on average costs – rather than marginal costs – then the incumbent's fixed costs are covered also. Calculations based on actual costs guarantee, by definition, cost coverage, whereas calculations based on efficient costs involve cost coverage only if the incumbent is as efficient as assumed. Cost-based methods do not, however, take into account the incumbent's opportunity value of displaced use, and may therefore lead to sub-optimal prices. Clearly, if the incumbent could make better use of the same resource, it pays for society to let the access prices rise to reflect the true costs to society. Price-based methods (Price-minus methods) lead in principle to full cost coverage and also compensate the incumbent for any lost profits of providing access. Basing calculations on the incumbent's prices in the market for consumer services, implies that the method also (implicitly) takes into account possible demand effects. However, it may be that actual profit margins overestimate the incumbent's actual loss from providing access; if so, the method leads to inefficiently high access prices. This could be the case if the incumbent possesses market power and is able to charge retail market prices with a mark-up. Having determined the best method for calculation the efficient prices, the prices must be calculated. There are basically two choices, depending on what is conceived to be more problematic: too high or too low prices. Low prices give new competitors a competitive advantage. This stimulates competition and leads to a pressure on consumer prices. Strict regulation will, on the other hand, give the incumbent incentives to limit access by other means. It may also induce less efficient competitors to gain market share at the expense of the incumbent. Low access prices also do not stimulate investment in infrastructure; the incumbent will refrain from undertaking new investment – especially if it is risky – while other firms will be encouraged to utilise those investments that have already been undertaken by the incumbent rather than invest themselves. Methods that disentangle prices from underlying costs provide the incumbent with incentives for cost efficiency. Such regulation also reduces incentives for shifting costs between activities and to engage in cross subsidisation. They furthermore allow the regulated firm to adjust prices according to demand conditions. On the other hand, prices that are not directly related to costs expose the regulated firm to risk. Such risk may undermine incentives for investment and for maintaining and improving product quality. Methods that are based on simple, verifiable rules, and which delegate decisions to the regulated firms themselves, involve relatively light regulatory burdens and low regulator costs for the regulated firms. By restricting the regulator's ability to exercise discretion, one also limits the danger of regulatory capture. On the other hand, such a regulatory regime may leave the regulated firms too much leeway to adjust their activities in ways that undermine competition. It may also lead to high risks of regulatory errors and inadequacies. Having calculated the efficient prices, the question remains how these prices should be implemented. There are at least three main models: caps on rate of return, price and income, respectively. The three mechanisms may be distinguished by the incentives to which they give rise. Under rate-of-return regulation (given that the rate is set sufficiently high) the regulated firm has strong incentives to invest and maintain and develop product quality; however, rate-of-return regulation provides poor incentives for cost efficiency. Price-cap regulation has the opposite effects: incentives for investment and quality improvements are poor, while incentives for cost efficiency are correspondingly strong. Revenue-cap regulation has similar effects to price-cap regulation, but in addition provides firms with incentives to restrict supply. Nevertheless, the differences should not be exaggerated. Even though price caps are not directly related to costs over the period in which they remain fixed, costs will play a role when revisions take place. In practice, therefore, price caps (as well as revenue caps) also have elements of cost-plus regulation. Furthermore, the cap will normally be combined with other regulations. Rate-of-return regulation is typically combined with cost control, to ensure that only those cost increases that are deemed reasonable are allowed to be passed through. Price caps (and revenue caps) are combined with quality requirements and the possibility of passing through costs increases that are out of the regulated firm's control. The choice of implementation mechanism is therefore to some extent a matter of taste. If the cost structure is relatively easily observed, and it is important to provide incentive for investment and product development, then rate-of-return regulation is preferable. If the major concern is to provide stimulus for increased efficiency, while quality may be regulated directly, then a price cap is more natural. In other cases, it may be necessary to combine elements from various mechanisms, so as to control costs, prices, quality and rate of return. Fourth, we can combine the above insight with our previous analysis of the potential for competition between networks on the three markets under consideration. First, we can summarize our insight as follows: If we want to stimulate competition within networks, choose access regulation giving rise to low access prices, for example methods based on the cost-plus principle. If we want to stimulate competition between networks, choose methods giving rise to high access prices, for example methods based on the price-minus principle. Secondly, we have previously argued that the potential for creating competition between networks in fixed voice is limited, while it is much more promising in mobile voice and fixed broadband. A straightforward combination of the two lines of arguments would require regulators to go for light access regulation, that is, relatively high access prices in mobile voice and fixed broadband, but tough access regulation, that is, relatively low access prices in the fixed voice market. Dynamic access pricing has been proposed to achieve a compromise between stimulating competition within and between networks. Canoy et al. (2003) describe dynamic access pricing as follows:
Regulators in Canada and the Netherlands have implemented dynamic access prices for unbundled local loops. Initially, the price of unbundled local loops is set at a level appreciably below cost but then increased step by step over a longer period until it reaches a cost based level. The initiative is a straightforward form of entry assistance that promotes the investments that users of local loop unbundling must make. However, it does so in a manner which makes it clear that prices will rise. As a result, there should be no significant danger that such an offer will encourage inefficient entry. In general, it has been argued that dynamic access pricing could be a bridge allowing companies to gather experience, customer basis and capital as independent service providers on the PSTN network before transforming themselves into full blown network investors and owners (OPTA, 2003). They distinguish between investments in known and already existing networks, e.g. PSTN, and investments in new networks, e.g. PLC. In the former case, the incumbent typically has a strong first mover advantage making entrant's investments in complete new networks rather unlikely. Instead complementary investments in local loop network and DSLAMs should be promoted by initially setting low tariffs for access to the unbundled local loop. This is in line with our earlier reasoning. The rate could be increased over time reaching cost-based levels when the market is more mature. In the latter case, the incumbent has not necessarily a first mover advantage making entrant's investments in new networks more likely. The costs and risk of rolling out new networks are so high, that high tariffs, cost-based tariffs with a mark up, is a necessary requirement for investments to take place. However, if the incumbent with its financial strength and easy customer access decides to invest in the new network technology it may in any case deter new entrants. Also for this reason it may be important to limit the incumbents' possibility to own or control new network technologies that could be a source of potential competition in the future. Alternatively, shared ownership on new technologies could be promoted, even though it creates problems of its own. It is still too early to gauge whether dynamic access pricing is effective in promoting more entry than would otherwise occur. Empirical evidence from the US seems to indicate that low access prices are not inconsistent with high investment levels (Cave et al., 2001). However, it is not clear whether these investments are complementary or substitutes. Empirical evidence from Europe is very limited. Cave et al. reports some bi-variate results that has the same flavour as the US results, but both the empirical method employed and the quality of data call for caution. However, the dynamic access pricing seems to be a clever compromise between conflicting goals and as such seems worthwhile investigating further. Footnotes52 Various degrees of unbundling are available: Accounting, legal, management and ownership unbundling. Accounting unbundling is the least radical type of unbundling, ownership the most radical. 53 We use the term "market" in a broad sense covering both retail and wholesale aspects and not as a synonym for the relevant market. 54 Broadband access requires an access line with at least 128 kb/s. Accordingly, access using modem on the PSTN network is not part of the broadband market 55 Measured by the share of outgoing international calls, competition seems to be even more intense as the incumbent's market share is lower, about 50 percent. 56 The lower market share for the incumbent in Denmark is not due to the existence of alternative networks, but to Wholesale Line Rentals, that is, service providers' pure resale of subscription services rented from the incumbent. The same possibility exists in Norway. 57 European Commission, 2003, Barriers to competition in the supply of electronic communications networks and services, Report prepared by OVUM. Version 1.0 October 2004 • © Danish Competition Authority. |
