Prepared for the Competition Commission by James Hodge & Nicolas Theopold School of Economics, University of Cape Town March 2001
Table of Contents
INTRODUCTION................................................................................................................................................. 1 1. THE NATURE OF THE TELECOMMUNICATIONS INDUSTRY .......................................................... 2 1.1 THE TELECOMMUNICATIONS MARKET ........................................................................................................... 2 1.1.1 Provision of Network Infrastructure..................................................................................................... 3 1.1.2 Provision of Services ............................................................................................................................ 9 1.1.3 Features of Networks.......................................................................................................................... 10 1.2 FORMS OF INTERNATIONAL TRADE .............................................................................................................. 11 1.2.1 Provision of network infrastructure.................................................................................................... 11 1.2.2 Provision of services........................................................................................................................... 11 1.3 ROLE IN THE ECONOMY ............................................................................................................................... 12 1.3.1 Role as Intermediate Input ................................................................................................................. 12 1.3.2 Demand for Intermediate Inputs......................................................................................................... 13 1.3.3 Medium for Content Providers ........................................................................................................... 14 1.3.4 Household Expenditure Item .............................................................................................................. 15 2. INTERNATIONAL REFORM OF TELECOMMUNICATIONS REGULATION................................. 17 2.1 THE RATIONALE FOR REGULATION ............................................................................................................. 17 2.1.1 Normative Theory of Regulation ........................................................................................................ 17 2.1.2 Positive Theories for Regulation ........................................................................................................ 18 2.2 DIFFERENT TYPES OF REGULATION ............................................................................................................ 19 2.2.1 Regulation under Perfect Information................................................................................................ 19 2.2.2 Regulation under Asymmetric Information......................................................................................... 19 2.3 INTERNATIONAL TRENDS IN TELECOMM REGULATION ............................................................................... 20 2.3.1 The move to incentive Regulation....................................................................................................... 20 2.3.2 Liberalising Network Industries: a Three Stage Process ................................................................... 21 2.4 CONVERGENCE IN TELECOMMUNICATIONS ................................................................................................. 25 3. CURRENT REGULATORY FRAMEWORK............................................................................................. 26 3.1 BACKGROUND AND OBJECTIVES OF CURRENT REGULATION ...................................................................... 26 3.2 INDEPENDENT REGULATORY BODY SATRA TO ICASA ......................................................................... 27 3.3 FIXED LINE NETWORKS ............................................................................................................................... 29 3.3.1 Local Access ....................................................................................................................................... 29 3.3.2 Long-distance ..................................................................................................................................... 31 3.3.3 Value-added Network Services (VANS).............................................................................................. 31 3.3.4 Private Networks (voice) .................................................................................................................... 32 3.4 MOBILE COMMUNICATIONS......................................................................................................................... 32 3.4.1 Cellular............................................................................................................................................... 32 3.4.2 Satellite ............................................................................................................................................... 33 3.5 INTERCONNECTION, FACILITIES LEASING AND NUMBERING ....................................................................... 33 3.6 PUBLIC INTEREST CONSIDERATIONS ........................................................................................................... 35 3.6.1 Universal Service................................................................................................................................ 35 3.6.2 Human Resource Development........................................................................................................... 35 4. MARKET STRUCTURE AND CONTESTABILITY ANALYSIS............................................................ 36 4.1 PROVISION OF CUSTOMER PREMISES EQUIPMENT (CPE) ............................................................................ 36 4.1.1 Relevant Markets ................................................................................................................................ 36 4.1.2 Market Structure................................................................................................................................. 36 4.1.3 Contestability...................................................................................................................................... 37 4.2 FIXED LOCAL ACCESS ................................................................................................................................. 37 4.2.1 Relevant Market.................................................................................................................................. 37 4.2.2 Market structure ................................................................................................................................. 39 4.2.3 Contestability...................................................................................................................................... 39 4.3 LONG DISTANCE ......................................................................................................................................... 42 4.3.1 Relevant Market.................................................................................................................................. 42 4.3.2 Market structure ................................................................................................................................. 42
i
Competition and Regulation in Telecommunications 4.3.3 Contestability...................................................................................................................................... 43 4.4 MOBILE ....................................................................................................................................................... 45 4.4.1 Relevant Market.................................................................................................................................. 45 4.4.2 Market structure ................................................................................................................................. 45 4.4.3 Contestability...................................................................................................................................... 45 4.5 VALUE-ADDED NETWORK SERVICES (VANS) ............................................................................................ 47 4.5.1 Relevant markets ................................................................................................................................ 47 4.5.2 Market structure ................................................................................................................................. 47 4.5.3 Contestability...................................................................................................................................... 48 4.6 OTHER MOBILE SERVICES ........................................................................................................................... 48 4.6.1 Relevant Markets ................................................................................................................................ 48 4.6.2 Market Structure................................................................................................................................. 48 4.6.3 Contestability...................................................................................................................................... 49 5. REGULATORY ISSUES FOR THE SA TELECOMMUNICATIONS INDUSTRY AND THEIR POTENTIAL EFFICIENCY/PUBLIC INTEREST IMPACT ....................................................................... 50 5.1 STRUCTURE ................................................................................................................................................. 50 5.1.1 Vertical separation of networks.......................................................................................................... 51 5.1.2 Vertical separation of services from networks ................................................................................... 52 5.1.3 Horizontal separation of mobile services ........................................................................................... 52 5.2 ENTRY......................................................................................................................................................... 53 5.2.1 To Restrict Entry or Not ..................................................................................................................... 53 5.2.2 Dealing with Entry Deterrence........................................................................................................... 55 5.2.3 To Actively Assist Entry or Not........................................................................................................... 55 5.2.4 Asymmetric regulation........................................................................................................................ 56 5.3 PRICING....................................................................................................................................................... 56 5.3.1 Final consumer prices ........................................................................................................................ 56 5.3.2 Interconnection (Access) pricing........................................................................................................ 57 5.4 QUALITY ..................................................................................................................................................... 58 CONCLUDING REMARKS.............................................................................................................................. 59 BIBLIOGRAPHY ............................................................................................................................................... 60
ii
List of Tables
Table 1.1: A breakdown of the telecommunications market _________________________________________ 2 Table 1.2: Cost characteristics of local access technologies for telecommunications _____________________ 6 Table 1.3: Direct forward linkages - Telecommnuications share of intermediate input costs for sectors of the economy (1997) percentage of total intermediate input costs _____________________________________ 13 Table 1.4: Structure of capital requirements for a fixed line network provider and an Internet service provider (2000) _________________________________________________________________________________ 13 Table 1.5: Direct backward linkages intermediate inputs into the communications industry (1997) percentage of total intermediate input costs ____________________________________________________ 14 Table 2.1: State of European Deregulation (1998)_______________________________________________ 21 Table 3.1: Rollout Targets for Telkom in terms of License_________________________________________ 30
List of Figures
Figure 1.1: A telecommunications network _____________________________________________________ 4 Figure 1.2: Household expenditure patterns (1995)______________________________________________ 15 Figure 2.2: Path of Regulatory Intensity ______________________________________________________ 22 Figure 3.1: Proposed path of deregulation_____________________________________________________ 31
iii
Introduction
The telecommunications sector in South Africa has undergone a partial deregulation in the past 5 years. It is at this point that there needs to be a decision on how to proceed with the process, while complying with liberalisation commitments for the sector in the WTO. Further, the Competition Commission recently received joint jurisdiction over a number of the publicly regulated sectors, including telecommunications. The purpose of this paper is to review the industry and its deregulation from a competition policy perspective. This involves understanding the various horizontal and vertical markets in the sector the regulation of entry & conduct, and how this regulation has evolved the level of actual and potential competition in the industry the impact of different regulatory choices on market structure, competition and performance in the industry
The paper consists of five parts. Part one examines the vertical structure of the telecommunications industry and its changing nature. Part two looks at the rationale for conduct regulation in the past and how this practice has evolved internationally. Part three describes the current regulatory practice in South Africa, including some of the public interest issues. Part four takes a hard competition analysis view of the different vertical stages of production, drawing out market structure, potential competitors, barriers to entry and the potential impact of entry. Part five then draws on the lessons of the other four chapters to put forward a position on the major regulatory issues that need to be decided.
basic voice vs. VANS vs. broadcast Fixed vs. mobile national long distance vs. international Business vs. residential
Telecommunications production can roughly be divided into the provision of network infrastructure and the provision of services on that infrastructure. Typically a public monopoly is vertically integrated and so provides all parts of the production chain and all horizontal markets. However, in the context of deregulation it is important to focus on the various stages of production and not what individual firms do. In this way it is
2
easier to identify essential facilities and anti-competitive behaviour stemming from vertical integration. It is also easier to track where competition can feasibly survive without regulation.
There are three different components to the network that can be seen as three different stages of production. These are local access, long-distance and international. Figure 1.1 below demonstrates how these three components interconnect to form a complete international telecommunications network. It looks at a mixed fixed line/mobile example. The local access network connects the customer premises to the local switch through the local loop. From the local switch there is an interoffice transmission facility to either other local switches or long-distance (national or international) points-of-presence. The long-distance networks then transmit the signal to another long-distance point-of-presence where it is distributed to a local switch and onto the other customer premises. What follows is a more detailed discussion of each component of the supply chain and the various technological platforms used.
FIXED LINE
Customer premises
Customer premises
Local Loop
Local Switch
MOBILE
Customer handset
Base Station
Customer handset
Base Station
Customer premises equipment (CPE) The concern here is the retail supply of customer premises equipment and not the manufacture of the items. CPEs include fixed line telephones, mobile handsets, and PABXs (private exchange equipment for business use). Each of these can be considered a separate market. In the heyday of public monopoly, the customer premises equipment had to be supplied by the incumbent PSTN. This monopoly position, on what is essentially a service with no natural monopoly features, represented an opportunity for abnormal profits to be made. Almost without fail the first step in any deregulation process is to liberalise this part of the supply chain. Local Access Local access can now be provided by a number of different technologies, each with different cost structure and therefore different degrees of substitutability with the traditional public switched telephone network (PSTN). The focus below will be on fixed line (PSTN, data networks, cable TV) vs. wireless (fixed or mobile) options. Although local access is being handled as one stage of production, it should be noted that this stage can be broken down further. In particular, connection of premises to local loop local loop local exchange (switch)
In the search for greater competition at the local access level, it may be desirable to further unbundle the production process and find ways to inject competition. For instance, the SA White Paper on telecommunications suggests allowing community groups and SMMEs putting in local loops that then connect to the Telkom PSTN. In the USA, Competitive Access Providers (CAPs) have emerged to connect large customers directly to the long distance networks points-of-presence (PoPs),
4
effectively bypassing the PSTN (these access lines are point-to-point and have no switching capacity) (Laffont & Tirole 2000). Fixed Line Within the group of network providers that use fixed line infrastructure, networks were historically built to focus on one of three different types of transmission voice, data or image. However, increasingly networks are offering all three. This is feasible because almost all networks now use a digital format. In voice, the traditional PSTN consists mainly of copper or coaxial cable transmission equipment at the local access level and switching equipment that allows two-way transmission between two individual points on the network by establishing a dedicated line between two points for the duration of a call. The transmission mediums have low capacity or network speed (measured in bits per second) making them inadequate for video transmission and a slow but adequate medium for data transmission. However, recent technological developments in data compression have substantially increased the speeds available on these wires (in particular the asymmetrical digital subscriber line systems ADSL and other DSL technologies). Upgrading the networks to provide greater speed does require additional investments in the local access infrastructure. An alternative is the use of fibre optic, which has a far greater network capacity. Due to the higher cost of fibre optic, it is only cost effective with large businesses and not residential homes in the local access component of the network. It is used extensively in the long-distance and international networks. Either way, it is increasingly possible for PSTN networks to be upgraded to offer video services too. In fact, growing use of capacity for data transmission (i.e. the Internet and intra-corporate transfers) make the investments in fibre optic or ADSL increasingly worthwhile, offering the springboard to broadcasting. Cable TV (CATV) networks (video networks) use a coaxial cable with broadcast quality network speeds as their transmission medium. Initially these networks were only focused on broadcast and so were established with one-way, one-to-many transmission (tree-and-branch structure), disqualifying one-to-one communication. However, where cable operators have entered telephony, they have upgraded their networks to handle two-way transmission. This has enabled the creation of the new pay-per-view television services, and also allowed them to actively enter the Internet providers and telephony market. Dedicated data networks may be private or public (Internet). Data can be run over any network and in fact, most local access to dedicated data networks occurs through either the PSTN, the cable TV network or a mobile service. As such, the discussion of data networks is largely a long-distance issue. However, larger businesses are increasingly being connected directly to these long-distance data networks making it worth a mention here. The switching equipment of data networks does not establish a dedicated line and instead routes packets of data on common usage lines from one point to another through the path of least resistance. This is efficient for data but as line traffic congestion can delay packets, this has previously been unsuitable for quality voice or image transmission. However, recent International Telecommunications Union (ITU) agreements on a common gateway protocol for passing multimedia from telephone networks to data networks should greatly improve the quality in the near future (ITU 1999). This will see priority being given to voice and image packets, avoiding the problem of congestion.
Table 1.2 examines the cost structure of different network types. As it reveals, the cost structure in fixed line networks (PSTN or cable) is geared towards sunk and fixed costs making it largely insensitive to traffic volumes. For this reason it is often noted that the marginal cost of making a phone call is negligible. In fact, the marginal cost/revenue decision for the PSTN operator is whether to connect a new customer or not (Laffont & Tirole 2000). If the new customer is expected to generate more revenue than it costs to install the line, then the PSTN will install the line. The result of this cost structure is that economies of scale and density are reasonably large and there is a large initial investment required to begin operating local access. The sunk costs are largely the trenches and ducts for wire that can make up to 40% of annualised costs (Caves 1995). Allowing the distribution through poles above ground does lower costs. These sunk costs are the reason that lower cost entrants are companies with a local transmission network already in place e.g. cable TV companies or electricity companies. In these cases there are economies of scope as many of the additional costs of rolling out a network are already incurred. The difficulty of course is allocating these costs between the different services. The other service that cable can enter once upgraded for telephony is data transmission. Cave (1995) notes that in the comparison of costs between cable and PSTN, the following conclusions may be drawn: Cost per home passed (the local loop) is higher for a CATV network The incremental cost of provision of a telecommunications service by a CATV operator with an established local loop is a small fraction of the total cost faced by a PSTN For the subscriber already connected to the CATV local loop the cost disparity with a telecommunications provider is even greater.
Table 1.2 assumes that the cable provider already has a local loop in existence and is providing both video and telephony (hence exploiting their economies of scope). The conclusion one can draw is that cable can in all likelihood be considered in the same market for telecommunications once established.
Table 1.2: Cost characteristics of local access technologies for telecommunications
Fixed line Fixed/variable ratio Sunk/salvageable ratio Traffic sensitive/non-traffic sensitive Economies of scale Economies of density Economies of scope Initial investment required Requires special handset High High Low CATV network Medium Medium* Medium* Fixed wireless Low Low High Mobile wireless Low Low High
If the PSTN were to exploit their potential economies of scope into video then part of the sunk costs allocated to telephony would be shared with other services. The PSTN has already lowered some of the costs by exploiting its ability to provide access to two-way transmission data networks (i.e. Internet or private data networks). The exponential growth in traffic volume has been associated with data and not voice. This enables PSTNs to operate at greater levels of capacity utilisation and so lower the unit costs. This is a strong reason for the better productivity performance of US telephone companies. Wireless Networks Wireless networks differ from fixed line in their use of the radio frequency spectrum for transmission. The local access process involves a handset for the subscriber transmitting to and from a base station using a specific spectrum that the network provider is licensed to use. The base stations are usually connected to each other or another network through a fixed line infrastructure. In local access there are two types of wireless networks fixed and mobile. The fixed wireless local loop is a recent addition and is being used to provide a last drop to the consumer for fixed line voice or data networks. It is similar to two-way radio where the physical coverage is very limited, and the receiving device (a telephone) is often fixed in location. The same receiving device as fixed lines is used. In terms of a market boundary for competition analysis, it is not designed to compete with the cellular networks but rather to provide a cheap alternative to using fixed wire as the last drop to the home. Therefore it would usually fall into the fixed telephony market (in fact it is usually rolled out by the PSTN as part of a mixed technology strategy Telkom is no exception). The cellular networks provide local access but also the added advantage of mobility. The subscriber is required to invest in a handset that cannot be used for fixed line or fixed wireless access (a switching cost). The current mobile networks are constrained in their network speed to offering voice and data services only. However, the socalled 3rd Generation or Universal Mobile Telecommunications Services (UMTS) hope to achieve networks speeds that would enable the transmission of video too. The first networks of this kind are due to be operational in Japan in 2002. The cost structure of mobile is different to that of fixed lines (see table 1.2). The spacing of the base stations (and therefore the number required) is dependent on the traffic volumes. The result is that the initial investment required to establish a local access infrastructure is lower than that of fixed lines. It also means that there are lower economies of scale and density, making more network providers viable. A key question is the extent to which mobile competes with fixed line in the local access market (i.e. is a close enough substitute to be considered in the same market). This is crucial for competition and regulatory analysis. From a use and quality of product perspective, they clearly offer the same potential group of basic and advanced voice products but with mobile having the added feature of mobility. What may therefore determine whether they are in the same product market is the difference in price between the two or whether other services are bundled in too (in particular data services). If voice is bundled with data services then fixed line currently holds a broader array of products to mobile, where data is available but in a far more limited way due to screen size. The actual cost differential depends on the
7
relative density of subscribers. Wireless is cheaper at lower densities due to the low fixed costs, while fixed line is cheaper at higher densities. The cross-over point is estimated to be in the range of 200-800 subscribers per km2 from US and UK studies (Cave 1995). This makes mobile more expensive in urban and metropolitan areas with high subscriber rates. However, in many developing countries with low telephony density of use, cellular is proving to be a good substitute for fixed line (South Africa may well be a case in point). Rapid technological developments in mobile communications could make it even more substitutable. Mobile telecommunications also includes satellite personal communications services. Satellite makes use of more powerful devices to transmit to one of a number of earth stations, which in turn link to each other via one or more satellites. The greater distances the receiving equipment must transmit over, means they are larger and more expensive than cellular. However, the use of satellites enables the network to minimise the number of earth-based transmission stations. This will not be discussed because a) the first global start-ups have all failed, and b) the cost is so much higher than fixed line or mobile that it cannot be considered remotely in the same market. Other wireless communications also include the broadcasting group (radio, free-to-air TV and Pay TV) which are one-to-many operations without two-way capacity and so cannot offer telecommunications. Finally, one can include two-way radio and radio trunking. Two-way radio is often transmitted over very short distances and includes networks for emergency services and communication within a defined corporate space. Often the distances are short enough that no transmission stations are required, and the hand-held devices are sufficient. The longer distances are covered by radio trunking, which operate in a fashion similar to cellular networks. Neither are suitable for public local access telecommunications and so will not be discussed further. Long Distance (National and International) Local access networks connect to a long distance network through a point-ofpresence (PoP). The long-distance network is made up of these PoP exchanges and a transmission network. Given that these networks draw on a large pool of local access subscribers, they are able to get greater density of use in their transmission networks. As a result, most long distance networks use fibre optic when using a fixed line solution. Alternative transmission mechanisms include satellite and microwave (national only). Despite the choice of technologies available, a single market is defined as one for national long-distance and for international long distance. In addition, as most of these networks have sufficient network speed for broadcast quality, there is no real need to differentiate at the infrastructure stage between voice, data and video. As with local access, there are economies of scope in rolling out fibre optic longdistance networks. A substantial proportion of the costs are in paying royalties to landowners to pass the transmission cable through and putting in place infrastructure to carry the transmission medium (Cave 1995). If a new entrant already has a national infrastructure (e.g. railways, electricity grid), then some of these costs will be spared allowing a faster and cheaper rollout.
linking to a server farm that would include a router, authentication server, firewall, mail hosts, proxy servers and local content servers. Some form of customer management systems this would include customer information, billing and customer support (call centre) Content some advanced services may have content that will be received by the customer (e.g. Internet, video-on-demand)
A local example of separation of network and service provider is the cellular industry in South Africa. The network providers do not retail to the public but instead wholesale network access to a group of approved service providers. These providers in turn offer retail outlets to access customers, they stock and sell the phones, do the credit checks, link the customer to the network and perform all billing and debt collection. Horizontal market divisions in the service component can be taken from the divisions in the network markets and include: Local vs long distance vs international standard voice products Business vs residential the business market has greater capacity usage and concentration changing the cost and product package provided. Data vs. voice vs. video separation of voice products from Internet and other VANS from broadcast type products. Fixed vs mobile this would apply across all of the other divisions e.g. mobile Internet may be separated from fixed line by the more limited range of applications it supports (for now).
A key question when examining the horizontal market divisions is product bundling and technological change. In terms of product bundling, most voice services offer the more advanced features as part of the service. Are these additional products or are they part of a basic voice service as most companies provide them? In terms of technological change, it needs to be recognised that product boundaries will continually change and so need to be reassessed periodically.
10
callback services have sprung up in countries like the USA whereby a customer from South Africa places a call as if they were in the USA through the callback operator calling them and connecting them to the number they wanted to call. Finally, in mobile telephony one can arrange international roaming agreements for mobile phone users to capture their consumption. Any calls that they make abroad will be routed back through the domestic network, allowing the domestic mobile network provider to capture some of the charges on that call. However, for basic voice products in the local access market, the competition must come from domestically situated firms. As such, provisions around foreign ownership limits will serve to restrict competition.
12
Table 1.3: Direct forward linkages - Telecommnuications share of intermediate input costs for sectors of the economy (1997) percentage of total intermediate input costs
Sector Primary Manufacturing Natural resource intensive Labour intensive Scale intensive Knowledge intensive Other Services Producer services Consumer services Infrastructural services Community/social services % of intermediate costs 0.2 0.4 0.3 0.5 0.3 0.6 0.7 4.4 7.2 3.4 0.7 11.8
However, there are larger multipliers for ongoing operations. Table 1.5 below presents the first-round direct and indirect output multiplier effects of a R1 increase in network provision1. The multiplier effects are negligible for the primary sectors but
1
Data limitations prevent putting together a multiplier analysis for service providers.
13
significant for the manufacturing sectors - metal products and equipment (transmission and receiving devices); chemicals, rubber and plastic products (plastic casings for equipment); paper and printing (marketing and office materials). The largest gains are in other service industries, in particular, other communication or business service providers, distribution, transport and utilities. A second round of output and employment creation can be expected stemming from additions to wage and capital income created by the first-round effects. These figures are not high compared to agriculture or manufacturing. Part of the reason for this is a low participation of SA industry in communications equipment, with much derived demand going to imports from other countries. This may change with time and raise the average multiplier effect.
Table 1.5: Direct backward linkages intermediate inputs into the communications industry (1997) percentage of total intermediate input costs
Sector Primary products Manufacturing Natural resource intensive Labour intensive Scale intensive Knowledge intensive Other Services Producer services Consumer services Infrastructural services Community/social services Total intermediates/Total input costs % of intermediate input costs 0.0 20.0 0.6 2.6 4.5 12.1 0.2 80.0 44.3 20.2 13.9 1.6 22.9
for television or education broadcast - from advertising to dramas. Finally, data content would be associated with those firms providing content for the Internet or private data networks - from online news corporations to electronic commerce. In all these industries, expansion of the network and lowering the costs of transmission expand the number of people connected to the network and therefore expand the market for content. Enhancing the speed and quality of the networks open up the opportunities for new content products to be introduced which also expands the market for content. An important question remains - how big are the multiplier effects on content providers stemming from improvements to the network and its expansion.
15
food housing income tax transport cloth/foot furniture/equip health insurance read/recr/hols drink/cigs personal care communications invest/savings education house operation pensions domestic workers fuel/power other
18.2 16.1 14.7 10.2 5 4.1 4 3.3 3.3 2.7 2.5 2.4 2.1 2 1.8 1.6 1.6 0.6 3.8
10
12
14
16
18
20
Percentage of Expenditure
Source: Statistics South Africa, Income and Expenditure Survey, 1995
16
Having identified the different types of efficiency, the ways in which market failure arises have to be identified. Here, we identify three basic types: Externalities: any transaction has a cost and a benefit to it. These costs and benefits may well differ for the individual and the society. A car driver, for example, will not take into account her effects on pollution when planning a trip. When the social costs exceed the private costs, or the private benefits exceed the social benefits, the level of activity is too high. When social benefits exceed private benefits, allocation will be lower than desired. An example for the first would be car usage, an example for the second investment by firms who do not take into account the beneficial spill-overs for other companies. Taxes, in the case of over allocation, and subsidies in the case of under allocation may be used to rectify the situation.
17
Natural Monopoly: occurs when it is more expensive to have several companies providing the same product than it is to only have one. Public utilities, such as water and electricity and most other network industries are classic examples of natural monopoly. Single product monopolists will derive their advantage from economies of scale, whereas multi-product natural monopolies may also derive their advantages from economies of scope2. Large specific investments and hold-up problems: when private activity invests large amounts in assets that may only be used for one purpose, there is a potential for the extraction of rents from a contract partner. If a pipeline company invests in a pipeline from an oil field to a refinery, this asset is only good for transportation between these two points. Once it is built, the investment is sunk. The oil field operators might be able to extract the sunk cost from the pipeline builders should they have other possibilities for the transportation of their oil. Private contracting may be costly in this case and regulation the preferred tool.
Regulation must hence be used in sectors where competition is not sustainable due to one of the three reasons mentioned above. The success of regulation is to be measured by the efficiency criteria outlined.
When the cost of producing two products together is less than producing them separately
18
monopoly has had greater costs to a group of industry users who have become harsh critics of the continued monopoly.
system is that the firm has no incentive to continually reduce cost. Inflated cost is created through the attraction of capital rather than efficiency, an incentive to inflate costs before a rate hearing and the need for very detailed regulation. Incentive regulation: has been created to deal with the inefficiencies inherent in rate of return regulation. There are two basic types of incentive regulation: o Price-cap regulation: imposition of a price ceiling on a basket of goods, adjustment of the ceiling with the formula CPI-X, where CPI is the consumer price index and X is the average increase in efficiency expected. The potential problem with this mechanism is the uncertainty about the true level of cost and hence the crudeness of the mechanism. Errors in the index may lead to windfall gains for the firms or heavy losses. Furthermore the incentive to minimise cost can lead to decreased product quality and to the elimination of goods which do not show potential for efficiency increases. o Earnings Sharing: whilst price cap regulation achieves cost efficiency, earnings sharing attempts to strike a balance between cost and allocative efficiency. There is a sharing in the risk and rewards from any deviation in the rate of return between consumers and producers. The adjusted rate of return is determined by a sharing of the deviation from the target rate of return ( ra = rt + h(r * rt ) , where 0<h<1, r* is the target rate of return, rt the rate of return at prevailing prices, and ra the return after adjustment). In rate of return regulation, h=1, in price cap regulation, h=0. The responsiveness of the system depends on the preferences of the regulator. This approach has been used in numerous cases in the USA as it is deemed to be fairer than price cap regulation while providing a good proportion of incentives to firms.
The results are robust in the way that different model specifications regarding input and output prices, as well as using slightly different data sets do not change the basic results that incentive regulation does have beneficial effects on efficiency. A further result from the modelling is that those regulated companies with the highest expenditures on research and development benefited most from the change to IR, a result that is consistent with intuition. Bergman et al (1998) note that more and more European countries have started to use price cap regulation, initiated by the United Kingdom, which introduced the measure in 1988 due to a report by Stephen Littlechild, pointing out the productive and dynamic inefficiencies inherent in ROR. Among other states, France, Germany and Sweden also introduced PCR in the early 1990s. Despite the positive effects on productive efficiency of PCR, there has been criticism of the system relating to strategic behaviour and allocative efficiency. The strategic behaviour component relates to the timing of rate reviews. To maximise the cost reduction incentives for the firms, the regulator has to increase the time between hearings, such as to allow some appropriation of the gains made by savings. This, however, increases the scope for the firm to act strategically: before adjustments, cost savings are not fully exploited, so as to signal a lower savings potential to the regulator. Once the new rates have been installed, maximum savings are being exploited as they can be benefited from for the maximum possible time. As to allocative effects, the problem is that the X-factor is backward looking rather than future oriented. The nature of negotiations between the regulator and the firms dictates that X-factors are based on historical experience rather than future potential. It is the superior knowledge managers have in comparison to regulators that makes it possible for companies to exploit extra profits. Large segments of the telecomm industry are potentially competitive, though, and it is thus the aim of European policy to liberalise these markets, as effective competition can replace regulation and ensure better market outcomes.
Ibid. p.80
21
This describes the different states that network industries were in at that moment in time. As the system evolves towards a more liberalised system, there are different demands on regulation in the different phases. The book differentiates between three phases: 1. Monopoly 2. Monopoly and Competition in some Markets 3. Competition
Figure 2.2: Path of Regulatory Intensity
Phase 2
Phase 3
Time Phase 1 The problems associated with an industry in phase 1 of liberalisation are best dealt with through the measures outlined in the previous sections: close regulation of the monopolist to ensure productive and allocative efficiency. Ramsey pricing would be the preferred measure under perfect information, but is not possible due to the impossibility to access accurate data for effective regulation. ROR has the problems associated with over investment in capital expenditure to increase the absolute level of profits allowed and the lack of incentives for cost reduction. Regulatory lag could, however, provide limited incentives for cost reduction, as the rate of return would be higher for the period between rate reviews. PCR, on the other hand leads to efficient production, but lacks on the allocative side due to information asymmetries about the true efficiency potential of the firm. Phase 2 Here, there is introduction of competition in some parts of the operations. It is possible that effective competition will never be possible in all parts of the operations, as duplication of some network components may not be viable in all fields. In telecommunications, for example, fibre-optic backbones are existent on different networks in some countries, but there hardly ever is any duplication of the last mile, i.e. the connection from the users premises to the local exchange. The viability for duplication of these parts is not given. In the second phase of liberalisation, there are three strands of activities that have to be pursued: firstly the separation of the regulator from the monopolist, secondly the regulation of incumbent pricing and thirdly the regulation of interconnection. As to the independence of the regulator, it has to be ensured that there is no capture of the regulator for any of the parties involved in the market. This is important to instil
22
confidence in any companies wishing to enter the market, providing them with the certainty that they will have a level playing field. On the front of price regulation, the monopolist has the possibility to use rents from monopoly markets to subsidise activities in the competitive markets and thus gain an unfair advantage. The importance therefore lies in analysing the level and structure of prices set by the monopolist. Network industries have a large part of their cost in fixed components, making the allocation of costs to services very difficult and arbitrary. Considerable effort thus has to be expended towards the analysis of the cost structure and to the development of a fair level of prices to be set by the incumbent in the competitive market. Deviations in both directions are potentially damaging: setting prices too low will discourage entry by firms even it they were more efficient than the incumbent. Should prices be set too high, then entrants will be able to enter even if they are less efficient than the incumbent. Furthermore, it will jeopardise the profitability of the company and thus the cross subsidies to unprofitable market segments necessary for universal service obligations. Even in the most competitive telecomm markets, such as Germany4, not all aspects of the network have been subjected to competition, such as the last mile. There is strong controversy about the correct level of interconnection pricing between the network components of Deutsche Telekom and its competitors. Once again, the incentives of the incumbent and competitors are opposed. If interconnection prices are set too high, the monopolist will receive an overly high amount for the interconnection services and thus deter entry by efficient companies. Should the prices be set too low, however, this will also attract inefficient entry. Furthermore, it will discourage investment by the incumbent into new technologies, as entrants consequently do not carry a fair share of costs and risks of the investments and benefit disproportionately from the incumbents investments. When interconnecting networks, there are different possibilities to design access prices: Efficient Component Pricing Rule (ECPR): this mechanism warrants entry as long as the entrant is at least as efficient as the incumbent. It is assumed that retail prices are fixed, i.e. that allowing the entrant access does not have an effect on the incumbents price. The cost structure is defined by the average incremental cost of access, b, the average incremental cost of providing the competitive part of the service, c, and the retail price of the service, p. The optimal access price is thus defined as: a = b + [ p - (b + c)]. In the context of a long-distance telephone context, b would be the cost of connecting the customer from her home to the local exchange and c the cost of routing the call on backbone infrastructure. The problem associated with the ECPR is the fact that allocative effects are being ignored, due to the assumptions of a homogenous fixed product, a fixed coefficient of technology and the inability of competitors to bypass the incumbent. Ramsey access pricing: in the spirit of optimal pricing for society, Ramsey access pricing also takes some allocative effects into account. The formula for the calculation of the access price changes to: a = b + (p (b + c)). is what
Purton (2000), SURVEY - FT TELECOMMS: Former monopoly loses significant market share
23
is called the displacement ratio, namely the change in incumbents retail prices as access prices vary divided by the change in the incumbents sales of access services to its rival as the access charge varies. When the three assumptions quoted above hold, then =1, otherwise <1. Hence the optimal access price is likely to be lower than under the use of the ECPR. The system once again suffers from informational limitations, since it will be very difficult for the regulator to obtain data on bypass, factor substitutability and supply and demand elasticities. The amount of intrusion in the market would most likely present a hindrance rather than supporting the notion of liberalisation. Global price caps: developed by Laffont and Tirole, these allow the firm far larger discretion about the setting of prices, as long as a global cap on access and retail prices is not exceeded. w1 * p + w2 * a p . This can achieve the same outcome as Ramsey access pricing, as long as weights are exogenous and proportional to the quantities realised under Ramsey access pricing. Despite the reduction of informational needs in comparison to Ramsey access pricing, the selection of the correct weights for global price caps remains a difficult task for the regulator. Thus far, there have been no global price caps introduced by regulators. Accounting methods: a very widely distributed form of setting access prices is that of using accounting methods for the determination: the price is set at the direct cost of providing the access plus some fair share of common costs allocated to the access price. These have been used extensively due to their familiarity and ease of implementation, but have also been heavily criticised for their inability to identify optimal access prices. The outcome from accounting access price creation will be largely dependent on information sets available to the regulator and on negotiating outcomes between incumbents, entrants and regulators.
Access pricing is one of the crucial ingredients in the second phase of liberalisation. Experiences from European countries show us that it can be very difficult to achieve agreement between the parties involved, due to the stakes riding on the pricing of access, the complexity of access arrangements in modern multi-level networks and the opposing objectives of incumbents and entrants. Phase 3 By definition, when the market has reached phase three of liberalisation, competition is sustainable. In reality, however, some network industries may never reach this stage due to natural monopoly characteristics of some parts of the network. The need for regulation decreases markedly in phase three as competition enforces pricing discipline on players and the incumbents market power is insufficient to price predatorily. Light-handed regulation is needed to ensure that fair trading practises are adhered to. Secondly, an important part of regulation will revolve around universal service regulation. These usually involve serving uneconomic areas or customer groups. It is necessary to distribute the burden fairly to ensure that no firm has an unfair advantage from these obligations. Equity considerations dictate that basic services should be provided at equal prices to different groups, independent of the true cost attached to it. Cream-skimming should be avoided as competitors without universal service obligations undercut the incumbent in profitable markets and leave the incumbent stranded with unprofitable markets.
24
In the three phases of liberalisation, there are different challenges for the regulator. In phase 1, classical regulation challenges are to be dealt with, i.e. achieving productive, allocative and dynamic efficiency. Once competition is introduced in phase 2, the regulator has to deal more with the behaviour of incumbents and entrants toward each other and the abuse of market power by incumbents, and the provision of fair access prices for entrants to non-competitive parts of the network. In phase 3, regulation becomes more light-handed and revolves around the fair distribution of universal access obligations and the adherence to fair trading practises, since competition takes care of the other aspects of regulating price and entry.
25
26
encourage the development of human resources in telecommunications industry; promote small, medium and micro-enterprises within telecommunications industry;
the the
These are the public interest considerations that permit the deviation of regulation from full promotion of competition. The most important economic goals expressed in the objectives were: promote the provision of a wide range of telecommunication services in the interest of the economic growth and development of the Republic; encourage investment and innovation in the telecommunications industry; encourage the development of a competitive and effective telecommunications manufacturing and supply sector; ensure fair competition within the telecommunications industry; ensure efficient use of the radio frequency spectrum;
27
The purpose of SATRA/ICASA is to regulate the industry in terms of the Telecommunications Act (1996) and pursue any new policy directions that are issued by the Minister of Communications as long as they are consistent with the broad objectives expressed above and if prior consultation has taken place. Many aspects of the Telecommunications Act provided only limited guidance and regulatory principles and so much of the initial work of SATRA/ICASA was to put some concrete guidelines together for the purpose of regulation. The regulatory aspects left to the discretion of the regulator included interconnection, facilities leasing, GPMCS (global personal mobile communications services) and licence conditions. Despite this show of independence, the Act permits the Minister of Communications to amend, withdraw or substitute any decision by the regulator. This provision has already been used by DoC to withdraw SATRAs guidelines on interconnection and facilities leasing. Further, the implication of the Minister setting policy means that the shape of deregulation is left to the Minister. Although the White Paper provides guidelines for deregulation, it is subject to review by the Minister who also is entrusted with setting the dates for which segments of the market are opened for competition, and the number of new entrants. The regulator role is merely to call for tenders for licences and issue licences according to fair and transparent criteria. SATRA/ICASA has lost credibility in the industry as both independent and effective through failure in a number of crucial regulatory episodes. The first incident occurred shortly after the passage of the Act, when Telkom claimed that the Internet was a basic service and therefore fell under its exclusivity agreement. Although SATRA eventually ruled that this was not the case, Telkom took the challenge to the courts ensuring that the matter dragged on another 6 months. The second major incident was the as yet unsuccessful issuing of a third cellular licence through a beauty contest process. SATRA made a recommendation for Cell C to be awarded the licence, going against the external consultants recommendations. One of the losing bidders, Nextcom, challenged the outcome and has succeeded in delaying the process in the courts for over a year. In the process, it has been alleged that there was political interference in the licensing process, eroding the credibility of SATRA as an independent body. Finally, Telkom accused VANS operators of illegally providing telephony to their clients in contravention of the Act. The action it took was to deny further capacity to them and in some cases withhold services. SATRA ruled against the unilateral action by Telkom but was duly ignored by the operator. The issue has been partly resolved after AT&T threatened to take the issue to the WTO where South Africa has committed to an opening of the VANS sector and to a telecommunications reference paper that upholds action against anticompetitive practice by industry players.
28
This coincided with the need to restructure Telkom itself to face competition such as improving efficiency and rebalancing its tariffs to remove crosssubsidisation. It is also felt that the granting of an exclusivity period helped to raise the market value of Telkom allowing for a better price on the equity sale. For these reasons, the Telecommunications Act gave Telkom a regulated monopoly for five years, extendable to six, in the local access market. To ensure that the exclusivity period fulfilled the goals of infrastructure rollout and preparing Telkom for competition, strict licence conditions were placed on the network provider. The licence conditions included rolling out 2.81 million new lines over the exclusivity period, of which 2/3rds will be in under-serviced areas and for priority customers. It is estimated that this will require capital investment in the region of R53 billion. The specific rollout targets are presented in table 3.1 below. The only part of local access network that is not covered by the exclusivity agreement is customer premises equipment (CPE), which was opened to full competition immediately on passing the Act. Telkom wins an extra year of exclusivity if by the end of the fourth year it has achieved a roll-out of 90% of its cumulative five-year total line target and 80% of its five-year under-serviced line target. This will be granted if Telkom accepts a new five-year total of three million new lines and a proportionate increase in its under-serviced line target. There are financial penalties for failing to reach these targets. Telkom pays penalties of R450 per line for the first 100,000 and R900 per line for each extra line missed. If it misses Priority Customer targets the penalty per unit is R4,500, schools R900, public payphones R2,250 and villages R1,125.
29
In addition to these rollout targets, Telkom is committed to upgrading the network and its service record. This involves upgrading the entire network to digital by 2000, lowering the waiting period for installation, lowering the number of faults per 100 lines and reducing the time to fix faults. The licence also allows for restrictions on price increases introduced by Telkom to ensure that its monopoly position is not abused during the exclusivity period. These are set for a bundle of services (including local, long-distance and international) and are restricted to the CPI -1.5% for the first three years, after which SATRA will set price caps. The White Paper on telecommunications suggested that once exclusivity has ended, then initially the local loop should be open to entry with operators connecting to the PSTN of Telkom (see figure 3.1). The thinking was that entry was unlikely anyway but co-operative organisations or SMMEs might be willing to connect specific communities and so be encouraged. After a further year, it was envisaged that a full switching network could be licenced in competition to Telkom, providing the local exchanges too. The exact shape of deregulation is currently under review. However, the timing of issuing a second PSTN licence is bound in the WTO and must be complete before the end of 2003.
30
1
Cellular Networks CPE VANS
Local Loop
Business LL Telephony LL
3.3.2 Long-distance
Long distance was bundled together with local access in the granting of exclusivity to Telkom until a review of policy and a subsequent call from the Minister to proceed with licensing. The initial thinking in the White Paper was to introduce the licenced resale of long distance facilities operated by Telkom by 2000. Two years later it was envisaged that the regulator could licence long-distance carriers with their own facilities but who connected to the Telkom PSTN. Finally, a full service PSTN would be licenced by 2003. Currently the resale has not materialised and we await the decision on other licencing. Any licence holder would be eligible for universal service fund contributions, human resource contributions and a licence fee of 1% of turnover. There may be additional conditions attached but these are unlikely to exceed any that Telkom would have to bear.
31
32
MTN and Vodacom were required to install 7,500 and 22,000 community service telephones respectively to under-serviced areas over a period of five years Achieve a population coverage of 60% within 2 years and 70% within 4 years. Price increases limited to the CPI or below in any one year The conditions for a new licence have similar requirements but have been relaxed to attract international investors. In particular, the new entrant will be allowed domestic roaming on the MTN and Vodacom networks to provide them with extensive coverage while their network is being rolled out and they will be protected from further new entrants for 5 years. There is also no specified shareholding size for the required empowerment partner. The Act also recognises that in order to provide a cellular service, the licence holders will need to provide long-distance and international services, often through alternative technologies (satellite or fixed line). To protect the Telkom monopoly, the cellular providers are not allowed to make use of any fixed lines other than those from Telkom, and all international services must be routed through the Telkom network. They are of course free to negotiate international roaming services for when the consumer is out of the country. The key structural issue for the cellular industry is the separation of network from service provider. The network providers provide wholesale airtime to retailers who then retail this airtime to customers on a per minute basis. This separation has been the regulatory path in the UK and makes regulation easier as the focus becomes regulation of the network provision and not any potential abuse of dominance in the service supply component stemming from vertical integration.
3.4.2 Satellite
Satellite networks can be used for transmission of national long-distance and international traffic, as well as Global Mobile Personal Communications. As Telkom does not have any satellite network itself, the network provision is open to competition. However, in order to protect the Telkom monopoly on international traffic, all signal distribution for incoming and outgoing calls to SA through earth stations must take place through the Telkom service provider. In the event that the Satellite network is able to bypass the Telkom facilities through satellite-to-satellite links or earth stations positioned outside the borders, a bypass fee must be paid. This fee has yet to be negotiated. This arrangement will end when Telkom exclusivity ends.
Under non-discriminatory terms, In a timely fashion under conditions that are transparent and fair, and At rates determined by ICASA. These will be set at the long run incremental cost (LRIC) of building and maintaining a network by an efficient telecommunications provider, including an average return on capital. The information required for setting such rates, will be determined by the Chart of Accounts and Cost Allocation Manual (COA/CAM) that is obligatory for all network providers to submit. ICASA reserves the right to settle disputes over interconnection and enforce the minimum requirements established above for interconnection agreements. Similar guidelines have been established for facilities leasing, another important practice for furthering competition in the telecommunications industry. The numbering plan has yet to be finalised by ICASA though a proposed policy document has been released which provides us with an insight into the principles that ICASA is working with. For local access, the competition concerns revolve around number portability. Consumers face a cost of switching local access providers if they have to change their telephone number. This represents a barrier to entry and limits effective competition from any new entrant. The switching costs include informing all associates of the change and, for businesses, adjusting any letterheads, business cards and informational material that has the number on it. These costs are significant enough to prevent numerous changes occurring and so initially provide the incumbent with some market power. ICASA has not made a decision on this or provided any direction. The key problem is that it does come with costs to the network provider because their networks need to be upgraded to operate number portability. ICASA is putting aside a block of numbers in case number portability does not materialise and the new entrant requires its own numbers. For long-distance network providers the preferred outcome is that a customer can pre-select a long-distance network provider and so there is no need to dial additional digits first. Again, ICASA has not made policy on this but suggests that pre-selection may be a requirement under the new regulatory structure (SATRA 1999). An interim measure is to use as short as possible carrier access code until pre-selection is available. The recognition is that the shorter the access code, the less inconvenience to customers and so the less discrimination against the entrant. In order to allow consumers the flexibility to select an alternative provider for individual calls, it is preferable to have preselection override codes too. The issue of offering pre-selection concerns the upgrading of Telkoms networks. For long-distance service providers who resale leased capacity, the access codes will be necessary. Maintaining these as short as possible is a means to make them more convenient to customers. If regulation allows service providers to be pre-selected too, then these would not be necessary.
34
35
In terms of competition analysis, there may be scope for more narrow definitions depending on the features of each piece of equipment. This would need to come out in a specific enquiry.
In the UK, the incumbent BT still had over 50% of the market after 10 years of liberalisation. Mobile handsets There are likely to be no dominant players in the provision of mobile handsets. The crucial difference to fixed line is that the market for CPE supply was liberal from the start and that the network providers were not involved in retailing them. The service providers who did provide them and who would have some vertical integration advantage are sufficiently numerous not to have market power. PABXs This is likely to be a more competitive market than fixed line phones as businesses are likely to source directly from manufacturers.
4.1.3 Contestability
The market for supply of CPE is an open one with no real barriers to entry. There are also a large number of firms in the market. However, there are advantages for vertically integrated service providers (fixed or mobile) that need to be monitored. This problem may be particularly acute with fixed line where there is only one service provider (Telkom) and a history of monopoly supply (so starting at 100% coverage).
37
on a common national pricing system, cellular is more expensive than fixed line access. Many customers make use of both suggesting they are not substitutes but complements The two are regulated separately in most countries which suggests that they are considered separate markets
Also, the current level of substitutability may have been enhanced by two temporary factors: Excess demand for telephony that existed when cellular operators were licenced - excess demand raises the effective price of fixed line for those without access (because they incorporate the cost of waiting) making a higher priced cellular a closer substitute. Public monopoly Telkom is less efficient than many international players in liberalised markets and may also be pursuing monopoly pricing. The combination of possible inefficiency and monopoly pricing means that prices may be inflated making them more substitutable with the more expensive mobile technology.
It is difficult to determine whether the level of substitutability in South Africa is sufficient to have mobile and fixed categorised in the same market. It is also unclear whether the level of substitutability may actually decrease as soon as more competition is introduced into fixed line market and saturation of local access demand is reached. What is clear is that they are usually regulated separately and so considered separate markets. It may therefore be wise to adopt a more cautious narrow market definition for now, separating fixed line local access from mobile. However, it is going to be important to review this periodically. In particular, given the more rapid rate of technological change in mobile, specifically impacting on costs and bandwidth (UMTS), the market boundaries may well change and the two considered sufficiently substitutable. The second case is to see if fixed access should be split further. The division into cable, fixed wireless and PSTN does not seem a suitable division as all are merely alternative means of delivery of what is essentially a fixed local access product. Regulatory practice also usually places them in the same market, suggesting that this is the case. The one dimension on which to differentiate may be bandwidth. However, for telephony this is not an issue, it is only an issue for VANS. Further, the PSTN can provide broadband access at a higher monthly rental fee. The conclusion would therefore suggest that cable and fixed wireless should be included in the same market and merely considered different technological options but still offering the same service. The provision of local access can be split horizontally into business and residential customers. The reason for this is the very different usage patterns and the density of demand. These serve to make business customers much more profitable, as there are lower installation costs and higher usage of capacity.
38
4.2.3 Contestability
The primary barrier to entry is of course the regulated monopoly. A key question for competition analysis is whether the deregulation of the sector would result in entry in South Africa. Entry can be defined as both complete (offering services to the entire market) or partial (servicing only a component of the market). This requires a focus on the other barriers to entry that exist in the fixed local access market. The discussion of entry barriers also examines how specific regulatory decisions may impact on the size of the barrier faced by potential entrants. Entry Barriers High sunk infrastructure costs the high level of sunk costs (as discussed in section 1) are why local access is considered to be the only natural monopoly component in the telecommunications production chain. The excess capacity at the local access level also implies that it is socially wasteful to duplicate resources, adding to the natural monopoly concerns. The high sunk costs imply low marginal costs, allowing the incumbent space to pursue a credible predatory pricing strategy to deter entry. These costs can be reduced if the entrant has an existing local transmission infrastructure that they can piggyback on. It is for this reason that the majority of entrants at the local level in other countries are cable operators who have incurred much of the fixed costs of putting in a local loop already. However, there are currently no cable operators in South Africa. Another possibility is the fixed wireless approach where costs are more closely related to traffic volumes and so less investment is required initially. Finally, the regulator could grant an entrant permission to piggyback on the incumbents infrastructure (poles and ducts). An alternative strategy, and one being mooted by ICASA, is to grant a single entrant protection from further competition in all sections of the market (incl. long-distance) i.e. licence a second national operator (SNO) only. The rationale is to provide the entrant with higher profits to incentivise them to make the sunk investments. The problem with this strategy is that it usually favours the incumbent more than the entrant, as they are the ones with a greater market share over which to make abnormal profits. Other problems include the lack of incentives for the entrant to serve low-usage customers (resulting in a cherry-picking strategy), and a questionable impact on long-term structural change and competition. This was the UK experience, with Mercury (the entrant) concentrating on business customers and selected households. In fact, Armstrong et al (1995:241) note of the British duopoly policy, (t)he deliberate restrictions on competition contained in the duopoly policy, together with insufficient attention paid to overcoming BTs incumbency
39
advantages, acted to preserve the essentially monopolistic character of the old system in the core area of network operation. Customer switching costs customer switching costs include the cost of changing numbers only, as equipment will have a common standard. Costs are usually considered higher for business than residential. These costs are seen to be significant enough for customers not to switch often. Given the history of monopoly, switching costs are likely to provide the incumbent with a significant advantage. Further, full number portability is unlikely to be granted by an incumbent interested in maintaining market power and would need to be regulated. Armstrong et al (1995) note that this has proved important in inducing cable providers to enter the local telecommunications market in the UK. Interconnection pricing the reality of competition at the local access level is that calls may originate or terminate at the competitors network. When a call is made from a subscriber of one network to a subscriber of another network, then an interconnection charge is imposed to route and terminate the call. The implication is that interconnection costs will have a significant influence on the price of calls for both networks. This applies equally to interconnection charges with foreign networks and with domestic mobile networks. The key issue is whether the pricing of interconnection can affect the ability of a competitor to compete in the market. Clearly, if interconnection affects the price of their product, then it must be able to affect their competitiveness. There are two possible scenarios that all favour the incumbent. First, a common interconnection price is agreed. In this case it would favour the incumbent to overprice interconnection. The reason is simple. The primary problem faced by an entrant is that on balance there will be more traffic terminating at the incumbents network than on their own, due to the fact that the incumbent has more subscribers. High interconnection charges will therefore affect a far greater amount of calls made by subscribers of the entrant, than those subscribers of the incumbent. The result is a higher average price for local calls for the entrant relative to the incumbent. Second, an interconnection price for each network is negotiated. The incumbent again has far greater market power and so will be able to use this to get better terms out of the entrant than they are willing to offer themselves. Again the result is higher average call costs for the entrants subscribers. Another strategy for the incumbent would be to delay the reaching of an agreement, which would prevent the entrant from launching their service. Although the ICASA rules would then call on the regulator to break the deadlock, the decision on what the interconnection level would be could ensure further delays. These time delays would impose heavy costs on the entrant. Regulation can overcome these problems to an extent by stipulating a level of interconnection charges that come into effect if better terms cannot be negotiated. This lowers the incentive for the incumbent to delay
40
proceedings and ensures that further regulatory delays do not occur if no negotiated terms are agreed. The problem is estimating the interconnection rate. The nature of network industries is that many costs are common costs and so there is discretion in how they are allocated. It is in the interests of the incumbent to allocate as much of its costs to the service where they expect the least competition, and where they are able to negatively influence the price of their competitors the most. This is of course local access. The incumbent is therefore likely to try pass off as much of the costs into local access to justify high interconnection prices that penalise both other local access providers and all other services that interconnect (long-distance, VANS, mobile). Interconnection pricing can also be used to compensate for other incumbency advantages to facilitate entry. The usual strategy is to force the incumbent to offer interconnection rates that are below cost to provide the entrant with some advantage of their own. Either way, failure to resolve the interconnection issue satisfatorily will result in greater uncertainty for potential entrants in all other services as local access is the essential facility in telecommunications. Uncertainty can only lead to lower investment incentives, especially when there are high sunk costs. Current network coverage and universal service subsidies given high sunk costs and customer switching costs, there are low incentives to enter directly into competition with the incumbent where the incumbent has a local loop in place. The exception to this is high usage customers (especially long distance usage) that have high density i.e. business customers. However, for lower usage customers, the entrant may have more incentive to enter areas that are not currently served by the incumbent. The exception would be areas that are not serviced because their usage is very low (usually poor households) and/or where installation costs are very high per subscriber (i.e. rural areas). This can be overcome if the entrant has access to the universal service subsidy to part fund the installation of local loops in these areas. If the unserviced areas are a large part of the market, then access to universal service subsidies and rapid entry into the market are important in establishing a viable local access base for the entrant. Regulation can help by favouring the entrant in the provision of universal service subsidies and servicing new property developments. Likelihood of Entry The high barriers to entry at the local access level have made this the last natural monopoly component of the telecommunications system. Entry that has occurred in other countries has tended to be in the business component of the market as high usage and density serve to lower installation costs, and allow investments to be written off quickly. Their high use of long-distance services provides the capacity to make investments in the more lucrative longdistance market. Entry takes the form of establishing long-distance networks and local bypass of the PSTN to these networks. Competition that has emerged in the residential component of the market has generally been from cable companies using their economies of scope. As
41
South Africa does not have a cable TV operator, this is not a source of entry now. It is also unlikely to be the case in the future, as a cable TV company would only be interested in a small group of high income earners, plus the choice of technology for pay TV in South Africa is satellite (Multichoice and Sentech). The low teledensity in South Africa suggests that the entrant may be tempted by the number of unserviced areas. However, the reality is that they are unserviced because they are the least viable group of potential subscribers. Therefore, unless the universal service funding was sufficient to make these customers profitable, they are unlikely to be served by the entrant. Clearly entry at the residential level is unlikely to be significant unless there is significant incentives from the regulator for entry, including putting rollout conditions in the licence, providing access to universal service subsidies, having favourable interconnection agreements and possibly lowering competition in long-distance through a duopoly policy to allow them to cross-subsidise services. Either way, the entrant is unlikely to get a significant share of the local access market to act as a disciplining effect on the incumbent. Rate regulation of the incumbent will be the primary tool for ensuring the protection of consumers against abuse of dominance.
42
4.3.3 Contestability
The primary barrier to entry for network and service provision is of course the regulated monopoly. Again the key question is how will the market change if entry was liberalised. In answering this question for long-distance, it is important to separate the network provision from the service provision, as the barriers to entry are considerably different. Entry can also be partial or complete. Partial may include business or metropolitan areas only, while complete would include a national service. Barriers to entry The barriers to entry for network providers include: Sunk costs the long distance network has fewer sunk costs to a full local access network. In addition, long distance is not considered a natural monopoly because excess capacity is not a problem. The collecting of local traffic and funnelling it down a single long distance network ensures higher capacity utilisation and scope for more than one network to profitably exist. The level of sunk costs for long-distance networks can be more easily lowered than in local access. First, there are more established networks delivering other services making economies of scope possible (include railway, electricity). Also, satellite is a viable alternative to fixed line given high capacity utilisation rates. Finally, in South Africa there exist private networks with a national long-distance network already installed that has spare capacity. The incremental cost of providing the service to local access subscribers involves providing a point-of-presence (PoP) switch and extending the network to areas that are not covered yet. The sunk costs can be lowered for initial entry if the entrant is able to lease capacity from the incumbent on entry while they roll out a more complete network (those that intend getting a full network). This is the equivalent issue to allow a new mobile network operator to roam on existing networks during the rollout period. Customer switching costs customer switching costs between long distance networks are focused on the dialling process. If there are more than one long-distance providers, then it is necessary for the subscriber to make the choice. Given that the incumbent is going to be the natural default if no specific choice is made (because they own the local access network), this offers the incumbent a distinct advantage over competing networks. The worst possible scenario is if the customer has an incumbent default and needs to tap in an access code to use a competing long distance provider. The longer this code is, the higher are the switching costs. The least distortionary means is to have the customer pre-select the longdistance provider of choice and then any long-distance call without an access code will be routed through their network of choice. If the subscriber wishes to override their selection, then the access code can be used. However, it is preferable that as short as possible access code is used to limit switching costs.
43
Interconnection pricing a long distance call must originate and terminate in a local access network. As such, the cost of interconnection will impact on the overall cost of the long-distance call. Unless the incumbent is vertically separated (as with AT&T), there is a danger that the incumbent can set interconnection rates at higher levels for competing long-distance networks and so raise the costs of their rivals. As discussed in the local access section, there are also dangers of delaying interconnection agreements to raise rivals costs. The other danger with interconnection is that the incumbent offers a lower quality of connection to competing longdistance providers, making their service less attractive to subscribers. Facilities leasing as with interconnection pricing, the price at which facilities are leased will influence the price of rivals costs. If the entrant needs to lease facilities as a short-term measure, then the price of leasing will be an important influence on average costs. It is clearly in the interest of the incumbent to overprice these services. The preferable arrangement is for the leasing price to at least reflect long run costs of providing the facilities. However, it is also feasible that the incumbent price below cost in off-peak periods to get rid of excess capacity.
For long-distance service providers the barriers to entry are substantially less. Service providers lease network capacity from network providers and then resell it to subscribers on a per minute basis. The infrastructure they require is a billing system and a switch connection to the long distance providers pointof-presence (they are routed through the local access network like all others). However, service providers face the same customer switching costs as competing network providers. In liberated markets the pre-select option is rarely made available to service providers because of the highly contestable nature and proliferation of suppliers. Service providers also face the problem of interconnection pricing and facilities leasing. The facilities leasing is more of a problem for service providers than network providers because they lease all facilities and not just where they have not rolled out. Likelihood of entry Entry into long distance network and service provision is highly likely in South Africa if the market were liberalised. On the network side, Transtel and Esitel already have networks in place, and international network providers already exist but cannot sell directly to the subscriber. Given the growth in data traffic, there is considerable scope for other entrants into the market. Opening resale and network competition at the same time makes it highly likely that initial competitors will lease facilities while they establish their own networks. This will accelerate the process of competition. Competition in service provision is even more likely given the low costs of entry. However, the regulatory decisions around facilities leasing, customer switching and interconnection costs will have a considerable impact on the entry of firms and the coverage of their networks.
44
4.4 Mobile
4.4.1 Relevant Market
The discussion of the relevant market in fixed local access determined that mobile access was not a substitute for fixed local access, though it may be considered as such in the near future. Also on the horizon are the so-called 3rd generation cellular services (UMTS). These can be separated from the current 2nd generation GSM networks in that they have broadband capacity and so are able to provide a range of product options that the GSM networks cannot. Although these can be seen as two distinct markets, there are economies of scope in their production. Finally, network provision can be separated from service provision.
4.4.3 Contestability
The primary barrier to entry for network provision is regulation the need for a licence (currently restricted to three). In service provision there is less regulatory intervention but service providers need to be approved by the network providers for reasons of protecting their reputation and customer service levels.
45
Again the key question is how will the market change if entry was liberalised. In answering this question for mobile, it is important to separate the network provision from the service provision, as the barriers to entry are considerably different. Barriers to entry In the network component the barriers to entry include: Sunk costs if the network provider is required to roll out a complete nationwide network (as is the case in South Africa), then the sunk costs will clearly be high. However, with mobile, as opposed to fixed line, the level of infrastructure required depends on the level of traffic (see section 1). The result is that there is a lower investment required to launch a service, but which can rise in proportion to the growth in subscribers. This makes entry easier and more likely that the industry can support more than one or two network providers. A further means of easing the entry of a new network is to allow the use of infrastructure from competing network providers until a full network is up and running. The reduction in time to enlist subscribers means that the entrant can start generating a return on their investment sooner, thereby lowering costs of establishment. Interconnection pricing as with any network, a mobile network needs to connect to all other networks in order to provide their subscribers with the chance to call anyone else. The key interconnection prices are fixed local access, other mobile networks and long-distance. There are similar concerns about overpricing of interconnection, especially when there is a dominant player in the mobile networks or a player that is vertically/horizontally integrated into other markets (long-distance or local access). The dominant player concerns were discussed in detail under local access. Problems are non-standard pricing, delaying interconnection, or overpricing interconnection. For the vertically integrated/horizontally integrated firm there are opportunities to engage in transfer pricing from the competitive segment to the uncompetitive segment. If there is a local access monopoly with a mobile network, it may be in their interests to raise local access interconnection prices for all mobile networks (including their own). There is no fear of retaliation in the local access market where there is no competition, and overpricing enables them to extract rents from other mobile providers and also affect the overall substitutability of mobile with fixed line. Facilities leasing similar concerns to interconnection arise on facilities leasing. The difference is merely the nature of the contract, but the problems of strategic pricing of facilities remains the same. Customer switching customer switching costs include the usual number switching issues, but also would include the use of long-term contracts by the current service providers (linked to specific network providers). Longterm contracts are part of the product range offered by service providers (including pay-as-you-go) and serve the customer by providing a chance to hire a phone as part of the contract thereby saving them from the purchase price. However, the length of contracts would need to be
46
examined more closely to see whether they constitute significant barriers or not. The entry of service providers is far easier. The sunk costs include a retail network and customer billing system. However, they also need to be approved by the network providers so some quality assurance is required. This is justifiable on business grounds. Only if the network provider has vested interests in another service provider, is there a potential cause for concern around discriminatory behaviour to favour their service provider. Likelihood of entry The lower initial sunk costs involved in mobile network establishment means that the market is able to sustain far more network operators. The existence of scale economies does mean that a limited number of competitors will exist, especially if they are limited to provide a national service. The level of competition is likely to be fiercer in mobile networks because there was no initial monopoly.
47
There are also a large number of service providers who lease capacity from the networks and retail to the public at their own tariffs. There are currently over 100 Internet Service Providers (ISPs) in South Africa, of varying size. At the end of 2000 there were a few large ISPs - M-web, Worldonline and Xsinet in dialup access, and Internet Solution in leased line access. The difficulty in determining market structure in this sector is that the market changes dramatically each year. The most recent development has been the introduction of Absa freemail that is apparently signing users at a rate of 8000 per week (Business Report March 2001). Private data services The industry was launched when VANS service provider licences were issued in 1991. There are currently about 60 license holders with approximately 12 000 customers and a total annual revenue of around R180-200 million. The top three service providers have a market share of 75% and are Trafix, EDS/Vanco and FirstNet. Other significant networks include GOVNET, Info Van, and Denel Informatics.
4.5.3 Contestability
The VANS market has scale economies and sunk costs but not on a scale near that of a PSTN or mobile network. This makes it a relatively contestable market, reflected in the greater number of market players. The key position that Telkom holds as the only provider of lines to all these operations, raises concerns over whether it is favouring its own companies in these markets. There is also scope for strategic behaviour from other players in the market as they exploit customer switching costs or market power.
million. The dominant operators are Radiospoor (40%), Autopage (25%), Paging Plus (10%). A handful of these operators do not provide services to the public but operate their private corporate paging services only. Radio Trunking The first radio trunking licences in South Africa were only issued in 1993, a year before the introduction of cellular networks. The expansion of the cellular networks has stunted the growth of radio trunking as they are direct competitors in mobile communications services. There are currently 11 licensed operators sharing a market of between 5000 and 7000 radios. The sector is relatively small with revenues of R9 R12.5 million per annum. The sector is dominated by 3 operators Fleetcall, Q-Trunk and One-to-One. Wireless Data Services There are currently only 2 wireless data network providers, Swiftnet and WBS.
4.6.3 Contestability
As with VANS, entry barriers are low to these mobile offerings. The fact that the same services are feasible over the cellular networks make this a highly contested market.
49
5. Regulatory issues for the SA telecommunications industry and their potential efficiency/public interest impact
The current regulatory regime in telecommunications has to change by the end of 2003 in compliance with South Africas commitment to the WTO. In deciding how the regulatory regime will change, one needs to take account of the range of objectives laid out for telecommunications policy and not just focus on regulating for efficiency only. The full list of objectives were discussed in section 3, but are repeated here. They include promote the universal and affordable provision of telecommunication services; ensure that, in relation to the provision of telecommunication services, the needs of the local communities and areas are duly taken into account; encourage ownership and control of telecommunication services by persons from historically disadvantaged groups; encourage the development of human resources in the telecommunications industry; promote small, medium and micro-enterprises within the telecommunications industry; promote the provision of a wide range of telecommunication services in the interest of the economic growth and development of the Republic; encourage investment and innovation in the telecommunications industry; encourage the development of a competitive and effective telecommunications manufacturing and supply sector; ensure fair competition within the telecommunications industry; ensure efficient use of the radio frequency spectrum; There is no specific priority ordering of objectives and different interest groups will rank them differently. The regulatory outcome will depend on the bargaining process amongst the different interest groups. This section examines the various regulatory options on the table. Rather than put down various possible models, the discussion is structured around four critical components of the regulatory regime namely structure, entry, pricing and quality. The interconnection between policies on each component will be made clear.
5.1 Structure
The key structural questions in telecommunications deregulation are: Whether to separate the natural monopoly component (or essential facilities) fixed local access from the parts of the networks that offer potential for competition CPE supply, long distance (vertical separation) Whether network provision as a whole should be separated from service provision - VANS or even basic voice (vertical separation)
50
Whether mobile networks should be separated from fixed networks (horizontal separation)
Forcing customers to dial an access code to use the competing long distance networks, thereby raising switching costs. Armstrong et al (1995) note that another reason for avoiding the breakup was that BT management was opposed to the idea and the regulator needed their support for a rapid privatisation.
51
the table for discussion at the moment and that the models considered are just the number of full facilities based competitors, supports this reasoning. If there is to be vertical integration, then naturally there needs to be at least separate accounting to facilitate regulation of local access and competition analysis. However, given Telkoms history of conduct since the partial privatisation7, there may be good reason to ensure that there is close competitive oversight of the interconnection process and agreements.
In particular their dealings with VANS operators, denying them additional capacity after alleging they were providing clients with telephony and not just data services.
52
separate the markets to ensure maximum competition. The case against is the economies of scope argument. The second argument for horizontal separation is a potentially very powerful argument in the South African scenario. As discussed in section 4, there is unlikely to be a cable competitor on a broad scale in South Africa due to low average incomes and the current technology choice of satellite for Pay TV. It is also unlikely that a significant residential local access competitor will emerge unless there is some form of exclusivity granted (and even then there is doubt). The most likely disciplining force from industry players will come from mobile and so it would be advisable to avoid corrupting this competition by allowing cross-ownership. It is also relatively easy to implement as the current cross-holding is exactly that a share holding and not an integrated organisation. Restructuring would come at little cost.
5.2 Entry
The key entry issues in telecommunications deregulation are: Whether or not to restrict entry into any parts of the market How to deal with possible entry deterrence if entry is permitted Whether or not to actively assist the entrant(s) in competing with the incumbent How to deal with asymmetric regulation of the incumbent and entrant(s)
53
Clearly this approach has a stifling effect on competition that is why the incumbent, main contender and revenue-maximising government support the approach, and the rest of the industry does not. The approach has some merit if it is able to permanently alter the market structure of local access in the short exclusivity period allowing the regulator to exit control of incumbent pricing in the medium term. Evidence from other countries suggests that this is unlikely to be the case. The UK pursued this model and after a seven year duopoly period the incumbent still had 97% of the residential local access market, 87% of the business local access market and 76% of the Internet market. Clearly this policy had no significant long-run impact on the market structure and no reduction in regulatory activity and hence regulatory cost. The approach may also has merit if it is able to significantly raise access to telephony by residents of underserviced areas (social goal). This is also a questionable assumption. The primary tool for universal service will be the non-distortionary USF, which raises money from a tax on all telecommunications operators. The duopoly approach is an attempt to supplement this with internal cross-subsidy an inefficient alternative that will have business local access and long-distance supporting residential access. Raising the penetration of telephony is probably better served through raising the USF tax rather than distorting competition, especially if the adjustment has to come eventually. Open Competition Approach Evidence from most countries suggests that in the residential local access market there is a considerable natural monopoly element resulting in the incumbent remaining highly dominant even with completely open entry. If duopoly is not going to succeed in changing this either, then tight regulation of local access will remain a part of regulatory life for the foreseeable future regardless of entry strategy. If this is the case, it seems that one should reap the gains from competition in other parts of the telecommunications market now rather than later, while supporting some long-term market structure change in local access. This approach would open up the market for long-distance network and resale to make this a competitive component. This is a potentially competitive market and so is likely to see entry. It also offers significant gains for business and the Internet, both of which make considerable use of the long-distance network. In this way the policy supports the intermediate role of the sector in growing business in South Africa. In the local access market there are two possible approaches. One approach would be to open entry completely. In this scenario, entry is likely to occur in the business component, linking businesses to the long-distance networks (as in the USA). If the incumbent is cross-subsidising residential then the amount of bypass will be greater. To ensure that the incumbent is not overly exposed to bypass, the universal service approach should be entirely through the USF and not built into contracts. Entry could also be encouraged into residential local access through bidding for USF funds to install local loops in underserviced areas (a successful USA approach).
54
The alternative is to maintain a cross-subsidy approach and allow entry into the business local access on condition of certain residential obligations. This appears less attractive as it requires considerable regulatory oversight and is distortionary and so less preferable to a raising of the USF contribution. Under the more open competition approach, the likely main competitor for local access will be mobile. This being the case, the regulator would need to ensure that no cross-holdings exist to limit scope for strategic behaviour by the incumbent. With both the duopoly and open competition model, the incumbent will have considerable market power and scope to limit entry. The regulator will need to have tight control of prices by the incumbent (but not the entrant), and adopt measures to limit entry deterrance and promote entry. These are discussed below.
All these policies may be desirable in South Africa. The last option may be politically sensitive to implement if an open competition model is taken, because the rationale is to rather raise the contribution to the USF to support universal service rather than use distortionary cross-subsidisation.
5.3 Pricing
Long after the market is opened to competition, the incumbent will have sufficient market power to warrant strict price regulation. The entrants will usually not warrant such regulation as they lack the market power. This section looks at how final consumer prices and interconnection prices should be regulated.
56
One of the price cap issues is what to include in the bundle. As already mentioned, it is advisable to remove monopoly components into a separate price cap to avoid transfer pricing by an integrated incumbent. This would leave one price cap for local access and one for all other prices. Some countries have then left certain prices outside of the cap under the reasoning that those markets were competitive. However, the trend now is to include all other prices as the scope for strategic anti-competitive behaviour by the dominant incumbent is enormous (the USA is a case in point with progressively greater price control of AT&T). There may also be consideration of an earnings sharing approach (discussed in section 2). Under this scheme, any under or overshooting of the target is shared between incumbent and consumer. The incentives are less high powered but they are politically more manageable. Another pricing issue is the structure of prices. Local prices are ordinarily made up of a monthly access charge in order to cover the fixed costs of establishing the network and a usage charge to cover marginal costs (which are very low in the case of telecommunications). Optimal pricing suggests raising the access charge well above marginal cost as demand is inelastic and lowering the usage charge closer to marginal cost as demand is elastic (Ramsey pricing). Universal service suggests the opposite, with low access charges to allow low-income households to get access. As each user needs to cover the costs of linking them to the network, an alternative is to have a menu of tariffs to suit the preferences of each consumer (this is not unfair price discrimination as the consumers have a choice from a bundle of options). The high-usage customer (business or heavy Internet user) would opt for a high access charge, low usage charge option while the poorer household would opt for a low access charge, high usage charge option. An optimal range of pricing options is unlikely to emerge in a monopoly scenario, but will emerge under competition (note that in cellular in SA there are many such pricing options from pay-as-you-go with high usage rates to subscription options, while in fixed line there are less options). This suggests that while the incumbent remains dominant, the regulator may wish to impose pricing structures as well as price caps to get optimal pricing in the market. The precendent for such action is the UK where early into the privatised model options for low-usage customers was enforced8 (resulting in 1.5million users taking this option) and more recently a high usage option enforced (for heavy Internet users).
Armstrong et al (1995)
57
a) if two networks do not compete in any markets (as is the case of two national providers from different countries) then the outcome may resemble costs more closely, b) If there is uneven traffic flow, one provider will deliberately overprice in order to gain an effective subsidy from users of the other network (a common practice amongst African national providers and one which Telkom could employ in a duopoly scenario) c) If the two networks do compete in other segments of the market (such as provision of long distance communications) then there is incentive to deliberately overprice in order to ruin the competitiveness of the other network in the other market. As Telkom will have a dominant position in the local access market regardless of model chosen, it is likely that they will have the incentive to overprice access. An alternative is to bundle interconnection pricing with the price cap bundle a global price cap as suggested by Laffont & Tirole (2000). This has the advantage of the incumbent viewing access as just another product. If it overprices local access then it needs to lower other prices to compensate. This may mean that long distance prices are not overly high in the end. The level of price for interconnection needs to recover a portion of the access deficit from other firms if there is no scope for cross-subsidy (i.e. the open market option). In this case, the regulator needs to ensure that there is no cost transfer from the long-distance operations of the incumbent.
5.4 Quality
The existence of price controls often necessitates the use of quality controls too. The reason is that a price-regulated firm has incentives to lower the quality of service to meet price controls instead of lowering costs. The current contract with Telkom for the exclusivity period has such measures. As Telkom is likely to still have some price controls in the post-exclusivity period due to their dominance, quality measures would be appropriate too. There are a range of quality measures that can be used, including: call completion rate faults per 100 lines speed for repairing faults speed of installing a line public phone box maintenance and repair
Quality can be incorporated through financial penalties (current method), incorporating a quality index in the price cap, or an informal warning and discipline through more stringent price caps in the next round.
58
Concluding Remarks
There are many industry analysts who view the past 5 year exclusivity period of Telkom to be a mistake by continuing to hamper downstream businesses. These analysts will argue for complete liberalisation. Other groups of society will view the universal service rollout and low loss of jobs as success points for the policy, and will look to continue this in the form of an exclusive duopoly. The WTO commitment ensures that we need to at least move to a limited entry policy. However, in the end, the policy framework for deregulation of the telecommunications industry is going to be a political process. Ayogu and Hodge (2000) note that the distribution of gains and losses in the full competition model favour the wealthy in society the businesses and the heavy users of telecommunications. The low-income groups are likely to bear all the short-term adjustment costs (job losses) for some uncertain future reward (jobs from downstream users of telecommunications and lower prices on future telephone use). Given the constituency of the government, this suggests a gradual liberalisation approach may be preferred, where revenue maximisation occurs and the gains can be transferred to the losers through fiscal policy. If the full competition model is to be adopted, it needs to be structured to provide stronger universal service support and deliver downstream benefits to low-income groups rapidly. This suggests maybe raising the USF levy and pursuing strict competition and regulatory enforcement in the sector to lower prices and benefit downstream job creators.
59
Bibliography
Armstrong M, Cowan S. Vickers J. 1995. Regulatory Reform: Economic Analysis and British Experience. MIT Press, London Ayogu M, Hodge J. 2000. The socio-economic context and institutional foundations of privatisation and regulatory reform in South Africa, GDN Conference, December 10-13, Tokyo, 2000 Bergman L, Doyle C, Gual J, Hultkrantz L, Neven D, Roller L, Waverman L. 1998. Europes Network Industries: Conflicting Priorities. CEPR, UK. Cave, M Alternative Telecommunications Infrastructures: Their Competition Policy and Market Structure Implications, 1995 Hodge J. 2000. Liberalising Communications Services in South Africa, Development Southern Africa, vol. 17 (3), Sept. ICASA. 2000. Notice in respect of a review of fees and charges in the public switched telecommunications sector Government Gazette Number: 21925 General Notice Number: 4715 ITU, Trends in Telecommunication Reform 1999: Convergence and Regulation, 1999 Kridel D, Sappington D, Weisman D. 1996. The Effects of Incentive Regulation in the Telecommunications Industry: A Survey. Journal of regulatory Economics, vol 9, pp. 269-306 Laffont J, Tirole J. 2000. Competition in Telecommunications. MIT Press, Cambridge, MA. MediaAfrica.com. 2000. South African Internet Market 2000. M-Web. 2000. Annual Report Posner R. 1971. Taxation by Regulation, Bell Journal of Economics and Management Science. Vol. 2(1). Pp. 22-50 Purton, P. 2000. Former Monopoly loses significant market share, Financial Times Survey Resende M. 2000. Regulatory Regimes and Efficiency in US Local Telephony. Oxford Economic Papers. Vol 52 pp. 447-470 RSA. 1995. White Paper on Telecommunications Policy. RSA. 1996. Telecommunications Act 103 SATRA. 1999. Interconnection Guidelines. SATRA. 1999. Telecommunications Numbering into the Twenty-first Century. Consultative Document July. Statistics South Africa. 1995. Income and Expenditure Survey Stigler, G. 1971. The Theory of economic Regulation. Bell Journal of Economics and Management Science. Vol. 2(1). pp. 3-21 Telkom Annual Reports 1995-2000
60
Telkom. 2001. Accomplishments in the Telecommunications Sector during South Africas Democratisation Process, http:www.telkom.co.za/ government_ regulatory/index.shtml WEFA. 1997. Social Accounting Matrix
61