3.5.1 The Regulation versus Investment Debate

In designing policies to foster long-term, facilities-based competition, regulators are tasked with balancing the objective of promoting competition and entry with the need to maintain incentives for investment in new infrastructure and innovation. This entails identifying facilities that are not easily duplicated (that is, bottlenecks) and determining if they are capable of affecting competition in downstream (that is, services) markets. Such a determination would call for the regulation of such bottlenecks to give access to competitors on a nondiscriminatory basis and at cost-based prices, as fostering their duplication would either deter entry or result in a socially wasteful expenditure of resources. The success of such policies ultimately tends to pivot on the regulated prices and terms of access to bottlenecks.

In the absence of functioning market mechanisms, getting access prices just right is a huge challenge for regulators and will affect the incentives of both new entrants and incumbents. If prices are too low, entrants will have no incentive to invest in their own infrastructure, even when it is economically viable and efficient for them to do so. If access prices are too high, competitors either will not enter the market or will choose to deploy their own networks, resulting in inefficient duplication of networks. Conversely, incumbents may refrain from future investment in their networks if their facilities are open to competitors at low rates, as any advantage derived from these investments would be available to rivals, while risks associated with such investment would be borne exclusively by the incumbent.

3.5.2 Regulating Bottlenecks in the Broadband Supply Chain

Supplying broadband services involves a combination of network elements, processing, and business services that can be thought of as the broadband supply chain. More fully described in chapter 5, this supply chain can be divided into four main components: (a) international connectivity, (b) domestic backbone, (c) metropolitan connectivity, and (d) local connectivity. Bottlenecks in any of the links of the chain will stifle competition and the development of broadband. Hence, effective regulatory frameworks must identify and address such instances of market failure in a timely and effective manner.

3.5.2.1 International Connectivity

As electronic communications traffic—particularly Internet traffic—enters and leaves a country, it is typically routed through one or more international facilities, including submarine cables, cable landing stations, and international gateways.16 Traffic can also be routed using satellite connectivity, which may be the only alternative in many developing countries, including landlocked countries and SIDSs. However, satellite links have certain drawbacks such as limited capacity, are more expensive, and experience delays in transmission. Since international facilities provide the entry and exit point for voice, data, video, and other broadband services, they can become bottlenecks if access and traffic are restricted or prices are set above costs.

As the adoption of broadband services and applications increases, demand for international bandwidth also rises. Between 2002 and 2009, international bandwidth usage increased by 60 percent a year, with the strongest demand growth taking place on links to Africa, Latin America, and Middle Eastern countries, which experienced annual growth rates of over 74 percent during this period.

The most efficient way to lower costs and keep pace with demand is through liberalization and promotion of competition among facilities that provide international connectivity, in particular, international gateways, submarine cables, and landing stations. As such, it is important to ensure that there is more than one international carrier and international gateway and, where possible, that there are redundant international cables and other facilities linking a country to competitive global communication networks. For example, Nigeria supported facilities-based competition in the international connectivity market through the introduction of a unified access service license in 2006, which allowed licensees to “construct, maintain, operate, and use an international gateway” and networks consisting of any type of technology, including wireless or wireline systems (Singh and Raja 2010, 58). While it could be argued that the Nigerian Communications Commission’s (NCC) hands-off approach led to a long period of monopoly control by the incumbent provider, NITEL, over the only submarine cable landing in Nigeria, the NCC recently found that a highly competitive market with multiple cable systems is developing (box 3.3; see NCC 2010).

Box 3.3 Competition Analysis in the International Internet Connectivity Market in Nigeria

Source: Telecommunications Management Group.

In its 2010 review of competition in the international Internet connectivity market, the NCC found that this market was sufficiently competitive on a forward-looking basis and therefore did not require ex ante regulatory intervention. This determination was based on an expected increase in facilities-based competition by 2012, stemming from the landing of four additional submarine cables, one of which is to be operated on an open-access basis.

In its analysis, the NCC recognized that for the better part of the last decade the market had been dominated by NITEL, which since 2011 was the monopoly operator of Nigeria’s only submarine cable, the South Atlantic 3/West Africa Submarine Cable (SAT-3/WASC). During this time, competing providers added only limited extra capacity of their own, mostly via satellite links and limited terrestrial links. At the time of the market analysis, four new submarine cables were scheduled to commence service in Nigeria: two in 2010 (Globacom-1 and Main One) and two more within the next two years (the West Africa Cable System in 2011 and the Africa Coast to Europe in 2012). The NCC noted that the new cables would result in a thirty-three-fold increase in Nigeria’s international bandwidth and significantly change the competitive dynamics in the market. As a result, it concluded that any market power NITEL had been able to exercise in the past should be resolved as competitors enter the market.

Facilities-based competition in the international connectivity markets may not be feasible in all developing countries, especially those that generate small amounts of traffic. Also landlocked countries or isolated small island developing states (SIDSs) may not have access to submarine cables and may have to rely on the use of alternative technologies, such as satellites, that often carry a higher price premium.

For countries without a well-functioning international connectivity market, targeted ex ante regulation may be required to address market failure (Hernandez, Leza, and Ballot-Lena 2010). Some countries, such as India, Colombia, and Singapore, have adopted various obligations on international gateways, landing stations, and submarine cable systems (for India, TRAI 2007a; for Colombia and Singapore, IDA 2008). In Colombia, for example, after conducting a review of wholesale inputs for broadband Internet access, the regulator found that cable landing stations constituted essential facilities and required landing station operators to provide access to their facilities on nondiscriminatory terms and to publish a reference access offer.17 Colombia, Comisión de Regulación de Comunicaciones, Resolution no. 2065 (February 27, 2009).

Self-regulation can also be a tool for reducing costs and increasing access to facilities required for international connectivity. Consortium agreements for submarine cable systems, for example, are progressively including nondiscrimination and open-access clauses, whereby third parties are guaranteed access to facilities and capacity at terms comparable to those offered to the facilities’ owners or subsidiaries. For instance, the Eastern African Submarine Cable System (EASSy), which runs from South Africa to Sudan with connections to all countries along its route, includes such safeguards. Launched in 2010, EASSy allows any consortium member to sell capacity in any market in the region to licensed operators on nondiscriminatory terms and conditions (Williams 2008, 42).

3.5.2.2 Domestic Backbone

Constituting the second level of the network element supply chain, a country’s high-capacity domestic backbone network is essential for broadband connectivity since it provides the link from international gateways to local markets as well as domestic connectivity between major cities and towns. However, backbone networks require extensive investments. A major impediment to reducing these costs, particularly in many developing countries, relates to vertical integration in which the backbone network providers are vertically integrated with the local access network operators. This results in a single end-to-end provider that can wield great market power. As such, other service providers may not have access to the backbone or may face high costs for interconnecting, a problem addressed in growing debates on open network access.

From a regulatory perspective, the first step toward facilitating competition in vertically integrated networks is to ensure a liberalized market. In some countries in Sub-Saharan Africa, for example, mobile operators are prohibited from using the incumbent’s network for backbone services, resulting in slow growth in broadband infrastructure. The second step toward increasing competition may entail targeted, ex ante regulations requiring the backbone network provider to offer network capacity on a wholesale, open-access, and nondiscriminatory basis to downstream providers. Alternatively, some countries are setting up national backbone operators that only provide wholesale broadband services on an open-access basis in order to prevent any vertical integration. This scheme is being implemented or proposed in countries such as Australia, Brazil, Colombia, Singapore, and South Africa. However, public financing of national backbones should not crowd out private investment or distort competition. Moreover, where a public subsidy is provided to a backbone broadband network, open-access obligations should be imposed.

Cross-sector coordination is also relevant to the efficient deployment of national connectivity. Fiber optic networks are usually built along existing infrastructure networks such as roads, railways, pipelines, or electricity transmission lines. Most of the cost of constructing fiber optic cable networks along these alternative infrastructure networks lies in the civil works. These costs represent a major fixed and sunk investment, increasing the risks faced by network operators. By lowering the cost of access to these infrastructure networks and reducing the risk associated with it, governments can significantly increase incentives for private investment in backbone networks. One way to reduce costs is to make rights-of-way readily available to network developers by simplify­ing the legal process and limiting the fees that can be charged by local authorities. Additionally, governments can provide direct access to existing infrastruc­ture that they own or control. For example, a railway company could partner with one or more operators to build a fiber optic cable network along the railway lines. In January 2011, for example, Serbian Railways and PTT Srbija agreed to construct telecommunications infrastructure jointly along Serbian Railway’s corridors, totaling 2,031 kilometers.18 Serbian Railways, “PE Serbian Railways and PTT Serbia Signed the Contract on Telecommunications Infrastructure Construction along the Lines,” January 13, 2011, http://www.serbianrailways.com/system/en/home/newsplus/viewsingle/_params/newsplus_news_id/26885.html The United States, for example, has had a policy since 2004 that assists telecommunications providers seeking access to rights-of-way on federal lands (United States, White House, Office of the Press Secretary 2004).

3.5.2.3 Metropolitan Connectivity

Metropolitan connectivity, also referred to as the “middle-mile” or “backhaul” infrastructure, connects towns to the backbone infrastructure or remote wireless base stations and then to the operators’ core network. Competitive and well-functioning wholesale markets for backhaul capacity (for example, leased lines) are a critical component of broadband diffusion and adoption. Developing countries are beginning to focus on core backbone and backhaul networks as a means to increase broadband deployment. For example, South Africa established a state-owned fiber-based infrastructure provider, Broadband Infraco, to provide national backhaul connections on a wholesale basis.19 Independent Communications Authority of South Africa, “Infraco ECNS [Electronic Communications Network Services] License,” October 2009, http://www.infraco.co.za/Legal/ECNS%20License.pdf. Brazil has also begun focusing on backhaul by entering into an agreement with five wireline operators to build out broadband backhaul networks to 3,439 unserved municipalities in exchange for being relieved of existing obligations to install 8,000 dial-up-equipped telecenters.

Particularly for rural and remote areas, wireless technologies may be the most practical solution for high-capacity backhaul for mobile broadband. A study from ABI Research notes that the global revenues from wireless backhaul leasing are expected to increase fivefold between 2009 and 2014 as operators upgrade to Long-Term Evolution (LTE) and traffic demands on mobile networks rise.20  Recognizing the importance of backhaul for mobile broadband, the Telecommunications Regulatory Authority of India (TRAI) recommended to the Ministry of Communications that license conditions be amended in order to allow service providers to share their backhaul links from base transceiver stations (BTSs) to base station controllers (BSCs), noting that such sharing should be permitted via wireless and optical fiber links (TRAI 2007b, 19–20). TRAI maintained that, particularly where traffic from BTSs to BSCs is low in rural and remote areas, backhaul sharing would boost coverage, reduce maintenance efforts, and lower costs.

3.5.2.4 Local Connectivity

Local access networks, also called the “last mile,” refer to the links between the local switch and the consumer. This last link in the broadband supply chain has garnered much attention in recent years, as countries seek to expand service into unserved or underserved areas and to promote competition between operators at the retail level. Unlike other parts of the supply chain, local access regulation can be divided into two distinct areas of policy based on technology: wireline and wireless. Although the goals of policy makers are the same in each case—expand network availability and promote competition—the approaches must be tailored to the unique opportunities and constraints entailed in each technology.

Wireline Networks

The local access segment (the “local loop”) of the wireline network has historically been built and controlled by the incumbent provider of the PSTN. For many years, it was assumed that the local loop services were a “natural monopoly” because they tend to be the most difficult and costly part of the network for alternative operators to replicate. However, as cable networks and commercial wireless services began competing with traditional telecommunications operators, policy makers began reexamining the possibility of facilities-based competition or otherwise promoting service-based competition in the local loop. The degree and extent of regulatory intervention in access networks, particularly on the wireline side, depend on the legacy endowment of infrastructure of each country. In more developed markets, regulation has ranged from a light-touch approach to more extensive restrictions and obligations, such as local loop unbundling (LLU; see chapter 5 for a technical description of how LLU works). However, in developing countries without significant wireline (broadband) infrastructure at the local level, such obligations may have limited impact.

LLU obligations require the incumbent to provide access to exchanges and the physical local loop network so that new market entrants can offer services directly to customers without having to reproduce the incumbent’s network. LLU may be used as a surrogate for infrastructure competition or as a way of inducing price competition between facilities- and services-based competitors. The main advantage of LLU is that it permits much faster market entry than would be possible if entrants were obliged to construct their own networks. The main disadvantage is that it can be a disincentive to fresh infrastructure investment by the incumbent operator (for instance, in deployment of a fiber optic network), especially in developing countries where the local loop is not yet fully built out.

LLU has been widely implemented in Europe, where it was initially required by a regulation of the European Commission in 2000 (European Union 2000). It has been credited with stimulating intramodal competition in some countries. Many other countries around the world have also adopted LLU obligations (Berkman Center for Internet and Society 2010; see also Cohen and Southwood 2008), including Japan, Korea, Nigeria, Norway, Saudi Arabia, South Africa, and Turkey.21 infoDev and ITU, “ICT Regulation Toolkit, Module 2: Competition and Price Regulation,” sec.7.6.5, Competition and Sharing, www.ictregulationtoolkit.org/en/Section.3486.html. LLU has been applied mainly to wireline telephone networks for DSL services, although in theory it could also be applied to other wireline broadband technologies such as cable modem and fiber to the premises (FTTP). Several countries, including the Netherlands, Sweden, and Slovenia, have proposed or implemented fiber unbundling policies.

LLU has not been widely implemented in developing countries. One reason is that the base of installed wireline telephone lines is generally much lower in developing than in developed nations. Considering the limited regulatory resources in some developing nations, efforts might be better spent in encouraging full, open, and technology-neutral infrastructure competition, particularly in wholesale markets, rather than devoting scarce resources to LLU when there are only a limited number of loops to unbundle.

Wireless Networks

Commercial wireless networks have been an important local access technology for more than a decade and have become the predominant means of providing local access to voice and now broadband services in many developing countries. Wireless networks can help to overcome the last-mile wireline bottleneck by giving consumers multiple options for broadband access. For governments seeking to promote greater broadband connectivity, wireless offers some notable advantages, such as a lower cost structure in rural areas and faster rollout, since it is easier to deploy a series of cell towers than to connect each household with a physical wire. With the introduction of 3G and 4G technologies, wireless networks are expected to compete directly with, and be substitutes for, wireline broadband within the next decade. In Austria, for example, the telecommunications regulator (RTR) determined in 2009 that DSL, cable modem, and mobile broadband connections for residential consumers are substitutes at the retail level. The range of policy options and regulatory changes that could be made to improve wireless broadband development is set forth below:

  • Allocate additional spectrum. To support the expected increase in demand for advanced services requiring faster download speeds and the greater use of such services, regulators are implementing policies that promote the most efficient and effective use of spectrum resources, including freeing up spectrum bands that are either unused or underutilized.
  • Flexible allocations. Another major tool for promoting wireless broadband development is for governments to allow flexible use of spectrum so as not to constrain technology or service developments. This will help providers to meet the rapidly changing demands of their customers.
  • Technology neutrality. Technology neutrality refers to the concept that operators should be allowed to use whatever technology or equipment standard they wish in order to meet market demands. Thus, rather than having regulators mandate that a specific technology must be used in a certain band, operators are allowed to choose whatever technology they wish, subject to technical limitations—to prevent interference, for example.
  • Service neutrality. With the transition to digital technology and better processing capabilities, advanced systems are now capable of transmitting all kinds of services. Wireless operators can now provide voice, high-speed data services, and video over their networks. Government regulators should modify service and licensing terms to allow operators to realize the benefits of this flexibility.
  • Greater use of market mechanisms. The move to market mechanisms can be seen in two important trends: assigning spectrum to operators using some sort of competitive mechanism (for example, auctions) and charging market-based prices for acquiring or using spectrum. Having a competitive, transparent means of assignment can also give service providers greater access to spectrum. In conjunction with a regime that allows flexible use of spectrum, such competitive assignment can enable new models of service provision.
  • Spectrum trading. Once spectrum has been assigned, spectrum trading (secondary market license transfers) allows later entrants to a market to access spectrum by paying a market price for it. This improves competition by allowing companies who want (new or additional) spectrum to acquire it from those who may have excess spectrum in specific areas.
  • Mobile virtual network operators (MVNOs). Another way to introduce additional competition into the market is for governments to permit MVNOs to contract with existing mobile carriers to gain access to capacity and network services that they then use to establish their own services and brand. The MVNO model, however, has not been universally successful, as its impact appears to depend on the specifics of a country’s mobile market structure.
  • Coverage obligations. Governments can promote wireless broadband availability by establishing coverage obligations at the time of initial licensing. License requirements tied to coverage obligations, however, must be carefully considered. Requirements that are too easy to meet run the risk of not significantly expanding broadband coverage. Conversely, overly strict requirements are unlikely to be met and could result in either no interest in a license or lower payments.

3.5.3 Infrastructure Sharing

As governments seek ways to expand broadband networks and promote competition in broadband services, they inevitably encounter difficulties. In some areas, low population densities may make it unlikely that the market will support multiple competing wireline or wireless infrastructures. In addition, for some buildings in urban areas, there may not be sufficient physical space to run multiple sets of fiber or copper cables to each potential user or to place wireless towers and other equipment. In such cases, policy makers and regulators have begun to encourage—or even require—parties to share the physical infrastructure used to deliver broadband services.

Two types of infrastructure sharing are generally being considered today. “Passive” sharing includes common use of support structures, such as towers, masts, ducts, conduits, trenches, manholes, street pedestals, and dark fiber. “Active” sharing involves electronics, switching, power supplies, and air conditioning, among other elements. Infrastructure sharing can take many forms, with the most common being collocation (the sharing of physical space in buildings), tower and radio access network sharing, access to dark fiber for backhaul, and backbone networks and physical infrastructure sharing (ducts and conduits).

Infrastructure sharing is rapidly becoming an important means of promoting universal access to networks and offering affordable broadband services by reducing capital expenditures and ongoing operating expenses associated with the rollout and operation of networks. In recent years, a noticeable trend has been toward voluntary sharing of active and passive network facilities around the world, especially in the mobile sector. A push to upgrade and expand networks for mobile broadband is resulting in service providers searching for ways to cut costs and raise capital. For example, service providers may create joint ventures that manage the combined infrastructure assets either for shared use by its owners or on an open-access basis. This allows for network optimization and for avoidance or decommissioning of redundant sites, leading to significant cost reductions for the parties involved. The joint venture in the United Kingdom between Hutchison 3G and T-Mobile, now joined by Orange after its merger with T-Mobile in the United Kingdom, and the pan-European agreement between O2 and Vodafone to share infrastructure in Germany, Spain, Ireland, and the United Kingdom highlight this trend toward increased voluntary sharing in the sector.

The trend of sharing mobile infrastructure also extends to developing countries. In India, for example, the regulator, TRAI, proposed sharing rules for the mobile sector in 2007, both for active and passive components. Since then, Bharti Group, Vodafone Group, and Aditya Birla Telecom (Idea Cellular) have created Indus Tower, a joint venture that controls over 100,000 towers and provides passive infrastructure service to its shareholders and other third parties. Also in India, the drive to raise capital for 3G auctions and deployment during 2010 led to significant divestiture of mobile towers to independent companies that operate them on an open-access basis. For example, in January 2010 an Indian tower company, GTL Infrastructure, acquired 17,500 towers from Aircel, making GTL one of the largest independent tower companies in the world. American Tower, another independent tower company, has also been acquiring towers in countries such as Chile, Brazil, Ghana, India, Mexico, Peru, and South Africa, with the aim of providing open access to such infrastructure.

Many other regulatory authorities, including those of Bangladesh, Nigeria, and Pakistan, have adopted policies to promote infrastructure sharing, especially in the mobile sector. Carefully crafted policy measures can increase time to market, introduce new forms of competition, and foster take-up for ICT services. Sharing also addresses the environmental impact of ICT infrastructure, reducing duplicative mobile towers that affect a city’s skyline, for example. However, close ties and information exchanges between providers that participate in sharing agreements may create concerns with regard to competition, as they could facilitate collusion and reduce competition at the retail level if sufficient control over the network and services is not maintained and the provider’s ability to differentiate retail offers and innovate is curtailed. When promoting voluntary sharing, regulatory authorities and policy makers must balance the potential benefits and costs of such measures, in order to achieve the desired objective of promoting more competitive markets and increased rollout of services.

On the wireline side, several governments are promoting a variety of shared infrastructure approaches. In the most interventionist cases, such as Australia, New Zealand, and Singapore, policy makers have directed the establishment of a single, open-access network that will provide infrastructure services on a wholesale basis to a variety of downstream service providers. Rather than establish an entirely separate network, France has taken a more regulatory approach by setting up sharing requirements and obligations for firms building out fiber networks to more rural areas and to apartment buildings.22 ARCEP has promulgated a series of regulations that cover fiber deployments in the country. Different rules apply to installations in rural as opposed to urban areas. In addition, the Law on Modernizing the Economy (August 2008) introduced the idea of fiber “mutualization,” whereby the fiber installer must make the fiber available to other companies. ARCEP also contemplated (but ultimately did not adopt) a requirement that multiple strands of fiber be installed initially to accommodate multiple providers. Other countries are also considering regulations that will require incumbent operators (usually those that hold significant market power or are former monopoly providers) to make their infrastructure available to alternative carriers. This concept might also be extended to other, often government-owned, entities, such as power companies that maintain towers for electricity distribution.

3.5.4 Opening Vertically Integrated Markets

3.5.4.1 Benefits and Costs of Vertical Integration

Vertical integration, in which a single firm controls multiple levels of the supply chain, is commonly found in ICT markets around the world and often involves the same firm owning and operating network infrastructure as well using this infrastructure to offer retail services to end users. Two main advantages for a vertically integrated firm is the ability to achieve higher economies of scale and lower costs of production by reducing the costs of coordinating upstream and downstream activities. In a competitive market, these efficiencies can benefit consumers through lower retail prices. However, vertical integration may create barriers to entry for new competitors, particularly in the telecommunications sector, where a dominant operator may control essential infrastructure (Crandall, Eisenbach, and Litan 2010, 494–95). In such cases, a dominant, vertically integrated operator may strategically discriminate against competitors and stifle competition.

3.5.4.2 Remedies to Anticompetitive Conduct by a Vertically Integrated Operator

To address competitive concerns associated with vertical integration, some regulators have required dominant operators to separate vertically to some degree through accounting separation, functional separation, or, in extreme cases, structural separation.

3.5.4.2.1 Accounting Separation

The least intrusive and most prevalent remedy, accounting separation makes transparent the vertically integrated operator’s wholesale prices and internal transfer prices, enabling regulatory authorities to monitor compliance with nondiscrimination obligations or to ensure that there is no cross-subsidization. Generally, accounting separation requires the vertically integrated operator to maintain separate records for its upstream and downstream costs and revenues in order to allow the regulator to set wholesale prices for the regulated upstream services. These records are typically subject to independent audit and may also be made publicly available. Although the operator must make its costs transparent, under this remedy it is able to continue benefiting from the efficiencies of vertical integration.

In 2004 the Info-communications Development Authority (IDA) of Singapore issued accounting separation guidelines to allow monitoring of the ICT sector for potential anticompetitive behavior (IDA 2004). These guidelines established two levels of accounting separation: detailed segment reporting (applicable to dominant service providers and entities they control) and simplified segment reporting (certain other entities). This two-tiered approach is intended to provide the IDA with the necessary information, without unduly burdening operators, to ensure that no dominant provider is engaging in cross-subsidization or discrimination. Currently, incumbent SingTel is the only operator designated as dominant in any market, and it is subject to detailed accounting separation obligations.

3.5.4.2.2 Functional Separation

Obligations under functional separation range from simply requiring the operator to establish separate divisions for upstream and downstream activities to requiring the operator to separate the wholesale and retail divisions physically. This may involve the separation of employees (for example, physical separation of offices and prohibitions on the same employee working for both divisions) and the separation of information (for example, limitations on the type and amount of information that may be shared between divisions). Since there is no actual change in ownership or ultimate control under functional separation,23 infoDev and ITU, “ICT Regulation Toolkit, Practice Note: Structural and Functional Separation of Mobile Network Operators,” http://www.ictregulationtoolkit.org/en/PracticeNote.aspx?id=3268. the operator can continue to enjoy many of the benefits of vertical integration (European Regulators Group 2007). More intrusive than accounting separation, regulators may implement functional separation in “exceptional” cases where there has been persistent failure to achieve effective nondiscrimination in relevant markets and where there is little or no prospect of effective competition within a reasonable period after less intrusive remedies have been attempted (European Union 2009b, para. 61).

The 2009 EU Telecoms Reform formally granted national regulatory authorities explicit authority to require network operators holding significant market power to separate functionally their communication networks from their service branches, but only as a last-resort remedy (European Parliament and Council of Ministers 2009). Prior to requiring functional separation, the national regulatory authority must first find that all less intrusive, market-based remedies have failed to achieve effective competition.24 Under the 2009 Telecoms Reform, the market-based remedies are obligations of transparency; nondiscrimination; accounting separation; access to and use of specific network facilities; and price control and cost accounting. See European Union (2009a, art. 13a). Next, it must submit a proposal of functional separation to the European Commission, with evidence justifying the regulatory intervention and an analysis of the likely market impacts. Among the provisions that must be included in the proposal are the precise nature and level of separation, the legal status of the separate business entity, identification of the separate business entity’s assets and the products or services to be supplied by that entity, governance arrangements to ensure the independence of the staff, rules for ensuring compliance with the obligations, and a monitoring program to ensure compliance, including the publication of an annual report.25 Amended EU Access Directive, art. 13a.

To date, no EU member state has mandated functional separation. In some cases, such as that of the United Kingdom, dominant operators have voluntarily implemented functional separation. There, British Telecommunications (BT) agreed to establish a separate division for access services called Openreach, which provides most of BT’s wholesale products. According to the European Commission, BT’s functional separation led to a surge in broadband connections, from 100,000 unbundled lines in December 2005 to 5.5 million by 2008 (European Commission 2009b).

3.5.4.2.3 Structural Separation

Structural separation involves full disaggregation of the vertically integrated operator’s wholesale and retail divisions into separate, individual companies, each with its own ownership and management structure. All benefits associated with vertical integration are eliminated. Regulated structural separation is considered a last-resort measure and is typically used only if other regulatory interventions have failed and a comprehensive cost-benefit analysis has been conducted.26 infoDev and ITU, “ICT Regulation Toolkit, Practice Note: Structural Separation Explained and Applied,” http://www.ictregulationtoolkit.org/en/PracticeN ote.3149.html. Structural separation is extremely difficult to reverse and can dramatically affect the market, such as by increasing regulatory uncertainty and affecting infrastructure investment. Additionally, it is difficult to allocate the separated firms’ assets and liabilities in order to ensure the ongoing viability of both entities. As a result, regulatory authorities rarely impose structural separation as a remedy.

In 2010 the Australian Parliament passed the Telecommunications Legislation Amendment (Competition and Consumer Safeguards) Act 2010 (Australian Government 2010). The act and implementing regulations set out the procedures by which the dominant fixed-line operator, Telstra, must structurally separate control over its copper and hybrid fiber coaxial network infrastructure as well as its provision of wholesale access services, from retail fixed voice and broadband services (Australia, Department of Broadband, Communications, and the Digital Economy 2011). In August 2011 Telstra submitted to the Australian Competition and Consumer Commission its structural separation undertaking plan, which commits Telstra to full structural separation by July 1, 2018.27 Telstra, “Telstra Lodges Structural Separation Undertaking and Migration Plan with ACCC,” Press Release, August 1, 2011, http://www.telstra.com.au/abouttelstra/media-centre/announcements/Telstra-lodges-Structural-Separation-Undertaking-with-ACCC. Telstra’s structural separation is set to occur through the progressive migration of its fixed-line networks to the National Broadband Network (NBN) Company, which is rolling out a national broadband network to be provided on a wholesale-only basis. Additionally, the plan sets out various measures by which Telstra will ensure transparency and equivalence in the supply of regulated services to its wholesale customers during the transition to the NBN. In exchange for structurally separating and providing the NBN Company with access to its fixed-line infrastructure, Telstra will receive compensation in the amount of $A 11 billion.28 Telstra, “Telstra Signs NBN Definitive Agreements,” Press Release, June 23, 2011, http://www.telstra.com.au/abouttelstra/media-centre/announcements/telstra-signs-nbn-definitive-agreements-2.xml.