Module+2.+Competition+and+Price+Regulation-ICT Regulation Toolkit

Module 2.
Competition and Price Regulation
Downloaded Tuesday, April 06, 2010

This document is a downloadable report from the infoDev/ITU ICT Regulatory Toolkit. Please note that the version that appears on the
website at www.ictregulationtoolkit.org is the most updated version and has extra functionality (e.g., links, practice notes, reference
documents etc). If you find any problems, please contact Feedback@ictregulationtoolkit.org.
Module 2. Competition and Price Regulation

Executive Summaries
1 Overview: Putting ICT Regulation in Context
1.1 Key Developments in the ICT Sector
1.2 Trends in ICT Regulation
2 Competition Policy and the ICT Sector
2.1 Forms of Competition
2.1.1 Perfect Competition
2.1.2 Effective Competition
2.1.3 Market Contestability
2.1.4 Sustainable Competition
2.2 Why Focus on Competition?
2.2.1 Benchmarking Competition by Sector
2.2.2 Comparison Table: Competition by Sector and Region
2.3 Competition Policy and Regulation
2.3.1 Competition Policy
2.3.2 Regulation
2.3.3 Ex Ante and Ex Post Regulation
2.3.4 Advantages and Disadvantages of Ex Ante versus Ex Post Regulation
2.3.5 Regulatory Forbearance
2.4 Key Concepts in Competition Policy
2.4.1 Markets and Market Definition
2.4.2 Market Power
2.4.3 Barriers to Entry
2.4.4 Essential Facilities
2.5 Common Forms of Anti- Competitive Conduct
2.5.1 Abuse of Dominance
2.5.2 Refusal to Supply
2.5.3 Vertical Price Squeeze
2.5.4 Cross- Subsidization
2.5.5 Misuse of Information
2.5.6 Customer Lock- In
2.5.7 Exclusionary or Predatory Pricing
2.5.8 Tying and Bundling
2.6 Remedies for Anti- Competitive Conduct
2.6.1 Remedies for Abuse of Dominance
2.6.2 Remedies for Refusal to Supply and Price Squeezes
2.6.3 Remedies for Cross- Subsidization
2.6.4 Remedies for Misuse of Information
2.6.5 Remedies for Customer Lock- In
2.6.6 Remedies for Predatory Pricing
2.6.7 Remedies for Tying and Bundling
2.7 Mergers, Acquisitions, and Joint Ventures
2.7.1 Horizontal Mergers
2.7.2 Vertical Mergers
2.7.3 Joint Ventures
3 Regulating For Interconnection
3.1 Overview of Interconnection
3.1.1 What is Interconnection?
3.1.2 Why is Interconnection Important?
3.1.3 Why Regulate Interconnection?
3.2 Key Concepts
3.2.1 Forms of Interconnection
3.2.2 Unbundling
3.2.3 Asymmetric Interconnection Regulation
3.2.4 Issues Dealt with in Interconnection Agreements
3.3 Setting Interconnection Prices
3.3.1 Pricing Principles
3.3.2 Long- Run Incremental Cost Modelling
3.3.3 Commonly Used Cost Models
3.3.4 Benchmarking Interconnection Rates
3.4 Mobile Interconnection
3.4.1 Forms of Mobile Interconnection
3.4.2 Mobile Termination Rates
3.4.3 Retentions for Fixed- to- Mobile Calls
3.4.4 Modelling Mobile Network Costs
3.4.5 Mobile Roaming
3.4.6 Social Issues and Universal Service
3.5 Challenges and Opportunities for Developing Countries
3.5.1 Infrastructure Challenges
3.5.2 Transparency and Access to Information
3.5.3 Regulating State- Owned Operators
3.5.4 Free Trade Negotiations
3.5.5 Dispute Resolution
3 . 6 C r o s s - Border Interconnection
3.6.1 The Accounting Rate System
3.6.2 Regional Interconnection Clearing Houses
4 New Paradigms: Voice Over IP and IXPS
4.1 About the Internet
4.1.1 Overview of the Internet
4.1.2 The Seven Layers of Internet Interconnection
4.1.3 Evolution of the Internet
4.1.4 Current Internet Market Developments
4.2 About VoIP
4.2.1 Types of VoIP
4.2.2 Comparison of VoIP and Conventional Telephony
4.2.3 Protocols that Support VoIP
4.3 Arbitrage Opportunities in the ICT Sector
4.3.1 Common Arbitrage Strategies
4.4 VoIP and Regulation
4.4.1 Implications of VoIP for Regulators
4.4.2 Trends in VoIP Regulation
4.4.3 Differential Regulation of VoIP and Conventional Telephony
4.5 Interconnection Pricing for VoIP
4.5.1 A Comparison of Telecommunications and Internet Cost Recovery
4.5.2 Models for Internet Interconnection
4.5.3 Implications of VoIP for Interconnection Pricing
4.5.4 Pricing Mechanisms for VoIP Interconnection
4.5.5 Criteria for a New Interconnection Regime
4.6 VoIP Over Wireless Networks
4.7 Benchmarking Rates for Network Access
4.8 Internet Exchange Points
4.8.1 The Role of Internet Exchange Points
4.8.2 Supporting IXPs in Developing Countries
4.8.3 Internet Exchange Points in Africa
5 Regulating Prices
5.1 Why Regulate Prices?
5.2 Economic and Accounting Measures of Cost
5.3 Useful Economic Concepts
5.3.1 Economic Efficiency and Pricing
5.3.2 Economies of Scale and Scope
5.3.3 Single- and Multiple- Service Firms
5.4 Pricing Principles for the ICT Sector
5.5 Setting the Level and Structure of Prices
5.5.1 Fixed and Variable Costs and Price Setting
5.5.2 Determining Mark- U p s o v e r T S L R I C
5.6 Tariff Rebalancing
5.7 International Benchmarking of Prices
5.8 Rate of Return Regulation
5.9 Incentive Regulation
5.10 Rate of Return Regulation versus Price Caps
5.11 Implementing Price Caps
5.11.1 Price Cap Baskets
5.11.2 Assessing Price Variations
5.11.3 Calculating the Productivity Factor
5.11.4 Service Quality Factors
5.11.5 Exogenous Cost Factors
5.12 Towards a Double Price Cap
5.13 Price Regulation and Multiple Play Offerings
6. Infrastructure Sharing
6.1 Key Concepts
6.1.1 Passive and Active Infrastructure
6.1.2 Essential or Bottleneck Facilities
6.1.3 Open Access
6.2 Policy Issues
6.2.1 Promoting rapid and efficient network deployment
6.2.2. The efficient deployment of Next- Generation Networks
6.2.3. Environmental Considerations and Local Land Use Planning
6.2.4. Competition
6.2.5. Innovation
6.3 Forms of sharing
6.3.1. Infrastructure sharing and collocation
6.3.3 Interconnection
6.3.4 Unbundling
6.4 National Fibre Core Networks
6.4.1 The importance of national fibre backbones and access to broadband
6.4.2 Obstacles to the deployment of broadband networks
6.4.3 Key players
6.4.4 Passive Infrastructure Sharing
6.4.5 Active Infrastructure Sharing
6.4.6. Public- Private Partnerships and Open Access Networks
6.4.7. National broadband policies
6.4.8 Regulatory and legal imperatives related to sharing
6.4.9 Commercial and technical considerations related to implementing sharing
6.4.10 Regulatory Best Practices
6.5 Mobile Network Sharing
6.5.1. Policy reasons for supporting mobile network sharing
6.5.2. When might mobile network sharing be appropriate?
6.5.3. Options for passive mobile network sharing
6.5.4. Site Sharing Arrangements
6.5.5. Encouraging passive mobile sharing
6.5.6. Active mobile network sharing
6.5.7. Active Mobile Network Sharing: Extended Site Sharing
6.5.8 Active Mobile Network Sharing: Sharing the Radio Access Network
6.5.9 Active Mobile Network Sharing: Core Network Sharing
6.5.10 Active Mobile Network Sharing: Backhaul Sharing
6.5.11. Competition and Active Mobile Network Sharing
6.5.12 Regulatory and legal issues
6.6 Regulatory Issues Related to Sharing
6.6.1 Selecting an appropriate approach to sharing
6.6.2 National Infrastructure Sharing Policies
6.6.3 Licensing
6.6.4 Pricing
6.6.5 Competition and sharing
6.6.6 Regional Initiatives
Competition and Pricing: Index
Reference Documents on Competition and Price Regulation
This module of the ICT Regulation Toolkit was developed by NERA Economic Consulting in association with
Castalia Strategic Advisors and Kalba International.
Overview: Putting ICT Regulation in Context

The history of interconnection, price regulation and competition policy in the ICT sector, and the implications of
changes in technology and market structure for regulation.
Competition Policy and the ICT Sector

Key concepts of competition policy and regulation: why competition is important; a comparison of competition policy and regulatory
approaches; common anticompetitive practices and remedies; and analysis of business acquisitions.
Regulating for Interconnection

Tools and good practice in regulating interconnection, including: estimating costs and setting interconnection prices; commonly used cost
models; benchmarking interconnection prices; mobile interconnection; interconnection dispute resolution; and cross-border interconnection.
New Paradigms: Voice Over IP and IXPs

Challenges arising from Voice Over IP (VOIP) and the Internet: VOIP interconnection; pricing Internet interconnection; and Internet Exchange
Points (IXPs).
Regulating Prices
Approaches to retail price regulation in ICT, including: the rationale for price regulation; key pricing principles; measuring tariffs and costs;
determining the structure and level of prices; benchmarking retail prices; price caps and rate of return regulation; and non-price
considerations in regulating prices.
Infrastructure Sharing
This section explores policy and regulatory dimensions of sharing ICT infrastructure and capacity, with a particular emphasis on core
infrastructure sharing and mobile network sharing. This section provides information related to key concepts and terms associated with
sharing and the policy objectives frequently linked to sharing approaches. It provides a broad overview of various forms of sharing and a
detailed examination of the sharing of national fibre core networks and the sharing of mobile networks.
Executive Summaries

Executive Summary
1 Overview: Putting ICT Regulation in Context
This section discusses some key trends in ICT regulation. These trends provide important context for the specific issues of competition
policy, interconnection, and pricing discussed in this Module.
Specifically, this section provides an overview of:
■ Key technological and market shifts that are taking place in the ICT sector, and
■ Recent trends in ICT regulation, particularly the shift from government owned monopoly provision to a more competitive model.

The ICT sector is changing rapidly. The nature and pace of change create challenges for both regulators and regulated firms. This section
provides an overview of key developments in the ICT sector, across four categories:
■ Technological changes,
■ The emergence of new services,
■ Changes in market structure and the level of competition, and
■ Financial forces that are impacting on the sector.
Technological Changes
ICT technology is continuing to develop rapidly. Five key areas of technological change are having a significant effect on the structure
of ICT markets, and will continue to do so for the foreseeable future. They are:
■ The shift from analog to digital technologies,

■ The shift from voice to data
■ The shift from circuit switching to packet switching,
■ The shift from narrowband to broadband,
■ The migration of intelligence from the core of the network to the edge, and
■ The increasing role of wireless.
Key implications for network operators and regulators are summarized here.
The Shift from Analog to Digital
Human beings use analog techniques for listening to and viewing content. Historically, technologies for communication have also used
analog signals (for example, conventional telephony, music cassettes and records). More modern technologies convert analog signals into
a digital format for processing, storage, and transport. This has a number of advantages. Digital signals do not degrade when duplicated,
nor do they accumulate noise and other interfering signals when amplified. Devices such as computer hard drives can quickly store,
retrieve, and duplicate digital files.
Digital signals can also be transmitted more efficiently than analog. Because digital signals do not readily spill over into adjacent frequencies
many different signals can ride together, piggy-back style, on a broadband carrier. Substantial operating efficiencies accrue when a
network can aggregate traffic going in the same direction, and transport it all over a single “pipe”. Compression technologies make it
possible to reduce the size (bandwidth) of the pipe needed to send multiple or highly complex signals. Modern digital networks can
simultaneously transport bitstreams representing many different types of communication (such as data, text, audio, video, and voice).
Data Replaces Voice
Until quite recently carriers designed and operated the telecommunications infrastructure primarily with voice communications in mind.
Voice communications generated most company revenues and were the primary mode of communications.
While voice services are still dominant in revenue terms (in many cases voice accounts for more than 80 percent of company revenues),
the importance of voice compared to other information services appears to be declining. There is evidence that local exchange telephony
subscriptions, long-distance minutes of use, and revenues for conventional dial-up services are declining. Demand for information services
and data transmission is continuing to grow steadily. This shift reflects a number of factors, including:
■ Increased business data requirements,

■ Growing consumer demand for “bandwidth hungry” services such as online music, movies, and games,
■ The proliferation of private lines services, that can be used for voice and/or data transmission, and
■ Migration of some voice calls from conventional networks to VoIP.
On the transmission side, digital technologies now play a central role. In addition to digital services, much voice traffic is now digitized, for
at least part of the transmission circuit.
The shift from voice to data has important consequences for network design. Voice communications and data communications have quite
different technical requirements. Table 1 highlights some of the key differences.
Table 1: Comparison of Network Requirements for Conventional Voice and Data
Conventional voice service Data / information services
Small but constant information delivery rate. Typically “bursty” traffic patterns, with short term peak demand requirements.
Requires a dedicated narrowband channel. Generally requires a broadband, high speed network.
Little tolerance for delays, echoes, noise, Users have diverse requirements. Network should be able to handle both time sensitive
and sound distortions. and delay tolerant traffic, and peak traffic bursts.
Two-way, roughly symmetrical Often asymmetrical traffic requirements (for example for Internet services).
communications.
Packet Switching Replaces Circuit Switching
In response to growing demand for data and digital voice, ICT networks are shifting from circuit switching to packet switching.
Networks configured primarily for analog voice communications typically use a technology called circuit switching to set up and break
down links between the caller and call recipient. Circuit switching provides a dedicated, narrowband link available for use only by the caller
and call recipient. This provides for a highly reliable and good quality link appropriate for short duration calls with a small but constant two
way traffic flow.
Circuit switching does not work well for data communications such as Internet traffic, because it cannot handle bursts of high throughput,
and cannot provide switching and routing for other users when the initial parties have a temporary pause in communications. Configuring a
voice line for Internet access ties up a line for the duration of the call even though the data communications requirements may make up only
a small portion of the total time.
Packet switching is a superior, more efficient, way to manage data traffic. The network breaks traffic down into small packets that can be
routed over any available network link. Network links not needed by one set of communicators can become available for others, in effect
making it possible for many communicators to use a network link at the same time. The digital nature of packet switching means that a single
network can handle a variety of different packet-based services including voice, data, text, images, sound, and video.
However, packet switching can result in higher latency than circuit switching. When packets are routed over multiple networks and across
large distances, some packets may be delayed or lost, or packets may arrive out of order. Packet switched networks do have some ability
to reassemble and reorder packets, but where packets arrive too late for processing this will result in some lost information (for example
lost words in a conversation, or lost sound and picture in a videoconference).
Broadband Replaces Narrowband Networks
A further consequence of the shift from voice to data services is a shift from narrowband networks to broadband networks. As Table 1
(above) shows, data services typically require broadband networks. Broadband networks are able to transmit more information, faster.
Unlike narrowband networks, broadband can accommodate the “bursty” nature of data traffic, and transmit large and complex files.
Intelligence Migrates from the Core to the Edge
In conventional circuit-switched telephone networks the “intelligence” is located centrally on telephone company premises. User devices
are generally quite simple (for example telephone handsets). The hierarchy of switches in the telephone network provides the intelligence
to route calls, generate billing information, and provide additional services (such as caller identification, voice mail, and so on).
Centralized intelligence is efficient where most users have similar requirements, as in a telephone network. However, this approach offers
limited opportunities to customize services, and optimize the network, for individual user requirements. The architecture of the Internet
moves intelligence from the core network to users operating at the edges of the network. Users can use on-site information processing to
configure their own services, through software and other customized applications, and can treat the network as a generic information
transport service.
The Increasing Role of Wireless

Consumers have readily and quickly embraced wireless telecommunications. Wireless service removes the need to be “tethered” to the
network, and provides greater mobility for users. Wireless connections are often easier to install and can involve lower capital expenditure
than wired ones. In many rural and remote areas, wireless provides a cost-effective alternative for achieving universal service.
Some of the traffic growth in wireless networks is due to customer migration from wireline networks, (rather than from new
telecommunications customers), as wireless and mobile services become increasingly effective substitutes for wireline services.
Implications of these Trends
Table 2 identifies some of the implications of these technological shifts for regulators and network operators.
Table 2: Regulatory and operational implications of technological shifts
The Shift from

Analog to Digital The shift to digital communications makes it easier for users to tailor services to meet their individual needs, and to
reduce costs. For example users of separate voice and data services can combine them using a single leased line,
typically at a reduced price. This means that carriers can no longer engage in price discrimination by charging
different rates for interchangeable services. In a digital environment users can switch to cheaper services, and
services that trigger lower regulatory fees.
At the same time, new types of Quality of Service issues arise, including the possibility of carriers offering
different levels of service priority for different types of data at different price levels. This in turn has raised
concerns about lack of "net neutrality" from third-party service providers, worried that their access to carrier
networks will be limited by price discrimination.
Data replaces In the past, many incumbent carriers have relied on voice traffic to generate higher per unit revenues and profits
voice than data and leased line services. This approach is becoming less sustainable. In particular, as more users
migrate to lower cost Internet telephony services, this is likely to adversely impact on revenues from conventional
voice telephony.
Packet Network operators seeking to provide packet switched services will need to overhaul existing networks, and
switching install new equipment. This will involve significant costs for incumbent telecommunications operators. Regulators
replaces circuit may have to find ways to support the migration from analog to digital equipment, for example by recognizing the
switching lowered economic value of existing equipment and authorizing shorter depreciation lives for this equipment.
Broadband Incumbent operators will have to make sizable investments in new facilities, in order to offer broadband service.
replaces For operators with limited access to funds, this will involve a trade-off between improving the quality of service
narrowband and coverage provided by existing facilities, and installing next generation infrastructure that can provide both
networks narrowband and broadband services.
Intelligence Digitization and the development of the Internet give users greater control over networks and more flexibility to
migrates from optimize the services they consume. Regulators and carriers have less control over how consumers interact with
the core to the and manipulate networks. This means that neither carriers nor regulators may be able to fully limit what network
edge users can (and cannot) do with the lines they lease, or the services they acquire.
The increasing
role of wireless The migration from wired to wireless services underscores the need for flexibility in how we use radio spectrum.
National governments allocate spectrum, and often specify the uses for each portion of spectrum. These decisions
may adversely affect the flexibility, cost, and accessibility of spectrum for ICT services. In particular, historical
spectrum allocation decisions may limit the scope for service providers to roll-out new, innovative wireless
services. Mobile wireless ventures offering next generation features, such as broadband Internet access, may
need more spectrum than was previously made available. However, in many developing countries, making
spectrum available may be less costly than in developed economies as it typically is used less intensively.
As wireless telecommunications increasingly becomes a substitute for wireline services, regulators may have to
rethink how to achieve universal service goals. This may involve a combination of wireline and wireless
technologies.
Emerging Services
Ongoing technological changes in the ICT sector are driving the emergence of new services, and changing the way in which existing
services are delivered. For example:
■ The Internet provides a platform for new information services, and an alternative delivery mechanism for existing services. Initially a
medium for traffic associated with research and education, the Internet now delivers a wide range of information services including
(but not limited to) text, images, music, audio, movies, and news services (text and video clips).
■ The shift from analog to digital is driving convergence of communication services. Many different types of communication (such as
data, text, audio, video, and voice) can now be transported over the same networks and, in some cases, sent and received over the
same user equipment. For example users can now download movies from the Internet and view them on a personal computer, or can
send e-mails from a digital television unit.
■ Voice over the Internet Protocol (VoIP) is a converged service that has important implications for conventional voice operators and
regulators. VoIP provides voice service over data networks, rather than conventional voice networks. Consequently, VoIP providers
are able to bypass traditional interconnection and regulatory arrangements.
Market Structure and Level of Competition
Recent technological changes are changing the shape of the ICT sector.
Service convergence is blurring the boundaries between sectors. For example, the boundary between broadcasting and
telecommunications is no longer clear. Many content providers broadcast material over the Internet as well as (or instead of) over
conventional broadcasting networks. Broadcasting over mobile television networks is emerging as well. As networks migrate to digital
technologies, broadcasting networks are able to carry a range of services including, potentially, voice telephony. This has important
consequences for sector regulators and competition policy. Co-ordination across regulatory areas (between broadcasting, data services,
and telecommunications) will be important to avoid regulatory arbitrage. Mergers between entities in previously separate sectors may now
raise competition concerns.
Similarly, mobile and other wireless services are becoming effective substitutes for wireline telecommunications services for some users.
Indeed, in some developing countries, wireless minutes of use exceed wireline minutes of use. This trend brings into question the long held
assumption that an incumbent telecommunications operator will necessarily be the dominant operator in the market.
As the quality of VoIP services improves, it is becoming a more effective substitute for conventional voice service. VoIP is already placing
competitive pressure on prices for long distance calls. As more users switch to VoIP, this will have significant implications for regulators
and operators.
Financial Forces
In the past, telecommunications operators have been viewed as stable, monopolistic utilities. The main challenge for regulators has been to
prevent excessive pricing by incumbent operators.
With increasing competition from new providers and new services, the telecommunications sector is becoming more volatile. Revenues
from traditional telecommunications services are under pressure from competing providers and modes of delivery. The role of speculative
financial markets has grown.
These changes are threatening the financial health of many incumbent operators. Historically, many incumbent operators have maintained
high prices for long distance and international services, and used the proceeds to support below-cost prices for basic services. This
pricing structure has generally been supported by regulatory arrangements. As competitive pressures erode long distance and
international prices, regulated low prices for basic service will become unsustainable.
In addition, incumbent operators are facing potentially significant investment costs to upgrading existing infrastructure to keep up with
technological change. Failure to do so will further erode competitiveness over the long term.
This scenario has the potential to erode the value of (often state-owned) incumbent utilities. In the face of increasing competitive pressure,
regulators need to rethink the justification for pricing policies aimed at keeping prices for basic service low, and consider rebalancing tariffs
to better reflect economic costs.

This section draws on the ITU series “Trends in Telecommunication Reform”.
Internationally the way telecommunication services are provided is shifting. Many countries are replacing traditional state-owned service
provision with a competitive, market-based model.
The World Trade Organisation Basic Telecommunications Agreement has been an important catalyst for reform in the telecommunications
sector. This agreement commits countries to put in place transparent regulatory structures, laws, and procedures for the sector. As of
April 2009, 108 governments had made commitments to liberalize their telecommunications markets to various degrees. At the time the
agreement was reached in 1997, the signatories represented more than 90 per cent of international telecommunications traffic.
This section summarizes some key trends in telecommunication reform, covering:
■ Ownership reform,
■ Legislative reform,
■ Liberalization and the introduction of competition, and
■ Regulatory reform.
Ownership Reform
The latter part of the 1980s saw the beginning of dramatic changes in the ownership of many state-owned operators. This wave of
privatization followed the sale of both British Telecom and Cable and Wireless in the United Kingdom in the early 1980s.
In the following decade private sector participation in the telecommunications sector increased dramatically (although but not to the extent
that had been predicted by industry analysts). Rather than full privatization, corporatization of state-owned operators has emerged as the
dominant model for ownership reform.
In general, ownership reform is driven by the goals of improving efficiency, productivity, and service quality. Other factors also come into
play:
■ The political philosophy of the government in power at the time. This was behind some of the early privatizations, such as in the United
Kingdom and Chile,
■ Financial crises. This was behind privatizations in several developing countries in the late 1980s, for example Argentina, Mexico and
(to some extent) Malaysia,
■ The need to raise capital to expand infrastructure. For example this was an important factor in Singapore,
■ The desire to improve technology and management expertise. This has been a driver of reform in several African countries,
■ The intention of the government to open the market to competition, as in the case of Brazil, or
■ A combination of the above factors, as in Bolivia and Peru.
Legislative Reform
Telecommunication specific legislation has a number of benefits. Once written into law, the process of implementing policy objectives
becomes easier. Further, a sound legal infrastructure helps to add clarity and certainty to the sector, and helps to attract foreign
investment.
Early approaches to sector reform in the mid to late 1980s saw the introduction of complementary but piecemeal elements of legislation to
address reform objectives over a period of time. By contrast, countries that began their reforms in the mid 1990s (for example Botswana
and Zambia) issued single new telecommunication laws that overhauled their entire regulatory framework. This is still the most common
approach to legislative reform. See Module 6: Legal and Institutional Framework.
Liberalization and the Introduction of Competition
Historically, many countries have maintained a statutory monopoly in the telecommunications sector. That is, competing providers were not
permitted to enter telecommunications markets. This approach was based on the belief that economies of scale in telecommunications
meant that it would be wasteful for more than one firm to operate in the market.
Now, the international trend is to open telecommunications markets up to competition. Concerns about economies of scale do still exist in
some areas, for example where such economies mean that essential facilities cannot feasibly be duplicated. Where this is the case,
regulators or competition authorities usually require the incumbent to make the facility available to downstream competitors on efficient,
non-discriminatory terms.
Until recently, most countries that introduced competition did so by stages. Competition was first permitted in the “periphery” — services
that were marginal to the incumbent’s operations. For example, this might include markets for telecommunications equipment, such as
handsets and facsimile machines. The market for long distance services was generally opened up to competition before basic local
services. In recent years an increasing number of countries have introduced competition for basic services.
Competition is more common in markets for “new” services, such as mobile telephony and Internet services. For example, the majority of
countries with mobile cellular services have either partial or full competition. (For a more detailed comparison of the level of competition by
sector and region, click here.)
Regulatory Reform
Until recently, the majority of telecommunications operators were state-owned, self regulated entities. Sector reforms have usually been
accompanied by the establishment of separate regulatory authorities for the telecommunications sector.
A separate regulatory authority is important to support liberalization. For competition to develop, new entrants need confidence that the
institution that sets the rules will be independent, and not biased in favour of any particular market participant(s). This is especially the case
where the incumbent continues to be state-owned.
2 Competition Policy and the ICT Sector
This section introduces key elements of competition policy, as it applies to the ICT sector. The section provides information on the following
topics:
■ What we mean by competition, and why it is important

■ The roles of competition policy and regulation
■ Key concepts in competition policy
■ Common forms of anti-competitive conduct
■ Remedies for anti-competitive conduct
■ Approaches to analysing mergers, joint ventures and acquisitions.

There are many different forms of competition.
Many people think about competition in terms of the textbook model of perfect competition. Perfect competition is an ideal model of a
competitive market, but is unlikely to occur in practice.
Markets that are not perfectly competitive can still deliver significant benefits for buyers and sellers. A useful standard for analyzing real
world markets is workable or effective competition. The concept of contestability is also useful for analyzing markets in which there are
few players but market power is constrained by the potential for entry.
The textbook case of perfect competition is an ideal model of a competitive market. Perfect competition rarely (if ever) occurs in practice. It
is more an ideal than a market reality, and so is not useful as a standard for analyzing the performance of real world markets.
Perfect competition requires a number of conditions:
■ The product concerned must be “homogeneous”. That is, the product must have identical attributes and quality regardless of who
buys or sells it;
■ There must be a large number of buyers and sellers for that product;
■ Buyers must be homogeneous and perfectly informed;
■ No single consumer or firm must buy or sell anything more than an insignificant proportion of the available market volume of that
product;
■ All buyers and sellers must enjoy the freedom to enter or exit the market at will and without incurring additional costs;
■ There must be no economies of scale. Economies of scale arise where the average cost of production falls as the volume of
production increases. Where economies of scale exist it is more efficient for a single firm to produce a given volume than for two or
more firms that between them produce the same total volume, as the larger firm;
■ There must be no economies of scope. Economies of scope arise when different products have significant shared fixed costs, so
that a single firm can produce them using a common facility. Where economies of scope exist it is cheaper (and more efficient) to
produce different products out of a common plant or facility than to produce them separately;
■ There must be no externalities. An externality is an unintended side effect (either beneficial or adverse) of an ordinary economic
activity that arises outside the market or price system so that its impact is not reflected in market prices and costs;
■ There must be no regulation of the market or franchise obligations; and
■ There must be no restrictions on capital.
Effective competition occurs in economic markets when four major market conditions are present:
■ Buyers have access to alternative sellers for the products they desire (or for reasonable substitutes) at prices they are willing to pay,
■ Sellers have access to buyers for their products without undue hindrance or restraint from other firms, interest groups, government
agencies, or existing laws or regulations,
■ The market price of a product is determined by the interaction of consumers and firms. No single consumer or firm (or group of
consumers or firms) can determine, or unduly influence, the level of the price, and
■ Differences in prices charged by different firms (and paid by different consumers) reflect only differences in cost or product
quality/attributes.
In effectively competitive markets, consumers are protected to some degree from exploitative prices that firms, acting unilaterally or as a
collusive bloc, could charge. Likewise, firms are protected from manipulation by large individual consumers (or groups of consumers) and
from disruption or interference from other firms.
Competition occurs on the basis of both price and the quality or features of the product. Products are often differentiated, that is they are
not identical across firms. One form of a product is usually a reasonable substitute for another form of that product. This is often referred
to as “functional equivalence”. Sellers may also offer product combinations or bundles that appeal to specific consumers or consumer
segments.
Effective competition can occur even in markets with relatively few firms that differ substantially in size, market share, and tenure.
However, for such markets to be competitive, it is important that there are no barriers to entry and exit.
High firm concentrations in a given market may not translate to market power. Even in markets where only one or a few firms can
efficiently operate (for example due to economies of scale), it is possible for competition to work.
A market is said to be contestable when barriers to entry and exit are so low that the threat of potential entry prevents the incumbent
from exercising market power.
In perfectly contestable markets there are no barriers to entry or exit. With free entry into and exit from the market, the threat of potential
entry will constrain the behaviour of incumbent firms. Should an incumbent firm increase prices above the normal level of profits, then new
firms will enter the market and force prices down again.
Contestability requires that there are no sunk costs for market entry. That is, should an entrant fail, it can recover its fixed costs (for
example by selling assets or reusing them elsewhere).
Competition is a desirable goal not for its own sake, but because of the benefits from competition. These benefits derive from the pressure
competition places on firms to be efficient, innovative and customer focused in order to thrive and survive. They include lower prices,
higher productivity, more service choices, and greater connectivity.
The overall aim of competition policy is to achieve sustainable competition, where competition occurs on a “level playing field” and
consumers and operators are not subject to anti-competitive practices.
This aim faces a number of challenges in the ICT sector.
Competition policy is typically developed and implemented in the context of residual regulations from the monopoly era. In addition, it must be
weighed against other policy objectives, such as consumer protection and the development of a viable telecommunications industry.
In addition the telecommunications marketplace is increasingly volatile. In many developed countries the industry has experienced ups and
downs of financing and development during the last 10 years. This has resulted in spurts of growth in facilities and services deployment,
followed by reductions in service operators and consumer choices and a slowing down of connectivity expansion. This has in turn slowed
down the financing of some viable communications projects in developing countries.
Against this background regulators’ task of fostering the transition to sustainable competition is a complex one. Regulators may be tempted
to micromanage the market to ensure that competition (or a particular form of competition) takes place. Alternatively, they may decide
prematurely that the market is fully competitive. Neither of these paths is likely to result in sustainable competition.
Regulators are faced with a complex balancing exercise. Individual regulatory decisions need to balance:
■ The long term objective of ongoing, sustainable competition,

■ The resolution of immediate short-term concerns, and
■ Conformance with the regulatory and legislative provisions under which regulators operate.

Competition policy and economic regulation are based on the premise that the “public interest” or “social good’ is best served when markets
work efficiently. This generally occurs in a competitive environment.
Competition is the most efficient and equitable mechanism available for organizing, operating, and disciplining economic markets. Competitive
markets distribute resources efficiently and fairly without any need for a single centralized controlling authority. Competition maximizes
benefits to society at large by:
■ Ensuring that resources, products, and services are allocated to the person or persons who value them the most (allocative
efficiency)
■ Forcing market participants to use scarce resources as productively as possible (productive efficiency)
■ Encouraging market participants to innovate, and to invest in new technologies at the best time (dynamic efficiency).
There are numerous examples internationally of the benefits of competition in the ICT sector. For example, in Jamaica, liberalization of the
mobile phone sector led to large increases in the accessibility of telecommunications to consumers, and a jump in total teledensity. Similar
results have been seen in other countries, including Morocco.
RELATED MATERIALS
Module 6, Section 4, Legal and Institutional Impacts of Regulation: Impact of Convergence
This section benchmarks the level of competition in key sectors, worldwide and by region. The analysis compares the level of competition
in:
■ Local service
■ Domestic long distance
■ International long distance
■ Mobile
■ Internet services, and
■ Leased lines
View a summary table of benchmarking data here. This analysis uses data provided by the ITU.
Worldwide Comparison by Sector
The trend towards liberalization is evident in the data. Over 60 percent of the 184 economies for which data are available have either full or
partial competition in the sectors analysed here. Competition in mobile and Internet services is extremely common — 90 percent of countries
have either partial or full competition in the mobile sector, and 93 percent in the Internet services sector.
textHowever, in many countries the provision of fixed services is still a monopoly. Between 35 and 38 percent of countries have
monopolies in the provision of international, domestic long distance and local call services.
Figure 1: Competition by Sector Worldwide

Key: M = Monopoly, C = Competition
Source: ITU World Telecommunication Regulatory Database, 2009. The most up-to-date information can be found on the ICT Eye.
Back to top
Local Service
Monopoly provision of local service is still prevalent, particularly in Africa and the Arab States, where 41 and 52 percent of countries
respectively have a monopoly local service provider.
The data show significant competition in Europe and the Commonwealth of Independent States (CIS), where 87 percent of countries report
full or partial competition in local service. This reflects the significant impact of the European Union’s competition policy and
telecommunications requirements.
Figure 2: Competition in the Local Services Sector

Back to top
Domestic Long Distance
The picture for domestic long distance is very similar to the local service sector. Approximately 40 percent of African countries in the
sample and 50 percent of Arab States have a monopoly in the provision of domestic long distance services. Approximately 80 percent of
countries in Europe and the CIS, and two-thirds of countries in the Americas, report full or partial competition in this sector.
Figure 3: Competition in the Domestic Long Distance Sector

Back to top
International Long Distance
Competition is more widespread in most regions in the international long distance sector than in the local and domestic long distance
sectors. In Africa and the Asia Pacific region respectively, 60 and 61 percent of countries have introduced full or partial competition for
international calls. In the Americas 71 percent of countries have introduced competition, and in Europe and the CIS 85 percent of countries
have full or partial competition in this sector. The Arab States show a high level of monopolization in this sector compared to other regions
(52 percent of Arab States report a monopoly).
Figure 4: Competition in the International Long Distance Sector

Back to top
Mobile
All regions show a high degree of liberalization in the mobile sector. Across the regions surveyed, between 78 and 96 percent of countries
have introduced full or partial competition in the mobile sector. Competition is most widespread in Europe and the CIS and Africa (with
competition in 96 percent and 93 percent of countries respectively).
Figure 5: Competition in the Mobile Sector

Back to top
Internet Services
Unsurprisingly, the Internet services sector is by far the most competitive of the sectors surveyed. Only 12 of the 184 countries in the
sample have not introduced competition in this sector. Over 90 percent of countries in Africa, the Americas and the Asia Pacific region, and
all of Europe and the CIS have either full or partial competition in the Internet services sector. Over 80 percent of Arab States have
introduced competition.
Figure 6: Competition in the Internet Services Sector

Back to top
Leased Lines
Leased lines show a similar pattern to the local, domestic long distance and international sectors discussed above. In most regions, the
majority of countries have introduced some degree of competition. Over half (57 percent) of the Arab States have either full or partial
competition in the provision of leased lines. Again a large majority (87 percent) of countries in Europe and the CIS have introduced full
competition.
Figure 7: Competition in the Leased Lines Sector

RELATED MATERIALS
Module 2, section 2.2.2., Comparison Table: Competition by Sector and Region
The following table summarizes the level of competition in six key sectors, analyzed by region. For each sector and region, the table shows
the number of countries with a monopoly, duopoly, partial competition and full competition. (The total number of countries in each region is
shown in brackets). View benchmarking of competition by sector here.
Table 1: Competition by Sector and Region
Local Service Domestic Long Distance International Mobile Internet Services Leased Lines
Africa (42)
Monopoly 16 17 16 3 3 13
Partial competition & Duopoly 11 11 12 15 4 8
Full competition 12 13 13 23 32 16
Americas (31)
Monopoly 11 12 10 5 3 10
Arab States (21)
Monopoly 11 10 11 4 3 8
Asia Pacific (38)
Monopoly 14 14 15 4 1 10
Europe and CIS (52)
Monopoly 7 12 9 2 2 7
World
Monopoly 59 65 61 18 12 48

Source: ITU World Telecommunications Indicators Database, 2009. The most up-to-date information can be found on the ICT Eye at
http://www.itu.int/ITU-D/ICTEYE/Default.aspx.
RELATED MATERIALS
Module 2, section 2.2.1., Benchmarking Competition by Sector

In practice, many markets do not exhibit all the conditions necessary for workable or effective competition. Market failures occur in many
forms. The two most associated with the need for regulation are:
■ Monopoly, including natural monopoly, and

■ Externalities.
When market failures arise, it is necessary to consider whether the problem is likely to correct itself. If market failures will not correct
themselves, then there may be a need for additional tools to foster effective competition or to prevent socially undesirable outcomes.
This section of the Toolkit introduces two broad approaches to promoting competition in the ICT sector, namely competition policy and
regulation. Competition policy and regulation are not mutually exclusive. Many countries use a mix of both. However, care is required to
ensure that sector regulation and competition laws and policies are developed and applied consistently.
This section discusses the following topics:
■ Competition Policy
■ Regulation
■ Ex Ante and Ex Post Regulation
■ Advantages and Disadvantages of Ex Ante versus Ex Post Regulation
■ Regulatory Forbearance
RELATED MATERIALS
Module 2, section 2.2, Why Focus on Competition?
Module 2, section 3.2.3, Asymmetric Interconnection Regulation
Training on Competition Policy and Interconnection in Telecommunications Market of Mongolia, 4-7 July 2003 - Ulaanbaatar,
Mongolia (contains numerous presentations, reference documents, and summary reports)
Competition policy provides a set of tools to promote sustainable competition and to preserve a market environment in which such
competition can flourish. Competition policy may be implemented through general competition laws or through competition enhancing rules in
specific sectors. In the ICT sector, such rules might include:
■ General prohibitions on anti-competitive behaviour and mergers or acquisitions that would reduce competition (as in the case of Hong
Kong), or
■ Specific rules designed to encourage competition in the sectors, such as interconnection requirements or unbundling policies.
Competition laws (or “antitrust laws”, as they are called in the US) aim to promote efficient competition by penalizing or undoing conduct
that reduces competition in a market. Competition laws generally include provisions to:
■ Prevent competing firms from banding together (“colluding”) to increase prices or reduce quantities of goods and services, or to
exclude other firms from a market,
■ Prevent firms with a dominant position, or “significant market power”, from using their market power to exclude competitors from the
market, or otherwise reduce competition,
■ Stop mergers or acquisitions that would reduce competition.
With the exception of provisions for mergers and acquisitions, competition laws are generally ex post regulation. They give the competition
authority or the courts powers to respond to anti-competitive behaviour once it has occurred.
RELATED MATERIALS
Module 2, section 2.3.2, Regulation
Module 2, section 2.3.3, Ex Ante and Ex Post Regulation
Module 2, section 2.3.4, Advantages and Disadvantages of Ex Ante and Ex Post Regulation
2.3.2 Regulation
Regulation is useful where the market by itself would produce undesirable or socially unacceptable outcomes.
Regulation attempts to prevent socially undesirable outcomes and to direct market activity toward desired outcomes. For example, ICT
regulation is widely used to promote prices that reflect efficient costs and promote universal access to basic services.
However, regulation has potentially high costs. The regulatory process is inherently time consuming to administer and requires
considerable expenditure of resources. In addition, regulation can have unintended consequences which may be detrimental to customers
and the "public interest". No matter how capable and well intentioned regulators are, they will never be able to produce outcomes as
efficient as a well-functioning market.
Accordingly, regulation should only focus on those parts of the ICT sector where there is a clear need for regulation (that is, where
effective competition is not feasible) and should only be a temporary measure. Over time, regulators should aim to establish or restore the
conditions that provide for effective competition on a sustained basis. This entails, for example, removing or reducing barriers to entry and
exit. It also involves enabling the market itself to prevent the incumbent from abusing its market power, for example, through the entry of
additional competitors.
RELATED INFORMATION
Competition Policy
Advantages and Disadvantages of Ex Ante versus Ex Post Regulation
Ex Ante and Ex Post Regulation
Regulatory Forbearance
Asymmetric Interconnection Regulation
Practitioners commonly distinguish between “ex ante regulation” and “ex post regulation.” Various countries have adopted competition
policies that rely, to varying degrees, on mixing elements of these two approaches.
Ex ante regulation
Ex ante regulation is anticipatory intervention. Ex ante regulation uses government-specified controls to
■ Prevent socially undesirable actions or outcomes in markets, or

■ Direct market activity towards socially desirable ends.
Ex ante regulation is mainly concerned with market structure, that is the number of firms and level of market concentration, entry
conditions, and the degree of product differentiation.
Ex ante regulation often takes the form of sector-specific regulation.
Ex post regulation
Ex post regulation addresses specific allegations of anti-competitive behavior or market abuse. Ex post regulation aims to redress proven
misconduct through a range of enforcement options including fines, injunctions, or bans.
Ex post regulation is mainly concerned with market conduct — the behaviour of a firm with respect to both its competitors and its
customers.
Ex post regulation often takes the form of competition laws (or anti-trust laws, as they are called in the US).
RELATED INFORMATION

The table below compares the advantages and disadvantages of ex ante and ex post regulation. In general, competition laws (or antitrust
regulation, as it is referred to in the US) tend to be ex post regulation, while sector regulation is generally ex ante.
Advantages Disadvantages
Ex Ante
Regulation Sets forward looking expectations for firm behaviour. Prevents all conduct of a certain type, regardless of
Avoids damage from anti-competitive behaviour by whether it would actually be harmful. May prevent
anticipating and preventing it potentially beneficial behaviour
Can provide certainty for market participants, by setting Often uses the perfect competition model as a
out clear rules in advance. (This requires good regulatory benchmark, which can lead to unnecessary or
and institutional design that prevents the government or excessive intervention
regulator from changing the rules unpredictably)
Can introduce unforeseen distortions in the operation of
Promotes transparency the market. Asymmetric regulation can encourage
service providers to focus on exploiting opportunities
Eases dispute resolution, as the competition framework is for arbitrage
already established
Imposes high informational requirements on regulators
Regulators and affected parties know in advance the
types of information required for regulatory proceedings, Can be costly. Inevitably involves lengthy regulatory
and can collect it accordingly proceedings
Regulatory processes can be captured by regulated

entities
Ex Post
Regulation Competition laws specify in advance which forms of Ex post regulation is triggered after alleged anti-
conduct are prohibited competitive conduct has already occurred. Does not
prevent harm to competition, only ameliorates it
Attempts to only stop conduct that is shown to be harmful
to the social good. Temporary departures from competition Securing the information needed to enforce ex post
benchmarks (for example due to innovation) are not regulation, from the accused firm, can be difficult
punished without investigation
General competition laws may be unsuitable for
Lower informational and monitoring requirements than ex identifying and penalizing anti-competitive conduct
ante regulation, and therefore lower costs. Competition specific to a certain market
authorities can limit monitoring and information gathering to
firms that are the subject of investigations When applied alongside industry-specific ex ante
regulation, general competition laws can cause
Ex post competition laws apply the same rules across all inconsistencies in regulatory outcomes
sectors, and so should produce consistent outcomes
across sectors Can create uncertainty for firms, particularly firms with
market power. At what point do they cross the line
Ex post regulation is the least disruptive form of regulation between aggressively competitive behaviour and anti-
for emerging markets competitive use of market power?
Competition authorities are less susceptible to capture than

sector specific regulators
RELATED INFORMATION
Competition Policy
Regulation
Ex Ante and Ex Post Regulation

Regulation is not a panacea. While it may address market power concerns, regulation comes with costs. Where it is possible, effective
competition will generally deliver better outcomes than regulation.
Where regulation is necessary, regulatory forbearance is the key to good outcomes. Regulatory forbearance is about focusing
regulation to where it is needed, and withdrawing regulation in those parts of the market where it is no longer necessary. In other words,
the concept of regulatory forbearance rests on the goal of a gradual removal of ex ante regulation and an accompanying increase in the
use of general ex post competition regulation.
The concept of regulatory forbearance has two elements:
■ A regulator may refrain from applying certain regulatory conditions or from intervening in certain markets. For example, the Canadian
Radio-television and Telecommunications Commission has explicitly stated that it will forbear from regulating certain services.
■ A regulator may reduce the scope of regulation or withdraw entirely from regulating specified markets.
RELATED INFORMATION
Regulation

This section introduces some key concepts that underpin competition policy and sector regulation.
The aim of competition policy is to promote sustainable competition. Competition analysis generally asks the question: Will a given
practice, transaction, or business acquisition reduce competition or increase market power in a given market?
Competition analysis follows the following steps:
■ Define the relevant market or markets,

■ Assess the level of competition in the market, without the behavior or act in question, and
■ Assess the level of competition in the market, with the behavior or act in question.
The level of competition in a market depends on the structure of the market, and whether it meets the conditions for effective competition.
Important considerations include:
■ Whether any firms in the market have market power, and the impact of the trade practice or business acquisition in question on market
power,
■ In particular, whether any firms in the market have a dominant position or significant market power in the market,
■ Any barriers to entry or exit and the potential for competition from new entrants,
■ The role of any essential facilities.
The first step in any competition analysis is to define the relevant market.
The purpose of market definition is to determine the boundaries of a given market. Only then will it be possible to analyze the prospects for
competition in the market, opportunities for particular firms to acquire and exercise market power, and implications for consumer welfare.
A market exists where buyers wishing to buy a good or service come into contact with sellers wishing to sell that good or service, so that
transactions occur. For competition purposes, a market includes all those suppliers, and buyers, between whom there is close competition,
that is:
■ All those goods or services that are close substitutes in the eyes of buyers, and
■ All those suppliers who produce (or could easily switch to produce) those goods or services.
The “SSNIP” or “Hypothetical Monopolist” Test
The “SSNIP” or “hypothetical monopolist” test defines a market as:
The smallest group of products and the smallest geographical area in which a hypothetical monopoly could successfully
implement a “small but significant and non-transitory increase in price” (or "SSNIP").
For example, imagine that a hypothetical firm has a monopoly over the supply of the all widgets within a defined geographical area. Could
that firm increase the price of widgets, for example by 5 or 10 percent, and sustain the increased price in the future?
If such a price increase would cause consumers to switch to alternative products or to suppliers in neighboring areas, then the relevant
market includes those products or areas. Similarly, if the price increase would cause other suppliers to start selling widgets in the
geographic area being considered then the relevant market includes those suppliers.
Market Definition and Substitutability
Market definition focuses on the substitutability of differentiated products or services. Whether two differentiated products should be
considered to be in the same market depends on the extent to which they are reasonable substitutes:
■ From the point of view of consumers (are they “functionally equivalent”)

■ From the point of view of suppliers (how easily can firms not already supplying the product or service in question start doing so?)
As well as considering whether products are substitutes based on their product attributes, market definition must also determine the
geographic boundaries of the market. The test for assessing the geographic scope of a market is:
Can a SSNIP for a product in one location substantially affect the price of the same product in another location?
If the answer is “yes”, then the relevant geographic market includes both locations.
Market definition in the ITC sector can be difficult. Effective substitutes may not be only those services supplied by similar
telecommunications carriers (or by carriers at all). For example:
■ Voice and data services are now available from conventional wireline or wireless networks, using either circuit-switched or packet-
switched technologies,
■ Voice mail services are available from telecommunications networks, answering machines, or manned answering services.
Other Dimensions of Market Definition
Market definition may consider other dimensions of the product or service in question, where they are relevant. Other dimensions include:
■ The functional dimension: The relevant level of the production or distribution chain. For example is the market at the wholesale or retail
level?
■ The temporal dimension: The timeframe or timing within which the market operates
■ The customer dimension: The different customer types within a market. For example should large business customers and residential
customers be viewed as separate markets?
2.4.2 Market Power
This section covers the following topics:
■ What is market power?

■ Testing for market power,
■ Dominance and Significant Market Power (SMP).
Defining Market Power
Market power is:
The ability of a firm to raise prices above competitive levels, without promptly losing a substantial portion of its business to
existing rivals or firms that become rivals as a result of the price increase. [1]
Market power is only damaging if the firm concerned abuses its power. Should a firm with market power raise prices above competitive
levels, this can dampen consumer demand, generate efficiency losses, and harm the public interest.
In addition, firms with significant market power or dominance may be able to implement a range of strategies to reduce competition, and
enhance their position in the market.
Testing for Market Power
The starting point in looking for market power is the competitive price level. Pricing above the marginal, or incremental, cost of a service
cannot be regarded per se as evidence of market power. In real world markets the competitive price level will often be higher than
incremental cost. In industries with high fixed costs, such as telecommunications, prices must include mark-ups over incremental costs for
firms to break even across their whole business.
Regulated prices may also be an inappropriate starting point for detecting market power, as they may differ from competitive price levels.
For example, in many countries prices for certain “basic” telephone services are set below their economic cost, to meet universal service
goals. In these circumstances market power cannot, and should not, be inferred by comparing any given firm’s price to the regulated price
level.
For a finding of market power, the price increase must be sustainable. Firms may be able to temporarily increase prices above
competitive levels, for example due to opportunistic behavior or based on innovation. However, in the absence of market power, such price
increases are unsustainable. True market power requires that the firm be able to profitably implement the price increase for a significant
period of time.
A high market share does not necessarily infer market power. Firms may gain high market shares through means other than market
power. A firm’s market share may increase, at least temporarily, due to a successful new invention or better customer service.
Alternatively, a firm may have a high market share for historical reasons. For example, incumbent telecommunications firms were once
monopoly franchises in most countries and have high market shares as a result. As competition emerges, an incumbent's market share
cannot guarantee it the ability to charge prices higher than its competitors.
Market share in itself is neither necessary nor sufficient for market power. Firms with high market shares may be constrained from raising
prices by a range of factors, including:
■ Competition from other suppliers already in the market

■ The potential for competition from new entrants, and
■ The “countervailing power” of customers in the market, for example their willingness to do without the service if the price increases.
Several quantitative measures exist that can help to assess whether a firm may have market power. These indexes include measures
of market concentration (such as the Hirschman-Herfindahl Index), and measures of price such as the Lerner Index.
Dominance and Significant Market Power (SMP)
The mere fact that a firm possesses dominance or Significant Market Power does not by itself imply abuse of that dominance or market
power. However, such firms can raise prices above competitive levels, and may be able to hinder competition.
There is no universally accepted definition of dominance. In general, a firm is considered to be dominant based on its market share. In
some jurisdictions additional factors are also considered in assessing dominance. For example the European Commission also takes into
account:
■ Firm size,
■ The role of any essential facility,
■ Any technological advantages, or privileged access to financial resources,
■ The strength of the countervailing power of consumers,
■ Economies of scale and scope,
■ Barriers to entry,
■ Product differentiation,
■ Potential competition, and
■ The type and availability of sales channels.
The European Commission introduced the concept of Significant Market Power to bring an element of ex ante regulation to competition
policy in telecommunications. The concept of SMP has since been adopted in other jurisdictions.
The European Commission defines Significant Market Power as the ability of a firm to act independently of competitors and customers.
Under the European model, firms that are found to have SMP are subject to additional ex ante regulatory obligations. This allows
telecommunications regulators to impose ex ante regulatory obligations on firms with SMP, such as:
■ Obligations to align interconnection prices with costs,

■ Accounting separation requirements, and
■ Mandatory publication of reference interconnection offers.
Endnotes
[1] See Robert Pitofsky, "New Definitions of Relevant Market and the Assault on Antitrust", Columbia Law Review, 90(7), 1990. Having a
dominant market share, however, is not sufficient for being able to exercise market power.
RELATED INFORMATION
Barriers to Entry
Essential Facilities
In a competitive market, the threat of potential entry is an important constraint on firms already in the market. Should an incumbent firm
increase its price above competitive levels, potential competitors would respond to this opportunity for profit by entering. Competitive entry
would force prices down again. High barriers to entry prevent such competitive entry, and so increase incumbent firms’ market power.
A barrier to entry (typically in the long run) is a cost that a new entrant incurs, but that incumbent firms avoid. This cost asymmetry can
prevent the potential entrant from competing with the incumbent even if its other costs are exactly the same as the incumbent’s, and both
face identical prices. Thus, barriers to entry may prevent entry by otherwise equally efficient competitors.
A barrier to exit is a cost (typically experienced only when exiting the market) that is so prohibitive that it can reduce, or destroy
altogether, a firm’s incentives to enter the market in the first place. Therefore, a barrier to exit may pose a barrier to entry as well.
Barriers to entry may arise due to:
■ Legal barriers: Prior to liberalization it was common to prohibit entry into telecommunications markets. This is still the case in some
countries,
■ Economies of scale and scope: For example, in the telecommunications sector, a new facilities-based entrant may have no choice but
to start out at a relatively large scale of operations, in order to achieve unit costs close to the incumbent’s,
■ High fixed or sunk costs: If an entrant must incur high sunk costs to enter the market, then the entrant must be prepared to absorb
those sunk costs in the event that it fails. However, at the time the new carrier is weighing its prospects and incurring sunk costs, the
incumbent carrier faces none of the same risks or costs (even if it did so at an earlier point in time). This basic asymmetry in their
positions may pose an entry barrier for the prospective new carrier,
■ Essential facilities: If an entrant needs access to an essential facility that is controlled by one of its competitors, this creates a barrier
to entry. The entrant must incur the cost of purchasing access to the facility — a cost not faced by the firm that owns the essential
facility.
RELATED INFORMATION
Market Power
Essential facilities are resources or facilities that have the following properties:
■ They are critical inputs to retail production. Essential facilities are located at the wholesale level of the production chain, and are
essential inputs in the production or supply of the retail product or service,
■ They are fully owned and controlled by vertically integrated incumbent firms. The owner of the facility participates in the retail as well
as the wholesale stage of the market,
■ They are a monopoly. Retail competitors can only acquire an essential facility from the incumbent firm that owns and controls it,
■ It is not feasible, either economically or technologically, for retail competitors to duplicate the essential facility or develop a substitute
for it.
At the wholesale level the incumbent supplies other firms with a critical input, and those firms are dependent on the incumbent for that
input. At the retail level, the incumbent competes with those same firms (see Figure 1). The owner of an essential facility may seek to use
its position to prevent or impede competition, by implementing a “price squeeze” or even refusing to supply the facility.
Figure 1
RELATED INFORMATION
Market Power
Barriers to Entry
2.5 Common Forms of Anti-Competitive Conduct

Telecommunications firms with market power may try to use their position to reduce competition. This section gives an overview of some
common forms of anticompetitive conduct:
■ Abuse of dominance,
■ Refusal to supply,
■ Vertical price squeezes,
■ Cross-subsidization,
■ Misuse of information,
■ Customer lock-in and restrictive agreements,
■ Exclusionary and predatory pricing,
■ Tying and bundling of services.
RELATED INFORMATION
Remedies for Anti-Competitive Conduct
Abuse of dominance occurs when a dominant firm adopts predatory or exclusionary business practices with the aim of eliminating or
substantially lessening competition and excluding competitors. Abuse of dominance may entail:
■ Refusals to deal, for example a refusal to supply an essential facility to a competitor,

■ Exclusive dealing arrangements, in which a seller prevents its distributors from selling competing products or services,
■ Tying and bundling, where a firm sells makes the purchase of one product or service conditional on the purchase of a second product
or service,
■ Predatory pricing, where a firm sets prices below cost in order to force a competitor out of the market,
■ Non-price predation, where a firm adjusts the quality of its product offering to customers with the aim of harming its competitor. For
example, a incumbent might offer an improved level of service to just those customers served by a new entrant.
In 2003, Deutsche Telekom (DT) was found to have abused its dominant position by committing a price squeeze, contrary to Article 82 of
the European Commission Treaty. DT offered local access services at the retail level to end-users and at the wholesale level on an
unbundled basis to competitors. DT was thus active in both upstream and downstream markets. Beginning in 1998, DT had been legally
obligated to provide competitors with wholesale access to its local loops. In its decision finding that DT had abused its dominant position,
the European Commission found that DT charged new entrants higher fees for wholesale access to the local loop than what DT charged
its retail subscribers for fixed line subscriptions. The Commission assessed the margin between DT’s wholesale access prices and the
weighted average price of its corresponding retail services for access (analog, ISDN, and ADSL). Given that wholesale access prices
were higher than the weighted average of the corresponding retail prices charged to end-users, the Commission determined that the price
margin was insufficient for new entrants to compete with DT. The Commission concluded that DT’s pricing practices constituted a price
squeeze. The Commission further concluded that DT’s pricing for local access services deterred new competitors from entering the local
access market and reduced the choice of telecommunications service providers for consumers and suppressed price competition. DT
unsuccessfully appealed this decision to the European Court of First Instance (CFI). For more details about the DT abuse of dominance
case, please see the Practice Note "Vertical Price Squeeze Charge against Deutsche Telekom". A link to this Practice Note is set out
below.
Article 82, European Commission Treaty

Any abuse by one or more undertakings of a dominant position within the common market or in a substantial part of it shall be
prohibited as incompatible with the common market in so far as it may affect trade between Member States.
Such abuse may, in particular, consist in:
(a) directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions;
(b) limiting production, markets or technical development to the prejudice of consumers;
(c) applying dissimilar conditions to equivalent transactions with other trading parties, thereby placing them at a competitive
disadvantage;
(d) making the conclusion of contracts subject to acceptance by the other parties of supplementary obligations which, by their nature
or according to commercial usage, have no connection with the subject of such contracts.
A firm does not need to be dominant (in the sense of possessing a high market share) in order to implement these strategies. However, the
consequences for competition can be particularly severe when the firm concerned is dominant.
RELATED INFORMATION
Remedies for Abuse of Dominance
Incumbent firms often control access to facilities that are essential inputs in the supply of services at the retail level. Competing retailers
depend on the incumbent for access to the essential facility.
In the telecommunications sector, for example, the local loop connecting end customers to the network is often regarded as an essential
facility.
Incumbent firms may attempt to prevent competitors from entering the market by refusing to provide access to an essential facility (see
Figure 1). To encourage competition, many jurisdictions require firms with control over essential facilities to provide access to retail
competitors. Rules may also determine the way in which access prices will be agreed, and procedures for resolving any disputes.
Figure 1: Refusal to Supply an Essential Facility

The figure shows a vertically integrated incumbent firm (red) and a downstream entrant (blue). The incumbent firm controls an essential
input, on which the downstream entrant depends in order to provide services to its customers. The incumbent also competes with the
downstream entrant at the retail level. By refusing to supply the essential input, the incumbent can prevent the downstream entrant from
competing.
RELATED INFORMATION
Remedies for Refusal to Supply and Price Squeezes
To be able to implement a vertical price squeeze, a firm must be vertically integrated, and control an essential wholesale input to the retail
service. A firm implementing a price squeeze offers to supply this essential input to its retail competitors only at a price greatly in excess of
its costs.
The key elements of a price squeeze are:
■ The firm demands a price for the essential facility that is so high that it is not possible for an equally-efficient retail-stage competitor to
operate profitably (or even survive) given the level of retail prices, and
■ The firm does not charge its own downstream operation this high price.
In the extreme, the firm might demand a price for the essential input that is higher than the full retail price of the service.
A vertical price squeeze can only succeed if the essential input has no effective substitutes. If such substitutes are available, the price
squeeze will simply encourage entrants to use the substitute to produce competing retail services.
A price squeeze has a similar effect to a refusal to supply an essential facility. By charging a high price for the essential input, a vertically
integrated firm can reduce the effectiveness of its competitors, or in the extreme force them out of the market (see Figure 1).
Figure 1: Example of a Vertical Price Squeeze

In the figure, an incumbent firm owns an essential input, on which an entrant depends in order to provide service to its customers. Both
firms have the same costs at the retail stage of the market. The incumbent obtains the essential input at incremental cost, but charges the
entrant a price substantially greater than incremental cost. As a result, the entrant’s total costs exceed the retail price for the service, and it
is forced to exit the market.
In 2003, Deutsche Telekom (DT) was found to have abused its dominant position by committing a price squeeze, contrary to Article 82 of
the European Commission Treaty. DT offered local access services at the retail level to end-users and at the wholesale level on an
unbundled basis to competitors. DT was thus active in both upstream and downstream markets. Beginning in 1998, DT had been legally
obligated to provide competitors with wholesale access to its local loops.
In its decision finding that DT had abused its dominant position, the European Commission found that DT charged new entrants higher fees
for wholesale access to the local loop than what DT charged its retail subscribers for fixed line subscriptions. The Commission assessed
the margin between DT’s wholesale access prices and the weighted average price of its corresponding retail services for access (analog,
ISDN, and ADSL). Given that wholesale access prices were higher than the weighted average of the corresponding retail prices charged
to end-users, the Commission determined that the price margin was insufficient for new entrants to compete with DT. The Commission
concluded that DT’s pricing practices constituted a price squeeze. The Commission further concluded that DT’s pricing for local access
services deterred new competitors from entering the local access market and reduced the choice of telecommunications service providers
for consumers and suppressed price competition. DT unsuccessfully appealed this decision to the European Court of First Instance (CFI).
For more details about the DT abuse of dominance case, please see the Practice Note "Vertical Price Squeeze Charge against Deutsche
Telekom". A link to this Practice Note is set out below.
RELATED INFORMATION
Remedies for Refusal to Supply and Price Squeezes
2.5.4 Cross-Subsidization
In the ICT sector, it is common for firms to supply a number of services. Network operators generally sell services in both competitive and
non-competitive markets.
A firm with market power in one area may charge a high price for non-competitive services and use the proceeds to subsidize low prices
for competitive services.
If the firm breaks even overall, a given service receives a subsidy if it does not generate sufficient revenue to cover its total service long
run incremental cost (TSLRIC).
For example take an incumbent firm with market power in the provision of long distance calls. The incumbent could use its market power to
charge high prices to long distance customers, and use the excess revenue to support low prices for internet access and undercut
competing internet access providers.
By cross-subsidizing competitive services, a telecommunications firm can:
■ Ensure that it covers its overall costs, including fixed costs, and
■ Strengthen the firm's competitive position where it matters most, namely in the supply of its more competitive products.
Cross-subsidization will only maximize the firm’s profitability if the resulting gain in market share in the competitive market outweighs the
loss in revenue from the reduced price. This is because the firm could still increase prices for the non-competitive service, even if it did not
subsidize the competitive service. So its next best option would be to increase the non-competitive price and keep the resulting revenue.
RELATED INFORMATION
Remedies for Cross-Subsidization
It is common for vertically integrated firms to sell wholesale products ("essential facilities") to other firms, while competing against those
same firms in retail markets. In this situation the vertically integrated firm can obtain sensitive commercial or business information through its
wholesale transactions that gives it a competitive advantage in its retail activities.
For example, suppose a vertically integrated incumbent firm is the sole source of dedicated access lines needed to provide retail private
line services. Other firms may have no choice but to acquire wholesale dedicated access lines from the incumbent. To complete the
wholesale transaction, the incumbent needs information about the identity, size, and other characteristics of end-users being targeted by
its competitors. It could use this information to target the same end-users with superior service offerings, placing its competitors at a
considerable competitive disadvantage. This would constitute a misuse of information (see Figure 1).
Figure 1: Misuse of Information
RELATED INFORMATION
Remedies for Misuse of Information
2.5.6 Customer Lock-In
Service providers may attempt to “lock in” customers to prevent them from switching to alternative products, technologies, or suppliers.
Customer lock-in involves raising customers’ switching costs to the point that the cost of switching outweighs the potential benefits from
switching.
Switching costs may be:
■ Transactional, for example the cost of replacing existing equipment and technology in order to move to a different service provider,
or
■ Contractual, for example penalties for breaking an existing contract with one service provider, in order to switch to a new service
provider.
Contractual provisions that increase switching costs are not necessarily anti-competitive. Service providers may use contractual
provisions that ensure customer loyalty to recover legitimate underlying costs over a period of time, for example:
■ Service providers may incur substantial upfront fixed costs to acquire and serve customers. For example, it is common for mobile
service providers to subsidize the cost of mobile handsets and recover the cost of the subsidy through service charges over time,
■ Service providers may have incentives to spread non customer-specific fixed costs over as many customers as possible. In order to
do this, a service provider may use contractual provisions to ensure customer loyalty and maintain its installed customer base.
Where the customer’s switching cost is less than the present value of the expected revenue from the customer, competing firms may offer
to pay the customer’s switching cost. In this case, switching costs are not effective as a means of locking in customers.
RELATED INFORMATION
Remedies for Customer Lock-In
Predatory pricing is a pricing strategy used by an established firm to eliminate competition from equally efficient firms, and secure a
monopoly position in a previously competitive market.
A firm practicing predatory pricing lowers its price below cost and maintains it there until equally efficient competitors are forced to incur
unsustainable losses and exit the market. The firm then raises its price to a monopoly level in order to recoup its lost profits.
Predatory pricing is a risky strategy. The firm involved incurs high up-front losses, with no guarantee of future gains from monopolization.
The strategy will only be profitable if, once all competitors have been forced out of the market, the incumbent is able to raise its prices to a
monopoly level and keep them there. If the firm is subject to either direct price regulation or some other form of control, predatory pricing is
unlikely to succeed.
Predatory pricing requires high barriers to entry. If firms are able to enter the market easily, then each time the incumbent increases its price
this will attract new entrants into the market, forcing the incumbent to drop its price again.
A less aggressive type of exclusionary pricing is limit pricing. Limit pricing occurs when a firm with low costs sets prices above its own
costs, but below a potential competitor's costs. This can discourage new firms from entering the market, but may not force existing
competitors out of the market.
For it to succeed, limit pricing may require tacit collusion from all or most existing firms. Existing firms must be willing to reduce the market
price below profit maximizing levels, to the point that any higher cost entrants have no prospect of making a profit.
Limit pricing may only discourage entry by less efficient firms. So even though limit pricing may deter new entry, it does not necessarily hurt
customers or reduce social welfare.
RELATED INFORMATION
Remedies for Predatory Pricing
Tying of services occurs where a service provider makes the purchase of one product or service over which it has market power (the
"tying good") conditional on the purchase of a second, competitively supplied, product or service (the "tied good"). By tying services, a
service provider can try to use market power in one market to give itself an advantage in another, competitive market. Customers who opt
to buy the tied good from a competitor cannot find a feasible substitute for the service provider's tying good.
Tying is primarily a strategy to maximize profits. It can be profitable:
■ Where the demands for the two products are complementary, such that end users consume both products together (for example a
network subscription and local calls), or
■ If the tying good is regulated and the reglated price is below the service provider's profit maximizing level. In this case a successful
tying strategy would enable the service provider to increase its overall profitability by increasing the price of the tied good.
Tying will not be profitable where:
■ The demands for the two products are independent, so that end users are unlikely to consume them jointly,
■ The price of the tying good is already at the service provider's profit maximizing level. In this case there is no room to increase profits
further, or
■ The two products are consumed in fixed proportions. To maximize its profits, all the service provider needs to do is set the price for
the product over which it has market power at its profit maximizing level.
A tying strategy is only likely to exclude competitors from the market for the tied good if competitors are unable to overcome the loss of
sales to customers who have been successfully tied. For example this might be the case if:
■ Competitors face economies of scale, so that a loss of sales causes their average costs to increase, or
■ The tied good is associated with network externalities, so that a loss of sales to some customers causes other customers to drop off
as well.
Even where tying does have an exclusionary effect, this may be an unintended consequence of a strategy to maximize profits.
Service bundling occurs where a service provider offers two or more services separately, but gives a discount to customers who
purchase the services as a combined bundle. Bundling is typically pro-competitive and consumer friendly.
Bundling is common in telecommunications and other multiproduct industries, reflecting both cost savings from producing services jointly,
and consumer preferences for service bundles. In telecommunications, local and long distance services are often bundled with services
such as call waiting, call forwarding, voice mail, or Internet access.
RELATED INFORMATION
Remedies for Tying and Bundling
2.6 Remedies for Anti-Competitive Conduct

The section of this module on common forms of anti-competitive conduct discusses various forms of anti-competitive behavior that may
arise in ICT markets. This section provides an overview of the remedies available to governments and regulators for responding to:
■ Abuse of dominance,
■ Refusal to supply and vertical price squeezes,
■ Cross-subsidization,
■ Misuse of information,
■ Customer lock-in and restrictive agreements,
■ Exclusionary and predatory pricing,
■ Tying and bundling of services.
RELATED INFORMATION
Common Forms of Anti-Competitive Conduct
Abuse of dominance occurs when a firm uses its dominant position in a market to lessen competition in that (or another) market.
The first step in any investigation of alleged abuse of dominance is to determine whether the firm in question has a dominant position, or
significant market power, in the relevant market.
The second step is to consider whether the behaviour in question constitutes an abuse of the firm’s dominant position. Is the behaviour
harmful to competition and to consumers? It is important to distinguish between aggressively competitive behaviour that harms individual
competitors but benefits customers (for example by reducing prices), and behaviour that is anti-competitive.
Figure 1: Responding to Abuses of Dominance

A range of possible remedies exists. Which remedy is appropriate will depend on the specific nature and seriousness of the behaviour,
and the likelihood that the firm may repeat the behaviour in the future.
Directive Remedies
Directive remedies, such as injunctions or bans, require the firm to:
■ Cease its abusive behaviour, or

■ Make specific changes to its behaviour so it is no longer damaging to competition.
Directive remedies may require ongoing monitoring, to ensure that the behavioural change is sustained.
Punitive Remedies
Punitive remedies include:
■ Fining the firm,

■ Ordering the firm to pay compensation to its competitors and/or customers,
■ Fining company officers with direct responsibility for the behaviour.
Punitive remedies are intended to discourage abusive behaviour in the first place by making such behavior unprofitable. However, this
objective must be weighed against the potential to “chill” dominant firms’ behaviour. If the cost of being found to have abused a dominant
position is very high, then dominant firms will “err on the side of caution”. They may not engage in aggressively competitive behaviour, in
case such behaviour is found to be anti-competitive.
Accounting Separation
Accounting separation aims to separate out the competitive and non-competitive parts of the firm’s business, without going to the extreme
of full structural separation.
For example, this can be achieved by requiring the dominant firm to publish a set of regulatory accounts for the non-competitive part of its
business. The objective is to make the costs of non-competitive services transparent so that regulators and others can more easily detect
possible abuses. New Zealand used this approach as part of its “light handed” regulatory regime, prior to 2001. New Zealand's current
regulatory regime also obliges the Commerce Commission to require the incumbent service provider to undertake accounting separation and
to publish information related to its accounts.
Accounting separation is ex ante regulation—it is more concerned with preventing future anti-competitive behaviour than punishing past
abuses.
Structural remedies
If the anti-competitive behavior is very damaging and there is a high probability of repetition, structural separation may be necessary. For
example, this might involve breaking the firm into two competing firms with smaller individual market shares, or separating monopoly and
competitive elements of the firm. A landmark example of structural separation is the United States break up of AT&T in 1984. In the United
Kingdom, Ofcom announced that BT will restructure elements of its business to ensure equality of access to critical infrastructure.
RELATED INFORMATION
Abuse of Dominance
Where a vertically integrated incumbent firm controls a facility that is an essential input to its retail competitors, this can create a
“bottleneck” to competition. The vertically integrated firm may prevent competitive entry by refusing to supply the essential input. Or it may
charge a price for the input so high that it is not possible for competitors to operate profitably, given the level of retail prices.
There are two possible remedies for this essential facility problem:
■ The market may provide a technological solution, by developing feasible substitutes for the facility.
■ The government may require the vertically integrated firm to provide equal access to the essential facility to any firm that requests
access, including competitors. Typically, this means imposing non-exclusion and non-discrimination obligations on the owners of
essential facilities.
Even if the vertically integrated firm agrees to supply the essential facility to its competitors, it may still attempt a vertical price squeeze. A
number of remedies for vertical price squeezes exist, including:
■ Ex ante resale obligations;

■ Ex ante price floors; or
■ Ex post structural remedies.
Figure 1: Remedies for Refusal to Supply and Vertical Price Squeezes
Ex Ante Resale Obligations
Resale obligations require the vertically integrated firm to make its retail services available for resale by any competitor. Competitors gain
access to the wholesale components of the service when they resell the vertically integrated firm’s retail services. This approach is used
in the United States, under the Telecommunications Act 1996.
The generally accepted price rule for resold services is “retail minus” or “avoided cost discount”. Under this rule, the price paid by resellers
is equal to the retail price of the service, less the cost resellers avoid by substituting their own retailing functions for the vertically
integrated firm’s.
Not all competitors are interested in using resale as their retail market strategy. Alternative protections against price squeezes may be
needed.
Ex Ante Price Floors
A price floor sets a minimum retail price for the incumbent’s retail service, with reference to wholesale prices. A price floor should ensure
that competitors that as efficient as the vertically integrated firm are able to cover their costs. The rule for setting a price floor, or
"imputation rule" can be stated in a number of ways:
The retail price must be no less than the wholesale price plus the direct incremental cost of the vertically
integrated firm’s pure retailing functions.
The retail price must be no less than the vertically integrated firm’s wholesale price, plus the direct incremental
cost of the vertically integrated firm's pure retailing functions, plus the difference between the firm’s direct
incremental cost to provide the wholesale facility to itself and its direct incremental cost to provide that same
facility to its competitors.
The retail price must be no less than the vertically integrated firm’s direct incremental cost to supply the product,
plus the profit margin it could earn from selling the essential input to its competitors.
The profit margin on the vertically integrated firm's price for the retail product must be no less than the profit
margin it earns from selling the essential input to its competitors.
The above imputation rules are equivalent, but provide different insights into the conditions that must hold for a vertical price squeeze to be
impossible.
Ex Post Structural Remedies
Structural remedies seek to separate the wholesale and retail operations of the vertically integrated firm, to remove the opportunity for a
price squeeze, through:
■ Functional or accounting separation of the firm’s wholesale and retail operations, or

■ Full structural separation of the firm’s operations (by divesting either the wholesale or retail operation).
These measures may achieve the objective of preventing a price squeeze, but they can have substantial costs. In particular, under
structural separation the firm would lose any efficiencies or cost savings from vertical integration. This loss would ultimately fall on
customers, through higher prices.
RELATED INFORMATION
Refusal to Supply
Vertical Price Squeeze
2.6.3 Remedies for Cross-Subsidization
A firm with market power in one market may charge a high price for non-competitive products and use the proceeds to subsidize low
prices for competitive products.
The remedies for cross-subsidization are preventative in nature:
■ Implement and enforce a price floor,

■ Require accounting separation of the costs of the firm’s competitive and non-competitive products.
Figure 1: Remedies for Cross Subsidization
Price Floor
For a firm that at least breaks even across all of its products, any single product receives a subsidy if the revenue it generates fails to
recover its total service long run incremental cost (TSLRIC). Thus, the effective price floor in a test of whether a product receives a
subsidy is:
TSLRIC of the service / number of units produced
For a multiproduct firm, the rule for preventing cross-subsidization requires that, for a firm that at least breaks even, every product must
satisfy this price floor test.
Accounting Separation
The objective of accounting separation in this context is to separate the costs of the firm’s competitive and non-competitive products. This
can be achieved through price regulation (either direct regulation, or a “price cap”). Such regulation can prevent cross-subsidization by
allocating competitive and non-competitive products to separate “baskets”, with separate controls or rules for the each basket.
RELATED INFORMATION
Cross-Subsidization
It is common for a vertically integrated firm to supply an essential wholesale facility to other firms against which it competes at the retail
level. The firm may obtain commercially sensitive information in the course of providing the wholesale service, which it may use at the retail
level for marketing purposes. This can place a potential entrant at a substantial competitive disadvantage.
Figure 1: Responding to misuse of information
Remedies for misuse of information are generally ex ante in nature, and include:
■ Establishing strict rules or procedures governing the use or disclosure of commercially sensitive information, and setting limits on the
sharing of sensitive information between a carrier and its affiliates
■ “Win back” rules, limiting the extent to which the vertically integrated firm may directly market to customers that choose to switch to a
competitor.
RELATED INFORMATION
Misuse of Information
2.6.5 Remedies for Customer Lock-In
High switching costs and customer lock-in tactics do not necessarily cause problems for competition or exclude competitors. Most service
agreements that seek to lock-in customers do not warrant regulatory interference. Indeed, in some cases, high switching costs may trigger
market responses that improve efficiency.
Cases of lock-in need to be considered on a case by case basis, taking account of:
■ The degree of competition in the market,

■ Whether the firm in question has market power, or a dominant position, and
■ The effect of the locking-in arrangements on competition. Are the arrangements blocking efficient competitors?
RELATED INFORMATION
Customer Lock-In
A firm engages in predatory pricing by temporarily pricing below cost in order to force its competitors out of the market.
Predatory pricing is notoriously difficult to prove. It can be difficult in practice to distinguish predatory pricing from aggressively competitive
below-cost pricing (such as “loss leaders” and promotional activities).
Establishing whether predatory pricing has taken place requires that two tests be met (see Figure 1):
■ Is the firm pricing below cost? And

■ Whether the firm has an “objectively reasonable expectation” of being able to recover the losses it must incur by pricing at below
cost.
Figure 1: Remedies for Predatory Pricing
Is the Firm Pricing Below Cost?
There is no universally accepted test to determine whether a firm is pricing below cost.
Under the Areeda-Turner rule, prices must be below a firm’s short run marginal cost to qualify as predatory pricing. Recognizing that short
run marginal cost is very difficult to measure, alternative short run measures of cost may be used (short run average variable cost,
SRAVC, or short run incremental cost, SRIC).
Many economists promote the use of long run incremental cost (LRIC) as the appropriate cost threshold for predatory pricing. If two firms
are equally efficient, they must have the same long run incremental cost. When one of them sets a price below LRIC, the other firm cannot
match that price without incurring a loss.
Regardless of the measure used, calculations of firm-specific costs for individual services can be highly contentious.
Does the Firm Expect to Recover its Losses?
Many practitioners are skeptical about the prospect that a firm could know in advance all of the information needed to implement a
predatory pricing strategy. In order to have a reasonable expectation that the strategy will succeed, the firm must know:
■ How long it must price below cost before it succeeds in forcing its competitors out of the market,
■ The size of the loss that it must withstand while predatory pricing is in effect, and
■ The probability that it will recover its losses once it has achieved a monopoly.
Remedies
Ex post antitrust remedies, such as fines or compensation, may be available for proven instances of predatory pricing. However,
predatory pricing is difficult to prove with sufficient certainty to justify punitive measures.
A more useful remedy for predatory pricing is an appropriate price floor for the affected product or service. This is a preventive remedy,
requiring ex ante regulation.
RELATED INFORMATION
Exclusionary or Predatory Pricing
There are few circumstances in which tying can be profit-enhancing for the firm concerned. Accordingly firms with market power will
often have no incentive to engage in a tying strategy.
In recognition of this, the courts in the United States have developed a four-part test for analyzing allegations of tying (see Figure 1).
Figure 1: Test for Alleged Tying
In addition to the tests illustrated in Figure 1, some courts require that the alleged harm exceed any efficiencies produced by the alleged
tying, before allowing a complaint to proceed.
Bundling is generally a pro-competitive, and customer friendly, strategy. As such bundling does not call for regulatory intervention.
RELATED INFORMATION
Tying and Bundling

Mergers, acquisitions, and joint ventures are all different ways for two or more firms to integrate or coordinate their operations:
■ A merger is a structural fusion of two firms that results in a common ownership and management structure. Mergers usually happen
through stock swaps.
■ An acquisition is a type of merger in which a firm with more resources and greater market strength may acquire another firm. The
acquiring firm usually uses some combination of stocks, debt, and cash to finance the transaction.
■ A joint venture is a strategic alliance between two firms that share resources, equity, revenues, expenses, and management to
pursue a common goal. Each firm usually retains its own corporate identity.
There are three types of mergers: horizontal, vertical, and conglomerate. Conglomerate mergers occur between firms operating in separate
markets. As such they do not generally raise competition concerns and are not covered further in this section.
Mergers, acquisitions, and joint ventures are motivated by a range of factors such as cost savings from synergies between the firms or
economies of scale and scope, efficiencies from vertical integration, or geographical diversification or cross-selling of products.
This section discusses common approaches to analyzing:
■ Horizontal mergers,
■ Vertical mergers, and
■ Joint ventures.
The Role of Competition Authorities and Regulators
Provisions governing mergers and acquisition are generally included in competition or antitrust laws, where these exist. In this case,
investigation of proposed mergers is usually the responsibility of a competition authority.
Some countries with no competition law have included sector specific merger provisions in their telecommunications laws (for example
Hong Kong).
In countries with both a competition authority and a telecommunications regulator, both agencies may have a mandate to investigate
mergers in the telecommunications sector. For example, in the US the Federal Trade Commission and the Justice Department have a general
responsibility to investigate potentially anti-competitive mergers. However, the Federal Communications Commission may also investigate
horizontal mergers between telecommunications firms to determine whether or not the merger is “in the public interest”.
A horizontal merger brings together firms that produce the same product within the same market.
Horizontal mergers can be either beneficial or detrimental overall. By definition, horizontal mergers reduce the number of actual competitors
in the market. Horizontal mergers may also produce cost savings and other benefits. If these benefits outweigh any reduction in
competition, then the merger should be allowed to proceed.
Analysing Horizontal Mergers
Competition authorities commonly take a two-stage approach to analysing horizontal mergers (see Figure 1).
Figure 1: Two Stage Process for Analysing Mergers
The first stage uses measurable thresholds or “safe harbors” to determine whether a merger is likely to raise serious competition
concerns. If a merger falls within the specified threshold then it is considered to be “safe”, and may proceed without further investigation.
For example, in the United States, antitrust authorities set thresholds based on the change in market concentration from a proposed merger.
In Europe, the Merger Control Regulation applies only to mergers, acquisitions, and joint ventures that satisfy thresholds based on the
turnover of the firms involved.
The purpose of these thresholds is to focus resources on investigating those transactions that are most likely to raise serious competition
concerns. Those mergers that do not fall within specified safe harbors are investigated in depth.
A full merger investigation should consider a range of factors to determine whether the merger would increase market power, and to
evaluate any offsetting benefits. Relevant factors include:
■ Technological change and dynamic efficiencies that would result from the merger,
■ Cost savings and other efficiencies claimed by the merging firms,
■ The ease of market entry, or existence of any barriers to entry,
■ The potential for collusion among firms in the market following the merger,
■ The possibility that the merged entity may act anti-competitively,
■ Whether one or both of the merging firms are likely to survive or fail if the merger does not proceed,
■ Whether the merger would eliminate any potential competitors,
■ Whether customers in the market have “countervailing power” that would constrain the merged entity.
Remedies
If a merger is found to substantially reduce competition, or give the merged entity a dominant position in a market, the first step is to evaluate
any benefits from the merger. If the merger is likely to generate benefits that outweigh the damage to competition, then it should be allowed
to proceed.
In some jurisdictions regulatory authorities may impose ex ante obligations on a merged firm, where the merger would otherwise be anti-
competitive. For example, in both the United States and Europe, National Regulatory Authorities may impose conditions on a merger that
would otherwise be anti-competitive.
A vertical merger brings together firms in potential customer-supplier relationships, such as that between a firm that provides wholesale or
intermediate products to a firm that produces retail or final products.
Vertical mergers are generally considered beneficial. Vertical mergers can:
■ Reduce transaction costs by streamlining the process of acquiring and converting inputs into outputs,
■ Improve efficiency through more integrated production, and
■ Eliminate the potential for a “double markup”, which can occur where there is market power at both the wholesale and retail stage of
the market.
Vertical mergers may raise competition concerns in limited sets of circumstances.
A vertical merger may “foreclose” the market by preventing non-integrated retail competitors from staying and competing in the market.
Foreclosure generally requires pre-existing market power at one or more levels in the new vertically integrated firm. For example suppose
that a firm controlling an essential facility at the wholesale level merges with a retailer. The merged firm may withhold supply of the
essential facility to its retail competitors, preventing them from competing.
Alternatively, a vertical merger may be motivated by the goal of raising rivals’ costs. For example, suppose a retail firm merges with the
supplier of a wholesale input. By removing a source of supply from the wholesale stage of the market, the retailer is able to increase the
price of the input to its competitors (but not itself).
Analysing Vertical Mergers
Analysis of vertical mergers focuses around the two areas of concern above. Competition authorities in the Unites States typically pay
attention to three issues (see Figure 1). Could the merged firm:
■ Raise the costs of its retail rivals? If it can, the remedy is a requirement that the wholesale resource be made available at non-
discriminatory prices.
■ Misuse competitively sensitive information gathered about rivals when selling them the wholesale resource? If it can, the remedy is to
implement rules and procedures to prohibit information-sharing between the firm’s retail and wholesale operations.
■ Foreclose retail competitors from the market by exercising market power at the wholesale stage of the market? If it can, the remedy is
to require the merged firm to provide equal access to the wholesale resource to its non-integrated retail-stage competitors.
Figure 1: Analysing Vertical Mergers

Joint ventures can have many different objectives, and have different implications for competition.
Joint ventures with the purpose of fixing prices, restricting output, or allocating markets between firms reduce competition, and generally
should not be permitted.
Joint ventures may generate efficiency gains and cost savings. In this case, regulators or competition authorities should consider whether
the joint venture will increase market power sufficiently to cause a substantial lessening of competition. Will the joint venture lead to an
increase in prices or a reduction in output? If the potential gains from the joint venture outweigh any competitive damage, then the joint
venture should be allowed to proceed.
In some cases joint ventures include an agreement for the parties to acquire assets or voting rights in their respective firms. This type of
arrangement is more durable than a conventional joint venture, and so requires additional scrutiny. The investigation should consider
factors such as:
■ The level of competition in the relevant market,

■ The number and power of competitors in the relevant market,
■ The market power of the parties in the joint venture,
■ The background of, and the relationship among, the parties in the joint venture,
■ The setting in which the joint venture was created,
■ The relationship between the lines of commerce of the joint venture and of the individual parties in the joint venture.
Telecommunications Joint Ventures
Telecommunications joint ventures come in many forms. They may have one or more of the following objectives:
■ Integration of operations at one or more stages of the production process,

■ Pooling of diverse resources and talents in order to conduct research and development, or
■ Building efficient marketing and sales channels.
Telecommunications joint ventures raise three broad types of competition concern:
■ The potential for collusion among the parties in the joint venture,
■ A loss of potential competition, and
■ The potential for market exclusion and access discrimination.
Ultimately, regardless of the benefits they produce for the collaborating parties, joint ventures must deliver consumer benefits and limited (in
both duration and scope) integration in order to enhance the public interest.
3 Regulating For Interconnection
The purpose of this section is to set out the main issues related to interconnection. Topics covered include:
■ A general overview of interconnection: what is meant by “interconnection”, why interconnection is important, and why
interconnection regulation is needed,
■ Key concepts, such as one-way and two-way interconnection, unbundling, and asymmetric regulation,
■ Setting prices for interconnection,
■ Mobile interconnection,
■ Challenges and opportunities for developing country regulators, and
■ Cross-border interconnection.
Parts of this section are based on Timothy J. Tardiff, "The Economics of Access and Interconnection Charges in Telecommunications", in
Michael Crew and David Parker, eds., The International Handbook of Economic Regulation, Cheltenham: Edward Elgar, 2006, Chapter 13.
The Legal module also covers interconnection in Section 4.3.3.

RELATED INFORMATION
Economic and Accounting Measures of Cost

Useful Economic Cost Concepts

There are many situations in the ICT industry in which networks must be linked with each other in order to provide services to customers.
This section considers three questions:
■ What is interconnection?
■ Why is interconnection important? and
■ Why is regulation of interconnection sometimes necessary?
The World Trade Organization defines interconnection as:
Linking with suppliers providing public telecommunications transport networks or services in order to allow the users of one
supplier to communicate with users of another supplier and to access services provided by another supplier, where specific
commitments are undertaken. [1]
As technology has changed and competition has intensified, many forms of interconnection have developed. All involve the linking of
networks to enable customers of one network to communicate with customers of another network or to have access to services offered
by another network operator. Examples of interconnection include:
■ Two adjacent, non-competing telephone networks interconnect so that subscribers on one network can call those on the other (see
Figure 1).
■ Long-distance carriers obtain access to the facilities of a local service provider and compete against that provider in providing long-
distance services to a common customer base (see Figure 2).
■ Traditional wireline telephone and new wireless mobile carriers interconnect so that subscribers of the traditional phone service can
call wireless subscribers, and vice versa (see Figure 3).
■ New competitive local telephone carriers interconnect with the incumbent carrier so that they can attract subscribers in the common
service territory, and enable those subscribers to call subscribers on the incumbent’s network. Such competitive local carriers may
also lease specific network elements from the incumbent (see Figure 4).
■ Customers of the incumbent telephone carrier make calls to their dial-up Internet Service Provider, which in turn is a customer of a
competing local carrier (see Figure 5).
■ Firms offering a service in which part of the call is routed by Voice over Internet Protocol (VoIP) interconnect with traditional local
service providers to complete the call (see Figure 6).
Endnotes
[1] Telecommunications Services: Reference Paper, 24 April 1996. Downloaded from

http://www.wto.org/english/tratop_e/serv_e/telecom_e/tel23_e.htm
RELATED INFORMATION
Forms of Interconnection
Why is Interconnection Important?
Why Regulate Interconnection?
FIGURES
Figure 1: Adjacent Telephone Networks
Figure 2: Long Distance Carrier and Vertically Integrated Incumbent
Figure 3: Wireline Carrier and Mobile Carrier

Figure 4: Competitive Local Carrier and Incumbent Location Carrier (Common Service Territory)
Figure 5: Dial-Up ISP Connected to a Competing Local Carrier

Figure 6: Competing Operator Routes Calls Using VOIP
ICT service providers need access to networks owned by others in order to provide services to their customers. Without interconnection,
a customer cannot call subscribers on other networks, access Internet content located on another network, and so on.
Networks interconnect with each other for a number of reasons:
■ To provide a service that is not economically feasible without interconnection, for example calls to customers on another operator’s
network
■ To increase profitability. Where interconnection increases the value of telecommunications services, or the range of services
operators can provide, it can be in the mutual interest of the operators to interconnect
■ To expand or improve services that are valuable to customers.
Interconnection has been important for telecommunications providers since the invention of the telephone. Even before competition
emerged, adjacent carriers interconnected with each so that their customers could make long distance and international calls.
With recent technological developments the range of services that depend on interconnection has increased. Interconnection is an
essential input to local, long distance and international fixed voice calls, mobile voice and data services, satellite services, Internet access,
e-mail and message services, broadband data transmission, and a wide range of multi-media services.
RELATED INFORMATION
What is Interconnection?
Why Regulate Interconnection?
Telecommunications operators will interconnect voluntarily in some circumstances. If two operators are not in direct competition with each
other, then generally they will have an incentive to interconnect. This is because interconnection increases the value of a network to its
subscribers, by increasing the number of people they can call and the range of ICT services they can access (network externalities).
Sometimes incumbent operators will have little incentive to allow access to their network, or to allow access on reasonable terms. Where
the interconnection seeker is a potential competitor, an incumbent may seek to limit competition, and preserve its market power, by:
■ Refusing to interconnect
■ Offering interconnection at a price, or on other terms, that make it difficult for an efficient entrant to compete, or
■ Seeking to “sabotage” the entrant by providing a lower quality interconnection service to the entrant than the incumbent provides
itself.
In these cases regulatory intervention can lead to a more efficient outcome. The motivation for interconnection regulation is that efficient
competition in “downstream” markets would be difficult, or even impossible, unless entrants can access the incumbent’s network at
appropriate prices, terms and conditions.
For example, the European Union’s interconnection directives allow National Regulatory Authorities to impose interconnection or unbundling
obligations on carriers that have significant market power, in:
“situations where the national regulatory authority considers that denial of access or unreasonable terms and conditions having a
similar effect would hinder the emergence of a sustainable competitive market at the retail level, or would not be in the end-users’
interest.”[1]
In any market, regulation needs to be able to adapt to changing circumstances. This is especially important in the ICT industry, where
outdated regulation risks stifling market growth and innovation.
Endnotes
[1] “Directive 2002/19/EC of the European Parliament and of the Council of 7 March 2002 on access to, and interconnection of, electronic
communications networks and associated facilities (Access Directive)”, Official Journal of the European Communities, Article 12.
RELATED INFORMATION
Unbundling
Setting Interconnection Prices
Regulation
3.2 Key Concepts
This section introduces several key concepts in interconnection:
■ One-way and two-way interconnection

■ Unbundling, facilities sharing and co-location
■ Asymmetric interconnection regulation,
■ Other issues dealt with in interconnection agreements.
There are two broad forms of interconnection: one-way interconnection and two-way interconnection.
One-way and two-way interconnection can co-exist. For example, new entrants often obtain parts of their networks from the incumbent
carrier (one-way interconnection), and then exchange traffic with the incumbent (two-way interconnection).
One-Way Interconnection
One service provider or carrier must obtain inputs from another carrier in order to offer services to its customers (Figure 1 provides an
illustrative example). The carrier supplying the inputs may or may not compete with the firm purchasing the inputs.
Figure 1: Example of One-Way Interconnection
For example, prior to 1996, local exchange carriers in the United States were prohibited from offering long-distance services. Long-
distance carriers such as AT&T, Sprint and MCI obtained access from these local exchange carriers, to offer long-distance services to
customers on the local exchange network.
Payment for one-way interconnection is always from the interconnecting operator (in the example above, the long-distance carrier) to the
interconnection provider (the local exchange carrier).
One-way interconnection also occurs in other industries, for example when rail operators seek access to rail networks.
Two-Way Interconnection
Two or more carriers must connect their facilities (networks) so that customers of one carrier can call customers served by other carriers
(and vice versa).
Figure 2: Two-Way Interconnection

There are several approaches to structuring interconnection payments for two-way interconnection, which are discussed here.
Two-way interconnection also occurs in other industries. For example, credit cards such as VISA and MasterCard are provided over
interconnected networks of member banks and participating merchants. Cardholders, member banks and merchants pay fees to access a
credit card network.
The operator’s willingness to interconnect will vary depending on the form on interconnection and, in particular, whether the
interconnecting parties are competitors.
RELATED INFORMATION
WHAT IS INTERCONNECTION?
PRICING PRINCIPLES
3.2.2 Unbundling
This section addresses the following questions:
■ What is unbundling?
■ Why should regulators require unbundling?
■ How much unbundling should be mandated?
■ What are the costs and benefits of unbundling?
What is Unbundling?
Unbundling is the mandatory offering by network operators of specific elements of their network to other operators, on terms approved
by a regulator or sanctioned by a court.
Unbundling goes further than imposing an obligation on incumbents to offer interconnection services to entrants. It requires the incumbent
to allow entrants to lease certain individual building blocks that make up a telecommunications network.
Unbundling of network elements allows competing operators to enter the market and roll out services with considerably less sunk
investment in some or all components of a competing network. For example:
A new entrant might initially install switches in central business districts only, and lease those components of the incumbent carrier’s
network needed to directly serve customers in other areas, or
An entrant might lease just those network elements needed to offer competing retail services (such as DSL services). In this way the
entrant can offer competing services to customers without duplicating all components of the incumbent carrier’s infrastructure, and without
simply reselling the incumbent’s service offering.
Unbundling usually requires facilities sharing or collocation, where the incumbent operator houses the communications equipment of
competing operators to facilitate connectivity, or permits entrants to share infrastructure such as cell-site masts, cable ducts, or telephone
poles. One example of facilities sharing is the policy adopted by the Malaysian Communications and Multimedia Commission, whereby the
operators granted 3G licenses have agreed to share their infrastructure with mobile virtual network operators.[6] Infrastructure sharing is
intended to facilitate improved coverage and service by allowing operators to share the risks of investment into cheese grater economies in
the utilization of fixed network assets. However, operators are reluctant to share network assets that they view as strategic.
Many countries have implemented unbundling of their telecommunications networks. As of late 2004, 65 ITU member nations had required
local loop unbundling, up from just 23 in 2000.[1]
Why Require Unbundling?

The rationale for unbundling is similar to that for interconnection regulation more generally.
Some inputs are available only from certain network operators, and cannot easily be duplicated. Unless those inputs are available at
appropriate prices, competition in downstream telecommunications markets would be difficult or impossible.
The emergence of competition from alternative technologies — such as wireless, cable telephony, and VoIP — is eroding this rationale for
mandatory unbundling.
Unbundling can be an enormous task for regulators. The administrative costs of defining, and setting prices for, a range of network
elements can be high. In addition, unbundling can impose high compliance costs on incumbent carriers. Regulators should carefully
consider the merits of unbundling on a case-by-case basis, with a thorough assessment of the likely costs and benefits.
How Much Unbundling?
There are a range of options for unbundling interconnection services.
Under full unbundling, the incumbent must offer a separate fully unconditioned local loop service. This provides access to raw copper
local loops, and subloops.
Under shared access the incumbent must provide access to the non-voice frequencies of a local loop and/or access to space within a
main distribution frame where DSLAMs and similar types of equipment can be interconnected to the local loop.
Under bitstream access for high-speed access services, the incumbent must furnish and lease to other carriers links capable of
providing high speed services.
The extent of unbundling has significant effects on the development and nature of telecommunications competition. If there is not enough
unbundling, entry by efficient competitors may be inhibited. If there is too much unbundling:
■ Entrants may focus on arbitrage opportunities, by obtaining services at attractive wholesale prices and reselling them to customers,
instead of designing innovative product mixes that give customers greater choice
■ Entrants may delay investing in infrastructure and focus instead on expanding re-bundled services as quickly as possible
■ Incumbents may have fewer incentives to invest in unbundled parts of the network. This can lead to inadequate capacity, lower
quality, and slower development of new technology (such as high capacity broadband).
There is a recent trend towards unbundling only those elements of a network that can be considered part of a natural monopoly.
In the United States, the Telecommunications Act 1996 requires all telecommunications carriers to interconnect to exchange traffic. The
Federal Communications Commission’s initial approach was to require incumbent local exchange carriers to unbundle extensively. It has
since narrowed its approach to require unbundling of a more limited set of network elements. For example, incumbent local exchange
carriers are no longer required to unbundle switching equipment.
Some jurisdictions require incumbent operators to only unbundle network components that are essential facilities. For example, the
Canadian Radio-television and Telecommunications Commission used the essential facilities approach in when it required unbundling of
local loops but not end-office switching, as switches were competitively supplied.
The ITU has developed guidelines for the West African Common Market that recommend that dominant operators (typically incumbent
carriers with significant market power (SMP) should be required to provide new entrants with access to copper pairs (full local loop
unbundling).[2] The guidelines suggest that unbundling begin with shared access with full unbundling scheduled for a later stage. The
guidelines also note that bitstream access may be an attractive option for Internet Service Providers because it does not require collocation.
Costs and Benefits of Unbundling
There is considerable debate over the costs and benefits of unbundling. The table below summarizes the potential costs and benefits of
unbundling, as put forward by regulators and incumbent carriers. The magnitude of these costs and benefits will vary depending on:
■ The form of unbundling, and

■ Whether regulated prices for unbundled network elements reflect economic costs.
Benefits Costs
Increases, and brings forward, entry by reducing entry Potentially high administrative and compliance costs (costs increase with
costs the extent of unbundling)
Increases competition in the provision of services May reduce incentives for incumbents to invest in new infrastructure.
supported by the existing network Enables incumbents to obtain legislative and regulatory relief, by making
investment in next generation networks contingent on such relief
Can bring forward the introduction of new services that
rely on the incumbent’s network technology (such as May reduce incentives for entrants to invest in new infrastructure.
DSL services) and competition in those services Entrants may focus on reselling the incumbent’s services, instead of
designing innovative new service offerings
Functional Separation
One possible safeguard that has been discussed and proposed at various times and places over the years is to require a “functional
separation” be affected for operators that are required to provide wholesale inputs to competitors.[3] By “functional separation” is meant
that separate business units with separate accounting are created for the firm’s retail offerings and wholesale offerings. The wholesale
business unit would sell to the retail business unit on the same terms and conditions as to competitors for the retail services. This idea
could find application in situations where infrastructure competition is not likely to develop soon and, thus, the best hope for competition in
the near term is service competition. The main advantage of a functional separation safeguard is that it would be enable it to be seen more
clearly if the retail business unit is profitable while paying the interconnection or unbundled elements charges that the retail competitors
must pay. However, this advantage may be able to be obtained by less dramatic means short of actual separation, through the use of
accounting or imputation tests to see if retail services are profitable. A disadvantage of functional separation is that the wholesale entity
charged with operating the actual infrastructure that all competitors are using may not perceive itself to have strong incentives to invest in
greater coverage and better technologies. However, this disadvantage may come more from the requirement to share network elements
with competitors and not necessarily so much from the separation requirement itself.
One operator that has begun to implement this concept is British Telecom with its OpenReach subsidiary. OpenReach's website describes
the degree of separation it has from BT as follows:
■ Separate disclosure of financial results

■ No BT Group element to Openreach incentive plans
■ Headquarters team in separate accommodation
■ Introducing separate operational support systems
■ Strict Code of Practice to be followed by all employees
■ Strict rules about sharing information in an equivalent way with ALL Communications Providers
■ Own identity (the Openreach wordmark)[4]
A much earlier version was proposed in 1993 by Rochester Telephone in the United States but ended when the company was purchased
by Global Crossing in 1999.[5]
Endnotes
[1] ITU Trends in Telecommunications Reform 2004/2005, page 13.
[2] ITU West African Common Market Project: Harmonization of Policies Governing the ICT Market in the UEMOA-ECOWAS Space,
Interconnection, pages 31-32.
[3] See,
Viviane Reding, Member of the European Commission responsible for Information Society and Media, "From Service Competition to
Infrastructure Competition: the Policy Options Now on the Table", ECTA Conference 2006, Brussels, 16 November 2006. [pdf, 102 KB]
[4]
OpenReach’s code of practice [pdf, 329 KB]
[5] See,
Organisation for Economic Co-operation and Development, Working Party on Telecommunication and Information Services Policies, THE
BENEFITS AND COSTS OF STRUCTURAL SEPARATION OF THE LOCAL LOOP, November 2003. [pdf, 216 KB]
[6] More information on Malaysia's infrastructure sharing policy is available from the Malaysian Communications and Multimedia Commission
GUIDELINE ON REGULATORY FRAMEWORK FOR 3G MOBILE VIRTUAL NETWORK OPERATORS, 16 February 2005. [pdf, 169 KB]
RELATED INFORMATION
Overview of Interconnection

Interconnection regulation can apply equally to all telecommunications carriers (symmetric regulation) or to incumbent carriers only
(asymmetric regulation).
Asymmetric interconnection regulation is very common. The rationale for asymmetric regulation is to redress the consequences of market
power. Asymmetric regulation does this by placing additional requirements on incumbent or dominant operators that might otherwise be
able to prevent or deter competition.
For example, United States and Canadian regulators impose an interconnection obligation on all firms classed as telecommunications
carriers. However, only incumbent firms are required to unbundle and share network components.
In Europe, the European Union requires National Regulatory Authorities, in markets that are not effectively competitive, to impose regulations
on those operators with significant market power.
Asymmetric regulation can be useful in addressing existing imbalances in ICT markets. However, the need for asymmetric regulation should
be kept under regular review. As market conditions change, new firms enter the market, and new competitive services emerge, market
power can be eroded. Where this occurs, regulators need to reconsider the justification for asymmetric regulation and, if market power is
no longer a concern, remove the additional requirements.
RELATED INFORMATION
Market Power
Competition Policy and Regulation
Overview of Interconnection
Free Trade Negotiations
To have a successful interconnection, the following issues should be dealt with in the interconnection agreement or by rule or order from
the regulatory authority:
Prices and adjustment of prices over time. This includes the initial level of interconnection charges, a definition of the currency in
which interconnection charges are to be paid (this is especially complicated when retail prices are set in a local currency and
interconnection is set in another currency), and how prices will adjust over the term of the agreement to account for exchange rate
changes and inflation. The "ownership" of the call must be defined. For example, in mobile-to-fixed interconnection, one possible mode is
for the call to be "owned" by the mobile operator, who sets the retail price and pays for interconnection and billing and collection to the
fixed operator. Another mode would be for the call to be "owned" by the fixed operator, who would set the retail rate and pay the mobile
operator an origination charge. Liability for bad debt and uncollectable bills should be defined.
Points of interconnection. The physical locations where interconnection will take place and the technical standards to be employed in
the interconnection are defined. A process for requesting and obtaining additional points of interconnection should be established. This is
closely related to the issue of transport charges and traffic routing.
Transport (conveyance) charges and traffic routing. Some definition must be made for how calls will be routed. In other words, if
there are multiple interconnection points defined, what is the proper routing and hand-off point for each type of call. Otherwise, higher
charges may apply to misrouted calls. The applicability of transport charges in the receiving network for calls that must be carried beyond
the area local to the point of interconnection must be defined. If one carrier has requested interconnection in a particular area so as to
avoid paying the receiving network for transport charges, and the interconnection point is not made available, sometimes a virtual point of
interconnection is defined for that location whereby transport charges are not collected to bring calls to that area.
Frequently, incumbent operators desire to offer as few as possible points of interconnection so as to maximize transport revenues.
However, over time, entrants usually wish to build out their own networks and interconnect in more places so as to avoid paying the
incumbent's transport charges.
Quality of service standards. Quality standards are defined, particularly for time to provision circuits and for call blocking levels, and
remedies are defined for when those standards are not met. Often, an incumbent provider is required to provide at least as high a level of
quality to interconnecting carriers as he provides to his own retail customers. Testing opportunities should be provided each party.
Billing and collection. When and how to collect traffic data, when and how to exchange bills, and when and how to make payment
should be specified. A process for reconciling traffic data and for making inquiries to the other party and for handling claims also should be
incorporated. A procedure for resolving discrepancies is useful, which often involves seeking recourse to arbitration, the regulator, or to
the courts.
Traffic measurement and settlement. Sometimes specific trunk groups are identified to carry different types of traffic so that each
type of traffic can be billed for separately. However, these arrangements can be defeated and traffic will thus end up disguised as the
cheapest type of traffic. The responsibilities of each interconnecting operator to measure traffic are defined, as are settlement procedures
for when there are discrepancies over the amount of traffic measured. Obligations to cooperate in fraud detection and enforcement
activities should be specified.
Numbering resources. Access of each operator to the country's numbering plan and numbering resources must be defined. It is
particularly important that numbers be provided in a timely manner so that potential sales are not blocked. If number portability may be part
of the local regulatory regime and terms of participation should be defined.
Forecasting network needs. Part of providing interconnection is having the available capacity to deliver and receive the traffic that
flows between the interconnecting networks. To do so, a planning process must be followed between the interconnecting operators so
that investment for additional capacity can be agreed, budgeted, and installed in time to meet the forecasted demand. Procedures to
resolve differences over forecasts also must be defined as well as what constitutes a bona fide request for additional interconnection
capacity. At a minimum, a mutual obligation to notify the other party of network changes and upgrades well in advance is needed to avoid
disadvantaging one competitor over another.
Access to customer information. By necessity, when completing calls and billing for them, interconnecting operators pass back and
forth considerable information about each other's clients. Limits on the permitted uses of this information should be defined, particularly
regarding the temptation to engage in marketing activities in approaching another operator's clients based on information obtained through
interconnection activities. Safeguards are also necessary to protect customers' privacy.

This section of the module covers:
■ Broad objectives and pricing principles for interconnection

■ Specific pricing principles for one-way and two-way interconnection
■ Long-run incremental cost modelling, including “bottom-up” and “top-down” modelling methods
■ Commonly used cost models, and
■ Benchmarking of interconnection charges.
Why is the Interconnection Price Important?
There is a consensus among economists and regulators that interconnection prices based on cost are most likely to lead to desirable
outcomes. Measuring “cost” is challenging — there is no single correct interconnection price. However, if the interconnection price is set
“too low”:
■ Inefficient competitors may enter the market

■ Entrants may look for opportunities to profit by purchasing services at low regulated prices and simply re-selling them, instead of
developing innovative new product offerings
■ Incumbent operators may not invest in the network or maintain its quality.
For many new entrants, interconnection is one of their largest costs. If the interconnection price is set “too high”:
■ It will deter entry by efficient competitors

■ In the case of two-way interconnection, carriers may concentrate on maximizing payments from other carriers, instead of focusing on
providing services to retail customers
■ Customers will be paying more than they need to.
Interconnection charges have generally been designed following either the paradigm of (1) revenue sharing or (2) interconnection usage
charges. Revenue sharing means that the telecommunications operators involved in a call have agreed to share the revenues, on a
percentage basis or some other agreed basis. They thus share the risk of billing disputes and bad debts. On the other hand,
interconnection usage charges imply setting charges to compensate explicitly one operator for the costs imposed on him by the other
operator’s use of his network to originate or terminate a call. The operator paying the interconnection usage charge "owns" the call and
takes the risk of disputed and unpaid charges. In addition, retail charges may be in one currency and interconnection usage charges may
be in another.
RELATED INFORMATION

Useful Economic Concepts

This section of the module discusses:
■ General objectives and principles for interconnection pricing

■ Specific principles for one-way and two-way interconnection, and
■ Key trade-offs for regulators in determining their approach to regulating interconnection prices.
Interconnection Pricing Objectives
Access and interconnection prices have several possible, not necessarily compatible, goals.
In general, interconnection prices should promote economic efficiency. There are three forms of economic efficiency:
■ Allocative efficiency requires that resources, products, and services are allocated to the person or persons who value them the
most. For this to happen, consumers of final products or services (such telephone calls to other customers) should pay prices that
reflect the cost of the resources used to provide those products or services
■ Productive efficiency requires that market participants use scarce resources as productively as possible. This means that the most
efficient provider should not be precluded from serving customers, and
■ Dynamic efficiency requires that all firms (entrants and incumbents) should have proper incentives to invest in technologies that
reduce costs and/or expand product offerings.
Some countries have additional objectives telecommunications, such as:
■ Actively promoting competition, by making it easy for new entrants to obtain interconnection. This is sometimes takes the form of low
interconnection prices, to encourage new entry
■ Achieving universal service. Many jurisdictions have historically maintained charges for basic telephone services that are below cost.
This is to encourage widespread subscribership. Recently, some countries have mandated high charges for call termination by
wireless carriers. The aim is to keep charges to wireless subscribers low, in order to encourage rapid uptake of wireless services.
Interconnection Pricing Principles
There appears to be a general consensus that, where possible, interconnection prices should be based on the additional cost to the
incumbent from providing interconnection services. However, it is difficult to strictly align prices with the cost of interconnection.
Broadly, three broad principles, or “pricing rules” are used to set interconnection prices:
■ Incremental cost pricing. Interconnection prices are based on the forward looking, long-run incremental cost of providing
interconnection (usually TSLRIC or TELRIC). Incremental costs are estimated using a suitable cost model.
■ Retail minus pricing. This approach starts with the incumbent’s retail price for the downstream service, and subtracts retail costs.
The final interconnection price should also include any additional costs to the carrier that arise directly from providing interconnection
services.
A formal exposition of the retail minus approach is the Efficient Component Pricing Rule, ECPR:
Interconnection (Access) price = additional marginal cost of interconnection (access)

+ (Retail price – marginal cost of retail)
The ECPR results in interconnection prices that are higher than incremental costs. ECPR prices incorporate the opportunity cost to the
interconnection provider of customers lost to the entrant. This includes any contribution to shared and common costs and any
foregone profits. For this reason, ECPR is controversial. Although it does encourage productive efficiency, it does not necessarily
support the goal of allocative efficiency.
■ Bill and keep. Bill and keep only applies to two-way interconnection. With “bill and keep” the calling party’s network retains whatever
revenue it raises through retail usage charges. Neither the calling nor receiving parties’ networks pay each other — the
interconnection charge is effectively zero. One advantage of a bill and keep policy is that it can be adopted quickly without the need to
employ a cost analysis. This could be useful in the situation of a small, developing country needing an interim policy to facilitate
interconnection between competitors while developing a policy based on cost analysis.
When the traffic exchanged between networks is roughly in balance, the net payments in either direction would be relatively small,
approximating the result of a “bill and keep” regime. Accordingly, “bill and keep” has sometimes been limited to situations where such
approximate traffic balance occurs, with positive payments to the terminating carrier when traffic is not reasonably balanced.
Specific pricing and charging considerations vary between one-way interconnection and two-way interconnection.
Pricing Principles for One-Way Interconnection
The pricing principles listed here derive from the general pricing principles above.
The interconnection price should give the interconnection seeker incentives to purchase interconnection from the upstream carrier where
this is the least cost option (for the economy as a whole). For this, interconnection prices should not exceed the cost of providing
interconnection.
If the interconnection provider is vertically integrated, and competes with the interconnection seeker, then the interconnection price should
be set so that the most efficient downstream provider has a legitimate opportunity to compete successfully. (For example, the combination
of interconnection and retail prices should not result in a vertical price squeeze.)
Economic theory suggests that access prices can be set to offset imperfections in retail price levels, for example by:
■ Setting access prices higher (lower) than interconnection costs when retail prices are above (below) cost, or
■ Setting access prices below cost in order to offset market power in the downstream market (where market power would otherwise
lead to downstream prices that are above cost)
Pricing Principles for Two-Way Interconnection
There are several approaches to structuring interconnection payments for two-way interconnection:
■ Calling party pays (CPP): The calling party, or the calling party’s network pays the network of the party receiving the call. CPP is
commonly used for mobile services (including throughout Europe, South America, India, and Africa)
■ Receiving party pays (RPP): The receiving party, or receiving party’s network, pays the calling party’s network for interconnection.
RPP is less common than CPP, but is used in North America and Japan
■ Bill and keep: Neither the calling nor receiving parties’ networks pays the other. In many two-way interconnection situations, this
may be the best form of regulation once the full costs of regulation are taken into account. Bill and keep can reduce incentives to rely
on arbitrage to maximize payments from other carriers, and is significantly less costly to implement than cost based interconnection
pricing.
Models of two-way interconnection are very complex, and conclusions about how to charge for two-way interconnection tend to be
model-specific. Which approach is optimal depends on a range of factors, including:
■ Assumptions about the distribution of the benefits from the call between the calling party and the call recipient
■ Whether or not traffic between the two interconnecting networks is approximately in balance
■ Differences in costs between the two networks.
Trade-Offs in Regulating Interconnection Prices
Setting interconnection prices requires trade-offs between the complexity of pricing framework, its accuracy (how closely price tracks
cost), and transaction costs for affected parties. Theoretically optimal prices vary significantly depending on the assumptions made in the
economic model.
Governments and regulators need to be pragmatic about interconnection regulation for three reasons:
■ The direct regulatory costs of a detailed forward-looking cost regime may be significant: operators may hire engineers, economists
and lawyers to put forward their views; the regulator must have enough resources to assess competing claims about cost; and there
may be costly dispute resolution processes
■ As regimes increase in complexity, operators and potential entrants are more likely to focus on arbitrage opportunities than ways to
offer consumers genuinely new services
■ There is no guarantee that detailed cost estimation approaches will be accurate.
RELATED INFORMATION

Long Run Incremental Cost Modelling
3.3.2 Long-Run Incremental Cost Modelling
The economic cost of interconnection is generally the starting point in establishing economically efficient interconnection prices.
In many jurisdictions, regulators set interconnection prices based on long run incremental costs (LRIC). (Examples include Australia, the
United Kingdom, the European Community, and the United States.) The most common form of LRIC is Total Service Long Run Incremental
Cost (TSLRIC), known as Total Element Long Run Incremental Cost (TELRIC) in the United States.
There are numerous methods of estimating LRIC. Approaches to modeling LRIC can be broadly categorized as bottom-up and top-down
modeling approaches. Bottom-up models include scorched earth or scorched node methods. Click here for a comparison of bottom-up and
top-down modeling approaches.
“Bottom-Up” Modeling
Bottom-up modeling uses detailed data to build a hypothetical network that can supply telecommunications services, including
interconnection services. The costs of this network, including capital costs and operations and maintenance costs, are then allocated to all
the services provided. Bottom-up modeling has the following steps:
Step 1: Define the services to be modeled (for example local access services). This step includes gathering data on the number
and location of customers in the geographic area under consideration
Step 2: Determine the design of the network — what facilities are required to provide the service, and where should they be
located? Figure 1 illustrates a typical design of such a local service network and the facilities that are needed to provide
telecommunications services. As the figure shows, a PSTN generally includes: wires and support structures that connect
customers to telephone switches (loop facilities); end-office and high-level switches; and facilities that connect the switches
(transport)
Figure 1: Generic PSTN Network Structure
Source: Macv Sullivan, Presentation on “The Basics of Interconnection”, ITU Workshop on Telecommunication Reform, 3-5 May 1999.
Step 3: Determine the amount of each type of equipment needed to construct the network
Step 4: Estimate the costs of each element. For each type of equipment multiply the amount required by its unit prices to arrive at
the total investment cost. (TSLRIC models usually use current “best-in-market” costs)
Step 5: Convert the total investment cost, for each network element, into an annual (or monthly) amount. This amount equals
depreciation costs and cost of capital for the firm in question
Step 6: Estimate annual (or monthly) operations and maintenance costs and non-network costs. This includes direct out-of-pocket
operating expenses associated with the investment and indirect expenses, such as corporate overheads
Step 7: Estimate total costs for each network element by adding the annual (monthly) amounts calculated in Steps 5 and 6
Step 8: Divide the total costs of each network element by the relevant cost-driver, to arrive at unit costs. For example, use the
number of lines to derive the unit costs for subscriber loops, or the number of minutes to derive unit switching costs.
“Scorched Earth” and “Scorched Node” Models
Step 2 — designing the network to be modeled — requires the regulator to make choices about how much optimization to include in the
modeled network. These choices can be represented on a spectrum, as shown in Figure 2. (The distance between the points on the
spectrum is illustrative only.)
Figure 2: Approaches to Network Design in TSLRIC Models

The scorched earth approach represents one extreme. It assumes that nothing is fixed, not even the location of the nodes. The scorched
earth network is what an entrant would build if no network existed, based on the location of customers and forecasts of demand for
services. This approach would give the lowest estimate of LRIC, because it removes all inefficiencies due to the historical development of
the network.
At the other extreme, LRIC can be estimated from the current costs of the existing firm, using a top-down modeling approach. This will give
the highest estimate of cost because it does not allow for optimization.
The “scorched node” approach to LRIC estimations represents a compromise between the two extremes. It assumes that the location of
network nodes is fixed, and the operator can choose the best technology to configure the network around these nodes. Scorched node
models are common internationally. Regulators in Australia, New Zealand, the United States, the United Kingdom, Austria, Switzerland,
Denmark, the Netherlands, and Ireland have adopted the scorched node approach.
Regulators must make trade-offs between different objectives. Basing the estimate of LRIC on current costs would mean that entrants
would pay more than the efficient costs, potentially reducing entry. Basing the estimate on a scorched earth approach is also problematic. It
could deter the network operator from making investments that are efficient given the actual configuration of the network, since the
scorched earth approach ignores the existing network configuration.
“Top-Down” Modeling
“Top-down” modeling attempts to measure LRIC starting from the firm’s actual costs, as set out in its accounts. This method does not
involve detailed network modeling. Instead, a top-down model separates the firm’s assets and costs into service groups, and then adds the
costs associated with interconnection to arrive at an estimate of LRIC. This usually involves the following five steps:
Step 1: Identify the firm’s services and separate out interconnection services
Step 2: In the firm’s accounts, identify and separate all costs and assets
Step 3: If a cost item or asset is attributable to only one service, allocate it to that service
Step 4: Use allocation rules to allocate shared and common costs between services
Step 5: Calculate LRIC for each service by adding up the costs allocated to that services, including an appropriate return on those
assets allocated to the service.
“Top-down” modeling uses the firm’s current operating costs and historic capital costs. These are not forward-looking costs. It is more
difficult to take account of future changes in costs in a top-down approach than in a bottom-up approach that can incorporate explicit
assumptions about technological change and its impact on the firm’s choice of inputs.
It is possible to make adjustments to top-down approaches to remove inefficiencies in the firm’s current network configuration and costs,
but it is difficult to do so transparently. The incumbent firm will have more information about its historic performance and its accounts than
the regulator or new entrants.
Comparison of “Bottom-Up” and “Top-Down” Modeling
Bottom-up models Top-down models
Incorporate actual costs

Can model costs that an efficient entrant would face
Useful for testing results from bottom-up model
Flexible — can change assumptions readily
Advantages
May be faster and less costly to implement, but this
Transparent — much of the information used is publicly
depends on how well categories in the financial
available
accounts match the data required
Include the firm’s actual costs, and so are likely to

May optimize “too much” or omit costs. If this happens, the
incorporate inefficiencies
operator will be under-compensated and will reduce
investment in the network Less transparent — confidentiality issues mean other
stakeholders may not have access to the information
Modeling of operating expenditure is usually based on
Disadvantages simple margins instead of real-world costs used
The parties may dispute the cost allocation rules used
Data needed for the model may not exist
(the rules used to allocate shared and common costs
The modeling process can be time-consuming and among specific services)
expensive
Data may not exist in the required form
RELATED INFORMATION

Commonly Used Cost Models
Benchmarking Interconnection Rates
This section provides an overview of three commonly used cost approaches. The FCC Synthesis Model and World Bank Group Model are
both bottom-up TSLRIC models. The COSITU model is an example of a top-down cost model.
FCC Synthesis Model
The FCC Synthesis Model is an example of a “bottom up” TSLRIC model. The Federal Communications Commission used this model to
determine high-cost universal service support for non-rural areas in the United States. In addition, the model has been adapted by some
regulators internationally (for example by New Zealand’s Commerce Commission).
The original version of the FCC Synthesis Model and its documentation are available on the web site of the Federal Commerce
Commission. The model allows the user to change various inputs from their default values, to tailor the model to different requirements.
The FCC Synthesis Model follows the steps set out in the general discussion of bottom up cost models here. The specific steps in the
model are:
Step 1: The model starts with a database of geocoded customer locations.
Step 2: The model’s clustering algorithm then groups the locations into small geographic areas. This grouping must satisfy a number
of constraints. For example they must be consistent with the predefined maximum length for copper cables, maximum number of
lines (or households), that can efficiently be served by equipment such as remote terminals. The resulting geographic areas are
intended to approximate the distribution serving areas that a telephone engineer would design.
Step 3: Once geographic areas are identified, the model designs facilities routes:
■ Feeder routes between switches and either a remote terminal or serving area interface within the cluster,
■ Distribution routes between the remote terminal or serving area interface and customer locations, and
■ Interoffice routes between switch locations.
Step 4: The model determines the specific amounts of network facilities needed to provide service along the routes determined in
Step 3. Network facilities include:
■ Copper and fibre cables,

■ Support structures such as telephone poles and underground conduits,
■ Remote terminals, and
■ Switches.
Step 5: Estimate the investment that would be required to build the model network. This requires unit price inputs corresponding to
each type of facility. The FCC Synthesis model contains hundreds of such price inputs.
Step 6: The model determines the annual costs and corresponding unit costs for the form in question in the typical manner used by
bottom up TSLRIC models (see Step 5 to Step 8 of the bottom up modelling process outlined here).
World Bank Group model (WBG)
The World Bank Group’s Cost Model was designed to develop usage costs (including interconnection costs) for developing countries in
sub-Saharan Africa.
The model represents the components of a fixed (wireline) network suitable for African conditions, up to remote switching units. Unlike the
FCC Synthesis Model, the World Bank Model does not depict the local loop facilities used to connect end-use customers to the first
switching point.
The World Bank Group Model allows the user to change various inputs from their default values. Inputs that can be changed include:
■ Network equipment prices, and additional mark-ups that might apply to purchase and install equipment in developing countries,
■ Traffic and demand information, such as calls per line and routing tables,
■ Total lengths for important network facilities, such as metres of fibre cables of particular capacities, and metres of duct facilities, and
■ Cost of capital inputs.
COSITU
COSITU is an example of a top-down cost model. COSITU is based on enhanced fully distributed costing principles, as adopted in the ITU-T
D series of recommendations. [1]
COSITU requires the following input data:
■ Investment and expense data from accounting systems,

■ Current cost data to convert historical capital asset costs to current costs. For example, when accounting records report the
purchase price of a switch, the model calculates the cost of the switch at current purchase prices,
■ Inputs for depreciation and cost of capital. Where the inputs needed to estimate the cost of capital are not available, COSITU
benchmarks these to countries or firms of comparable risk, and
■ Traffic demand and routing data.
COSITU produces unit costs and prices for international, subregional, and regional calling. COSITU can account for the effects of universal
service funding, taxes, and any access deficit as mark-ups over current unit costs, to calculate interconnection prices.
To the extent that accounting and demand data are available, COSITU’s basic modelling framework can be used to model interconnection
costs for both fixed and mobile networks.
COSITU embodies the following principles:
■ Transparency: Information used in the cost derivation process should be openly available, so that external analysts can comprehend
the final rate,
■ Practicality: The demands of the costing methodology with respect to data availability and data processing should be reasonable, to
keep the costing exercise economical yet still useful,
■ Causality: The model should demonstrate a clear cause-and-effect relationship between service delivery, on the one hand, and the
network elements and other resources used to provide the service, on the other hand, taking account of relevant cost determinants
(cost drivers),
■ Contribution to common costs: The cost calculation should provide for a reasonable contribution to common costs,
■ Efficiency: The cost calculation should provide a forecast of cost reductions that are likely to result from more efficient use of
resources over time.
Endnotes
[1] The ITU-T D series of recommendations can be viewed by following the “D series” link under “Recommendations” at
http://www.itu.int/ITU-T/studygroups/com03/index.asp.
RELATED INFORMATION
Long-Run Incremental Cost Modelling
Benchmarking is the process of establishing interconnection rates based on rates in other jurisdictions. For example, the rate charged to
long distance carriers for terminating calls on a local network might be based on rates for this function in other jurisdictions.
Benchmarking has two main purposes in interconnection pricing. In situations where detailed cost models can be estimated, benchmarking
can be used as a common sense check on the results of the modelling. Alternatively, benchmarking can be used directly to set
interconnection prices.
Benchmarking can be very useful to regulators if undertaken carefully. Undertaking a full forward-looking cost modelling exercise is
challenging and time-consuming. In some markets the detailed information required may not be available. Regulators in many jurisdictions
have used benchmarking to set initial interconnection rates (for example Botswana, New Zealand).
Where benchmarked rates allow competition to develop satisfactorily, rates based on benchmarking may be used for extended periods.
In a benchmarking exercise, adjustments need to be made for differences among jurisdictions, for example exchange rates, traffic
patterns, or the cost of shipping network equipment.
For a discussion of practical issues in benchmarking prices click here.

RELATED INFORMATION
International Benchmarking of Prices

Long-Run Incremental Cost Pricing

The principles underpinning interconnection are broadly the same for mobile networks as for fixed networks. However, there are some
important technical, commercial, and regulatory differences between mobile and fixed networks that affect interconnection arrangements.
This section covers:
■ Key forms of mobile interconnection,

■ Issues in setting mobile termination rates, and retentions for fixed-to-mobile calls,
■ Additional considerations in modelling mobile network costs,
■ Mobile roaming, and
■ The role of mobile networks in achieving social goals, and implications for regulators.
RELATED MATERIALS
Section 6.5 Mobile Network Sharing
In many countries, mobile interconnection is regulated and priced differently, depending on the form of interconnection. There are three
broad forms of mobile interconnection:
■ Fixed-to-mobile interconnection: A mobile network terminates a call from a fixed network. The call might originate from a local
fixed operator, a domestic long-distance operator, or an international operator,
■ Mobile-to-fixed interconnection: A mobile operator interconnects with a fixed network in order to complete calls for the mobile
operator's customers. Again, the fixed network might be owned by a local fixed operator, a domestic long-distance operator, or an
international operator,
■ Mobile-to-mobile interconnection: A mobile operator interconnects with another mobile operator.
RELATED INFORMATION
Section 3.4.2 Mobile Termination Rates
There is no a unique treatment of mobile termination charges among countries. Some countries only regulate mobile termination charges
for fixed-to-mobile calls. In other countries, mobile networks are required to apply a single regulated termination charge regardless of
where the call originates.
This section discusses:
■ Calling Party Pays,

■ Regulation of mobile termination rates, and
■ Other pressures on operators to reduce mobile termination rates.
Calling Party Pays
Under Calling Party Pays (CPP) the calling party, or the calling party's network, pays for the call. The recipient of the call pays nothing.
CPP is used in many countries to structure interconnection payments for fixed-to-mobile calls. Under the "old" CPP model, the mobile
operator sets a fixed-to-mobile tariff. The fixed operator deducts specified charges from this fee (such as an origination charge, and billing
and collection charges), and passes the balance of the call revenue to the mobile operator.
In recent years, some regulators have decided to regulate fixed-to-mobile tariffs, rather than leaving this to the mobile operator to
determine. This generally reflects concerns that fixed-to-mobile tariffs are too high. This concern has also led regulators to control mobile
termination charges.
Receiving Party Pays (Mobile Party Pays)
A minority of countries, predominately developed countries such as the United States, use a system of receiving party pays or mobile party
pays for interconnection with mobile operators. Under this system, the mobile user pays airtime on received calls as well as calls that user
has initiated. This reduces the problem of setting interconnection charges to defining the costs of just the link between two networks,
which generally is low and easily defined. Thus, countries using receiving party pays have largely avoided the problem of high mobile
termination charges. This is a definite advantage of the receiving party pays system. Since a receiving party pays system requires the
mobile user to pay directly for network usage on the mobile network, its main disadvantage is that it makes it difficult commercially to extend
service to mobile users with very low income levels, precisely where the calling party pays system has been most successful.
Regulation of Mobile Termination Rates
Regulation of fixed-to-mobile rates and/or mobile termination charges is usually justified on the basis that those prices are "too high"
compared to a cost-based estimate, or to prices for outgoing mobile calls.
The premise is that mobile operators are able to sustain high fixed-to-mobile prices because they have market power in setting prices for
fixed-to-mobile calls. This market power derives from that fact that the fixed subscriber who places a call to a mobile subscriber has no
influence over which mobile network is used. Mobile subscribers make this decision when they decide to join a network. Under Calling
Party Pays mobile subscribers do not pay for fixed-to-mobile calls, so they may not take the price of these calls into account in selecting a
network.
Many regulators now control mobile termination charges. There are several forms of such regulation:
■ International benchmarking: In the absence of cost based data, regulators are increasingly relying on international benchmarking
to set regulated mobile termination charges in their own countries,
■ Rounding: Some regulators have introduced regulations requiring mobile operators to round each call to a lower unit of charging (for
example rounding to the second when the charging unit is to the minute). The effect of this requirement is to reduce revenue from
mobile termination,
■ Cost-based termination charges: Regulators are increasingly pressuring operators to base mobile termination charges on long
run incremental costs or fully allocated costs.
Other Pressures to Reduce Mobile Termination Rates
Market forces are also pushing down CPP rates and mobile termination charges. For example users are increasingly substituting mobile-to-
mobile calls for fixed-to-mobile calls, creating additional pressure on mobile operators to reduce fixed-to-mobile rates and mobile termination
charges.
United States international carriers, supported by the United States Government, are pressuring developing country operators to reduce
international mobile termination rates. Because United States carriers are net exporters of telephone traffic to developing countries, a
reduction in mobile termination charges would reduce their net interconnection payments to foreign operators.
RELATED INFORMATION
Pricing Principles
Retentions for Fixed-to-Mobile Calls
3.4.3 Retentions for Fixed-to-Mobile Calls
Under Calling Party Pays (CPP) for fixed-to-mobile calls, the fixed operator deducts specified charges from the fixed-to-mobile rate and
passes the balance of the call revenue to the mobile operator.
The fixed operator may retain charges for the following items:
■ Call origination: Call origination charges reflect the cost of the fixed network used to originate the call,
■ Billing and collection: The fixed operator may levy a contribution to the cost of collecting call revenue from its customers. This fee
may be expressed as a percentage of the fixed-to-mobile tariff, or as an absolute charge per minute, per call or per bill,
■ Bad debts: The fixed operator may levy a fee for bad debts, on the basis that fixed-to-mobile calls may make up a significant
proportion of customers' total bills,
■ Other fees: For instance in some countries fixed operator’s charge fees for managing complaints related to fixed-to-mobile calls.
The proportion of the call revenue retained by the fixed operator varies from country to country. Figure 1 shows the proportion of
retentions in selected Latin America countries in 2003. In Chile retentions amounted to 7 percent of the fixed-to-mobile tariff while in
Venezuela retentions were as high as 31 percent.
Figure 1: How Much of a Fixed-to-Mobile Tariff is Kept by the Landline Operator? Latin America 2003
RELATED INFORMATION
Mobile Termination Rates
The basic economic principles of forward-looking cost models apply to both fixed and mobile networks.
However, the importance and types of cost drivers in a mobile network differ from traditional fixed networks. (See Figure 1 for an example
of a typical GSM mobile network.)
The costs of both fixed and mobile networks increase with increases in:
■ The number of subscribers, and

■ The traffic produced by those subscribers.
The costs of a mobile network also increase with coverage — the geographic size of the network. Coverage costs are an example of
common costs. They do not increase with the volume units usually considered in wholesale or retail price structures (such as access
connections and usage). Accordingly, retail and interconnection prices for mobile usage need to contain mark-ups to recover coverage
costs.
For more information see Europe Economics, Cost Structures in Mobile Networks and their Relationship to Prices, Final Report for the
European Commission, Contract N. 48544, November 28, 2001 (“European Economics Report”) and its Annex.
Figure 1: Structure of a GSM Mobile System
Source: European Economics Report, page 17.
RELATED INFORMATION
Long-Run Incremental Cost Modelling
Roaming is the term used to describe the situation when a subscriber of one mobile operator’s service travels outside that service area and
obtains connectivity and service from another operator. Roaming can take place within a country or between countries, as long as it
involves a customer of one operator being connected to the mobile network of another operator.
For example, roaming enables a subscriber of Cabo Verde Telecom in Cape Verde (which operates using GSM technology) to travel to
Angola and obtain services from a GSM operator there.
Conceptually, roaming is similar to a call forwarding arrangement. Callers use the customer’s usual mobile phone number. The home
network hands the call over to the host network, which passes the call to the customer’s mobile phone (see Figure 1).
Figure 1: Mobile Roaming
Roaming charges are generally much higher than termination charges within the home area. Customers often pay a monthly fee to be able
to roam plus usage charges, the combination of which can be quite expensive.
Implementing Roaming
For roaming to be possible, the customer’s handset must be compatible with the host network. If the home operator and host operator use
different technologies, roaming can only accomplished using a different handset when in the host operator’s coverage area. This can be
expensive and cumbersome.
Even if network technologies are compatible, roaming cannot occur until the operators have agreed on the terms and conditions for
accepting each others' roaming traffic. “Roaming agreements” between the operators establish the commercial and technical basis for
implementing roaming.
RELATED INFORMATION
In many developing countries, mobile networks play an important role in meeting universal access goals. Mobile or wireless networks may
be more economic than conventional landline networks, particularly in remote or rural areas.
In many developing countries, there is a tension between ensuring low fixed-to-mobile rates, and encouraging greater penetration of mobile
services.
In markets with low mobile penetration there may be theoretical and practical justifications for above cost mobile termination charges.
Revenue from high termination charges may enable mobile operators to subsidize:
■ Access to the service, for example by subsidizing the cost of handsets, and
■ Usage charges (mobile-to-mobile and mobile-to-fixed calls).
Before intervening to reduce mobile termination charges, developing country regulators should weigh up the potential effect on access to
the network for poorer segments of population, against the purported benefits from reduced termination charges.
The Role of Prepaid Services
Prepaid services can play an important role in reducing barriers to access for poor customers, or those in rural areas. Under prepay
options, customers:
■ Do not need a fixed address in order to access the service,

■ Are not required to visit the service provider’s offices to pay their bills (particularly if the provider has wide retail distribution channels
for prepay cards), and
■ Do not need to commit to a fixed monthly payment, which can be a problem for customers with low or unpredictable incomes.
For example, in Jamaica, access to telecommunications services increased dramatically when competing mobile phone companies
introduced prepaid, no-deposit accounts. This was a major cause of Jamaica’s phenomenal increase in telephone penetration following
liberalization.
RELATED MATERIALS
Module 4. Universal Access

Establishing a regime to develop and implement interconnection rates, terms and conditions, and other provisions can place significant
demands on a developing country’s legal and administrative infrastructure. This section considers particular challenges that may be
significant for developing country regulators.
Many of these challenges apply to all countries, but are more difficult in countries with weak legal systems or no tradition of decision-
making by independent regulators.
Key challenges include:
■ The physical state of telecommunications networks in developing countries;

■ Transparency and access to information;
■ Regulating state-owned operators;
■ Free trade negotiations; and
■ Dispute resolution.
Compared to developed countries, ICT infrastructure in developing countries have a number of features that create both challenges and
opportunities:
■ Developing countries may not have extensive telephone network coverage, particularly outside main population centres.
■ Wireless and mobile operators often play a significant role, particularly in rural and remote areas. Typically, wireless demand in
developing countries exceeds wireline demand, sometimes by significant amounts.
■ Fibre-optic systems are often not widely rolled-out (or not all fibre is “lit” with the necessary electronics). Customers may have limited
or no access to broadband services, particularly in rural areas.
■ The technology in use, and network architecture, are often outdated.
These factors create a number of challenges. In particular, significant investment may be needed to achieve universal access goals or to
make broadband service widely available.
At the same time, developing countries that are designing interconnection regimes now have the opportunity to design better regulatory
regimes. The fact that traditional wireline technologies are not deeply embedded in many developing countries enables regulators to
implement interconnection policies that are more appropriate to wireless networks, VoIP, and other emerging technologies. For example:
■ The prominence of per-minute rates is a product of wireline technology. Per-minute rates may be irrelevant, or even
counterproductive, when applied to VoIP services.
■ Policies seeking to “unbundle” network elements assume that the wireline incumbent enjoys a near monopoly position in the provision
of critical ICT infrastructure. This assumption may not be valid in many developing countries.
The absence of a well-established interconnection regime may allow regulators in developing countries to bypass policies that are no
longer appropriate, in favour of arrangements that are sustainable, minimize opportunities for arbitrage, and are more in line with emerging
technologies.
In many developing countries, ensuring the transparency of interconnection arrangements and access to information are key challenges.
Transparency
Many countries require dominant operators to make the terms and conditions of interconnection transparent. In addition, the WTO requires
Members to ensure that agreements or model interconnection offers of major suppliers are made public.
The objective of such transparency is generally to prevent dominant operators from discriminating between different competitors or
otherwise acting to limit competition. For example, a dominant operator could enter into confidential interconnection agreements that provide
unfavourable interconnection arrangements with competitors, and favourable terms for affiliates.
Requiring operators to publish interconnection agreements enables regulators and other operators to monitor interconnection terms and
agreements and to identify discriminatory or potentially anti-competitive behaviour.
Transparency is also important in regulatory processes. For a regulator’s decisions to be credible, the regulated firm and other
stakeholders must have confidence in the decision-making process. Ways to achieve this include public consultation processes and
requirements for regulators to publish the reasons for their decisions.
Regulatory transparency may be difficult to implement in countries with weak legal and administrative structures and that have no tradition
of transparency. However, where an independent regulator has been recently established, there is an opportunity to introduce procedures
for regulatory transparency.
Access to Information
In order to regulate effectively, a regulator needs access to detailed information about the regulated firm. For example, regulators often
require detailed cost information and information on the regulated firm’s cost of capital.
In many developing countries such detailed information is simply not available. The incumbent firm may not have sufficiently detailed
network data to enable long run incremental cost modeling. Or the regulator may not have sufficient powers to require the regulated firm to
provide the information.
Where this is the case alternative, less data-intensive approaches can be taken. These approaches include:
■ Top down cost models: these models are based on the firm’s existing financial accounts. The ITU’s COSITU model is a top down cost
model that has been designed for use in developing countries.
■ International benchmarking: benchmarking can be used to estimate interconnection prices or individual inputs for costing exercises.
(For example, COSITU provides benchmark data for the inputs needed to estimate cost of capital, where this information is not
available).
3.5.3 Regulating State-Owned Operators
In countries where the incumbent telecommunications operator is government owned, it may be challenging for the regulator to be
independent, or to be perceived as independent. Entrants may have little confidence that their interests and those of their customers will be
given due weight.
Key questions for entrants and the government are whether the regulator will have sufficient influence over the state-owned provider to:
■ Effectively mandate access to the state-owned provider’s network, and

■ Be able to enforce terms and conditions for access that will allow equally efficient new entrants to compete.
Recent research has concluded that in the European Union, regulatory outcomes tend to favour government-owned incumbent operators.
[1] This favouritism is largely mitigated when the regulator is independent from governmental interference. This requires:
■ An arm’s length relationship with political authorities, and

■ Organizational autonomy, such as earmarked funding and exemption from restrictive civil service salary rules. Without organizational
autonomy, political actors may influence decisions indirectly, for example by threatening the regulator’s funding.
Regulatory independence can be difficult, or even impossible, in some countries. Even where formal structures separate regulators from
political authorities, Ministers may seek to use their personal influence to affect regulatory outcomes. Where this is the case it may be
better, in designing the regime, to explicitly recognise the role of Ministers in decision making. For example political decisions should be
explicitly made by policy makers. These would include:
■ Decisions to proceed with liberalization, which involve trade-offs between maintaining the value of the state-owned company on the
one hand, and introducing effective competition on the other, and
■ Decisions to rebalance retail tariffs, or otherwise remove implicit subsidies.
Endnotes
[1] See, for example, Geoff Edwards and Leonard Waverman, “The Effects of Public Ownership and Regulatory Independence on
Regulatory Outcomes” (2006) Journal of Regulatory Economics, Vol. 29, No. 1, pp. 23-67.
RELATED INFORMATION
Legal and Institutional Aspects of Regulation
The World Trade Organization (WTO) Agreement on Basic Telecommunications came into force in 1998. The Agreement includes
obligations relating to interconnection. WTO Members, or countries seeking to joint the WTO, must comply with these rules. If a Member
country fails to comply with its WTO obligations, other Members may take a dispute to the WTO.
The obligations of Agreement on Basic Telecommunications are summarized in a “Reference Paper”. As of April 2009, nearly 96
governments had committed to some or all of these obligations. (Countries’ commitments and exemptions are listed here.)
Key WTO obligations for interconnection are as follows:
■ Interconnection with “Major Suppliers” must be assured:

❍ At any technically feasible point in the network;
❍ In a timely fashion;
❍ On non-discriminatory and transparent terms (including quality and rates);
❍ Sufficiently unbundled to avoid charges for unnecessary components; and
❍ At non-traditional interconnection points if the requestor pays charges.
■ Procedures for interconnection to major suppliers must be made public.
■ Agreements or the model interconnection offer of major suppliers must be made public.
This topic is also discussed in Module 6, section 3.2.1., "Role of the World Trade Organization".
RELATED INFORMATION
Asymmetric Interconnection Regulation
Disputes pertaining to access, interconnection, and other aspects of regulation are common in the telecommunications sector. This section
covers:
■ Why interconnection dispute resolution is important,

■ The role of the regulator in resolving disputes,
■ Challenges for regulators, and
■ Ways to strengthen dispute resolution processes.
The Importance of Interconnection Dispute Resolution
Disputes pertaining to access, interconnection, and other aspects of regulation are common in the ICT sector. This can stall the
development of competition and the implementation of important national policy goals for infrastructure and economic development.
Reliance on the courts to resolve disputes between telecommunications firms is costly and can involve substantial delays. For example, in
New Zealand the first major interconnection dispute between the incumbent and a new entrant took over three years to resolve through
the courts and even then failed to deliver a conclusive resolution.
Without a mechanism to resolve interconnection disputes quickly and effectively, innovation and competition in the sector will be
threatened. Entrants will not commit resources unless they have confidence that their business will be viable and that they will be able to
resolve any disputes in a timely fashion.
The Role of the Regulator
The World Trade Organization Agreement on Basic Telecommunications includes obligations relating to dispute resolution.
Under the Agreement, Member countries must establish an independent domestic dispute resolution body, so that interconnection disputes
can be settled within a reasonable period of time. This need not be the regulator, but it often is.
Often a regulator will require the development of a Reference Interconnection Offer ("RIO") as part of opening the sector to competition.
The RIO sets forth the terms and conditions for interconnection services, and prices, that a competing operator can choose to accept
without further negotiations. The purpose is to avoid disputes and to shorten the entry time for a new competitor. The requirement to
develop a RIO is most usually imposed on an operator that is deemed to be dominant or have significant market power (often the incumbent
operator). A regulatory tool that accomplishes similar things is a "most favoured nation" or nondiscrimination requirement, whereby any
operator can choose to accept the terms and conditions that have previously been agreed or ordered to be in place for another competitor.
Many countries have adopted either or both of these measures. One example is described in the practice note on Jamaica's RIO.
Challenges for the Regulator
Dispute resolution presents a number of challenges for regulators, including:
■ Access to information: Operators usually have better information than the regulator on the details of interconnection disputes. This
makes it difficult for the regulator to come to a decision and be confident that it is the best one.
■ “Gaming” of the process: Either party may engage in anti-competitive gaming of the dispute resolution process. For example, an
incumbent may use delaying tactics to draw out the proceedings, in order to delay competitive entry. Or an entrant may not accept a
reasonable interconnection offer from the incumbent if it believes that it can persuade the regulator (or dispute resolution authority) to
mandate more favourable terms.
■ Capacity: Many countries face a shortage of people with the necessary legal, economic, and technical expertise to resolve
interconnection disputes.
Ways to Strengthen Dispute Resolution Processes
Options to strengthen dispute resolution process include[1]:
Improve information available to the regulator

Rationale: Enable the regulator to base its decision on better information.
■ Ask parties to define areas of agreement and dispute and to provide information to clarify disputed issues;
■ Require written submissions from operators on areas of dispute, supported by facts and research if necessary; and
■ Allow others (for example customer groups and other service providers) to comment on areas of dispute.
Obtain expert assistance

Rationale: Supplement the regulator’s in-house capability by drawing on external expertise.
■ Use external advisors (for example an experienced interconnection expert) to assist in resolving the dispute. The expert’s role could
include clarifying areas of agreement and dispute, identifying information needs, and providing advice.
■ Consider appointing an independent mediator (or, if the parties agree, an arbitrator).
■ Consult with other regulators on their approach in similar cases.
■ Review decisions and interconnection agreements approved by other regulators.
■ Use outside parties for informal mediation, arbitration, information gathering or other assistance. This can be particularly useful in
countries where the regulator lacks the legal authority to resolve the dispute, or may be biased.
Improve transparency
Rationale: Making more information publicly available should cause parties to consider their positions more carefully.
■ Make parties’ submissions available for comment by other parties and the public, with summaries to protect confidential information;
and
■ Publish a draft decision and give parties to the dispute and others an opportunity to make written submissions on it.
Endnotes
[1] Adapted from InfoDev, “Telecommunications Regulation Handbook, Module 3: Interconnection,” Washington, DC: The World Bank, 2000,
Table 3-2.
3.6 Cross-Border Interconnection

This section covers issues related to cross-border interconnection, specifically:
■ The accounting rate system and cross-border interconnection, and

■ Regional interconnection clearing houses.
RELATED INFORMATION
Mobile Roaming
The accounting rate system was developed as a way to allocate revenue for international telephone services. The system is a series
of arrangements between national operators in which the operators jointly provide international calls and divide the revenues from such
calls between them.
The accounting rate system provides a set of agreed prices for interconnection of international calls. The originating carrier charges the
customer making the call a retail rate, and is charged the accounting rate for terminating the international call. As their name suggests,
accounting rates do not always reflect costs.
If traffic flows along a route are balanced, the accounting rate system does not generate significant cash flows. However, for many less-
developed countries, traffic on international routes is unbalanced — more calls are terminated in these countries than originate from them.
As a result, the accounting rate system produced considerable revenue inflows to many less-developed countries.
Moving Away From Accounting Rates
The accounting rate system has come under pressure in recent years.
The presence of competitive long distance providers has made it necessary for providers in other countries to deal with more than one
correspondent. This has opened the gates to different arrangements, in search of lower prices.
Carriers exploit numerous arbitrage opportunities to offer customers rates that are well below international accounting rates.
The system has also come under regulatory pressure. In 1997, the United States Federal Communications Commission acted to reduce
these accounting rates by prohibiting United States-based carriers from paying rates above certain benchmark levels.
The accounting rate system has now been largely replaced by cross-border interconnection. Carriers directly negotiate rates to terminate
traffic, in some cases with long-term contracts, in other cases on a short-term or spot basis. Electronic exchanges have emerged that
enable trading of international voice, data, and mobile capacity. Arbinet is an example of such an exchange. Arbinet claims that more than
14.4 billion minutes of traffic were transacted through its trading platform in 2007.
RELATED INFORMATION
Regional Interconnection Clearing Houses
A promising area of activity in the exchange of international traffic is the development of regional interconnection clearing houses.
The idea of a clearing house is that multiple carriers connect to the same hub in order to facilitate international interconnection. This
approach contrasts with the traditional approach where each carrier must have interconnect arrangements in place with all other carriers
in order to complete international calls.
Incumbent landline operators traditionally have performed this clearing house function. However, the development of competition has led to
concerns that incumbent operators may seek to disadvantage new entrants. This concern has prompted proposals to establish
independent clearinghouses for interconnection. This idea is still largely at the inception stage.
RELATED INFORMATION
The Accounting Rate System

4 New Paradigms: Voice Over IP and IXPS
The ICT sector is developing rapidly. Technological advances are making new services, and new modes of service deliver, possible. In the
future, the Internet will be the primary medium through which converging voice and data services will flow. As a result, market structure,
business models, and commercial arrangements for interconnection are changing.
This section focuses on the implications of the Internet for interconnection, specifically on:
■ Background to the Internet and its development;

■ An overview of Voice over the Internet Protocol (VoIP);
■ Opportunities for arbitrage, that are creating pressures for change to existing regulatory and commercial models;
■ The implications of VoIP for regulators, and for interconnection;
■ VoIP over wireless technologies;
■ International benchmarking of charges for network access; and
■ Internet Exchange Points (IXPs).
RELATED MATERIALS
Module 7, section 1.2, "The Internet"

This section provides an overview of the Internet and its development, as background to other material in this section of the module. This
section covers:
■ An overview of the Internet,

■ Different layers of the Internet,
■ The evolution of the Internet, and
■ Current trends in the development of the Internet, that are acting to constrain the market power of Tier-1 ISPs.
RELATED INFORMATION
About VoIP
Internet Exchange Points
Module 7: New Technologies and Impacts on Regulation: Technology Trends -- The Internet
The Internet is a world-wide collection of interconnected networks. It is capable of switching, routing, and transmitting digital packets of
information corresponding to a variety of voice, data, text, audio, and video services. The Internet allows any computer (or other device)
with an Internet connection to communicate with any other device that is connected to the Internet (see Figure 1).
Figure 1: Overview of the Internet

The Internet uses packet switching to transmit information. All content to be transmitted is broken into a series of small “packets”. Each
packet can be transmitted separately, routed using any available network. The packets are then reassembled at the receiving end of the
communication.
Packet switching has a number of advantages. It can transmit any digital information, to provide voice, text, audio, video, and data services.
Packet switching is also more efficient than the circuit switching used by conventional telephone networks. Packet switching only uses
network resources as and when they are needed. In comparison, under circuit switching the entire link remains open for the duration of
the call, including any pauses in conversation. However, packet switching generally has higher latency than circuit switching.
RELATED INFORMATION
The Seven Layers of Internet Interconnection

Evolution of the Internet
Current Internet Market Developments
One can view Internet services, such as voice telephony as comprising seven horizontal layers of functions (see Figure 1).
At the bottom on the layered model lie the physical links, such as cables and radio spectrum, and the transmission technologies on which
Internet services ride.
In the middle lie several layers that identify and manage the delivery of traffic, for example the Internet Protocol, and the Transmission
Control Protocol.
At the top are software applications that shape the link to provide a particular service, such as telephony (or alternatively e-mail or multi-
media services).
Figure 1: The Seven Layers of the Open System Interconnection Model

Source: Abdus Salam International Centre for Theoretical Physics.
RELATED INFORMATION
Overview of the Internet

The evolution of the Internet has tracked four phases (see Figure 1):
■ Phase 1: Government administration first through the United States Defense Department, and later through the United States National
Science Foundation and universities and research institutes around the world,
■ Phase 2: Privatization of government financed networks and the ascendancy of former government contractors and other major
carriers,
■ Phase 3: The “dotcom boom”, which triggered irrational, excessive investment and overcapacity, and
■ Phase 4: The “dotcom bust”, followed by market re-entrenchment and resumed growth.
Figure 1: Evolution of the Internet

Phase 1 – Active Government Stewardship (1980s–1995)
Until 1995, the United States government through the Defense Department, and later the National Science Foundation, underwrote
development and maintenance of the core Internet backbone (NSFnet). National governments in other parts of the world pursued similar
network projects.
Internet access in this first phase sought primarily to achieve better geographical reach and greater numbers of users, with little regard to
the cost of access or who caused Internet Service Providers (ISPs) to incur access costs.
There were few ISPs, and these had roughly the same characteristics and traffic volumes. As a result, their routing assignments
generated approximately the same financial burdens. Because of this, and government funding of the backbone network, ISPs chose not
to negotiate interconnection agreements, or to meter traffic. Instead, all participating ISPs agreed to provide reciprocal, free, access to
each other’s subscribers. (This type of arrangement is known as “peering”, “bill and keep”, or “sender keeps all”.) ISPs exchanged traffic
at a few “public” Network Access Points.
In the early 1990s the National Science Foundation concluded that it could cease funding the Internet, and a commercial, privatized Internet
could evolve. In 1993, the National Science Foundation announced a plan to privatize the Internet.
Phase 2 – Entry Leads to Network Expansion (1995-1998)
The National Science Foundation privatized its very high speed backbone network through a tender process, which MCI won. The
privatization process also established four private Network Access Points, in:
■ Chicago (operated by the Ameritech Bell Operating Company and Bellcore),

■ Metropolitan New York / Philadelphia (operated by Sprint, the San Diego Supercomputer Center),
■ San Francisco (operated by the Bell Operating Company Pacific Telesis, and Bellcore), and
■ Washington DC (operated by Metropolitan Fiber Cities, which was subsequently acquired by MCI).
Following privatization, a hierarchical industrial structure developed. The Internet was dominated by a small number of large “Tier-1” ISPs.
These operators control the major backbone networks that provide transcontinental and transoceanic links, and have significant influence
over the terms and conditions for network access. Commercialization also opened the way for new, diverse, ISPs to enter the market.
Peering continued as the dominant interconnection model between Tier-1 ISPs. However, smaller ISPs had to pay the larger Tier-1 ISPs for
access. “Transit” arrangements developed in which larger ISPs sold access to their networks, their customers, and other ISP networks
with which they had negotiated access agreements.
ISPs in Asia-Pacific and Africa have borne the greatest financial burden under this commercialized model. In order to access larger ISPs
they must self-provision lines to and from Network Access Points in North America and Europe, and pay for transit. During this period
much of the content on the Internet was located in North America and Europe. Remote ISPs needed access in order to deliver the content
their subscribers wanted to access.
Phase 3 – Dotcom Boom Stimulates Overinvestment (1998-2001)
By the late 1990s demand for Internet services was growing rapidly. Investors responded by investing very heavily in Internet
infrastructure. Several hundred billion dollars were spent on installing network capacity to meet not only current demand, but optimistic
expectations of future demand.
While Internet use continued to grow, it did so much less quickly than anticipated. As a result, network capacity greatly exceeded
demand. This created substantial downward pressure on Internet transport charges. Many Internet businesses suffered financial
problems or went out of business. However, the excess network capacity created greater opportunities for smaller ISPs to obtain low
cost access, and for customers to obtain low cost service.
During the dotcom gold rush forecasters expected the demand for Internet services and telecommunications links to grow as extraordinary
rates, as much as doubling on a monthly basis. Both incumbents and market entrants responded by investing primarily in transoceanic and
transcontinental transmission facilities that link the developed world. A massive glut in fiber optic transmission facilities has resulted, but
the glut never included developing countries. Currently growing demand for full motion video and other high throughput applications has
begun to reduce capacity glut in developed nations even as it creates additional pressure on insufficient capacity in developing countries.
Phase 4 – Retrenchment (2001-present)
During this phase the extended boom in global share prices ended, with most major stockmarkets experiencing significant falls. Shares in
technology firms fell particularly sharply, in response to downward revision in the expectations of their future profitability.
In response, investors sharply cut back their investments in technology, including Internet infrastructure. However substantial surplus
capacity remains. Although prices for Internet transport are no longer falling, they have remained relatively low. The cost of Internet transit
now appears to be a relatively small proportion of overall business costs. [1]
Demand for bandwidth continues to rise, due to growth in:
■ Broadband access from residences and small businesses,

■ Bandwidth intensive applications, in particular peer-to-peer sharing of music and video files, and
■ Electronic commerce, streaming audio and video services, and wireless access.
The Internet today operates as a hierarchy (see Figure 2). Thousands of small, local, regional and small country ISPs operate at the bottom
of an Internet pyramid. These operators typically have to pay for access to the networks and customers operated by larger ISPs. At the
middle of the pyramid are several dozen Tier-2 ISPs that typically pay to transit the networks of the largest ISPs. Tier-2 ISPs seek to
interconnect on a "peering" basis with other, Tier-2 ISPs. At the top of the pyramid are a handful of Tier-1 ISPs that typically “peer” with
other Tier-1 ISPs.
Figure 2: Hierarchical Structure of the Internet
The diversification of ISPs has increased the number of locations where ISPs exchange traffic. When the Internet comprised a small
number of Tier-1 ISPs, these ventures could exchange traffic at relatively few locations. Over time the number of network interconnection
points has grown. These interconnection points are commonly referred to as Internet Exchange Points (IXPs).
The increased number of IXPs results from:

■ The proliferation of ISPs, and
■ The fact that many ISPs now operate networks within only a small geographic area.
With more networks lacking complete national and international coverage, more ISPs need to interconnect with and access the transit
services of other ISPs. IXPs enable even small, regional ISPs to offer global Internet access to their subscribers.
Next Generation Networks (NGNs) in Developing Nations
NGNs include wideband wireless networks capable of providing high speed data services including Internet access and video
conferencing. Other NGNs provide broadband via satellite and fiber optic terrestrial lines. Developing nations appreciate the need to
deploy next generation networks (“NGN”) even as teledensity by earlier vintage facilities lags the level achieved in developed nations. See
International Telecommunication Union, ITU World Telecommunications Indicator Database
NGNs satisfy the often complex and demanding requirements of business executives, tourists and information intensive applications, but
also can provide fundamental services more efficiently and cheaply than legacy technologies. See Michael Kende and Onuche Ochioli,
Leap-frogging the divide: next generation networks in developing countries (Aug. 31, 2006).
See also International Telecommunication Union, NGN Policy and Regulatory Resources Worldwide.
World Information Technology and Services Alliance, Next Generation Networks and the Policy Implications (2006)
Regulators may need to apply administrative tools that avoid delays in licensing NGNs and conflict between market entrants and
incumbents. Morocco has licensed NGN operators operating new, “greenfield” networks in less than one year. See Gihane Belhoussain,
LICENCES DE NOUVELLES GENERATIONS AU MAROC – PRINCIPE DE CONVERGENCE; see also Moroccan Telecommunications Regulatory
Authority (ANRT) Web site (in French).
Endnotes
[1] For example, see John Hibbard, John de Ridder, Dr George R. Barker, and Professor Rob Frieden, International Internet Connectivity and
its Impact on Australia, Final Report on an Investigation for the Department of Communication Information Technology and the Arts,
Canberra, Australia (May 31, 2004), p. 4.
RELATED INFORMATION

The Internet is continuing to develop. A number of trends appear to be shifting the geographic focus of the Internet (with respect to
content, servers, peering points and carriers) away from just North America and Europe. These factors are acting to constrain the market
power of Tier-1 ISPs.
■ Proliferation of telecommunications routing options: A significant legacy of the dotcom bubble is the vast proliferation of
international submarine cable facilities. Capacity fill rates remain well below 50 percent for many trans-oceanic cables. This has
created a larger set of transmission options for ISPs, and significant downward pressure on prices. As a result, long haul costs as a
proportion of total Internet access costs have dropped. Until demand increases to fill existing capacity at a much greater level, prices
will not rise significantly.
■ Lower transport costs reduce tromboning: Much lower trans-oceanic transport charges make self-provisioning of lines to
North American and European Internet Exchange Points less of a burden for developing country ISPs. However, for ISPs with no
nearby Internet Exchange Point, the need to “trombone” traffic continues to be a financial burden. (This is the case in many parts of
Africa, for example.) In many parts of the world, in-region transport costs have declined to the point that regional alternatives to
tromboning are viable.
■ Decline of the US dollar: The US dollar has depreciated by over 40 percent over the past few years. This compounds the drop in
price of US dollar denominated interconnection and transiting services.
■ European ISPs provide an alternative transit route: The larger set of transmission routing options has triggered market entry by
new Middle Eastern and European ISPs. These provide alternative routing options for accessing subscribers and content based in
North America and Europe.
■ In-region peering: Lower in-region transport costs, and declining Internet equipment costs, make it more feasible for ISPs in remote
regions to peer or to offer transit at access points in their own region. Companies such as Yahoo, America On Line, Google and
Amazon.com have established “public” peering points in different regions, closer to end users. With greater North-South connectivity,
in-region exchange of Internet traffic in developing nations is likely to increase.
■ More symmetrical traffic flows: In the early days of the Internet most of the content was located in North America and Europe,
creating a disproportionate demand for content housed in these two regions. Now compelling content is being developed in other
regions of the world. This is having a significant impact on traffic flows, which will become more symmetrical as a result.
■ Quality of service concerns: Quality of service concerns favour carriers that can reduce the number of network hops, and
transiting through unaffiliated carriers. To highlight superior quality, some ISPs now keep traffic on their networks as long as possible
by self-provisioning links instead of using transit lines. This is known as “cold potato” routing. (This contrasts with “hot potato”
routing, where an ISP hands traffic to another ISP as close as possible to the first ISP’s service territory.)
RELATED INFORMATION

4.2 About VoIP

Internet telephony, or “Voice over the Internet Protocol” (VoIP), is a category of services that enable users to make real time voice calls,
transmitted over the Internet (rather than using traditional circuit switched telephone networks).
VoIP enables network operators, service providers, and consumers make significant savings, by:
■ Reducing the underlying costs of a telephone call. VoIP uses network resources much more efficiently than conventional telephone
service, reducing the costs of providing a call (albeit with the loss of some call quality and service features), and
■ Creating opportunities for regulatory arbitrage that enable service providers and consumers to reduce or avoid call charges and/or
regulatory fees.
Currently the volume of voice telephony traffic is small compared to traditional, dial up, circuit-switched telephone services. However, the
very real potential exists for packet switched, Internet Protocol networking to become the primary medium for most voice and data
services. Should this occur, information services (including VoIP) will become the primary end user service provided by
telecommunications networks.
■ Different types of VoIP,

■ The similarities and differences between VoIP and conventional telephony, and
■ Some key protocols that support VoIP.
For a helpful powerpoint presentation on VoIP technology basics, see Peter Ingram, Voice Over Internet Protocol—An Introduction.
RELATED INFORMATION
About the Internet

Arbitrage Opportunities in the ICT Sector
VoIP and Regulation
Interconnection Pricing for VoIP
VoIP Over Wireless Networks
Module 7, section 4.4, "Hot Topics: VoIP"
Module 7, "Technology Trends: VoIP"
4.2.1 Types of VoIP
VoIP services differ depending on whether:
■ The service provides a competitive alternative to conventional telephone services,

■ A conventional telephone can transmit and receive calls,
■ Subscribers need to acquire and install additional equipment on their premises,
■ Traffic routes into or from the PSTN, and
■ Users pay for service.
This section describes three broad categories of VoIP service: Internet telephony via computer; Internet telephony that is partially
accessible from and to the PSTN; and Internet telephony that is fully accessible from and to the PSTN.
Internet Telephony via Computer
Internet telephony via computer offers a variety of voice communications services. These include:
■ Person to person voice calls,

■ Electronic commerce Internet sites that offer a “push to talk” icon that provides an opportunity to talk to a customer service
representative,
■ Video games that enable players to talk to other players.
Computer-to-computer Internet telephony services, such as Skype and Pulver.com require users to download software. Users can then
set up free voice conversations with other subscribers through the Internet. Calls are routed using a peer-to-peer arrangement that uses
the computer of any logged on subscriber as an intermediary for routing traffic on to the intended call recipient.
An end-to-end Internet telephony connection via computer requires that parties have personal computers equipped with:
■ Compatible software,
■ A sound card or similar device, and
■ Access to the Internet.
Figure 1: Internet Telephony via Computer
Source: ITU
Internet telephony via personal computer has several drawbacks:
■ Typically, calls do not access the PSTN (unless one of the computers accesses the Internet via a modem and conventional dial-up
telephone line),
■ Subscribers must log onto the service in order to make and receive calls,
■ The service does not provide caller identification and location information needed in emergencies,
■ Finally, the service does not offer the same sound quality and reliability as conventional circuit switched telephony.
For these reasons, most countries treat Internet telephony via computer as an unregulated information service, largely free of traditional
telephone carrier responsibilities.
Internet Telephony Partially Accessible from and to the PSTN
This category of VoIP calls includes:
■ Long distance telephone calls originated by subscribers of incumbent carriers, and by users of calling cards who call from
payphones and mobile phones. In both cases calls originate and terminate over the PSTN, but transit the Internet for all or a portion of
the long haul,
■ Internal corporate VoIP traffic that originates and terminates over an enterprise network. Some enterprise networks can route traffic
into the PSTN,
■ VoIP services that enable customers to make calls over the Internet. Such calls typically originate over a broadband Internet link and
terminate at the call recipient’s end without ever traversing the PSTN (see Internet Telephony via Computer). However, these services
can also deliver traffic to non-subscribers over the PSTN and a normal telephone handset (see Figure 2).
Figure 2: Internet Telephony Partially Accessible from and to the PSTN
Source: ITU
The Internet and conventional circuit switched telephone numbering systems use different addressing systems. Thus VoIP services in this
category must provide call processing software that can “map” Internet Protocol addresses to call recipients with conventional telephone
numbers. The software routes the call as far as possible through Internet networks, to a “gateway” or “point of presence” as close as
possible to the intended call recipient. At that point, the service converts the call to telephony traffic and hands it off to a conventional
telephone network.
To access this category of VoIP services, users need:
■ A subscription to a VoIP service,

■ Broadband Internet access,
■ A modem, and
■ An Analog Terminal Adapter, to configure VoIP onto the user’s DSL or cable modem link. This device converts the call signal from
analog to digital (and vice versa).
The ability of subscribers to access service from conventional telephones, or alternatively to call conventional telephone numbers, makes
this form of Internet telephony more attractive to customers (and therefore more commercially attractive) than Internet telephony via
computer.
Internet Telephony Fully Accessible from and to the PSTN
Many telephone companies already use Internet carriage to handle long distance calls (see Figure 3). The customer making the call may not
even be aware of this.
Most current VoIP services do not use the PSTN for both call origination and termination. In the future, almost all VoIP services will require a
broadband, digital Internet access link. Telephone companies and cable television companies will replace copper networks with optical
fibre. This will enable voice services to ride over a ubiquitous broadband digital network as a software application.
Figure 3: VoIP Fully Accessible from and to the PSTN

Source: ITU
RELATED INFORMATION
Comparison of VoIP and Conventional Telephony

Protocols that Support VoIP
This section compares the similarities and differences between VoIP and conventional dial-up telephony.
What Makes VoIP Similar to Conventional Dial-Up Telephony?
A number of factors indicate that consumers increasingly view VoIP as “functionally equivalent” [1] to conventional telephone service,
including:
■ Increasing numbers of consumers use VoIP as an alternative to conventional service. In making this choice, consumers are trading off
a reduction in quality and some loss of features, for a lower price,
■ Improvements in VoIP service have reduced the difference in quality between VoIP and conventional service,
■ Many carriers partially route calls over the Internet without their customers’ knowledge. In many cases, consumers are unable to
detect differences in quality between VoIP and conventional service,
■ VoIP customers can now obtain a telephone number and receive calls originated on the PSTN,
■ There is evidence that local exchange telephony subscriptions, total switched long distance minutes, and revenues for conventional
dial-up services are declining. This suggests that many consumers are switching to VoIP. (A number of other factors may also
contribute to this trend. For example migration from wireline to wireless services, the proliferation of private line and virtual private line
services that can access the PSTN, and the commingling of voice and data services on the same telecommunications link.)
What Makes VoIP Different from Conventional Dial-Up Telephony?
VoIP differs from dial up telephony in a number of significant ways, summarized in Table 1.
Table 1: VoIP versus Conventional Telephony
VoIP Dial-up telephony
Switching Packet switching. Circuit switching.
Quality Can be a lower quality service (echoes, High quality service.

temporary drop outs or even dropped calls).
Service VoIP is portable. However, the service typically Operator can provide information on the caller’s location.
features cannot provide information on the physical location
of the caller (e.g. for emergencies, or for national
security reasons).
Service Conventional voice service is generally a public, ubiquitous

availability May not be available on a public, ubiquitous basis. service. Many traditional operators are required to offer
Most VoIP operators have not positioned service to anyone that requests it.
themselves as “common carriers”.
Service often requires a broadband link, which

limits availability.
Cost structure Costs typically do not vary with the distance of a Although costs largely do not vary with usage, some
call or the time of day that the call occurs. VoIP regulators require network operators to recover non-traffic
providers are able to exploit differences in sensitive costs on a metered (per minute) basis. Despite
regulatory treatment to qualify for lower cost similar underlying costs, operators usually charge different
network access (arbitrage). rates for different call types / carrier types (domestic or
international, wireless or wireline, or Internet).
Billing / metering Packetized traffic measured on bandwidth used. Traffic metered. Billing based on minutes of use, distance
of call, time of day/week.
Interconnection Arrangements vary depending on negotiation and Local access interconnection charges usually based on
charges market power. Smaller ISPs may have to pay for forward looking costs of the network. For international calls,
the whole link. Zero fee “sender keep all” network operators generally share costs by splitting the
interconnection between very large ISPs. VoIP estimated cost of a link in half.
providers may configure calls to avoid paying any
local exchange access fee.
Regulatory Internet services are generally unregulated. As a Network operators, and in some cases other telephone
treatment result, VoIP providers are often exempt from service providers, must pay regulatory fees.
certain regulatory fees, such as universal service
contributions.
Endnotes
[1] Figure 3.1 in ITU, IP Telephony Workshop, (June 14-16, 2000); available at: http://www.itu.int/osg/spu/ni/iptel/workshop/iptel.pdf.
RELATED INFORMATION
Types of VoIP
Protocols that Support VoIP
VoIP uses a number of protocols to provide transmit voice calls using packet switching.
The Internet Protocol is one of several processing standards for routing Internet traffic. The Internet Protocol ensures that traffic can
reach the intended recipient even though it traverses different networks using different equipment.
Compression algorithms reduce the number of packets that must be transmitted by sampling the voice traffic and reconstructing a
digital replica.
The Real Time Transport Protocol provides procedures for loading packet headers with routing, signaling, and identification information
so that, for example, packets that arrive out of sequence can be rearranged.
The Session Initiation Protocol provides standardized call processing formats. This enables VoIP ventures to offer telephone service
features from ringing and busy tones to call forwarding.
The Transmission Control Protocol manages the complete link of sender and recipient through different networks.
RELATED INFORMATION
Types of VoIP
Comparison of VoIP and Conventional Telephony
Traditional network operators often charge different interconnection rates, depending on the type of call or type of service provider
involved. Often this reflects differences in regulatory treatment between service providers. This creates opportunities for service
providers to engage in arbitrage (either legally or illegally).
Arbitrage can cause marketplace distortions and reduce the effectiveness of regulation. If legislatures and regulators do not promptly
adjust the regulatory policy that triggered such arbitrage, the impact on the market can be substantial. Some common arbitrage strategies
are discussed here.
Not all regulatory arbitrage strategies violate laws and regulations even though they deviate from regulatory intent, or exploit loopholes.
Also, when network operators create arbitrage opportunities in the absence of a regulatory obligation, or if they fail to close a loophole
quickly once it is detected, this may indicate that they themselves expect to benefit. Operators will tolerate some loss of revenue if it is
outweighed by other benefits, such as regulatory relief or compensation.
Certain features of VoIP traffic create additional arbitrage opportunities. VoIP traffic can readily enter the Internet without traversing the
PSTN. Opportunities also exist for terminating VoIP traffic without traversing the PSTN, or through undetected transit of the PSTN. Even
when a PSTN operator is able to detect VoIP traffic, it may not be able to differentiate between local, domestic, and international VoIP calls
for billing purposes.
Arbitrage may involve:
■ Qualifying services as long haul transmission in order to avoid universal service surcharges,
■ Obscuring the origin of traffic to making international traffic appear domestic and long distance traffic appear local, in order to obtain
the most favourable access price,
■ Characterizing traffic as local instead of long haul, to generate a reciprocal payment obligation (instead of a one-way access charge),
■ Distorting or obscuring the origin of traffic and the method of transmission to reduce or avoid charges imposed by another carrier for
delivering the traffic to the intended recipient,
■ Offer telecommunications services as ancillary to, or a minor transport element for, an enhanced information service.
RELATED INFORMATION
About the Internet

About VoIP
VoIP and Regulation
A number of arbitrage strategies are sufficiently common that they warrant specific mention. These are:
■ Grey market strategies,

■ Leaky PBXs,
■ Resale of private lines,
■ International call reorigination (or "call-back"),
■ Refiling, and
■ Routing calls over the Internet.
Grey Market Strategies
These are primarily “self-help” strategies by which consumers can reduce telecommunications charges, usually with the assistance of a
business venture. Grey market strategies generally involve masking the true origination point of a call to qualify for a lower price.
Leaky PBXs
A business customer with an on-premises switchboard (Private Branch Exchange, or “PBX”) can “leak” traffic into the local exchange, for
example to avoid paying access charges on long distance calls.
For example, say a law firm has offices in Geneva and Zurich, with a private leased line linking them. The firm could use its PBX in each
office to provide local exchange access at both ends of the private line, while avoiding the normal local access charge (see Figure 1).
Figure 1: Example of Leaky PBX Strategy

Resale of Private Lines
Where there is a big gap between retail long distance or international calling charges and underlying costs, this creates an arbitrage
opportunity. Entrepreneurial ventures can market themselves as long distance or international service providers by reselling private lines.
The customer calls a “free phone” access number, and from there is linked with the private line for low cost calling.
In many countries this form of arbitrage has provided the first significant competitive pressure in the long distance calling market. However,
often this competition is limited, as resellers are poorly capitalized and have little incentive or ability to invest in new network facilities.
International Call Reorigination
This form of arbitrage, also known as “international call-back” takes advantage of the vast difference in international call charges across
countries. Call-back service enables callers in a country facing high international charges to trigger a dial tone in a low-cost country to
make an international call. The call appears to have originated in the low-cost country, and is billed accordingly (see Figure 2).
Figure 2: International Call-back
Some countries have sought to ban call-back services. However, in practice such a ban is difficult to enforce. Call-back has created
downward pressure on international call prices, and has contributed to pressure to move away from the accounting rate system for
international calling.
Refiling
Refiling involves cutting an international link into two or more shorter legs that collectively have a lower international accounting rate than
the single link. For example, the United States and the United Kingdom have achieved a competitive marketplace and lower accounting rates
than have existed in other European nations. Therefore, traffic from the United States to Portugal might be routed first between the United
States and the United Kingdom and then retransmitted for the leg between the United Kingdom and Portugal (see Figure 3)
Figure 3: Refiling
Routing Calls Over the Internet
In many countries regulators and network operators treat Internet traffic more favourably than traditional voice traffic. This creates
incentives for network operators and service providers to route long distance telephone calls over the Internet, to qualify for lower
charges. For example, Internet telephony may be exempt from certain regulatory fees, such as universal service contributions. VoIP
providers may also be able to avoid local exchange access charges, by acquiring inbound business telephone lines that their customers
can use to access the service, or by avoiding the PSTN altogether.

Technological innovations are continuing to make Internet telephony functionally closer to dial up telephone service. For example, it is now
possible to access Internet telephony services using an ordinary telephone handset. As VoIP becomes more similar to conventional
telephony, VoIP providers will compete more directly with incumbent telecommunications operators.
National legislatures and regulators will eventually have to decide what aspects of conventional telephony regulation should apply to VoIP
service. Once a significant volume of telephone traffic is carried over Internet networks, the differences between VoIP and conventional
traffic will have implications for universal service arrangements, telephone number management, public safety, and national security. For
example, VoIP services are not available on a public, ubiquitous basis. In addition, they are generally unable to provide access to
emergency service, or give location information in case of emergency.
This section reviews:
■ Key implications of VoIP for regulators,

■ Trends in the regulation of VoIP services, and
■ The effect of differential regulation of VoIP and conventional telephony.
RELATED INFORMATION
About the Internet

About VoIP
Interconnection
Legislative Approach
ITU Regulatory News: VoIP
Module 7, section 4.4, "Hot Topics: VoIP"
Module 7, "Technology Trends: VoIP"
VoIP presents a particularly compelling challenge to regulators. Decisions on the regulatory status, availability, and price of VoIP services
will directly affect the economic viability and future regulatory status of incumbent operators.
VoIP has the potential to erode the market share and profitability of incumbents. VoIP services can traverse the telephone network without
detection. Thus, even where regulators permit only limited or no VoIP services, incumbent operators will still face competition from this
source. Incumbent operators may no longer be able to expect voice traffic to generate lucrative revenues and profits.
In response to this competitive pressure, incumbents may seek regulatory relief. For example, incumbent operators may approach
regulators seeking:
■ Regulatory parity with new entrants, for example by removing asymmetric regulation not imposed on other operators, or
■ Protection from competition, for example by banning or seeking to limit VoIP services.
Finally, regulators will have to consider how best to encourage incumbent operators to retrofit their existing networks and install new digital
plant, optimized for switching and routing data (of which VoIP will be a significant component in the future).
RELATED INFORMATION
Trends in VoIP Regulation

Differential Regulation of VoIP and Conventional Telephony
In many countries Internet telephony qualifies for streamlined regulation on grounds that it is an “enhanced”, “value added”, or information
service (consistent with regulatory treatment of the Internet generally).
As VoIP becomes a closer substitute for conventional voice telephony, regulators may be less inclined to eliminate regulatory requirements.
This is particularly the case where VoIP services are close substitutes for traditional telephony, for example where VoIP operators seek
telephone number assignments and number portability.
Most of those countries that have developed a VoIP regulatory policy have adopted a “light handed” approach in general, and have targeted
regulatory interventions to specific matters, such as access to telephone numbers, number portability, access to emergency services,
universal service, and national security.[1]
In the European Union, VoIP can be classified as either an Electronic Communication Service, or a Publicly Available Telephone Service, and
is regulated accordingly. Which classification applies depends on variables such as:
■ Availability of the service to the public,

■ Use of telephone numbers, and
■ Access to emergency services.
Increasing VoIP traffic will undermine the profitability of incumbent operators, and sources of revenue such as international accounting rate
settlements and access charge payments for terminating voice telephony traffic. As a result many governments prohibit or try to limit VoIP
services.
In its 2004 Regulatory Survey the ITU reported that 37 nations restrict VoIP to licensed Public Telecommunications Operators, while 49
nations allow full competition (see Figure 1). Countries that attempt to limit VoIP to incumbent operators include Azerbaijan, Jordan, Costa
Rica, Cote D’Ivoire, Egypt, Ethiopia, and Ghana.[2] Countries that do permit VoIP (possibly subject to limited regulation) include Australia,
Canada, Columbia, the European Union, Hong Kong, Indonesia, Japan, South Korea, and the United States.
China initially banned Internet telephony, however despite the ban VoIP services have flourished. The Chinese government is reviewing the
situation. Still other countries, such as Kenya, allow independent VoIP offerings, but subject the service providers to licensing as well as
certain operating requirements.
In practice, a ban on VoIP services is not readily enforceable, because outbound traffic can originate via most broadband connections to
the Internet and inbound traffic can merge with permissible voice telephony. Some commentators estimate the volume of “gray market” VoIP
services at 30 to 50 percent of international voice traffic.[3]
To date, few countries have specifically addressed VoIP interconnection and pricing. VoIP operators have managed to secure satisfactory
arrangements on arm’s length commercial terms. The relatively low volume of VoIP traffic has not yet triggered demands for compensation
from most network operators. However, as VoIP traffic increases, the potential for disputes and outright refusals to interconnect will rise
and regulators will have to turn their minds to VoIP interconnection.[4]
Figure 1: Regulatory Status of IP Telephony, 2004
Note: Based on responses from 149 economies. “Prohibited” means no service is possible. “Restricted” means only licensed PTOs can
offer the service. “Partial competition” means non-licensed PTOs may use either IP networks or the public Internet. “Full competition”
means anyone can use or offer service.
Source: ITU World Telecommunication Regulatory Database (2005 questionnaire).
Endnotes
[1] Organisation for Economic Co-operation and Development, Communications Outlook 2005, ISBN 9264009507 (2005), pp 53-54.
[2] For background on VoIP Regulation in Africa see Commonwealth Telecommunications Organisation, Workshop on VoIP, Fourth African
Internet Summit and Exhibition, Abuja (Jan. 23, 2005) Power Point presentation of Marcel Belingue. See also, Commonwealth
Telecommunications Organisation, An Overview of VoIP Regulation in Africa: Policy Responses and Proposals (2005).
[3] Jerome Bezzina, "Interconnection Challenges in a Converging Environment—Policy Implications for African Telecommunications
Regulators", The World Bank, Global Information and Communication Technologies Department (June, 2005) , p.30.
[4] Organisation for Economic Co-operation and Development, Communications Outlook 2005, ISBN 9264009507 (2005).
RELATED INFORMATION
Implications of VoIP for Regulators

Differential Regulation of VoIP and Conventional Telephony
Many countries regulate information services and traditional telecommunications services differently.
Differential regulatory treatment creates opportunities for arbitrage. It also encourages incumbent network operators to:
■ Focus new investment into unregulated broadband networks, and

■ Migrate services (including voice telephony using VoIP) onto those new networks wherever possible.
This behaviour achieves operational savings, and also qualifies voice telephony traffic for a lower level of regulation.
The result will be an increase in the volume of information services, and a reduction in the volume of voice telephony minutes of use that
are subject to interconnection charges, or international accounting rate settlements. Network operators’ traditional sources of revenues will
erode, forcing regulators to rethink how network operators should be permitted to recover their costs.
RELATED INFORMATION
Implications of VoIP for Regulators

Trends in VoIP Regulation

As network operators migrate to next generation networks, voice services will become simply a software application riding over the
network. Converging technologies and markets make conventional approaches to interconnection charging unsustainable.
Many technology forecasters predict that in the future voice telephony will migrate completely from circuit switched telephony to VoIP.
Once this happens, Internet interconnection and pricing models may replace the current arrangements. In the interim, VoIP network
operators will need to interconnect with incumbent network operators’ PSTNs. This section addresses:
■ The differences in cost recovery between the Internet and conventional telephony,
■ Interconnection models by Internet Service Providers (ISPs), namely peering and transit,
■ Implications of VoIP for interconnection pricing,
■ Pricing mechanisms for VoIP interconnection, and
■ Criteria for a new interconnection pricing regime.
RELATED INFORMATION
About the Internet

About VoIP
VoIP and Regulation
Benchmarking Rates for Network Access
Cost recovery models in telecommunications and for the Internet differ substantially. As technologies and markets converge, these
differences are creating opportunities for arbitrage. This section compares the cost recovery models for telecommunications and Internet
interconnection.
Table 1: Telecommunications versus Internet Cost Recovery
Telecommunications Internet
Cost recovery subject to significant regulation and government oversight. Little or no regulatory oversight. ISP contracts are
Settlements are generally transparent. typically subject to non-disclosure agreements, making
it difficult for outsiders to determine access terms and
Network operators provide transmission, possibly with service conditions.
enhancements.
ISPs combine transmission and content, making it
Settlements based on traffic flows and charged on minutes of use. (May difficult to decouple the costs of each element.
include a fixed component to recover non-traffic sensitive costs.)
Cost recovery based on link capacity. Charged on
International traffic settled on measured traffic volumes, and a "half-circuit" bandwidth and derived throughput of the link.
approach to sharing the costs of the international link.
ISP network access provides onward transit to many
Settlements typically operate on a destination specific basis. other networks and destinations. In the extreme this
Under the accounting rate settlement model, the same system applies for all provides global reach. ISPs can exploit this access to
network operators, regardless of size, traffic volume, or geographical reduce their costs, using "hot potato routing."
reach. (As traffic moves away from the accounting rate model, larger ISPs use different charging models, depending on the
operators will be able to negotiate cheaper access arrangements.) characteristics (and bargaining power) of the ISPs
involved.
RELATED INFORMATION
Models for Internet Interconnection

Implications of VoIP for Interconnection Pricing
Pricing Mechanisms for VoIP Interconnection
Criteria for a New Interconnection Regime
Internet Service Providers (ISPs) use different models for interconnection pricing, depending on the specific characteristics of the ISPs
concerned. Broadly, ISPs can either:
■ Enter into “peering” arrangements, or

■ Enter into a transit arrangement.
Peering
Peering, also known as “Sender Keep All” or “Bill and Keep” is a zero compensation arrangement by which two ISPs agree to exchange
traffic at no charge. This kind of arrangement makes sense where the two ISPs have roughly the same characteristics and traffic volumes,
such that net financial burden from traffic flows between them is likely to be small.
The process by which an ISP qualifies for peering remains private. ISPs negotiate terms and conditions privately. They only rarely publicly
disclose the criteria they use to qualify for peering. However, several Tier-2 ISPs have posted general qualifications for agreements to peer
on their web sites. As an example, SBC’s conditions for domestic peering in the United States are summarized in the Box below. These
conditions emphasize network coverage, volume of traffic, and 24 hour a day network maintenance capability. These criteria are probably
more liberal than a Tier-1 ISP would require.
SBC Conditions for United States Domestic Peering
1. For domestic ISPs coast-to-coast nationwide OC-12 or larger public IP backbone network.
2. Presence at three or more public peering points (at least one on the East Coast, one on the West Coast, and one in the Mid West) for
domestic ISPs.
3. Presence at two or more public peering points for International ISPs.
4. A total minimum busy hour traffic exchange of 25 Mbps with SBC Internet’s Autonomous System Numbers will be required.
5. Must not have been an IP transit customer of SBC Internet in the past six (6) months.
6. Willingness to enter into a Bilateral Interconnection Agreement and Non-Disclosure Agreement with SBC Internet.
7. Operation of a 24x7x365 Network Operations Center (NOC) that proactively monitors all peering connections and provides an
escalation path to quickly identify and resolve network problems.
8. No requirement for a balanced traffic exchange ratio due primarily to the asymmetric nature of current broadband metallic
transmission systems such as ADSL and cable modems and of current Internet Data Centers.
9. Joint capacity planning reviews for interconnection augmentation to accommodate traffic growth and minimize the possibility of
latency or packet loss between both networks.
10. Consistent routes announcements at all public peering points.
Transit
Transit is an arrangement in which larger ISPs sell access to their networks, their customers, and other ISP networks with which they had
negotiated access agreements.
Under a transit arrangement, the sender pays the full cost of interconnection. Transit charges are set by commercial negotiation, and are
generally not disclosed.
Internet transit access arrangements provide a much greater geographical access than telecommunications transit arrangements. In
telecommunications, transit arrangements typically secure an indirect link to one carrier in one location (primarily because a small carrier is
unable to secure a direct link). Internet transit arrangements typically provide access to a vast array of networks, not limited to one country.
At the extreme, one Internet transit payment arrangement with one major Tier-1 ISP can provide a small, remote ISP with access to the Rest
of the World. This is because the Tier-1 ISP has ubiquitous access and so can provide extensive routing opportunities.
RELATED INFORMATION
A Comparison of Telecommunications and Internet Cost Recovery

Changes in how telecommunications services are delivered, including the emergence of VoIP will have significant implications for
interconnection pricing. In particular, the opportunities VoIP creates for arbitrage create pressures to:
■ Move toward cost-based pricing for interconnection (and other telecommunications services), and
■ Adopt uniform charges for access, regardless of the type of call, type of service providers, or other call characteristics.
Cost-based Pricing
Traditionally, telecommunications prices have been designed to keep prices for access and “basic” local service low, at the expense of
long-distance users. The resulting high long-distance prices have created numerous opportunities for arbitrage, which have placed
downward pressure on prices.
Recognizing that the traditional model is unsustainable and inefficiency, many regulators are now moving away from this model towards a
more cost-based model. This shift is often involves a long transition period, to avoid significant immediate jumps in prices for basic service.
Generally, pricing reforms are accompanied by a shift to transparent funding of universal service obligations, through explicit charges to
interconnecting service providers, or directly to end users.
Uniform Access Charges
It is common for network operators to charge different access prices depending on the type of call, the type of service providers, or the
distance involved. This creates opportunities for arbitrage.
In many cases it makes more sense to move to a uniform charging regime, for example:
■ Network operators, especially long-distance, average long and short haul traffic costs and charge a flat rate for calls. For example, a
single per minute rate for all calls in a wide geographic area (say, nationwide),
■ “All You Can Eat” pricing — a flat monthly rate for unlimited local and long distance calls. This form of pricing is already standard for
Internet access in many countries,
■ If the cost of measuring the distance between the call originator and call recipient exceed the cost difference in handling traffic of
different distance, then network operators should not bother to do so. In this case, charges should not differ based on distance.
To move to a more sustainable charging regime, regulators will need to:
■ Eliminate regulatory asymmetries that treat similar services differently based on the technology used to provide the services (for
example VoIP or conventional voice service), or the type of provider,
■ Decide whether VoIP providers offering equivalent service to conventional voice telephony should pay the same charges and
regulatory fees as other network operators.
Changes in technology and telecommunications network cost structures mean that per minute pricing may be becoming an inefficient cost
recovery mechanism. As more services are delivered as packets over digital networks, minutes of use are no longer an important cost
driver.
Technical developments are improving the ability of consumers to manage their own telecommunications services. As a result the premise
that the calling party is the sole cost causer may no longer be valid. The Calling Party Pays approach to call pricing (and interconnection
charges) may no longer reflect actual cost causation.
RELATED INFORMATION
■ The application of origination and termination payments to VoIP interconnection,

■ Cost drivers for VoIP,
■ Setting cost-based charges for VoIP interconnection, and
■ Reciprocal payment obligations between VoIP providers and conventional operators.
Application of Origination and Termination Payments to VoIP
VoIP providers require access to the PSTN to terminate calls to recipients who do not subscribe to the VoIP provider’s service, and for
some types of call originations. Such interconnection typically occurs between a VoIP operator’s gateway and the PSTN operator’s Tandem
Switch closest to the call originator or recipient.
For call terminations VoIP operators should pay PSTN operators for call switching and routing, in much the same way that other carriers
(such as mobile and long distance operators) do.
Call originations may require a different pricing and access mechanism. For many VoIP services, the caller originates the call over a high
speed, broadband Internet access link, such as DSL, a cable modem, or a wireless network. The call never has to transit the PSTN, and so
the local network operator typically has no basis for demanding compensation. Thus, originating VoIP traffic parallels a Calling Party Pays
financial model in the sense that VoIP subscribers must pay for and route telephone traffic through their Internet access link.
Cost Drivers for VoIP
Per minute cost recovery has a number of weaknesses in a VoIP world. Call duration has no meaningful relationship to the costs of a VoIP
call. Charging on a per minute basis creates opportunities for VoIP operators to engage in regulatory arbitrage, or to avoid interconnection
charges.
As VoIP traffic increases, interconnection charges based on bandwidth used would better reflect underlying cost drivers, and would be
more consistent with economic efficiency.
Setting Cost-Based Charges for VoIP Interconnection
An interconnection pricing mechanism for VoIP services should reflect the costs of the local network assets used to provide VoIP. If
interconnection prices reflect underlying costs, and appropriate cost drivers, opportunities for arbitrage will reduce. Similarly, where VoIP
operators provide a service that is functionally equivalent to conventional telephony, treating VoIP providers in the same way as
conventional service providers will remove arbitrage opportunities.
Cost reflective interconnection pricing for VoIP could involve:
■ End user payments: This would involve moving to full cost recovery from end users. For example, network operators would
recover most of the costs of the local loop from end users on a pro rata, per-line basis. In many cases this would involve a major
shift in charging, and would need an extensive transition period. In practice this might not work given existing arbitrage opportunities,
■ Unbundling: Unbundling the local loop used by VoIP operators from other local exchange carrier costs should enable regulators to set
charges that reflect the actual costs of the assets used by VoIP providers. However, unbundling does have some disadvantages. If
not done well, unbundling can distort competition further. In particular, the regulator must still determine how to allocate shared and
common costs between VoIP and other services — a challenging and contentious decision. The resulting interconnection price should
not be so low as to discourage entrants from investing in their own infrastructure,
■ Cost based VoIP origination and termination charges: Rather than a full unbundling exercise, this option would involve setting
cost-based interconnection charges. Under a regime of uniform access charges VoIP providers would pay the same origination and
termination fees as other service providers that use the same network facilities and services. These charges should reflect the
economic costs of access to the PSTN, and other switching and routing services provided by network operators.
Reciprocal Payment Obligations
VoIP operators currently do not receive any compensation from PSTN operators for terminating calls that originate on the PSTN. If VoIP
operators are treated in the same way as other service providers with respect to interconnection payments, then they should also have
the same rights to compensation. That is, VoIP providers should also be entitled to reciprocal compensation for terminating calls that
originate on the PSTN.
Effect of Convergence
The emerging trend in retail pricing for VoIP services is to offer to consumers "all you can eat" calling packages at a flat monthly or even
yearly rate. Many of the triple-play offerings (voice, broadband, television) are using this type of flat rate pricing for the voice component
of the service. As a consequence, this will create additional pressure for interconnection pricing for VoIP services to conform to the retail
pricing structure, further undermining the traditional system of charging for interconnection on a per minute basis.
RELATED INFORMATION

As more traffic migrates to VoIP, we will need a new approach to interconnection pricing. Any new approach to interconnection pricing
should:
■ Encourage efficient competition and the efficient use of, and investment in, telecommunications networks,
■ Preserve the financial viability of universal service mechanisms. Thus any proposal that would result in significant reductions in
intercarrier payments should include a proposal to address the shortfall,
■ Treat technologies and competitors neutrally,
■ Allow innovation,
■ Minimize regulatory intervention and enforcement, consistent with the general trend toward less regulation wherever possible.
This implies treating VoIP providers that provide service over the PSTN in the same way as other telecommunications service providers,
including with respect to:
■ Interconnection charges: VoIP providers should face the same payment obligations as other service providers that use equivalent
facilities and services. Similarly, VoIP providers should be entitled to the same reciprocal termination payments from PSTN operators,
■ Regulatory fees: Technology neutrality suggests that all providers (including VoIP providers) whose service accesses the PSTN
should be subject to the same regulatory fees, including universal service contributions,
■ Other regulatory requirements: Where feasible, VoIP providers should have similar obligations to other service providers that offer a
functionally equivalent service (for example with respect to emergency services, or obligations to support law enforcement call
intercepts).
NGN Challenges to Regulators
NGNs will present both recurring and new challenges to regulators. Regulators will have to address whether and how to resolve issues
relating to interconnection terms and conditions, but they may not have complete jurisdiction over all carriers. To the extend VoIP and other
Internet-mediated services do not fit within a telecommunications services classification, regulators may not have lawful authority to
mandate interconnection and to regulate rates. If VoIp and other Internet services fit within an enhanced, or information services category
carriers may lawfully refuse to interconnect. Similarly national regulatory agencies may not have access to privately negotiated, voluntary
interconnection arrangements.
VoIP Impact on Universal Service Funding
VoIP may have an adverse impact on universal service funding as consumers may have the ability to migrate from services that make
contributions to ones that do not. If VoIP traffic can “leak” into the public switched telephone network without triggering access charge and
universal service funding burdens, users of conventional services may have to bear a higher financial subsidy burden.
RELATED INFORMATION


Wireless networks will have a substantial impact on VoIP service development, particularly in developing countries.
As wireless and VoIP traffic increase, differences in the terms and conditions under which wireline, wireless and VoIP operators
interconnect networks will create opportunities for arbitrage, and distort markets. Differences in call termination rates and interconnection
arrangements can cause operators to adjust traffic flows to obtain the lowest possible rate, and to minimize regulatory fees.
Incumbent operators may seek to exploit “bottlenecks” and essential facilities, by imposing above cost termination charges to deliver calls to
wireless subscribers, or to deliver wireless traffic to wireline subscribers. This may encourage wireless carriers and VoIP providers to
avoid the incumbent’s network by seeking cheaper alternatives for originating and terminating traffic.
Examples of alternative delivery options are:
■ Third generation mobile networks: These promise to have sufficient bandwidth and operating standards to support high speed
data services, presumably including VoIP,
■ Wireless-Fidelity (Wi-Fi) networks: Wi-Fi can also support voice telephone calls. Wi-Fi is generally provided as unlicensed
broadband network access, on an stand-alone basis at homes, offices and public “hot spots” such as airport lounges and coffee
shops,
■ Voice Over Wi-Fi (VoWiFi): VoWiFi can integrate Wi-Fi access with licensed third generation mobile services. With seamless
roaming between the two networks, subscribers could use voice over a WiFi network (where available) and mobile connections
where WiFi is missing, or outside a WiFi network. VoWiFi has the potential to allow VoIP providers to completely bypass the PSTN. [1]
■ Worldwide Interoperability for Microwave Access (WiMax): WiMax is a wireless broadband technology, which has a range of
up to 30 miles and can be used for wireless networking like Wi-Fi, but at higher data rates over longer distances.
ENDNOTES
[1] See Organisation for Economic Co-Operation and Development, Directorate for Science, Technology and Industry, Committee for
Information, Computer and Communications Policy, Working Party on Telecommunication and Information Services Policies, Development of
Voice Over WIFI by Integrating Mobile Networks, DSTI/ICCP/TISP (2004) 9/FINAL (April 14, 2005), at p. 6.
RELATED INFORMATION
About the Internet

About VoIP
VoIP and Regulation

This section provides international benchmarks of voice telephony call rates and broadband data rates: [1]
■ Table 1 summarizes approximate interconnection fees and rate elements charged by carriers in several developed and developing
nations for local exchange access,
■ Table 2 provides rates for broadband services for selected countries.
The data benchmarked here should be interpreted with care. The terms and conditions of telecommunications carrier-to-carrier
interconnection vary on the basis of such factors as traffic density, routing distance, number of switches transited, equipment efficiency
and time of day. Regulatory policies also have a major effect, particularly if rates include non-cost factors such as universal service
contributions.
Some network access rates may be comparatively high, because they must compensate carriers for fixed costs that do not vary with
usage. Some nations require telephone subscribers to pay a monthly fixed fee for the privilege of having network access regardless of
usage. If subscribers do not pay such a fee per minute service charges may have to include a surcharge.
Table 1: Benchmarking Voice Telephony Call Rates
Nation or State Local Level Single Transit Double Transit
Fixed-to-Fixed EU 0.80 Euro-cents per

0.75 Euro-cents per MOU 0.63 Euro-cents per MOU
average 2003 MOU [2]
0.001147 cents per MOU originating; Tandem Switching 0.000481 cents

New York 2005 Port rate $2.57 flat rate
0.00111 cents per MOU terminating and 0.000203 common transport
1.85 Euro-cents per
Hungary 2004 2.4 Euro-cents per MOU 2.76 Euro-cents per MOU
MOU
Czech Republic
1.3 1.62 2.06
2004
Japan 2004 1.727 local switching yen per MOU 2.057 yen per MOU
0.20 - 1.10 Rupee

India 2005
based on mileage bands
Malaysia 2005 2.6 Sen per MOU 4.8 Sen per MOU 8.43 Sen per MOU
Table 2: Benchmarking Broadband Data Rates
Nation or 56/64 kilobits per second at 2, 1.544/2 megabits per second at 2, 155 megabits per second at 2,
State 50 and 200 kilometres 50 and 200 kilometres 50 and 200 kilometres
United Kingdom 11,964; 33,908; 71,795 Euro per

229; 460; 583 Euro per month 464; 1691; 3612 Euro per month
2004 month
New York 186; 410; 1009 Euro per month (56 488; 1578; 4635 Euro per month
n/a
2004 kbps) (1.544 mbps)
88; 233; 689 Euro per month (56 360; 1234; 3651 Euro per month
California 2004 n/a
kbps) (1.544 mbps)
Germany 92; 414; 478 Euro per month 340; 1979; 2504 Euro per month 1600; 7511; 12,161 Euro per month
7813; 13,727; 23,868 Euro per

Hungary 193; 337; 621 Euro per month 702; 2149; 5003 Euro per month
month
9302; 50,583; 122,998 Euro per

Japan 655; 1122; 1190 Euro per month 3164; 7301; 9046 Euro per month
month
Malaysia 2.6 Sen per MOU 4.8 Sen per MOU 8.43 Sen per MOU
ENDNOTES
[1] Benchmarking data is sourced from Teligen, the OECD, the European Commission, Intug, and the National Regulatory Research Institute.
[2] “MOU” stands for “Minutes of use”.
RELATED INFORMATION

Regional Internet Exchange Points (IXPs) play an important role in reducing the costs of ISPs and encouraging development of the Internet in
developing countries. This section discusses:
■ The role of regional IXPs,

■ Ways to support the development of IXPs in developing countries, and
■ The development of IXPs in Africa.
RELATED INFORMATION
About the Internet
Because the Internet offers access to content and users anywhere, each ISP has to secure network connections to all potential senders
and recipients of content, or suffer competitively suffer for the lack of global reach. Reciprocal interconnection — whether freely
provisioned or provided for a fee — makes it possible for an ISP to access the entire global Internet “cloud” for its subscribers.
The Internet operates almost free of regulation, so Tier-1 ISPs can largely dictate interconnection terms and conditions. ISPs in remote areas
(including most developing countries) must meet the entire cost of accessing larger “Tier-1” ISP networks, using expensive international
satellite links or submarine cables.
In some cases, where there is no local or regional facility for the exchange of Internet traffic, developing country ISPs must pay for
international transit facilities to deliver local traffic. This practice is known as “tromboning” (see Figure 1).
Figure 1: “Tromboning” Developing Country Internet Traffic
A key way to reduce Internet traffic costs for developing country ISPs is through the development of regional Internet Exchange Points
(IXPs).
IXPs provide a centralized hub and spoke network typology (see Figure 2). These enable ISPs to hand off traffic directly to other nearby
ISPs, and to aggregate long haul access. IXPs offer traffic switching and routing flexibility. By using an IXP, ISPs can individually and
collectively reduce their bandwidth and line transmission costs, provide more reliable service with lower latency, and operate more
efficiently.
Figure 2: IXP Hub and Spoke Networking

IXPs provide a centralized location for the exchange of traffic, close to both the originators and recipients of traffic and content. Having
access to an IXP has a range of benefits for local ISPs, and ISP subscribers. IXPs:
■ Reduce the ISP costs, and enable ISPs to manage traffic more efficiently,
■ Improve quality of service by reducing the transmission time, number of routers, and distance traffic must travel,
■ As a result, add value to an ISP service subscription. This creates new growth and development opportunities,
■ Provide a neutral, universally supported “clearing house” for the exchange of traffic, making it possible to keep local traffic local.
RELATED INFORMATION
Supporting IXPs in Developing Countries

Internet Exchange Points in Africa
Internet Exchange Points (IXPs) in developing countries are important for a number of reasons. They:
■ Enable efficient, cost effective management of Internet traffic,

■ Provide an interface between multiple ISPs. This enables ISPs to avoid tromboning local and regional traffic,
■ Should help stimulate market entry by new ISPs, web hosting and equipment co-location developers, and content creators.
This page reviews key challenges in establishing IXPs in developing countries, and suggests some elements of “best practice”.
Challenges for Regional IXPs
IXP development has lagged in developing countries. This appears to be due to regulatory, coordination, and management challenges. Key
barriers include:[1]
■ Legal restrictions: For example prohibitions on non-regulated telecommunications facilities, regulated monopolies on international
connectivity, restrictive licensing regimes, and unfavourable tax treatment,
■ Telecommunications sector regulation: Regulators may seek to extend their statutory mandate over telephony to Internet
infrastructure, and
■ Incumbent network operators and dominant ISPs: Dominant incumbents may seek to prevent effective competition by blocking
the development of IXPs.
Installing an IXP is not straightforward. It requires extensive coordination between various stakeholders including the incumbent
telecommunications operator(s), two or more local or regional ISPs, the national regulator(s), and sources of funding.
Most of the effort in developing an IXP is in building the necessary relationships and support. Within the country, competing ISPs must
understand that an IXP will not tilt the competitive playing field in favour of one or more operators. ISPs in adjacent nations need to
understand the value in routing traffic to the IXP, rather than attempting to develop their own facility.
IXP organizers will need to address the concerns of both incumbent telecommunications carriers and regulators. Incumbent network
operators may view IXPs as a threat, particularly in light of the ability of VoIP services to evade regulation, access charge responsibilities,
and universal service funding obligations. For similar reasons, regulators may see an IXP as a threat to telecommunications development
goals.
Finally, official legislative and regulatory support for IXP development is important to boost an IXP development project.
Best Practice for Regional IXPs
A “best practice” approach to IXP development would include:
■ An environment that views Internet development as beneficial, not a threat. This can take considerable effort to achieve — many
stakeholders must be “brought on board”, some of whom may see an IXP as a threat,
■ Neutral, transparent rules for interconnection. The interconnection regime should not favour any single ISP, or class of ISPs,
■ No unnecessarily burdensome regulatory and licensing barriers for the IXP operator,
■ Arrangements to share the costs of operating the IXP equitably between users. This might also apply to the cost of initial construction
and installation of equipment (unless a separate party is prepared to fund this).
Endnotes:
[1] Harvard Law School, Berkman Center for Internet and Society, Berkman Online Lecture and Discussion, Part 4 - Solutions in the
Architecture, Interconnection in Developing Countries (or "The Missing Links") (last accessed November 19, 2006). at pp. 1-2.
RELATED INFORMATION
The Role of Internet Exchange Points

Internet Exchange Points in Africa
Until recently the African region was especially disadvantaged by the absence of Internet Exchange Points (IXPs).
Compared to other continents, Africa had limited connectivity options and low initial traffic volumes. As a result, African ISPs often faced
high transmission costs, even when routing local and regional traffic, due to the need to “trombone” traffic. Tromboning increases delays
and can reduce the quality of the transmission.
In addition, African ISPs pay a substantial premium for overseas connections. International connectivity charges can be between 15 and 26
times greater than their equivalent local costs.[1]
In response to these pressures, IXPs are now emerging in the African region. Fifteen IXPs currently operate in Africa (see Table 1 below).
[2]
Despite the vast geographical separation between African IXPs, the development of IXPs has generated operational, cost, and quality of
service improvements. These improvements are primarily due to that fact that ISPs no longer need to route local traffic through an out-of-
region IXP.
For African ISPs, “tromboning” adds 200 to 900 milliseconds to each transmission. This type of latency can frustrate or foreclose the
development of new revenue sources (such as Internet telephony, streaming audio and video, video-conferencing, and telemedicine). With
a local IXP in place two ISPs (located near to each other) can route traffic to each other’s networks in 5 to 20 milliseconds.
Routing less traffic out-of-region reduces total international line lease expenses. Finally, the ratio of outbound traffic to inbound traffic has
also fallen. It is this ratio that most ISPs consider when deciding whether to peer with another ISP, or to require transit payments. Thus a
reduction in outbound traffic relative to inbound traffic has the potential to significantly improve an ISP’s competitive position.
Table 1: Internet Exchange Points in Africa
Country City Name Date Started Peers Traffic (Mbps)
South Africa Johannesburg JINX December 1996 15 45 [3]
Kenya Nairobi KIXP February 2002 13 8
Mozambique Maputo MozIX July 2002 7 4
DRC Kinshasa PdX November 2002 4 1
Eqypt Cairo CR-IX December 2002 9
Nigeria Ibadan IBIX March 2003 2 0.200

Tanzania Dar es Salaam TIX June 2003 10 1
Uganda Kampala UIXP July 2003 5
Swaziland Mbabane SZIX June2004 3 0.128
Rwanda Kigali RINEX July2004 6 0.400
Source: Association of African Internet Service Provider Associations.
Endnotes:
[1] Jerome Bezzina, Interconnection Challenges in a Converging Environment — Policy Implications for African Telecommunications
Regulators, The World Bank, Global Information and Communication Technologies Department (June 2005), p. 13.
[2] For more information on IXPs in Africa, see:
■ Global Internet Policy Initiative, Internet Exchange Points: Their Importance to Development of the Internet and Strategies for their
Deployment - The African Example (June 6, 2002, revised May 3, 2004), and
■ Russell Southwood, International Telecommunication Union, Via Africa Creating local and regional IXPs to save money and bandwidth,
Discussion Paper Prepared for IDRC and ITU for the 2004 Global Symposium for Regulators.
[3] Much of the local traffic exchanged in South Africa is via private peering links at JINX, rather than through JINX. This is an artifact of a
historical anomaly called "Equivalent Line Charges", which distorts the costs of using JINX.
RELATED INFORMATION
The Role of Internet Exchange Points

Supporting IXPs in Developing Countries
5 Regulating Prices
If effective competition is not possible in wholesale or retail markets, it may be necessary to regulate the prices dominant firms can charge.
Without price regulation, dominant firms can increase prices above competitive levels, harming their customers.
This section of the toolkit covers key issues in regulating prices:
■ The justification for ex ante price regulation – why regulate prices?

■ Economic and accounting approaches to measuring costs,
■ Determining the structure and level of regulated prices,
■ Benchmarking prices,
■ Methods of price regulation, specifically rate of return regulation and incentive regulation,
■ The relative merits of rate of return regulation versus price caps,
■ Issues in implementing price caps, including defining the basket(s), assessing price variations, calculating the efficiency factor, and
incorporating service quality and exogenous costs, and
■ Double price caps.
In addition, this section provides an overview of economic concepts that are particularly relevant to price regulation, and key pricing
principles.
RELATED INFORMATION

Regulation has potentially high costs. Among other things, it substitutes the regulator’s judgment for market interactions. No matter how
capable and well intentioned regulators are, they will never be able to produce outcomes as efficient as a well-functioning market.
Regulators should therefore forebear from interfering in pricing decisions unless regulation is justified. That is, unless the expected benefits
from regulating prices outweigh the expected costs from doing so. This requires that, without regulation, prices will either be:
■ Too high overall — if an operator or service provider has market power they may increase prices above competitive levels. This will
suppress demand for the service, leading to a loss of social welfare, or
■ Anti-competitive —an operator or service provider with market power may engage in pricing practices that hinder competition in a
market. Three important anti-competitive pricing practices are cross subsidization, price squeezes, and predatory pricing.
Regulatory Options
If there is a case for price regulation, a number of regulatory options exist. These include:
■ Rate of return regulation,

■ Incentive regulation, and
■ International benchmarking of prices.
Regulatory Criteria
The list below sets out common regulatory goals, which provide useful criteria for assessing regulatory options:
■ Prevent the exercise of market power: An important goal of regulation is to ensure that prices are fair and reasonable, where
competitive forces are insufficient. Any regulatory price control mechanism should encourage prices that reflect what one would
observe in a competitive environment,
■ Achieve economic efficiency: The regulatory mechanism chosen should improve economic efficiency. There are several
measures of economic efficiency:
❍ Technical efficiency (or “productive efficiency”) requires that goods and resources produced in the telecommunications
industry should be produced at the lowest possible cost. This ensures that society’s scarce resources are used efficiently and
are not wasted,
❍ Allocative efficiency requires that the prices one observes in a market are based upon and equal to the underlying costs that
society incurs to produce those services (generally the long run incremental cost of producing the service). This will ensure that
customers whose valuation of the service exceeds the cost of producing the service will purchase the service. Customers
who place a lower valuation on the service will forgo it. This ensures that the “optimal” amount of the service is consumed,
given cost and demand conditions. In the ICT sector prices must include some mark-up to recover shared and common costs.
Mark-ups should be set so as to minimize the impact on allocative efficiency, and
❍ Dynamic efficiency requires that firms should have the proper incentives to invest in new technologies and deploy new
services,
■ Promote competition: Many regulators operate under a legal framework where the goal is to permit and promote competition in
telecommunications markets. Where the legal framework permits competition, it is important that regulation (at a minimum) does no
harm to competition,
■ Minimize regulatory cost: All else being equal, regulators should choose a regulatory mechanism that is less costly to implement
over one that is costlier to implement,
■ Ensure high service quality: In addition to ensuring that the prices of telecommunications services are fair, regulators are also
concerned that consumers should receive a high quality service. In ranking alternative regulatory options, regulators should give
preference to mechanisms that result in higher quality service, all else being equal,
■ Ensure telephone prices are competitive with other jurisdictions: This is a relevant objective in countries, such as
Singapore, that use telecommunications infrastructure as a tool for competitive advantage. In these countries, telecommunications
infrastructure plays an important role in attracting foreign investment. It is therefore important that telecommunications prices are
competitive with other possible destinations for foreign investment,
■ Generate compensatory earnings: Any regulatory mechanism should provide the regulated company with the opportunity to earn
a reasonable profit and to achieve compensatory earnings. If not, the firm may be forced to reduce investment and quality of service
may decline.
RELATED INFORMATION
Pricing Principles for the ICT Sector

Setting the Level and Structure of Prices

Different cost concepts are useful for answering different questions about a firm and its activities. This section provides an overview of
cost measures that are particularly relevant to price regulation:
■ Historic costs
■ Sunk costs
■ Forward-looking costs
■ Fixed costs (service specific, shared and common costs)
■ Variable costs: marginal costs, incremental cost (including LRIC and TSLRIC)
■ Stand-alone cost, and
■ Short and long run cost concepts.
Figure 1 shows these cost concepts relate to each other.
Figure 1: Cost Concepts in Regulatory Economics

Historic cost is an accounting cost measure. The historic cost (or embedded cost) of an activity is the sum of the costs the firm
actually attributes to providing that activity in a given accounting period. Historic cost reflects what a firm actually pays for capital
equipment, its actual costs of operating and maintaining that equipment, and any other costs incurred to provide service during that
accounting period.
Sunk cost is an economic cost concept, but like accounting cost concepts, measures costs incurred in the past. Sunk costs are historic
costs that are irreversibly spent and independent of the future quantity of service supplied. An example of a sunk cost is the cost of a
marketing campaign for a new service. Once spent, this cost cannot be recovered regardless of whether the service continues to be
provided.
The economic cost of an activity is the actual forward-looking cost of that activity. This is the cost of accomplishing that activity in the
most efficient way possible, given technological, geographical and other real world constraints. Forward-looking costs are the costs of
present and future uses of a firm’s (or society’s) resources. Only forward-looking costs are relevant for making pricing, production, and
investment decisions in the present, or the future.
Costs can be broken into the fixed costs and variable costs of providing a given service.
Fixed costs do not vary as the volume of a service provided changes. For a firm that provides several services, fixed costs can be split
into:
■ Service-specific costs: Costs the firm must incur to provide a specific service. A firm supplying any level of the service would incur
service-specific fixed costs, but would avoid these costs altogether by ceasing production of the service.
■ Shared costs: Costs the firm must incur to provide a group of services. Shared fixed costs do not vary with the level of any
individual service in the group, and do not vary with decisions to produce or cease producing any service or subset of services
within the group. The firm can avoid shared fixed costs if it no longer provides any of the services in the group.
■ Common costs: These are fixed costs are shared by all services produced by the firm. The cost of the president’s desk is a
classic example of a fixed cost that is common to all services.
Variable costs vary with the volume of service provided. Two measures of variable costs are incremental cost and marginal cost.
Incremental cost is the additional cost of producing a given increment of output. How much does the firm’s total costs change if the
volume of a particular service increases (or decreases) by a given amount?
Marginal cost is the incremental cost of producing one additional unit of output. Marginal cost is a limiting case of incremental cost, where
the increment is a single extra unit of service in addition to the amount currently provided.
Incremental cost is usually considered over the long run — long-run incremental cost (LRIC) is the cost of producing a given increment
of output, including an allowance for an appropriate return on capital to reflect the costs of financing investment in facilities used for
interconnection, as well as the capital costs of those facilities.
Total-service long-run incremental cost (TSLRIC) is a special case of incremental cost, where the relevant increment is the total
volume of the service in question, and the time perspective is the long-run. TSLRIC is the additional cost incurred by a firm when adding a
new service to its existing lineup of services (holding the quantities of all those other services constant). For an existing service, TSLRIC
measures the decrease in costs associated with discontinuing supply of the service entirely, other things being constant. TSLRIC is
equivalent to the concept of Total element long-run incremental cost (TELRIC) used in the United States.
Stand-alone cost (SAC) is the cost that a stand-alone firm (producing no other services) would incur to produce a particular service. For
a single-service firm, TSLRIC and SAC are equal. For a multiple service firm, SAC will generally be greater than TSLRIC, because SAC
incorporates shared fixed costs and common fixed costs.
Firms incur costs in the short run, or the long run. Short run costs are the costs of providing a given service, assuming that the current
stock of capital is fixed. Over the long run, firms can vary their stock of capital, for example by investing in new plant. The long run cost
of a service therefore includes the cost of the capital plant required to supply that service.
RELATED INFORMATION
Fixed and Variable Costs and Price Setting

This section introduces some economic concepts that are particularly relevant to the task of price setting:
■ Economic efficiency,
■ Economies of scale and scope, and
■ Single and multiple-service firms.
In economics, the ideal of efficient pricing is often held up as a desirable social goal. Only efficient pricing can ensure that consumers
pay the true economic value of products they buy, and that society’s scarce resources find their best possible uses.
The following are two general principles pertaining to efficient pricing:
■ The economically efficient price of any increment of service is the price that exactly recovers the full economic cost that will be
incurred to provide that increment of service, and
■ In a perfectly competitive market, the price of any increment of service will be driven to the full economic cost of that increment of
service, and will therefore be economically efficient.
Unfortunately, in practice, perfect competition very rarely (if ever) occurs. Telecommunications markets are very different from a
hypothetical perfectly competitive market, as Table 1 illustrates.
This means that, even where there is strong market competition, certain industries cannot follow the simple pricing rules based on the
perfect competition model. When pricing services are provided by network operators, an alternative set of pricing principles apply. These
are described here.
In telecommunications, efficient prices typically consist of:
■ Recovery of the variable costs of the product, plus

■ Mark-ups to recover the product’s fixed costs, and any shared or common costs.
Table 1: Contrast Between Hypothetical Perfectly Competitive Firm and Real World Telecommunications Operators
Perfectly Competitive Firm Real World Telecommunications Operator
Single service Multiple services

Service differentiated by competitor (branding, different pricing plans,
Undifferentiated service provided by all competitors
packaging, customer service plans, and so on)
Large number of competitors. Each competitor has Fewer competitors, subject to different degrees of regulation and market
negligible market share and no control over price forces. Market shares may not be negligible
Economies of scale and scope prevalent. High fixed costs, often high sunk
No economies of scale or scope
costs
Varying degrees or terms of regulation. Franchise obligations common

No regulation, no franchise obligations
(universal service, carrier of last resort, below-cost pricing of local service)
No restrictions on capital. Depreciation determined purely Depreciation rates and cost of capital often below economic levels (subject
by technological and economic conditions (including risk) to regulatory approval) and may not reflect prospective market risks
Undifferentiated and perfectly informed customers Customer base with widely varying demand and usage characteristics
RELATED INFORMATION
Perfect Competition
The production process for telecommunications operators is characterized by economies of scale and scope. This is because
telecommunications operators generally have high fixed costs and high shared and common costs.
Economies of scale occur when a firm’s average cost decreases when it increases its volume of production.
For example, economies of scale occur where a firm has high fixed costs of production. By increasing production, the firm can reduce its
average cost per unit of output. (Provided that variable costs are relatively low, and/or do not increase quickly as production increases.)
Economies of scope occur when some of the fixed resources needed to produce one service can, at no extra cost, be shared to
produce another service. In this situation, it is more economical to produce the two services together and pay only once for the shared
resources, than to produce the services separately.
The practical significance of economies of scope and scale is that telecommunications operators with significant fixed costs can actually
experience lower costs per unit by sharing resources and becoming a provider of multiple services. Operators that start out by providing
only one service may benefit by diversifying and providing multiple services.
Customers also benefit because economies of scope translate into lower prices than under stand-alone production. By sharing resources
the operator only pays once for the resources concerned. As a result the total cost of providing all of the operator’s services is lower.
5.3.3 Single- and Multiple-Service Firms
A single-service firm is a firm that provides only one service to customers. A multiple-service firm is a firm that provides several
services to customers.
In a single-service firm there are no shared or common costs, and no need to attribute costs between services in order to calculate prices.
As Figure 1 shows, the service’s stand-alone cost is equal to the total cost of the firm.
Cost and price calculations are considerably more complex for a multiple-service firm. Some of the firm’s costs will be shared by groups of
services, or common to all services provided by the firm. For a multiple-service firm, total cost is the aggregation of the TSLRICs of the
individual services, the costs shared by various combinations of services and the costs that are common to all services (see Figure 1).
Shared and common costs cannot be directly attributed to individual services, but must still be somehow recovered through prices.
Figure 1: Cost Structure of a Single- Versus a Multiple-Service Firm

RELATED INFORMATION


Most firms in the ICT sector provide multiple services and operate in markets that are very different from a standard model of a perfectly
competitive market. When pricing services provided by multiple-service network operators, an alternative set of pricing principles applies:
Pricing Principle 1: In a competitive market, the efficient price of a service provided by a multiple-service operator need not be
equal to its TSLRIC. Instead, the efficient price must be equal to the full economic cost of the service which exceeds the TSLRIC.
In perfectly competitive markets, the efficient price is equal to the underlying incremental or marginal cost of producing a service. Pricing
Principle 1 recognizes that telecommunications firms are multiple services with substantial shared and common costs. Full economic costs
include incremental costs, and an appropriate contribution towards shared and common costs.
Pricing Principle 2: The TSLRIC of each service provided by a multiple-service operator is the price floor for that service;
incremental revenue from each service must cover the TSLRIC of that service. However, a price that is equal to the full economic
cost of a service will be efficient even if that price is above the TSLRIC.
In perfectly competitive markets, any price above incremental cost is inefficient. Pricing Principle 2 emphasizes that, for network operators,
a price that is above incremental cost is not necessarily inefficient.
Pricing Principle 3: The LRIC shall be the price floor for any additional increment of service provided by a multiple-service
operator; revenue from each increment of service must at least cover its LRIC.
LRIC is the appropriate price floor because the planned increment of service for which a price is set need not be the entire quantity of the
service. However, regulators increasingly use the (average) TSLRIC as the price floor in place of LRIC. The difference between LRIC and
TSLIRC is that TSLRIC includes service-specific fixed costs, while LRIC does not. As a result, TSLRIC usually results in a higher price floor.
In the United States, some regulators have established (average) TSLRIC as the price floor for the service as a whole so that the total
revenues received from the service must at least equal the service’s TSLRIC. Yet for the price floor for an additional unit of output,
regulators have accepted LRIC as the proper price floor.
RELATED INFORMATION
Economic Efficiency and Pricing
Setting the Level and Structure of Prices

This section discusses the task of setting prices for network operators and service providers. Click on the links below for information on:
■ The relationship between fixed and variable costs and efficient prices,
■ Methods for determining mark-ups over TSLRIC, and
■ Tariff rebalancing.
Pricing refers to the task of setting either a single price for an increment of service, or of determining a range within which that price
should fall. This means determining both the minimum acceptable price (the price floor) and the maximum acceptable price (the price ceiling).
The price for an increment of service should be set based on forward-looking costs. That is because the prime consideration in pricing is
the value of the resources that will be used to produce the increment of service, specifically:
■ The mix of technologies needed to produce the increment of service,

■ The prices of input resources, and
■ The future economic depreciation rates and cost of capital that will apply.
In other words, the price for an increment of service must at least cover the incremental cost of that increment. The incremental cost of the
service determines the minimum acceptable price for the service.
In order to determine a range of reasonable (and subsidy free) prices, regulators must also identify the maximum acceptable price. This is
usually the stand alone cost of the relevant increment of service. With free entry into the industry, no supplier could charge a price higher
than the stand alone cost without encouraging other suppliers to enter the market. (Note that, if the firm has no shared or common fixed
costs, the incremental cost will be equal to the stand alone cost.)
In practice, it is very difficult to reliably estimate the stand alone cost of services provided by a multiple-service firm, such as a
telecommunications operator. However, it is possible to determine whether a multiple-service firm is charging more than the maximum
acceptable prices for one or more of its services, using incremental cost information. The rule for ensuring that prices for all of a firm’s
services do not exceed stand alone costs is:
Provided that the firm just breaks even, the price of every service it provides must be no lower than the TSLRIC of that service.
A network operator cannot recover its total costs if it prices all of its services at exactly their respective TSLRICs. In order to recover
legitimate total costs network operators must mark up their prices above TSLRIC. For a discussion of ways to determine the level of such
mark-ups click here.
RELATED INFORMATION
Single- and Multiple-Service Firms

Determining Mark Ups Over TSLRIC
Efficient prices typically consist of:
■ Recovery of the variable (or incremental) costs for the product,

■ A mark-up to recover that product’s fixed costs,
■ An additional mark-up to recover any costs shared with other products,
■ Another mark-up to recover the firm’s common costs.
The mark-ups to recover fixed, shared, and common costs do not need to be uniform amounts or percentages. Mark-ups can vary,
provided that:
■ Total revenues for each product are sufficient to recover all variable and fixed costs for that product,
■ Total revenues for a family of products with shared costs are sufficient to recover all product-specific fixed and variable costs for
each service, plus the shared costs for that family, and
■ Total revenues for the firm are sufficient to recover the firm’s total costs.
Figure 1, below, illustrates the relevance of different types of cost for price setting.
In producing the total volume of a product, for example Product A in Figure 1, a firm may incur costs that are specific to that product. Prices
should generate sufficient revenues for that product to recover all variable and product-specific fixed costs.
If the firm charges a single price for all units of Product A, that price should include a mark-up above the variable (or incremental) costs, to
cover any fixed costs that are specific to Product A.
Some products may share costs (for example Products A, B, and C in Figure 1). In this case, economically efficient prices for these
products need to also include a mark up to cover the shared costs for that family of services.
Finally, prices for all services need to be high enough to recover the common costs of the firm.
Figure 1: Example of Costs of a Multiproduct Firm
Source: Paul Noumba Um, Laurent Gille, Lucile Simon, and Christopher Rudelle, A Model for Calculating Interconnection Costs in
Telecommunications, The World Bank and the Public-Private Infrastructure Advisory Facility, 2003, Figure 3.1 (“World Bank Group Model
Guidebook”).
RELATED INFORMATION
Determining Mark-Ups Over TSLRIC
5.5.2 Determining Mark-Ups over TSLRIC
A multiple-service firm that has economies of scale or scope cannot recover all of its costs if it prices its services at exactly their
respective TSLRICs. In order to recover its legitimate total costs, a multiple-service operators must mark up its prices above TSLRIC. If the
mark-ups are done right, the contributions from each service should enable the operator to fully recover its shared and common costs.
Although operators need to mark up prices above TSLRIC in order to recover their costs, such mark-ups can reduce social welfare below
its theoretical maximum. Mark-ups must be set with care, to minimize this distortion. This is known as second-best optimality. Two
approaches for second-best optimal pricing are value of service pricing and non-linear tariffs.
Value of Service Pricing
Value of service pricing (or “Ramsey pricing”) determines the appropriate level of contribution on a service-by-service basis, reflecting the
demand characteristics of each service.
In its simplest form, value of service pricing sets the mark-up for each service as follows:
The percentage contribution of the service is set in inverse proportion to the own-price elasticity of demand for the service.
Own-price elasticity of demand measures how sensitive the demand for a service is to its price. Own-price elasticity of demand is the
percent change in the quantity of the service demanded, for each percent change in the price. For example, an own-price elasticity of -3
indicates that if price were to rise by 10 percent, demand would fall by 30 percent. The negative sign signifies that demand moves in the
opposite direction to price.
Thus, under a simple value of service pricing rule, if a service has an own-price elasticity of -3, the price for that service should be marked
up by 33%. If a service has an elasticity of -1, the appropriate mark-up is 100%.
As the above example illustrates, under value of service pricing services with high own-price elasticities, for which demand responds
strongly to changes in price, will have small percentage mark-ups. Services with low own-price elasticities will have large percentage
mark-ups. In this way, value of service pricing seeks to minimize the effect on demand (and overall social welfare) from price mark-ups.
The main limitations of value of service pricing are:
■ It may not be possible to reliably estimate price elasticities for the services in question, and
■ Some commentators view this pricing method as unfair. This is because, under value of service pricing, “captive” customers who are
dependent on a service pay the highest mark-up. For example, under this approach basic telephone services in rural areas, where
customers have no other service option, would have a high mark-up. Similarly, this approach would allocate a high proportion of
shared and common costs to “bottleneck” inputs.
Non-Linear Tariffs
Non-linear or multi-part tariff pricing is an extension of value of service pricing, that seeks to allow market forces themselves to determine
the level each service contributes to shared and common costs.
The simplest form of nonlinear pricing is two-part pricing. Under two-part pricing, prices are made up of two components: a price for
access to the service and a price for usage. (For example this might comprise a monthly rental charge and a price per minute for usage.)
Consumers are offered a choice of pricing plans, each with a different combination of access and usage charges. For example, the
operator might offer two plans for a service:
■ A package with an access fee and unlimited usage at zero price,

■ An alternative package with lower access fee, and a per-unit charge for usage.
The idea of two-part pricing is that consumers will reveal the value they place on the service through their choice of plan. If enough
different plans are offered, actual customer choices will provide sufficient demand information for the firm to price optimally.
Two- or multi-part tariffs can be designed to appeal to very diverse consumers in ways that a single, uniform price never can. In the above
example, the high access fee/unlimited usage plan may appeal more to the high-volume consumer, while the low access fee/metered
usage plan may appeal to the low-volume consumer. In contrast, a single price may prove too high for at least some segment of consumers
and exclude them from the market.
Thus, two-part pricing can improve social welfare by broadening the appeal of a service to more consumers and encouraging greater
market participation.

Historically, telecommunications operators and regulators have set prices for network access as low as possible. Prices for other
services, such as long distance calls, have been kept high to subsidize low access prices.
The rationale behind such policies was to encourage customers to join the network, to realize network externalities. In practice, however,
subscribership in many developing economies has been extremely low. It is questionable whether low network access prices have in fact
led to economic gains.[1]
Whatever the benefits from subsidizing access prices, economists agree that rebalancing tariffs can produce significant economic gains.
Table 1 sets out estimates of economic gains from tariff rebalancing from four separate studies.
A policy of rebalancing seeks to increase access prices, and reduce prices for services that have traditionally subsidized low access
prices. The objective is to ensure that the price for each service reflects the underlying cost of providing that service.
Tariff rebalancing can improve social welfare by:

■ Stimulating demand for services such as long distance calling,
■ Providing improved signals to actual and potential service providers to invest in network access technologies, and
■ Improving incentives for competitors to compete for a broad range of customers.
At the same time, increased network access prices under tariff rebalancing generally have a relatively small impact on overall
subscribership levels. This is because demand for network access is not very responsive to changes in price. In addition, low prices for
usage can stimulate demand for access, helping to mitigate the effects of increased prices for access.
Table 1: Estimates of Welfare Gains from Tariff Rebalancing
Study Country / Service Type / Year Estimated Welfare Gains
Crandall and Waverman [2] United States / All / 1994 $6.42 Billion
Munoz [3] Spain / Local & National / 1996 2621.84 Million 1993 Pesetas
Lewis Perl [4] United States / All / 1988 $4,278 Million (1984 Dollars)
Griffin and Mayor [5] United States / Local / 1987 $685- $800 Million
While the economic benefits of rate rebalancing are clear from a theoretical point of view and empirical evidence supports the existence of
these benefits, rate rebalancing has been a very difficult policy to implement. The gains of rate rebalancing are spread across a large
group of individuals—those who make many calls and have moderate to high levels of income—while the costs of the policy usually fall
upon a small group of consumers that can be organized easily to apply pressure on policymakers. There are several tools policymakers
can use to mitigate the pernicious effects of rate rebalancing, however, including:
■ Avoid a flash-cut rate rebalancing program where the full subsidy is removed instantaneously and instead implement rate rebalancing
over several years;
■ Create specially targeted subsidies for those users who may be in jeopardy of dropping off the network in response to rate
rebalancing. Target subsidies can include a monthly discount for certain low-income or low-use customers.
Endnotes:
[1] For example, see Agustin J. Ros and Aniruddha Banerjee, “Telecommunications privatization and tariff rebalancing: evidence from Latin
America,” Telecommunications Policy 24 (2000) 233-252.
[2] See Crandall, Robert and Leonard Waverman, Talk is Cheap: The Promise of Regulatory Reform in North American Telecommunications,
pp. 90-91.
[3] Munoz’ study computed the welfare gains from a 20 percent increase in local rates accompanied by a profit-neutralizing fall of 7 percent
in national rates. See Teresa Garin Munoz, “Demand for National Telephone Traffic in Spain from 1985-1989: An Econometric Study using
Provincial Panel Data”, Information Economics and Policy 8 (1996) 51-73.
[4] See Lewis J. Perl, "Economic Consequences of Competition in Telecommunications," paper presented at the International
Telecommunications Society Seventh Bi-Annual Conference, Cambridge, Massachusetts, July 1988, p. 10.
[5] Griffin, James M. and Thomas H. Mayor, “The Welfare Gains from Efficient Pricing of Telecommunications Services”, Journal of Law and
Economics, October 1987, pp. 465-87.

International benchmarking the process of establishing the price of a service based on prices in other jurisdictions. Benchmarking can
be used as a common sense check on the results of cost models. Alternatively, it can be used directly to set prices.
For example in Singapore, the price SingTel can charge is based on the prices of telephone services in neighbouring Asian countries, New
York, and London.
Benchmarking involves:
■ Selecting a sample of countries or operators. Countries used in the benchmark should be at similar stages of socio-economic and
industry development as the country whose interconnection rates are being considered,
■ Gathering price data for the service(s) under consideration in each of the sample countries, and
■ Adjusting benchmarked rates to account for differences between the country being regulated and the benchmark countries.
Practical Issues in Benchmarking Interconnection Rates
Without appropriate adjustments, benchmarking can result in interconnection rates that make little sense. The goal of the adjustments is to
try to model interconnection costs without having enough detailed information on local cost inputs to carry out a full forward-looking cost
analysis.
Adjustments are often made for:
■ Population density: The number of inhabitants per square kilometre in each country can affect network development costs.
Countries with high population densities tend to have lower network costs than countries with lower population densities,
■ Local area size: This may affect the proportion of short and long distance calls and therefore the costs of interconnection,
■ Extent of urbanization: Network development costs are lower for urban areas than rural areas. Countries with a high degree of
urbanization tend to have lower network costs than countries with less urbanization,
■ Call duration: This may vary widely across countries for several reasons. For example, if customers pay a flat rate for unlimited
local calling, average call duration is likely to be longer than in countries where customers pay a per-minute rate. Networks with higher
call durations need more network capacity, and so will have higher costs,
■ Input prices: The costs of key inputs will vary across countries, and this will affect interconnection costs. For example, the cost of
capital will be significantly higher for most developing countries than for developed countries, due to higher risk in developing markets,
■ Scale economies: If a firm faces significant fixed costs, average cost is likely to decline as output increases. Markets with greater
scale generally have lower average costs. When attempting to extrapolate prices or costs from countries with scale advantages to a
country with a smaller market, it may be necessary to adjust the benchmarked data,
■ Exchange rates: Rates need to be converted to the local currency, or some other single monetary unit. This conversion can use
either market exchange rates or purchasing power parity (PPP) exchange rates. It makes sense to use PPP exchange rates when the
majority of the regulated firm’s costs are local currency denominated and locally sourced, such as staff costs. If the firm’s costs
largely consist of repaying foreign currency denominated loans and purchasing capital equipment on the international market, then
market exchange rates are generally more appropriate as a basis for comparing prices and costs. When using PPP exchange rates, it
is best to use rates estimated by recognized international institutions such as the World Bank, International Monetary Fund or OECD,
■ Taxes: Price data included in the exercise should either all include, or all exclude retail taxes,
■ Rounding effects: If tariffs are being compared based on a unit of time, the rounding effects of billing mechanisms have to be taken
into account. For example, if calls in one country are billed by the minute, and in another country by the second, charges must be
converted to a single unit (either per minute or per second) to allow a meaningful comparison.
RELATED INFORMATION:
Benchmarking Interconnection Rates

This section covers the following topics:
■ An overview of rate of return regulation,

■ Calculating the revenue requirement for regulated services, and
■ Setting prices for regulated services.
For an assessment of the advantages and pitfalls of rate of return regulation, compared to price caps, click here.
Overview of Rate of Return Regulation
Rate of return regulation is a way of regulating the prices charged by a firm. It restricts the amount of profit (return) that the regulated firm
can earn. Rate of return regulation has been used extensively to regulate utilities in many countries. It has been used in the United States
since public utility regulation began in the early 1900s.
There are two steps to implementing rate of return regulation:
■ First, determine the economically appropriate revenue requirement. This is based on prudently incurred expenses and a “fair” return
on invested capital, and
■ Second, set prices for individual services so revenue earned from all the regulated services is not greater than the revenue
requirement.
Calculating the Revenue Requirement
The revenue requirement is generally calculated using the following formula:

Revenue Requirement = Operating Expenses + Depreciation + Taxes + (Net Book Value * Rate of Return)
The rate of return used is the post-tax rate of return the firm is permitted to earn. This is also known as the opportunity cost of investor
capital. It is based on a weighted average of the cost of debt and equity financing.
Operating expenses should include only those expenses the firm has prudently incurred to provide the regulated services.
The net book value of the firm’s capital assets should include only those capital assets used by the firm specifically to provide the
regulated service. The formula includes an allowance for depreciation, so only the book value of the assets net of depreciation should be
included in this amount.
Setting Prices for Regulated Services
The regulator needs to set prices that allow the regulated firm to collect its revenue requirement. This requires that the sum of total
expected revenue for each regulated service is no greater than the permitted revenue requirement. This can be expressed mathematically
as:
Where Pi and Qi are, respectively, price and quantity of service i and N is the total number of regulated services. RR is the revenue
requirement. As the formula shows, in order to calculate prices under rate of return regulation the regulator first needs a reasonable
forecast of demand for the regulated services.
For a multiple-service firm, there is an element of discretion in allocating the revenue requirement amongst different services. As a guiding
principle, the regulator should ensure that prices of individual services are set at prices that minimize distortion of customer behaviour.
The costs used to determine prices under rate of return regulation are the actual embedded costs of the firm, not forward-looking economic
costs.
Under rate of return regulation, the firm can request rate increases if, for whatever reason, it believes revenues are not sufficient to
achieve a normal return on invested capital.
RELATED INFORMATION
Why Regulate Prices?

Incentive Regulation
Rate of Return Regulation versus Price Caps

The term “incentive regulation” refers to types of regulatory mechanism that seek to improve on the weak incentives for efficiency in
traditional rate of return regulation.
Incentive regulation includes:
■ Banded rate of return regulation,

■ Earnings sharing,
■ Revenue sharing,
■ Price freezes,
■ Rate case moratoriums,
■ Pure price caps, and
■ Hybrid price caps.
Banded Rate of Return Regulation
With banded rate of return regulation, the regulator specifies a range of authorized earnings for the regulated firm at the beginning of the
regulatory period. If actual company earnings fall within the range, the company’s prices are considered to be fair and the regulator does
not intervene.
If the firm’s earnings fall outside the permitted band the regulator intervenes:
■ If earnings are higher than the permitted ceiling, the firm must share these gains with its customers,
■ If earnings are lower than the floor, the company is permitted to increase rates.
Prices are thus initially set so that earnings fall within the permitted band, and price adjustments are required only if earnings fall outside the
defined range.
Banded rate of return regulation is not a common form of price regulation. This is because banded rate of return shares most of the
weaknesses of traditional rate of return regulation. It does not eliminate the need for frequent rate hearings and does little to provide
incentives for the regulated firm to reduce costs, unless the regulator defines a very wide band.
Earnings Sharing
Earnings sharing is very similar to banded rate of return regulation, but uses a more precisely defined mechanism for sharing excess
profits with customers. The regulator defines a band (referred to as a “deadband”) within which the firm is free to keep all earnings.
Earnings above or below some deadband are shared in various proportions between the company and the customer.
The deadband under earnings sharing tends to be wider than under banded rate of return regulation. As a result, the firm has greater
incentives to achieve productivity growth and increase efficiency.
Some regulators have used earnings sharing mechanisms when a price cap plan is first introduced, to reduce the risk to customers and
the firm of moving to a new form of regulation.
For example, earnings sharing plans were popular forms of incentive regulation and were a component of some of the initial price cap
plans implemented in the United States. However, earnings sharing does dilute the incentive efficiency properties that exist under a pure
price cap regime. Over the past decade, companies and regulators have moved away from this form of incentive regulation.
Revenue Sharing
Revenue sharing regulation is not common. Revenue sharing requires the regulated firm to share with customers any revenues over a
specified threshold. (This contrasts with earnings sharing regulation in which regulated firms are required to share earnings net of costs.)
Typically the regulated firm retains all of its revenue provided that total revenue does not exceed a specified threshold. The firm must share
some proportion of any revenue generated above that threshold with its customers.
Price Freezes
A price freeze specifies that a company’s prices cannot change within a defined period of time. At the end of the defined period, the
regulator may undertake a rate review. The ability to capture any additional profit during the period if a price freeze give the firm an
incentive to reduce its costs.
Regulators tend to use price freezes in conjunction with other forms of regulation, especially price cap regulation. In telecommunications,
price freezes in a price cap plan usually apply to basic residential service. These services have historically been set at low levels due to
universal service concerns and there is often a desire to maintain that policy under a price cap regime.
Rate Case Moratoriums
A rate case moratorium is an agreement between the regulator and the regulated company to abstain from general rate increases for
particular services. A rate case moratorium usually also suspends investigations of the firm’s earnings, guaranteeing the regulated firm that
profits made at current prices will not be taken away.
A moratorium imposes a regulatory lag. This is intended encourage the regulated firm to reduce operating costs, because the firm will be
able to retain the resulting increase in earnings. The length of a rate case moratorium is typically between two and five years, and is
usually specified in advance.
Pure Price Cap Regulation
Under price cap regulation, the regulator controls the prices charged by the firm, rather than the firm’s earnings. This focus on prices (and
not profits) is what provides for improved efficiency incentives.
The regulator determines an annual price cap formula. This formula determines whether prices should change in each annual period, and
by how much. The regulator usually specifies in advance how long the formula will apply for.
Under a typical price cap, the regulated firm is permitted to alter its average price for a basket of regulated services at the rate of the
general level of inflation minus an efficiency factor based on the regulated firm’s expected efficiency (the “X-factor”). Some regulators also
allow the firm to adjust for changes in costs beyond its control, by including an exogenous cost component in the price cap formula (the “Z-
factor”). A general example of a price cap formula is:
In the above formula, PCIt and PCIt-1 are the price cap index in the current year and the previous year, respectively. CPI is the Consumer
Price Index (or an alternative index of inflation). X and Z are adjustments for expected efficiency gains and for exogenous costs, as
discussed above.
Price caps have a number of advantages over other forms of regulation that focus on the firm’s realized earnings. The fact that the
regulated firm is permitted to retain any realized earnings creates strong incentives to improve efficiency and reduce costs, beyond the
level required by the X-factor. The infrequent reviews of the price cap formula reduce regulatory costs (by avoiding frequent rate cases),
and encourages the firm to implement strategies to reduce costs in future periods, as well as in the current year. Finally, under price cap
regulation, the regulated firm has much more flexibility in the prices that it can charge its customers as long as average prices do not
exceed the cap.
Regulators around the world have used price caps extensively in the telecommunications industry. The regulator in the United Kingdom
introduced price caps in 1984, and they are now increasingly common in the rest of Europe. In the United States, price cap regulation
began replacing traditional rate of return regulation for telecommunications carriers in 1989. By the mid to late 1990s, nearly every state had
a price cap regime in place for the telecommunications industry.
Hybrid Price Cap
Under a hybrid price cap scheme the regulator combines a price cap mechanism with a mechanism that uses realized earnings to
determine prices.
The most common type of hybrid price cap is one where the regulator sets a price cap formula and an explicit earnings sharing
requirement. If the firm’s regulated earnings exceed a certain threshold then it must share part of the gains with customers. Conversely, if
earnings fall below the threshold a share of the losses falls on customers. This provides the firm an incentive to improve its efficiency,
while also addressing concerns about excessive profits (for example, if the regulator sets an X-factor that subsequently appears to be too
generous).
RELATED INFORMATION
Why Regulate Prices?

Rate of Return Regulation
Rate of Return Regulation versus Price Caps
Implementing Price Caps

The following table compares the advantages and pitfalls of rate of return regulation and price caps, against the regulatory criteria
discussed here.
Rate of return (ROR) Price cap
Prevent exercise Yes. The regulated firm can only earn a normal rate of
of market power return. Yes. The CPI-X constraint in the price cap
formula prevents the firm from exercising market
power (if chosen with care).
The firm may exercise market power in prices for

individual services, provided that the average
price of the basket of services is within the cap.
Some regulators impose additional caps on
individual services to prevent this.
Technical No. The regulator directly controls profits. If the firm Yes. Firms are automatically rewarded with
efficiency lowers costs by becoming more efficiency, and so higher earnings when they reduce costs or
increases profits, prices will be lowered in the next expanding demand (and penalized when costs
rate case. The firm will not reap the benefit from increase). This encourages efficient behaviour
reducing costs and so has no incentive to do so.
Allocative
efficiency No. Prices usually based on embedded costs, not Yes. Firms have flexibility to set prices for
forward-looking costs. Prices for individual services individual services based on forward-looking
need not equal the costs of the service. costs.
It is possible for individual prices to deviate from
costs, particularly if the X-factor is set incorrectly.
Dynamic efficiency No. The firm does not retain any increase in profit from Yes. The firm has incentives to invest efficiently,
introducing new technology or services, and so has no because it must justify its investment on the
incentive to do so. profits it expects to earn from the investment (like
firms in competitive markets).
Promote
competition No. Does not generally permit pricing flexibility for the Yes. The firm is less likely to cross-subsidize
firm to set prices to reflect forward-looking costs in services. It is common to group regulated services
response to competition. into separate baskets for less competitive and
more competitive services, preventing cross-
Compared to price cap regulation, the firm is better able subsidization.
to misreport costs between competitive and non-
competitive services, in order to cross-subsidize The firm has sufficient pricing flexibility to respond
competitive services. to competitive pressures by setting prices that
reflect underlying costs and demand conditions
Minimize No. Rate proceedings are often lengthy and resource

regulatory costs intensive. Yes. Price cap proceedings are less costly than
rate proceedings, and are infrequent (once every
3 to 5 years). Between reviews, regulatory costs
are low.
Ensure high Yes. The higher the net book value of the firm’s No. Firms have strong incentives to reduce
service quality assets, the greater the return it is permitted to earn. operating costs, which may lead to reduced
There is a risk that service quality may be higher than service quality
efficient levels.
Prices competitive No. Prices are generally set with no reference to No. Prices are generally set with no reference to
with other prices in other jurisdictions. prices in other jurisdictions.
jurisdictions
Generate Yes. Rate of return regulation ensures that the No guarantee. If the X-factor is chosen
compensatory regulated firm generates sufficient compensatory correctly and the firm performs, the firm should
earnings earnings. generate sufficient compensatory earnings. A
sound price cap penalizes the firm for business
mistakes or poor performance.
RELATED INFORMATION
Rate of Return Regulation

Incentive Regulation
Implementing Price Caps

This section discusses key issues in implementing a price cap regime. The key features of a price cap regime are:
■ The regulator controls the average price of a basket (or baskets) of services. The firm has the flexibility to set prices for individual
services, provided that the average price for the basket does not exceed the cap. On an annual basis, the firm and the regulator
assess price variations to ensure that they conform to the price cap,
■ The price cap includes an efficiency factor, or “X-factor”. Usually, the efficiency factor requires real prices to reduce over time to
reflect expected efficiency improvements for the regulated firm,
■ The price cap formula may also include factors to account for other variables, such as service quality and exogenous changes in
costs.
In implementing price caps, regulators need to weigh the desirability of increased pricing flexibility for the regulated firm against the need to
protect customers and competitors.
Similarly, under a price cap scheme, the regulated firm may be able to implement anti-competitive pricing strategies, for example by cross-
subsidizing more competitive services from less competitive services. This can be addressed by:
■ Setting price floors for certain services below which the firm is not permitted to reduce prices, or
■ Placing more competitive and less competitive services in separate baskets.
A price cap plan typically limits the average price of a basket of services, allowing the firm to change prices for different services by
different percentage amounts. The first step in designing a price cap is to determine the number and scope of service baskets.
This decision requires the regulator to weigh the desirability of providing greater pricing flexibility to the regulated firm against the need to
protect customers, particularly “captive” customer groups, from high prices.
Pricing flexibility is an important advantage of price cap schemes. In general, the broader the basket, the greater the firm’s flexibility to
transition initial prices toward efficient levels (based on underlying costs), and to adapt prices to competitive changes in the market.
Where existing prices bear little relationship to cost, it makes sense to group regulated services in a single price-capped basket. This
ensures that overall customers continue to receive the benefits of technical change and cost reductions, while permitting the firm to move
towards a competitive rate structure.
Under some conditions a regulated firm could profitably reduce prices for more price-elastic services and raise them for less price-elastic
services, without breaching the price cap constraint.
For example, if residential and business customers were placed in the same basket, the regulated firm could lower business prices while
raising residential prices. This behaviour is generally consistent with competitive outcomes, but may create customer protection concerns.
Where regulators wish to limit this type of pricing flexibility, they can do so by placing the services in separate baskets.
Separate baskets can be used to give different degrees of pricing flexibility to different service categories, particularly where initial relative
prices do not reflect underlying cost differences.
For example, the initial price cap plan in the United Kingdom used separate price caps to address the problem of low residential basic
subscription rates. Residential access rates were placed in a separate basket, with an initial price cap index of inflation plus 2 percent. The
price cap index for the regulated business as a whole was initially inflation less 3 percent. This combination of caps allowed residential
access rates to increase towards costs, while ensuring that aggregate regulated service prices fell in real terms.
In the extreme, the regulated firm could seek to reduce prices for more competitive services below costs, increase prices for other
services to compensate. Again, using a larger number of smaller baskets will reduce the likelihood of anti-competitive cross-subsidization.
Alternatively, regulators can protect competitors simply by putting some type of price floor in place.
RELATED INFORMATION
Assessing Price Variations

Calculating the Productivity Factor
Service Quality Factors
Exogenous Cost Factors
Price cap regulation constrains the price changes a regulated firm is allowed to impose on its customers. In each period, the regulated firm
is responsible for calculating:
■ The Price Cap Index (PCI). This index is the overall constraint of the firm’s prices in a given period, and is calculated based on the price
cap formula,
■ The Actual Price Index (API). This index shows the actual level of prices, and should not exceed the PCI, and
■ A Service Basket Index (SBI) for each basket, where services are grouped into different baskets.
Regulators need to check the calculation of the API and the PCI to ensure that the company’s actual price changes do not exceed the
constraints of the price cap.
Calculating the PCI
The PCI is the overall constraint in a price cap plan. The PCI changes annually according to the price cap formula.
In the base period for the price cap (T-1), the PCI is initially set equal to 100.0. The PCI is allowed to change in each subsequent period
according to the values of inflation, X and Z. To illustrate, Table 1 gives an example of PCI calculations for a hypothetical price cap.
Table 1: Hypothetical PCI Calculation, X-factor of 3.5%
Period Inflation X-factor Z PCI
T-1 100.0
T 4.0% 3.5% 0.0% 100.5
T%201 3.5% 3.5% 2.0% 102.5
T%202 3.7% 3.5% -2.0% 100.7
T%203 4.2% 3.5% 3.2% 104.6
T%204 3.8% 3.5% -0.5% 104.4
Calculating the API
The API tracks the firm’s actual price changes, to ensure actual prices do not exceed the PCI. The API is normally assigned the value of
100.0 at the beginning of the plan, and changes when the regulated firm changes its prices.
Each time the regulated firm files new tariff revisions, it must calculate the API for each basket, and report this to the regulator to show that
the price changes do not exceed the PCI.
Most price cap plans use the following formula to calculate the API:
In the above formula:

■ APIt is the actual price index value following proposed price changes,
■ APIt-1 is the existing actual price index value without the price changes,
■ i is the number of rate elements in the price cap,
■ Pt equals the proposed prices for each of the i rate elements,
■ Pt-1 equals the existing prices for the i rate elements, and
■ vi is current estimated revenue weights for the i rate elements. This is calculated as the ratio of base period demand for the i rate
element prices at the existing rate, to base period demand for the entire basket of services priced at existing rates.
Table 2 provides a hypothetical example of an API calculation. The example assumes a basket with two services, and calculated the API
for two periods. The example demonstrates how the API can change even when one service price increases and the other decreases by
the same percentage amount. This occurs because the two services account for different shares of the total basket revenue.
Table 2: Hypothetical API Calculation

RELATED INFORMATION
Price Cap Baskets

The X-factor in the price cap formula is an efficiency target chosen to reflect the productivity growth potential of the regulated firm over the
(forward-looking) term of the price cap.
The inclusion of the X-factor ensures that prices change over time to reflect productivity gains. If the regulated firm’s actual cost reductions
exceed the X-factor, the firm is rewarded. On the other hand, if the firm’s actual cost reductions are not as great as required by the X-
factor, then the firm is penalized.
Care is required in setting the X-factor. The firm’s financial success will depend on its ability to reduce unit costs while there is a price cap.
This therefore be taken into account in setting the X-factor.
There are generally two ways to calculate the X-factor:
■ Using a forward looking financial model of the firm, or

■ Comparing the rate of increase in the firm's (or industry's) total factor productivity to the economy-wide rate.
In addition, some regulators include an additional component in the X-factor, to provide a consumer productivity dividend, or “stretch factor”.
Forward-Looking Financial Model
Under this approach, the regulator develops forecasts of the regulated firm’s costs, revenues and profits over a given period of time. The
regulator selects an X-factor that will ensure that (in the absence of better than expected efficiency improvements) the firm’s internal rate
of return is no greater than a normal return on profit, usually measured by the weighted average cost of capital.
If the firm’s rate of return is higher than its cost of capital (or significantly below the cost of capital) at the beginning of the price control
period, the regulator can:
■ Impose a one off price adjustment at the beginning of the period, to bring the rate of return in line with the cost of capital, and set X so
that the internal rate of return is no greater than the cost of capital (as above), or
■ Set X so as to bring the internal rate of return into line with the cost of capital by the end of the price control period.
The choice between these two options depends on:
■ Whether the regulator believes that “excess” profitability at the beginning of the price control period reflects monopoly power, or is
due to past innovation and cost reductions, and
■ Whether the regulator prefers to avoid large price adjustments at a given point in time.
Total Factor Productivity Differential Approach
Under this approach, the regulator calculates the firm’s total factor productivity (TFP), and compares it to the economy’s TFP. Ideally, the
regulator should use the TFP of the telecommunications industry, rather than the individual firm. This will maximize the good incentive
properties in the price cap, by encouraging the firm to outperform other firms in the sector.
The regulator must also examine the firm’s (or, ideally, the industry’s) input prices, and compare them to input prices in the general
economy.
The general formula for the X-factor under this approach is:
In words, X is calculated as:
■ The difference between the rate of economy-wide input price growth, and input price growth in the telephone industry (the input price
differential), plus
■ The difference between the rate of total factor productivity growth in the telephone industry, and economy-wide total factor
productivity growth (the TFP differential).
Consumer Productivity Dividend / Stretch Factor
Some regulators include a consumer productivity dividend or “stretch factor”, in price caps. A stretch factor adds an amount to the
productivity target to provide additional benefits to customers.
This additional factor recognizes that in many cases the efficiency gains from a price cap come about because the previous system
allowed inefficiencies to accrue. That is, compared to, say, rate of return regulation price caps should generate additional efficiencies.
A stretch factor is, in effect, a type of guarantee. It ensures that ratepayers will in fact be better off under a price cap regime, because it
directly passes any efficiency gains from price cap regulation to customers.
In general, a stretch factor is only warranted where the introduction of a price cap regime significantly improves efficiency incentives.
RELATED INFORMATION
Price Cap Baskets

A theoretical weakness of price cap regulation is that it could lead to a deterioration of service quality. Some price cap plans therefore
include a service quality component to ensure that service quality will not suffer under price cap regulation.
Regulators can set a service quality component by:
■ Selecting a key service quality measures, and

■ Monitoring the regulated firm’s performance against these measures.
If performance falls below the required standards the firm is penalized, for example through an increase in the X-factor. Conversely, the
firm should be rewarded for exceeding service standards.
RELATED INFORMATION
Price Cap Baskets

Many regulators include an exogenous cost component, or Z-factor, in the price cap formula. This allows the regulated firm to adjust for
changes in costs that are beyond its control.
By including a Z-factor in the price cap, regulators can allow changes in certain types of costs to flow directly through to the price cap
index, without affecting the regulated firm’s incentives to control its costs. Thus changes in the firm’s prices (or at least in the price cap
index) can more closely track changes in costs.
When establishing a Z-factor it is crucial that regulators treat only those events over which the firm has no control as exogenous. Some
exogenous factors are easy to identify, for example changes in taxation or in regulatory rules. However, other exogenous factors
changes are difficult to isolate.
As an example, the Canadian Radio-television and Telecommunications Commission adopted an exogenous component in its 2001 price cap
plan. Adjustments were considered for events or initiatives that satisfied the following criteria:
■ They are legislative, judicial or administrative actions which are beyond the control of the company,
■ They are addressed specifically to the telecommunications industry, and
■ They have a material impact on the Utility Segment of the company.
CONCLUSION
There are significant technological and competitive changes occurring in the telecommunications industry such as the advancement and
rapid growth of wireless services, the rapid development of packet-based communications, (VoIP), the emergence of cable companies as
strong competitors to traditional phone companies and the potential for convergence to obliterate remaining distinctions between fixed and
wireless communications.
As a result, regulators have begun the process of evaluating whether current price regulations still remain necessary for traditional
telecommunications carriers. While certain form of regulatory intervention will likely remain for the long run—such as numbering resources,
interconnection oversight, spectrum allocation, etc—some regulators have already decided that traditional telecommunications firms no
longer possess market power and accordingly these regulators have begun the process of deregulating telecommunications carriers.
RELATED INFORMATION
Price Cap Baskets


Wholesale and retail prices are interlinked. Regulators need to be aware this link in regulating prices.
One option to address the link between wholesale and retail prices is to implement a double price cap (also known as a “global price cap”.
Under a global price cap plan, the regulator includes wholesale interconnection services in the price cap plan, and treats them as any other
final good in implementing the price cap.
For example, in the United States, some state regulators control intrastate long distance interconnection prices through the same price cap
mechanism that regulates retail prices. These price cap plans are partial examples of a global price cap; they are not full global price caps
because local interconnection prices are regulated using cost-based price controls.
According to Laffonte and Tirole [1], a global price cap plan can incorporate a Ramsey pricing structure and has the following features:
■ The wholesale service (access) is treated as a retail service and is included in the computation of the price cap, and
■ Weights used in computing the price cap are determined exogenously and are proportional to forecast quantities of the associated
services.
As Laffonte and Tirole state, “global price caps thus enable regulation to be more light-handed, for global price caps reduce perverse
incentives and therefore diminish the need for regulatory oversight of the operator’s decisions.”
Endnotes:
[1] Jean-Jacques Laffont and Jean Tirole, Competition in Telecommunications, MIT Press 2000, page 174.
5.13 Price Regulation and Multiple Play Offerings

Significant questions about the applicability of price regulation arise with the advent of intermodal competition, that is, competition between,
say, traditional telephone networks and cable television providers whereby each provider offers voice telephony, broadband data, and
video content services. This particular offering is popularly described as a Triple Play. The addition of mobile services offers the possibility
of a fourth play in the mixture. Each provider uses a different network infrastructure. Generally, cable television providers seem to be
having an easier time upgrading their networks so as to offer broadband data and voice telephony than traditional telephone providers are
having in adding video content services to their offerings. Still, multiple play offerings are becoming more and more widespread throughout
the world.
The rationales described in section 5.1 whereby price regulation has been applied to a telecommunications provider are weakened
considerably when that provider faces competition from a competing infrastructure. Regulation is, after all, intended as a substitute for
competition and where competition itself exists or is emerging, the justification for continuing to regulate retail prices becomes less
relevant. Some regulators are responding to the emergence of intermodal competition by exempting multiple play offerings from price cap
regulation. The traditional voice telephony service is still available to be purchased at a regulated price, but the regulator has opted to allow
the market to determine the price for the bundled triple play offering. However, this is an emerging area and the policy questions are
evolving rapidly.
RELATED INFORMATION
Thomas M. Lenard and Randolph J. May (eds.). Net Neutrality or Net Neutering: Should Broadband Internet Services Be Regulated? (New
York: Springer: 2006).
6. Infrastructure Sharing
Note: This section is based in large part on two discussion papers prepared for the 8th ITU Global Symposium for Regulators[1] 2008. The
first paper, entitled Extending Open Access to National Fibre Backbones in Developing Countries was prepared by Dr Tracy Cohen and
Russell Southwood and the second one, entitled Mobile Network Sharing was prepared by Camila Borba Lefèvre.
This section explores policy and regulatory dimensions of sharing ICT infrastructure and capacity, with a particular emphasis on core
infrastructure sharing and mobile network sharing. Infrastructure sharing is rapidly becoming an important means of promoting universal
access to ICT networks and offering affordable broadband services by reducing construction costs. In light of under-developed markets
and the high costs associated with network deployment, carefully crafted sharing policy measures can introduce new forms of competition
into the market and stimulate demand for ICT services.
This section provides information related to key concepts and terms associated with sharing and the policy objectives frequently linked to
sharing approaches. It provides a broad overview of various forms of sharing and a detailed examination of the sharing of national fibre
core networks and the sharing of mobile networks.
[1] For more information on the GSR, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/index.html.

For a direct link to the GSR08 discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.1 Key Concepts

Note: This section is based in large part on two discussion papers prepared for the 8th ITU Global Symposium for Regulators [1] 2008. The
There are a number of concepts that are central to understanding the policy and regulatory framework governing sharing. These concepts
include: passive and active infrastructure; essential (or bottleneck) facilities; and open access. This section provides a brief overview of
these concepts, with links to further resources.

For a direct link to GSR 08 discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.1.1 Passive and Active Infrastructure
There are several different elements of ICT network infrastructure that can be shared. However, not all elements of the network
infrastructure can or should be approached in the same manner. In order to develop frameworks for regulating the sharing of network
infrastructure, it is helpful to conceptualize infrastructure as falling into two categories: passive and active infrastructure.
The easiest shorthand definitions of passive and active infrastructure are as follows:
■ Passive infrastructure includes all the civil engineering and non-electronic elements of infrastructure, such as physical sites, poles
and ducts (and also power supplies).
■ Active infrastructure covers all the electronic telecommunication elements of infrastructure like lit fibre, access node switches, and
broadband remote access servers.
Passive and Active infrastructure Sharing – Examples
Fibre Core Networks Mobile Networks
Passive Sharing Poles, ducts, power supplies Electrical cables, fibre optic cables, masts and
pylons, physical space on the ground, towers,
rooftops, or other premises, shelter and support
cabinets, electrical power supply, air
conditioning, alarm systems, and other
equipment.
Active Sharing Lit fibre, access node switches, broadband The Node-B (the base station next to an
remote access servers antenna), Radio Network Controller
Sources: Dr. Tracy Cohen and Russell Southwood, Extending Open Access to National Fibre Backbones in Developing Countries and
Camila Borba Lefèvre, Mobile Network Sharing.
For more information on the GSR, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/index.html.
For a direct link to the GSR 08 discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.1.2 Essential or Bottleneck Facilities
Essential facilities, or bottleneck facilities, are network elements or services that are provided exclusively or predominantly by a monopolist
or a small number of suppliers and that cannot easily be replicated or substituted by competitors for economic or technical reasons. These
types of facilities are critical inputs to retail service.
6.1.3 Open Access
InfoDev has provided the following definition of "open access":
"Open Access means the creation of competition in all layers of the network, allowing a wide variety of physical networks and
applications to interact in an open architecture. Simply put, anyone can connect to anyone in a technology-neutral framework that
encourages innovative, low-cost delivery to users. It encourages market entry from smaller, local companies and seeks to prevent
any single entity from becoming dominant. Open access requires transparency to ensure fair trading within and between the
layers, based on clear, comparative information on market prices and services."
Source: infoDev, Open Access Models: Options for Improving Backbone Access in Developing Countries (with a Focus on Sub-Saharan
Africa)
6.2 Policy Issues

There are a number of policy issues associated with sharing. Some of the policy concerns relate to why sharing has become an important
regulatory matter. These policy issues include:
■ promoting rapid and efficient network deployment;

■ the efficient rollout of Next-Generation Networks; and
■ minimizing the environmental impact of ICT infrastructure and harmonizing network rollout with local land use planning.
Other policy issues relate to concerns about how sharing is implemented in the ICT sector. These policy issues include:
■ preventing anti-competitive conduct,

■ reducing wholesale interconnection and charges (which should in turn lead to lower retail usage charges) and
■ ensuring that sharing does not inhibit innovation in the ICT sector.
For a direct link to the GSR 2008 discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.2.1 Promoting rapid and efficient network deployment

One of the most important policy concerns underlying the growing regulatory interest in sharing is the promotion of rapid and efficient
network deployment. In many developing countries, the network in question is the mobile network, which is increasingly becoming the
dominant form of infrastructure in these countries, as well as the backbone for the provision of universal access. In more developed and
industrialized countries, the emphasis is on national broadband core and access networks and Next-Generation-Networks (NGNs).
Although the modes of sharing differ and although each network raises particular policy concerns, broadly speaking, sharing facilitates a
rapid, less costly and less disruptive deployment of networks, whether the network is mobile, fixed broadband, or NGN.
Sharing helps to address three obstacles to efficient and timely network deployment: the high costs of network roll-out; restricted access
to bottleneck facilities; and poor investment incentives, particularly in un-served or under-served areas.
Reducing the costs of network roll-out
Sharing can reduce the cost of network deployment. For example, in the case of mobile networks, civil engineering costs can mount up
when the number of building sites is relatively high in the network roll-out. Site sharing allows operators to reduce their capital and
operating expenditures. Lower site-development costs can pay dividends when they result in networks covering larger areas, increasing
the likelihood of bringing wireless services to sparsely populated rural areas – and at more affordable prices.
Similarly, one of the most significant costs associated with the deployment of broadband fibre networks relates to the excavation of
conduits and the installation of fibre for the access part of the network. This entails actual construction and installation costs as well as the
cost of securing numerous permits such as digging permits and environmental permits. The shared use of ducts and poles, as well as
other infrastructure, reduces an operator’s physical deployment costs. Sharing is thus one dimension of creating an enabling environment
for national core and access broadband networks. For more information about the costs of rolling out broadband networks, see the
Practice Note entitled “The Cost of Broadband Fibre Network Roll-out”. A link to this Practice Note is set out below.
Facilitating access to bottleneck facilities
The control of bottleneck facilities by a single dominant infrastructure operator tends to impede the development of new infrastructure, the
expansion of competition, and market growth in general. The operator that controls these facilities (usually the incumbent) questions the
commercial rationale for providing access to its infrastructure to its competitors. Mandated sharing of bottleneck facilities is a key strategy
for opening up access to these facilities and thus for cultivating competition in downstream markets. Without mandated sharing, it is unlikely
that incumbents would willingly offer access to their bottleneck facilities on commercially fair terms. For more information on this topic,
please see the Practice Note entitled “Sharing and Access to Bottleneck Facilities”. A link to this Practice Note is set out below.
Low market investment
The high cost of deploying network infrastructure and low population density sometimes combine to impede investment in rolling out
network services in rural and remote areas. In sparsely populated areas, the returns on investment in high capacity network infrastructure
are often too low to sustain commercial operations. Sharing can assist regulators and policy-makers address this problem in a number of
ways.
First, as discussed above, sharing can reduce the cost of network deployment. Second, sharing can make a wider network roll-out more
affordable which, in turn, creates a greater “critical mass” of users. In combination, the lower costs of network roll-out and the larger
critical mass of users increase the return on expenditures, thereby generating incentives for investment. This is particularly critical at a time
when the costs of financing investment (interest rates) are rising. In the case of un-served or under-served areas, policy-makers usually
aim to create a greater “critical mass” of users by encouraging the roll-out of high-capacity, national infrastructure to a wider range of
places than the market alone might initially sustain. Allowing two or more operators to share (and therefore to pay for access to) a common
national infrastructure helps to finance a wider deployment, whereas traffic from a single operator would not sustain a widespread
network.
Sharing, network deployment, and universal access
By facilitating quick and efficient network roll-out, sharing advances universal access policy objectives. In developing economies, sharing
promotes network roll-out to un-served and under-served areas. In developed economies, sharing plays an important role in rolling out
FTTx access and expanding broadband access to under-served areas, such as rural communities. The relationship between the promotion
of universal access and sharing is discussed further in section 6.2.2 and in section 3.4 of Module 4, Universal Access.
6.2.2. The efficient deployment of Next-Generation Networks
Sharing is increasingly playing a central role in the development and deployment of Next-Generation Networks (NGNs). The transition to an
NGN environment requires significant investments as access providers and network operators must upgrade their equipment and build
new network infrastructure. At the same time, convergence and the move to an IP-based network allow a variety of different types of
services and applications to be provided over the same core and access infrastructure. Consequently, approaches to developing NGNs
frequently feature the deployment of a single core network, with competition occurring in other layers of the network, such as the access,
service, and application layers. Thus, these approaches typically are premised upon the sharing of the core NGN network infrastructure
and often feature sharing at other levels as well.
Box 6.2.3: Deploying Open-Access NGNs in Singapore
Singapore is currently seeking to roll-out wired and wireless NGNs by creating national core networks that are operated
by a single company but that are also open to access by operators and service providers active in other layers of the
network. Singapore’s strategic plan (the Next Generation National Infocommunications Infrastructure or “Next Gen NII”)
involves the creation of a wired, open access, and carrier-neutral Next Generation National Broadband Network (Next
Gen NBN) and an open-access Wireless Broadband Network (WBN). The Next Gen NBN and the WBN are to be built,
owned, and operated by the private sector. The government has made clear that the operation of the Next Gen NBN and
WBN will involve structural separation of the operator of the passive network infrastructure, the operator of the active
network infrastructure, and the retail services provider. The government of Singapore has indicated that it will provide
various amounts of funding to the operators of the passive and active infrastructure of the Next Gen NBN and WBN. The
funding is intended to kick-start the project and to ensure that the ultra high-speed broadband service provided over these
networks will be viable, affordable and sustainable in the long-term.
6.2.3. Environmental Considerations and Local Land Use Planning
Sharing policies respond to environmental and public health concerns, particularly with respect to mobile network transmission facilities,
and to issues related to local land use planning. People generally view wireless communication masts and antennas as negative additions
to the landscape. Local communities may object to the construction of new sites because of the visual impact or environmental
considerations. Also, residents may fear public exposure to electromagnetic fields around masts and antennas. [1] Site sharing can limit
such concerns and potential negative effects, since it limits the number of sites while achieving the required coverage. Another beneficial
aspect of site sharing is the amount of energy that can be saved when operators share electrical power, which is often in limited supply in
developing countries. [2]
Box 1: The EU Framework Directive and Facility Sharing
The EU Framework Directive contains a number of provisions related to facility sharing. These provisions are relevant to
policy objectives related to environmental considerations and local land use planning.
Recital 23:
“Facility sharing can be of benefit for town planning, public health or environmental reasons, and should be encouraged
by national regulatory authorities on the basis of voluntary agreements. In cases where undertakings are deprived of
access to viable alternatives, compulsory facility or property sharing may be appropriate. It covers inter alia: physical
collocation and duct, building, mast, antenna or antenna system sharing. Compulsory facility or property sharing should be
imposed on undertakings only after full public consultation.”
Recital 24:
“Where mobile operators are required to share towers or masts for environmental reasons, such mandated sharing may
lead to a reduction in the maximum transmitted power levels allowed for each operator for reasons of public health, and
this in turn may require operators to install more transmission sites to ensure national coverage.”
Article 12: Co-location and facility sharing
“1. Where an undertaking providing electronic communications networks has the right under national legislation to install
facilities on, over or under public or private property, or may take advantage of a procedure for the expropriation or use of
property, national regulatory authorities shall encourage the sharing of such facilities or property.
“2. In particular where undertakings are deprived of access to viable alternatives because of the need to protect the
environment, public health, public security or to meet town and country planning objectives, Member States may impose
the sharing of facilities or property (including physical co-location) on an undertaking operating an electronic
communications network or take measures to facilitate the coordination of public works only after an appropriate period of
public consultation during which all interested parties must be given an opportunity to express their views. Such sharing
or coordination arrangements may include rules for apportioning the costs of facility or property sharing.”
Source: Directive 2002/21/EC of the European Parliament and of the Council of 7 March 2002 on a common regulatory
framework for electronic communications networks and services (Framework Directive).
While sharing reduces the number of sites marking the landscape, it can also have adverse impacts. Because antennas generally have to
be separated from each other by a minimum distance in order to avoid interference, mast sharing usually requires taller (and more visually
disruptive) masts. (See Recital 24 of the European Framework Directive, outlined in Box 1, above.) Local planning authorities actually may
prefer several small towers to one large one. More discrete (or disguised) structures reduce visual intrusion, but cannot support more than
one operator’s antenna.
Sharing may also help operators address what can often be a very expensive and nettlesome aspect of network deployment: obtaining the
rights of way to lay fibre strands and obtaining local permission to breakup public roads and other places necessary to build trenches,
conduits, and ducts. With good planning and foresight, trenches, conduits, and ducts can be built large enough to accommodate more than
one operator. Sharing access to trenches, conduits, and ducts reduces the fees and “red tape” surrounding the obtaining of permission to
lay fibre and to undertake the related construction. From the perspective of local authorities, sharing access to trenches, conduits, and
ducts leads to less disruption on public roadways and other places since each operator does not have to undertake its own construction.
In addition to helping to address environmental concerns and local land use issues, sharing also offers a way of balancing the need for
communications infrastructure such as antennae and the preservation of historic sites. Maintaining the integrity of historic sites generally
requires minimizing external structures such as communications infrastructure. Strategies such as tower sharing and collocation help to
manage the coexistence of modern communications infrastructure and historic sites. In the United States, for example, the Federal
Communications Commission, the National Conference of State Historic Preservation Officers, and the Advisory Council on Historic
Preservation developed an agreement aimed at preventing the construction of unnecessary new communications towers by mandating
collocation. This agreement, the “Nationwide Programmatic Agreement for the Collocation of Wireless Antennas”, is linked below as a
reference document.
[1] Current scientific evidence indicates that exposure to radiofrequency fields, such as those emitted by mobile phones and antennas, is
unlikely to have negative health effects. In response to health concerns raised by certain communities, the World Health Organization
(WHO) established a project to assess the scientific evidence of possible health effects of electromagnetic fields (see www.who.int/peh-
emf/en/index.html). The International Commission for Non-ionizing Radiation Protection (www.icnirp.de) has established guidelines for the
maximum level of radiofrequency levels in areas of public access from antennas and for users of mobile handsets.
[2] Today’s standard 3G equipment consumes about 4,000 KWh of Grey energy per year per node, which corresponds to 2.5 tons of CO2,
or the equivalent need of 120 trees per node to compensate for the environmental effect. In a developing country with no or little alternative
Green energy, network sharing can significantly reduce the environmental impact.
6.2.4. Competition
The policy issues related to competition and sharing are complex. Sharing offers both the possibility of enhancing competition and the risk
of hindering competition.
On the one hand, sharing policies can help to increase competition in the ICT sector. One of the greatest impediments to market entry in the
sector is the cost of network deployment. Sharing allows operators to enter the market at a much lower cost than what they would
encounter if they were required to construct their own network infrastructure. Sharing also helps to overcome barriers to competition such
as the control of bottleneck facilities by dominant operators. Sharing is just as important as the principle of non-discrimination in lowering
the barriers to entry and to competition for new market entrants.
On the other hand, too much sharing undermines the incentives for investment in infrastructure-based competition. As the OECD has noted,
“There is widespread agreement that infrastructure-based competition provides the most sustainable and effective level of competition in
the communications market.” [1] It is important that there are adequate incentives within the regulatory framework for operators to invest in
infrastructure. Conditions must be created to spur continued roll-out of network infrastructure that is not easily replicable.
Sharing, however, can undermine incentives to invest in infrastructure based competition. A very permissive sharing regime makes it
possible for operators to become active without investing in their own infrastructure. If most operators rely on the same underlying
infrastructure providers, it is likely that there will be little ultimate differentiation in their services. The benefits of competition like lower
prices and consumer choice are reduced as a result.
While it is arguable that sharing creates incentives for some operators to become infrastructure providers, if terms and conditions such as
pricing for access are regulated, sharing policies like open access may entail some loss of return for the owner of the infrastructure. If too
much access is given to new entrants, the result can be just as bad as exclusivity: there will be little or no investment in infrastructure. An
example of this type of result is the market in the United States between 1996 and 2002. The reason that there was so little investment in
local loop infrastructure by new entrants was that they could not deploy new infrastructure at the regulated local service prices, which
were too low and acted as a disincentive to investment.
Ultimately, there is an inevitable tension between the equally important goals of reducing barriers to market entry and stimulating investment
in infrastructure. Both of these goals are relevant to maintaining healthy competition in the ICT sector. Striking the appropriate balance
between these goals is a delicate matter for policy makers and regulators.
For more information about competition-related aspects of regulating access, please see section 6.6.5 of this module.
[1] Organization for Economic Cooperation and Development (OECD), Convergence and Next Generation Networks, Ministerial Background
Report DSTI/ICCP/CISP(2007)2/FINAL (Paris: OECD, 2008). This paper was prepared for OECD Ministerial Meeting on the Future of the
Internet Economy, Seoul, the Republic of Korea, 17-18 July, 2008.
6.2.5. Innovation
One of the hallmarks of the ICT sector is innovation, both in service offerings and in technological advances. Sharing has the potential to
support ongoing innovation in the sector, but also risk impeding innovation if the regulation of sharing does not create the proper incentives.
On the one hand, sharing allows operators to spread the risk of investing in upgrades to equipment and new technologies, making it more
likely that operators will invest in innovative, though potentially risky, new technologies and service offerings. [1] Moreover, by facilitating
network deployment, sharing allows operators to focus on developing new service offerings and better customer service. [2] Sharing also
stimulates innovation in the types of commercial activities that occur in the ICT sector. For example, infrastructure sharing has led to the
emergence of at least two new types of industry players: infrastructure providers such as tower companies and resale-based service
providers such as Mobile Virtual Network Operators (MVNOs).
On the other hand, sharing can undermine the incentives that operators have to innovate. For example, if operators are readily permitted to
share infrastructure at very low rates, operators have little incentive to develop their own infrastructure. This, in turn, reduces prospects
for innovation since operators will not actively seek more efficient forms of infrastructure. Difficulties in accessing traditional bottleneck
facilities, for example, lead to the development of new technologies to get around these bottlenecks. If operators have too much cheap
access to bottleneck facilities, the incentives to develop new solutions to bottlenecks will erode.
Certain kinds of sharing also risk obstructing innovation in service offerings. For example, national roaming arrangements tend to lead to a
streamlining of terms and rates offered to end-users since operators rely on each other to provide services to customers when their
customers are outside of their geographic coverage area. Re-sale based sharing arrangements such as MVNOs also risk the
standardization of the terms and conditions under which customers receive services. In this regard, sharing is seen to reduce service
innovation due to the standardization of terms and conditions that often flow from sharing arrangements. Nevertheless, there appears to be
evidence that healthy competition and innovation on the basis of brand and other factors continues to exist.
At this stage, it is still too early to tell how sharing will impact innovation in the ICT sector. However, it seems clear that one factor that
regulators and policy makers should consider when they evaluate sharing arrangements is the potential impact that such arrangements
may have on innovation in the ICT sector.
[1] See Booz Allen Hamilton Inc," Telecom Infrastructure Sharing – Regulatory Enablers and Economic Benefits", November 2007, p. 3
available online at: www.boozallen.com/media/file/Telecom_Infrastructure_Sharing.pdf
[2] Ibid.
6.3 Forms of sharing

first paper, entitled Extending Open
Access to National Fibre Backbones in Developing Countries was prepared by Dr Tracy Cohen and Russell Southwood and the second
one, entitled Mobile Network Sharing was prepared by Camila Borba Lefèvre.
There are several different types of sharing. The most common forms of sharing include:
■ infrastructure sharing and collocation;

■ spectrum sharing
■ interconnection; and
■ unbundling.
More information on these forms of sharing can be found by following the links below.

For a direct link to the GSR 08 discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html
6.3.1. Infrastructure sharing and collocation
Infrastructure sharing refers to the sharing of passive and active infrastructure. It includes the sharing of support structures such as
towers, masts, ducts, conduits, trenches, manholes, and street pedestals, as well as the sharing of electronic power supplies, air
conditioning, and alarm systems. Infrastructure sharing also encompasses the sharing of the electronic telecommunications elements of
infrastructure such as lit fibre, access node switches, and Radio Network Controllers.
One interesting development related to infrastructure sharing is the emergence of service providers that construct, manage, and sell
access to passive infrastructure or who provide services related to active infrastructure on a wholesale basis only. In Pakistan, for
example, Infrastructure Licences and Tower Licences authorize licensees to establish and to maintain Telecom Infrastructure Facilities to
lease, rent out, or sell to Telecom Operators; holders of Infrastructure Licences may also provide end to end links to Telecom Operators.
However, neither Infrastructure Licensees nor Tower Licensees may provide telecommunications or broadcasting services. Thus neither
category of licensee may provide a service consisting of the emission, conveyance, switching, or reception of voice or data traffic.
Collocation refers to the sharing of physical space in buildings used to house radio and cable transmission facilities and related equipment.
The facilities of different operators that collocate within a premise are usually physically separated from each other (e.g., by wire mesh)
and are locked.
For information about regulatory issues related to promoting infrastructure sharing and collocation, please see the practice note, “Steps to
Promote Infrastructure Sharing and Collocation”. A link to this practice note is set out below. Further discussion of the regulatory issues
related to infrastructure sharing and collocation generally can be found in section 6.6 of this Module, a link to which is set out below.
Discussions about sharing in the ICT sector frequently centre on infrastructure sharing and collocation. Spectrum sharing is gaining in
importance, however. A link to a discussion of spectrum sharing is set out below. Interconnection and unbundling are also forms of
sharing, although they are often not thought of strictly in terms of sharing. Links to sections that discuss interconnection and unbundling in
terms of sharing are set out below.
6.3.3 Interconnection
Although interconnection is often thought of in terms of the exchange of network traffic to allow users of different networks to
communicate with each other, interconnection can also be viewed as a form of sharing. Interconnection facilitates network interoperability.
Network interoperability, in turn, allows customers to have the benefit of access to different service providers (e.g., local service, long
distance, and Internet access) and the ability to communicate with other users on different networks through a single local loop. In this
regard, interconnection facilitates “last mile” sharing. Network interoperability also promotes competition in the ICT sector.
Box 6.3.3. The World Trade Organization/Reference Paper Definition of “Interconnection”
The Reference Paper defines “interconnection” as: “linking with suppliers providing public telecommunications transport
networks or services in order to allow the users of one supplier to communicate with users of another supplier and to
access services provided by another supplier, where specific commitments are undertaken.”
Source: Section 2, WTO Regulatory Reference Paper being the Annex to the Fourth Protocol to the GATS Agreement, the
“Agreement on Basic Telecommunications”, February 1997, in effect on 1 January 1998.
Interconnection is one of the most established forms of sharing in the ICT sector. As a general rule, the introduction of competition into a
country’s ICT market is accompanied by regulatory prescriptions for interconnection. Guidelines for interconnection are typically contained
in regulations, tariffs, or standard interconnection offers (sometimes called Reference Interconnection Offers or “RIOs”), which usually
requires regulatory approval. Members of the World Trade Organization (WTO) are also subject to international trade rules related to
interconnection regulation. In particular, the WTO Regulation Reference Paper sets out a number of requirements for the regulatory
framework governing interconnection in member countries. For more information on the regulation of interconnection, please see the
section 3 of this Module and the Practice Note on the WTO Rules on Interconnection Regulation.
6.3.4 Unbundling
Unbundling is the mandatory offering by network operators of specific elements of their network to other operators, on terms approved
by a regulator or sanctioned by a court. When services are unbundled, they are offered to other operators on a stand-alone basis so that
operators are not forced to purchase services that they do not require when they obtain the services that they do actually need.
Unbundling thus goes further than imposing an obligation on incumbents to offer interconnection services to entrants. It requires the
incumbent to allow entrants to lease certain individual building blocks that make up a telecommunications network.
Unbundling of network elements allows competing operators to enter the market and roll out services with considerably less sunk
investment in some or all components of a competing network. For example, a new entrant might initially install switches in central business
districts only and lease those components of the incumbent carrier’s network needed to directly serve customers in other areas.
Alternatively, an entrant might lease just those network elements needed to offer competing retail services (such as DSL services). In this
way the entrant can offer competing services to customers without duplicating all components of the incumbent carrier’s infrastructure and
without simply reselling the incumbent’s service offering.
Box 1: Local Loop Unbundling as a Form of Infrastructure Sharing
Local loop unbundling (LLU) is one alternative to achieve some of the goals of infrastructure sharing. The various
ways to implement LLU include full unbundling; linesharing, and bit stream access.
Full unbundling: This method of LLU assigns the entire copper local loop to the leasing operator. New entrants then
install their own broadband equipment and collocate, i.e., the new entrants place all the equipment in the incumbent’s
premises or outside the incumbent’s premises depending on which collocation model is most appropriate.
Line sharing is where the incumbent and other licensed operator share the same line. From the Main Distribution Frame
(MDF), the wires are connected to a splitter (which separates the frequencies for voice telephony and those for higher
bandwidth services). The incumbent provides voice telephony over the lower frequency portion of the line, while another
operator provides DSL services over the high frequency portion of the same line.
Bit stream access provides access to the bit stream on the network side of the Digital Subscriber Line Access
Multiplexer (DSLAM). Bit stream access is a virtual form of LLU; the new entrant does not obtain access to actual network
infrastructure elements. In the case of copper circuit unbundled access, the DSLAM connected to the unbundled circuit is
always installed and operated by the new entrant. In the case of bit stream access, the DSLAM is installed and operated
by the incumbent who also configures the DSLAM and sets up the required technical parameters (speed and quality of
service (QoS) attributes) of each user's DSL access link. The output of the DSLAMs on the network side is configured as
an Asynchronous Transfer Mode (ATM) transmission system.
The European Commission’s Regulation on Local Loop Unbundling (EC/2887/2000) came into force on 2 January 2001,
requiring incumbent operators to offer unbundled access to their local loops upon reasonable request. The Regulation
also requires incumbents to offer shared access and sub-loop unbundling or bit stream access. [1] Several developing
countries – South Africa, for example – also are examining LLU as an option to increase competition in their markets.
Some, such as Morocco, are already seeing significant gains from LLU.
Source: Dr. Tracy Cohen and Russell Southwood, Extending Open Access to National Fibre Backbones in Developing.
For more information on the GSR, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/index.html. For a direct link to the
GSR 08 discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
Up until the present, unbundling has primarily been offered on copper networks. However, there is no reason why the range of unbundling
options currently provided on copper networks (raw copper, bit stream, line sharing, etc.) could not apply in a fibre environment. It is true
that slightly different considerations may arise from the much higher bandwidth capabilities. But if the fibre is deployed within a next-
generation network (NGN) – with new intelligence and guaranteed quality of service built in – software-based unbundling solutions can be
explored in addition to providing access to physical network elements. These intelligent solutions would boost initial access until there were
competitive options available to resellers.
With respect to fibre backhaul for broadband wireless networks in developing countries, it does seem evident that some backhaul is
replicable, dependent upon geographical and population density variables. Implementing this option might include the regulator applying the
principles of a standard interconnection access regime, as interconnection issues would remain pertinent.
[1] See http://europa.eu.int/eur-lex/en/lif/dat/2000/en_300R2887.html
6.4 National Fibre Core Networks

Note: This section is based in large part on a discussion paper entitled Extending Open Access to National Fibre Backbones in
Developing Countries by Dr Tracy Cohen and Russell Southwood. This discussion paper was prepared for the 8th ITU Global Symposium
for Regulators 2008. [1]
This section discusses national fibre core networks. It addresses sharing-related issues in the context of such networks – a matter that is
growing in importance as countries seek to roll-out broadband in order to reap the benefit of digital connectivity.
[1] For more information on the GSR, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/index.html. For a direct link to the GSR 08
discussion papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.4.1 The importance of national fibre backbones and access to broadband

The economies of developed countries increasingly rely on widespread access to broadband services and applications. These include not
only traditional applications such as e-mail, but also newer applications such as Skype (VoIP services). Broadband services and the
infrastructure on which they depend have become recognized as an essential input to business, education, healthcare and participation in
the information economy.
The growing scale of online media use can be seen by the fact that online advertising will shortly exceed television advertising in several
developed countries.[1] A developed broadband infrastructure is an important attraction for locating business and company operations. It is
a pre-requisite for increased investment in any community.
Many of the countries in the Organization of Economic Cooperation and Development (OECD) have achieved not only ubiquitous access to
basic Internet services, but also high penetration rates for broadband access. (See the Practice Note entitled “Broadband in the OECD and
Selected Developing Countries”, a link to which is set out below.) That, in turn, facilitates more complex and effective service provision and
delivery models for both government and the private sector.
In economic terms, access to a national broadband fibre network is as important a priority as building an effective national transportation
network. Given the central role that ICTs play in the information economy, many argue that broadband access is a “public good” similar to
roads and railways. Without broadband access, developing countries run the risk of enlarging the “digital divide” and becoming second or
third class nations within the global order. Having competitively priced national broadband access has become an important criterion of
global competitiveness. Broadband investment has also become a popular element of national and international programmes of economic
stimulus.
Although there are enormous obstacles to implementation in some developing countries, broadband access also offers the potential for
delivering government services more effectively and at a lower cost. It can help governments address poverty by minimizing the rural-
urban divide that is so common in developing countries. A broadband infrastructure can, for example, promote the economic welfare of
people living outside major cities and urban centres by relocating “back-office” jobs to rural and commercially depressed towns and cities
and by attracting work outsourced from developed countries.
Sharing infrastructure is one strategy for achieving a national broadband infrastructure more quickly than by simply letting the market take
its course. That is because the development of national communication infrastructure in developing countries is often blighted by a
recurring “chicken-and-egg” problem: without this improved Internet access, there will not be a “critical mass” of users, but without those
users, the customer base will not exist to support extensive network operations. Unless governments resolve this problem, the social and
economic benefits that national broadband access can deliver may never be realized. For more information on the obstacles to broadband
network deployment, please see section 6.4.2.
[1] The latest forecast suggests that this will occur in many developed countries in 2008. See “Burgeoning online advertising spend will
pass TV next year”, The Independent (London), 4 January 2008.
6.4.2 Obstacles to the deployment of broadband networks
Many countries in both the industrialized and developing worlds struggle on some level to achieve efficient and widespread deployment of
broadband networks. Rural and thinly populated areas are particularly vulnerable to under-development in terms of broadband access.
Two of the most common problems that impede rapid and efficient deployment of broadband networks are:
■ the control of bottleneck facilities by a single dominant infrastructure operator, and

■ a lack of investment in high-capacity infrastructure in un-served or under-served areas.
Both of these issues may be present in the same country and are often linked to the way infrastructure was constructed and maintained in
the past.
In the case of “bottleneck facilities,” the operator (usually the incumbent) has little or no commercial incentives to provide its competitors
with access to its infrastructure. The incumbent often has an unfair advantage over its competitors at all levels, but particularly in
downstream markets, due to its ownership of key network facilities. Most commonly, this shows up in the price advantages that a
vertically-integrated operator can retain unless it is constrained; it is both its own customer and competes with the other customers it
supplies. In these circumstances, the dominant infrastructure operator becomes the obstacle to both the development of new infrastructure
and, more generally, to the expansion of competition and market growth. The “bottleneck facilities” problem is the most fundamental of all
interconnection problems because it can prevent equitable sharing of a dominant operator’s network.
Sharing addresses the obstacle posed by the control of bottleneck facilities by requiring dominant operators to provide access to their
facilities on mandated terms and conditions. In some cases, the regulator has also responded to this issue by requiring the wholesale
infrastructure (network) business to be as operationally separate as possible from the retail business in order to allow completely
transparent trading between the wholesale and retail sides of the business. Another approach is to create a national, state-owned
enterprise to construct and to operate a national fibre core network. For more information on approaches taken by regulators to opening up
access to bottleneck facilities, please see the Practice Note entitled, “Sharing and Access to Bottleneck Facilities”. A link to this Practice
Note is set out below.
In the case of un-served or under-served areas, policy-makers usually aim to create a greater “critical mass” of users by encouraging the
roll-out of high-capacity, national infrastructure to a wider range of places than the market alone might initially sustain. Sharing facilitates a
wider deployment of a national fibre core network by reducing the high costs associated with rolling out the network infrastructure. In
essence, the argument for sharing national infrastructure is that two or more operators sharing (and paying for access to) a common
infrastructure will help finance a wider deployment, whereas traffic from a single operator would not sustain a widespread network.
Of course, different approaches are needed for different market dynamics. If multiple existing or potential players want to roll-out a
network, then a facilitating agency may succeed by offering passive infrastructure assets like rights of way and government land for sites.
In developing countries, however, the main challenge is often finding operators willing to go into under-served areas at all. The common
view is that network roll-outs do not make commercial sense in remote, thinly populated areas. Where nobody appears interested in
entering such markets, the government might need to take the primary risk by encouraging investment in a wider national network and then
devising a fair and efficient mechanism to share this resource with existing market players. The issue for governments is whether to
duplicate elements of existing networks or to “fill in” gaps in those networks. The latter might easily be seen as a suitable task for universal
access funding, but experience in some regions (Africa, for example) has shown that these funds often do not produce speedy network
roll-outs.
In markets where full liberalization has yet to occur, however, the slow progress of infrastructure investment should not necessarily be
interpreted as a lack of willingness on the part of the private sector to invest. In some cases, restrictive policy and regulatory environments
simply do not allow for commercial evaluation to mature into investment. As the chairperson of the South African Competition Tribunal has
noted, “The country's access deficit [the lack of broadband connectivity] was not due to market forces not working, but due to the fact that
we have not had a working market. If there is one market that responds to market incentives, it is the telecoms market.”[1]
[1] Paul Vecchiatto, “Telecoms policy a 'mistake'”, itweb.co.za, 15 June 2007.
6.4.3 Key players
There are a number of different players who may have a role in the establishment, management, and operation of an open access national
fibre core network. This section reviews the possible involvement of the following players:
■ Government;
■ Incumbents and dominant operators;
■ Non-dominant operators; and
■ Infrastructure owners and operators;
■ Users.
Government
Government has a key role to play in facilitating the most effective use of infrastructure assets, identifying parts of the country where
there are gaps, and getting coverage extended to them. Government may direct network roll-out through the imposition of universal access
obligations, for example. It may also create incentives for network roll-out through the award of universal access funding.
In a direct sense, government can make investment in rolling out services to a marginal location more desirable by becoming a key “anchor
tenant” in such areas. A remote border town, for example, might connect its customs post, local government centre and school through a
new broadband network. By becoming a significant customer, the government can increase the expected returns on investment in network
roll-out.
An important dimension of the role of government pertains to the creation of a regulatory and policy environment that facilitates sharing.
Policy makers and regulators must create a framework that permits or even mandates sharing, while also addressing concerns about anti-
competitive conduct and the safeguarding of consumer data.
Government may also have a direct role to play in the actual establishment and operation of a national core network. In some countries,
government has created national infrastructure companies or has partnered with private companies to fund the establishment and
operation of a national fibre core network. Although operators stand to save considerable costs by sharing infrastructure, there may be
other compelling, pragmatic reasons for not sharing. There is often insufficient trust between operators to enable them to cooperatively
examine how they might share national infrastructure. New operators often experience poor delivery of services from the historically
dominant infrastructure provider, so they often do not believe they can operate effectively except by providing for themselves. Moreover,
in some cases, operators may consider that it is the government’s responsibility to provide national fibre infrastructure or key parts of this
infrastructure that the operators are not providing.
For more information on the role of government in promoting the creation and operation of a national fibre core network, please see the
Practice Notes linked below and section 6.4.6 on Public-Private Partnerships.
Incumbents and dominant operators
Incumbents and dominant operators generally have weak incentives to allow open access to their networks. Given the high cost of
replicating the incumbent’s network and the control the incumbent has over bottleneck facilities, however, it is usually necessary to
mandate access to the incumbent’s network in order to promote competition and to extend network roll-out. Two of the most common ways
of providing access to the incumbent’s network are interconnection requirements and unbundling.
More recently, as Next-Generation-Networks grow in importance, there is increased regulatory emphasis on the operational separation of
the infrastructure (core network/wholesale) side and the service delivery (retail) side of the incumbent’s business. Operational separation
of the wholesale and retail sides of the business allows the incumbent to provide other operators with open access to its core network
while still competing in a fair way with these same operators in retail markets.
Non-dominant operators
Non-dominant operators are often the parties seeking access to the national fibre core network. However, there are synergies that can be
gained if non-dominant operators are open to collaboration in the construction and operation of core network infrastructure. Collaboration
may take the form of joint ventures to construct or to operate infrastructure, sharing arrangements, and interconnection agreements. It is
also possible that non-dominant operators may be reluctant to enter into these types of arrangements due to a lack of trust. Governments
and regulators may need to take steps to encourage sharing and to facilitate collaborative efforts to construct infrastructure. At the same
time, depending on the circumstances, regulators may need to monitor arrangements so collaboration does not become collusion.
Box 1. The German National Broadband Strategy: Capitalising on synergies from infrastructure projects
“…Up to 70 per cent of the costs of developing broadband infrastructure in the fixed-line network are excavation costs.
These costs may be reduced significantly if the various infrastructure providers become more open to collaboration and
allowing third-party access to their own systems. This openness would create many win-win situations for businesses,
while reducing costs for the economy as a whole. If there is broad support for the measures, thus reducing the
infrastructure costs for expanding broadband infrastructure within the fixed-line network by just ten per cent, we can
achieve a saving of approximately three billion euros in the next few years.”
Source: Germany, Federal Ministry of Economics and Technology, The Federal Government’s Broadband Strategy
(February, 2009), available at: www.bmwi.de.
Infrastructure owners and operators
The emergence of Next-Generation Networks has given rise to an interesting new type of player in the ICT market: infrastructure owners
and operators. Infrastructure companies focus on constructing, maintaining, and operating core and access network facilities for other
operators; these companies do not provide retail services to end-users. In some cases, such as in Singapore, the government has
provided subsidies to private companies to construct core network infrastructure.
Users
In order to promote the deployment of national broadband networks, governments may want to encourage a wider range of partners to
participate in the national infrastructure task. Users of broadband services can aggregate demand and develop partnerships to create open
access networks. There is a need for bodies like universities to participate through what is known as national research and education
networks (NRENs), which can be both users and bandwidth buyers. Having user voices in the governance of shared infrastructure
projects ensures a consistent focus on the overall objective of cost-effective delivery of bandwidth. The customers and participants in
such a joint venture might include: existing national operators (fixed and mobile), ISPs, large-scale corporate customers (like banks),
government funded services like public universities, hospitals and clinics and government departments. Depending of the state of
development in the market, access can be offered to users in just layer one or both layers one and two.
6.4.4 Passive Infrastructure Sharing
This section outlines some key considerations related to the sharing of various elements of passive infrastructure. Box 1 lists elements of
passive infrastructure that may be shared at the national level.
Physical ducts, masts/poles (in the case of power transmission lines), and rights of
Box 1: Elements of passive way are key passive network elements necessary for the roll-out of national fibre
infrastructure that may be shared in networks. Allowing access to these passive network elements allows for cost-
national fibre networks savings and a more rapid deployment of national fibre network infrastructure and
❍ Cables services.
❍ Ducts On a practical level, it is possible, but not desirable, for every operator to create its
❍ Splitters own system of physical ducts and conduits. There are several material reasons for
❍ Diesel electric generator favouring shared access to physical ducts and conduits and the sharing of rights of
way. First, requiring each operator to build its own system of ducts and conduits
❍ Battery
would delay network roll-out given the time necessary for this construction. Second,
❍ Electrical supply the process of network deployment would include considerable chaos and disruption,
❍ Shelters particularly in urban areas, since each operator would be forced to plough up roads
❍ Generators and undertake construction.
❍ Air-conditioning equipment Finally, sharing access to physical ducts and conduits, as well as the related rights of
❍ Technical premises way and easements, offers significant cost savings. National governments,
❍ Easements municipalities and state-owned enterprises frequently charge considerable sums of
money for rights of way that allow operators to carry out physical trenching of ducts.
❍ Pylons
Often, the actual laying of cable represents a relatively small part of the overall costs
Source: ITU and ARCEP of deploying a fibre network. Obtaining the rights of way is what adds to the overall
cost. If each operator had to buy access to rights of way separately, the wholesale
distribution costs would mount up rapidly, ultimately raising prices for consumers.
Moreover, the high cost of entering the market would reduce the number of new
entrants and limit competition. For more information on access to rights of way, please see the Practice Note entitled, “Sharing rights of
way”. A link to this Practice Note is set out below.
There is often opportunity to share passive infrastructure elements with operators active in other industry. Passive elements of a national
fibre core network such as the physical home of the fibre cabling, whether in ducts or physical sheaths, may be shared with operations
related to oil pipelines, power transmission facilities, and railways. Box 2 provides information about sharing passive infrastructure with
non-telecommunication infrastructure operators. Please also see the Practice Note, “Trends in regulating cross-industry infrastructure
sharing”, a link to which is set out below.
The physical home of the

Box 2: Sharing with non-telecommunication infrastructure operators cabling – whether in ducts
Cost-sharing infrastructure deals are often made because the infrastructure is being built for another or physical sheaths – is a
reason and the cost of adding more capacity is marginal. This is particularly true for fibre networks key part of achieving a
used to manage diverse operations such as oil pipelines, power transmission facilities, and railways. shared national network.
Each requires its own fibre for management purposes, but it is relatively easy to add fibre strands, Physical ducts are often
either before or after construction. The resulting additional capacity can then be shared either by the scarce, under-utilized and
operating company setting up its own wholesale fibre capacity sales operation or by selling the right to have long pay-back periods.
provide the capacity to an independent organisation. In Africa, there are several examples of this, Opening up access to them
including the Cameroon-Chad oil pipeline (known as the Doba-Kribi pipeline), Kenya Power and Light in a variety of different
and Tanzania’s TANESCO. In the case of the oil pipeline, 12 out of the 18 fibre cables installed will be ways creates incentives for
available for use by telecommunication operators. Arrangements of this kind clearly cut the cost of sharing among operators
building network infrastructure. and helps ensure maximum
use of these relatively
Source: IT Shared Infrastructure, paper prepared for the ITU Africa regional conference, Nairobi, 2007. scarce resources.
Once access to ducts or

power transmission masts is
allowed, operators wanting
national network capacity
between different physical points then can choose to either invest in their own dark fibre or buy capacity on a leased or indefeasible right
of use (IRU) basis. Lateral and mid-span splices may be available, giving greater flexibility. In this circumstance, operators can provide their
own equipment to connect to the network capacity they have bought. Of course, there must be sufficient space to accommodate all
potential network users, but since Internet Protocol (IP) network access equipment is relatively small, physical space is hardly an issue.
Source: Dr. Tracy Cohen and Russell Southwood, Extending Open Access to National Fibre Backbones in Developing Countries, at:
www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.4.5 Active Infrastructure Sharing
Sharing active infrastructure is a much more contested issue than sharing passive infrastructure, as it goes to the heart of the value-
producing elements of a business. Box 1 sets out some of the active infrastructure elements that may be shared in a national fibre core
network. This section provides examples of active infrastructure sharing that demonstrate the breadth of active infrastructure elements
that might be included in a sharing framework. These examples also demonstrate that while much is technically possible, operators often
have concerns about various forms of the sharing of active infrastructure elements.
At layer two, the transport layer, the operator sharing infrastructure

Box 1: Elements of active infrastructure that may be can provide a wholesale, point-to-point fibre service to other entities,
shared in a national fibre core network which can then use it to provide services across layer three - the
❍ Optical network unit (ONU) services layer - of the network. In this scenario, each service provider
and its associated customers are assigned to a separate virtual local
❍ Access node switches
area network. Or, if the provider is using Asynchronous Transfer Mode
❍ Management systems (ATM), it uses separately assigned permanent virtual circuits. Technical
❍ Broadband Access Remote Server (BRAS) and service characteristics may vary depending on the network
❍ Coarse or dense division multiplexing architecture, but there is no insurmountable obstacle to sharing in this
layer.
❍ Software (core network systems such as billing)
For video delivery (IP-TV and video-on-demand), IP networks will also
Source: Dr. Tracy Cohen and Russell Southwood,
be easily accessible, although there may be issues about provisioning
Extending Open Access to National Fibre Backbones in
the necessary capacity to deliver this kind of service. However, to
Developing Countries, at: www.itu.int/ITU-
date, video is far harder to deliver on a Passive Optical Network (PON),
D/treg/Events/Seminars/GSR/GSR08/papers.html.
except by using a video overlay. (Although it is theoretically possible,
there are no available examples of video delivery on a PON using this
kind of overlay.) Since vendors sometimes represent PONs as a way to
retain control over the core network, there may be an understandable,
but perhaps unwarranted, suspicion that the difficulty of creating
shared services is intentional. Not having such an overlay on an IP network can save electro-optical costs, but it raises transport and
switching costs. In sum, while there remain some technical issues related to sharing of IP networks for video delivery, the main
impediments remain policy and regulatory in nature.
Once there is a widespread fibre network, the next question is how that capacity will be delivered to the customer’s premises, whether
that is a home or an office. To encourage speed of network roll-out, it may be useful to encourage the infrastructure operator to provide
what has been termed a “fibre management point.” In effect, a fibre management point allows competitive service providers to take the fibre
capacity offered by the infrastructure operator and deliver it locally. Where and how these fibre facilities are provisioned will depend on
the population density, geographic characteristics, and the level of market development.
At a slightly more complex level, service providers can each transmit on their own wavelength using either Coarse or Dense Wave Division
Multiplexing over long national or international routes. Multiple service providers can be supported, with each one effectively operating its
own network and making its own commercial decisions. With this approach, service providers can treat their capacity as a component of
their own network offerings. However, as far as is known, this separate, shared network approach has only been used for international
cables.
Active infrastructure sharing is also possible on mobile networks. For information about mobile virtual network operators (MVNOs), please
see section 6.5.9.
Source: Dr. Tracy Cohen and Russell Southwood, Extending Open Access to National Fibre Backbones in Developing Countries.
For more information on the GSR, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/index.html. For a direct link to the GSR 08 discussion
papers, see: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
6.4.6. Public-Private Partnerships and Open Access Networks
One interesting development related to the goal of creating national fibre core networks is the emergence of new public-private
partnerships (PPPs). A variety of different PPPs have been created to facilitate the speedy roll-out of high capacity core and access
networks. These networks are designed to operate on the basis of open, non-discriminatory access. Many of the PPP initiatives focusing
on national fibre core networks feature the participation of national governments, while the PPPs that centre on establishing fibre access
networks (FTTx) often involve local, municipal, and regional governments.
In developing countries, a common challenge is finding operators willing to go into under-served areas. The common view is that network
roll-outs do not make commercial sense in remote, thinly populated areas. Where nobody appears interested in entering such markets, the
government may need to take the primary risk by encouraging investment in a wider national network and then devising a fair and efficient
mechanism to share this resource with existing market players. Although universal access funding could be used to promote the
deployment of national core networks to under-served areas, the experience in some regions (Africa, for example) has shown that these
funds often do not produce speedy network roll-outs.
The desire to create a more far-reaching national infrastructure has been the motivation of several African governments in creating national
infrastructure companies. The previous Kenyan government prepared a plan to build a fibre network designed to cover the whole country
based on open access, called FON (Fibre Optic Network). It envisioned a network either run on contract by a private sector provider or by
the former incumbent, Telkom Kenya, under clearly agreed terms. A similar approach has been adopted in Uganda, where the proposed
network architecture was amended to take into account fibre already laid down by the existing operators, MTN and UTL. In both cases the
governments’ proposals were presented as a way of offering a wider range of coverage and bringing national network costs down.
PPPs aimed at deploying fibre access networks exist mostly in developed countries, where basic access to services has already been
established. Municipalities and local governments have played leading roles in developing partnerships with private sector operators and
users to promote local FTTx. In a number of cases, local authorities own broadband access networks that pass by as many businesses,
schools and universities, hospitals, government buildings, and homes as possible. These networks are sometimes called “Metropolitan Area
Networks” or MANs. The local authorities contract with private sector operators, who provide open, non-discriminatory access to the
network facilities to retail service providers. The retail service providers then offer broadband, television, and telephony services to the
public. See Practice Notes (linked below) for a description of various PPP initiatives aimed at providing fibre access networks, including the
Amsterdam FTTH project, Ireland’s SERPANT Broadband Project, Sweden’s Stokab system, and the MBC Cooperative in Virginia, USA.
Box 1: UTOPIA: Open access municipal FTTH in the United States
UTOPIA was originally formed in 2002 by fourteen cities in the State of Utah in the United States. UTOPIA’s mission is to build and
maintain an FTTH open infrastructure network. The project is funded by the sale of bonds which are guaranteed by 11 of the cities
involved in the project. In 2004, USD 85 million in bonds were sold to fund the first phase of construction which involved the laying
down of fibre for six southern cities. The project is currently in its second phase, which involves rolling-out fibre in the remaining five
northern UTOPIA cities. To repay these bonds, UTOPIA will collect a wholesale fee from service providers. If such revenues prove to
be insufficient, however, the 11 guarantor cities will be required to honour UTOPIA’s bond commitments with monies levied from sales
taxes. Currently, several small service providers such as MSTAR, Veracity Communications and X-Mission Internet, as well as large
providers like AT&T, are offering voice, broadband and television services via UTOPIA’S network.
Sources: www.utopianet.org; Steve Cherry, "A Broadband Utopia Continued", IEEE Spectrum Online, May 2006, available at:
www.spectrum.ieee.org/may06/3434/3
The question remains, however, whether core and access networks are an appropriate targets for government investment. Should these
networks be operated “at cost,” as is currently being proposed, and if so, on what basis? Considering that it is unclear what costs will be
covered in “at-cost” operations. For example, will the costs cover capital replacement and maintenance? If not, the government provider
clearly will have an unfair advantage in the market. Therefore, it is important that regulators and policy-makers ensure that these initiatives
do not provide unfair advantages that dominant infrastructure operators have retained in the past.
6.4.7. National broadband policies
One interesting development related to the goal of creating national fibre core networks is the emergence of new public-private
partnerships (PPPs). A variety of different PPPs have been created to facilitate the speedy roll-out of high capacity core and access
networks. These networks are designed to operate on the basis of open, non-discriminatory access. Many of the PPP initiatives focusing
on national fibre core networks feature the participation of national governments, while the PPPs that centre on establishing fibre access
networks (FTTx) often involve local, municipal, and regional governments.
A growing number of countries are developing or have already developed national broadband policies. These policies are aimed at setting
goals for the roll-out of national broadband core networks and FTTx services, as well as identifying key policy and regulatory mechanisms
to promote this roll-out. National broadband strategies offer an opportunity to introduce a national infrastructure sharing policy. For example,
the German Federal Government’s Broadband Strategy specifically identifies infrastructure sharing as an important mechanism for
expediting the roll-out of faster high-speed networks and the connection of rural areas to broadband Internet. The German Broadband
Strategy also envisions sharing between electricity and energy suppliers and telecommunications operators.
A national infrastructure sharing policy gives regulators levers to overcome barriers and speed up broadband network development. Of
course, different circumstances require different approaches, and a light regulatory approach, using persuasion, is nearly always
preferable to imposing mandates. But there will be times when governments must legislate or regulate to overcome the intransigence of
some major stakeholders.
Box 1. Developing a national infrastructure sharing policy
Before framing a national infrastructure sharing policy, officials should consider two important questions:
(1) whether the creation of shared infrastructure promotes competition and lowers prices for consumers, and
(2) whether such sharing arrangements encourage operators to make technical innovations that lower costs and improve services.
Different approaches are needed for different market dynamics. If multiple existing or potential players want to roll-out a network, then a
facilitating agency may succeed by offering passive infrastructure assets like rights of way and government land for sites. In developing
countries, however, the main challenge is often finding operators willing to go into under-served areas at all. The common view is that
network roll-outs do not make commercial sense in remote, thinly populated areas. Where nobody appears interested in entering such
markets, the government might need to take the primary risk by encouraging investment in a wider national network and then devising a fair
and efficient mechanism to share this resource with existing market players. The issue for governments is whether to duplicate elements
of existing networks or to “fill in” gaps in those networks. The latter might easily be seen as a suitable task for universal access funding,
but experience in some regions (Africa, for example) has shown that these funds often do not produce speedy network roll-outs.
simply do not allow for commercial evaluation to mature into investment. In such cases, the development of a national broadband strategy
may offer an opportunity to advance the move to full liberalization.
Sharing infrastructure in a liberalized market can provide the catalyst for deploying costly nation-wide network infrastructure, while
simultaneously allowing operators to compete fiercely in other layers in the market. Infrastructure sharing also can be an additional tool for
policy-makers and regulators to entice new, green-field fibre backbone operators to compete in the same market as existing network
operators and service providers. These new operators may be merely authorized to enter the market or actively encouraged through tax
incentives and the creation of joint venture or co-operative vehicles – often involving governmental or quasi-governmental partners. In other
words, policy makers can act to maximize the number of potential backbone providers entering the market rather than limiting market entry
to one or two.
Where market entry is limited, policy makers and regulators then have to decide whether infrastructure sharing is a short-term tactic or a
long-term (or permanent) strategy for achieving policy goals. If sharing infrastructure is viewed tactically, a country may reach the point
where a national infrastructure is more or less in place, allowing regulators to create a level of infrastructure-based competition again. For
example, it may be useful to have competing infrastructure providers on key routes between cities. This approach would allow some price
competition and might open up additional capacity more quickly than would a single or shared infrastructure approach. Viewed
strategically, sharing mandates allow regulators to intervene to protect consumer welfare and the positive externalities that flow from
having a national broadband infrastructure. Regulators can use sharing to ensure that essential facilities are not used to limit competition.
Before framing an infrastructure sharing policy, however, officials should address two questions: (1) whether the creation of shared
infrastructure promotes competition and lowers prices for consumers; and (2) whether such sharing arrangements encourage operators
to make technical innovations that lower costs and improve services. The first question is likely to be answered in the affirmative. The
second question is slightly more complex. But both questions are directly linked to the type and quality of the access regime the policy-
makers or regulators implement.
6.4.8 Regulatory and legal imperatives related to sharing
Keeping in mind variations within and among countries at different levels of development, the general regulatory and policy imperatives for
infrastructure sharing include the following:
Investment Incentives – The key challenge is how to ensure that there are adequate incentives within the regulatory framework for
operators to invest in infrastructure. Conditions must be created to spur continued roll-out of infrastructure that is not easily replicable,
while at the same time providing access to these facilities for new competitors. The result is an inevitable tension between the equally
important policy objectives of open access and investment. Open access may entail some loss of return for the owner of the infrastructure
if prices are regulated.
Open Access – Along with the principle of non-discrimination, open access is important because it lowers the barriers to entry of new
market entrants. The key problem is to provide competition at all (or most) levels of access, as this will lead to direct benefits in terms of
technological innovation, lower prices, and broader market penetration. If licensing reform is required, regulators should ensure there is a
regime that is capable of addressing access at any of these levels.
Open access thus becomes a tool to motivate infrastructure investment and facilitate new market entry and competition. But if too much
access is given to new entrants, the result can be just as bad as exclusivity: there will be little or no investment in infrastructure. An
example of this is the market in the United States between 1996 and 2002. The reason that there was so little investment in (local loop)
infrastructure by new entrants was that they could not deploy new infrastructure at the regulated local service prices, which were too
low and acted as a disincentive to investment.
Figure 1: The Policy Challenge – Balancing Open Access and Exclusivity

Source: Dr. Tracy Cohen and Russell Southwood, Extending Open Access to National Fibre Backbones in Developing Countries, at:
www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
In Figure 1, the dotted line between open access and exclusivity represents the tension between the two options. The traditional approach
has been to provide the incumbent with an exclusivity period, during which it must roll out a prescribed amount of infrastructure. In nearly
every case this policy has failed to boost network deployment significantly. But there is also no point in a regulatory policy that promotes
the other extreme: full access under terms that undercut the incentive for investment in new facilities. Regulators must adopt a dynamic
approach that takes into account changing cost factors, technological innovation and the stage of competition at each level of access.
Regulatory Tools - The key challenge is how to avoid reinventing the wheel. Governments have been grappling with access regimes for
the last two decades as deregulation and liberalization trends have increased. The tools to open up access to essential facilities are well
ensconced in current practices, including rules on interconnection, pricing, local loop unbundling, dispute resolution and others.
While the tools to address access to essential facilities are well established, regulators and policy makers will need to ensure that there is
a clear definition of essential facilities that remains relevant in light of ongoing technological advances. Regulators and policy makers must
consider: where are the bottlenecks in the network, and is it possible to provide alternative infrastructure to bypass them? In other words,
is any given network element able to be duplicated in a way that is economically, environmentally or technically feasible? One of the most
commonly cited examples of essential facilities is the local loop. The conclusion in nearly all countries is that it is not feasible to duplicate
this network facility and, therefore, it should be declared an essential facility. Because of the historical role governments have played in
providing essential facilities, a political commitment to opening access to these facilities is essential.
6.4.9 Commercial and technical considerations related to implementing sharing
To move the debate on implementation beyond the broad principles of non-discrimination, regulators need to identify and remove any
technical constraints to infrastructure sharing. The following principles are important:
Commercial Imperatives: As far as possible, infrastructure sharing arrangements should proceed from commercial negotiations, within
the parameters of regulatory rules and guidelines. Regulators can set general principles and time limits for concluding agreements. The
Telecommunications Regulatory Authority of India (TRAI), for example, has required operators to set up programmes for passive
infrastructure sharing on existing and future mobile towers. TRAI also requires that that sharing be offered on a first-come, first-served
basis subject to commercial agreements. Policy makers/regulators might also simply list and identify critical infrastructure sites without any
further policy intervention.
Non-discrimination and Transparency: Price discrimination occurs when on operator sets infrastructure access prices in a manner
that precludes competition. Non-price discrimination is when an operator sets access terms and conditions that are less favourable than
those it provides to itself or its subsidiaries. Transparency of prices, terms and conditions is imperative to avoid both kinds of discrimination.
And ensuring transparency is a core regulatory duty. New market entrants and service providers should have access to the full range of
collocation and interconnection services, at regulated prices, if necessary. TRAI has not mandated precisely how passive infrastructure
sharing should take place, but it has required the process to be transparent and non-discriminatory. Licensees must publish on their
websites the details of existing as well as future infrastructure installations available for sharing by other service providers.
Technical Feasibility: Access at different layers in the network raises issues of technical compliance and feasibility. All too often,
regulators must resolve interconnection disputes over technical feasibility. Some of these claims of infeasibility may be valid. In the mobile
market for example, many networks were constructed specifically for certain operators, without any consideration of potential sharing.
There is also a view that fibre PON networks (point to multipoint) are not designed for sharing, particularly for video applications. But the
idea behind open access is that anyone should be able to connect to anyone else in a technology-neutral framework – at any level within
the network. Regulators must therefore guard against denial-of-access abuses, which often amount to a tactic to inhibit competition.
Countries that have yet to deploy fibre networks, meanwhile, should require deployment of networks that are capable of open access.
Moreover, as evidence from various countries illustrates, fibre networks are not only capable of open access, many of them are already
providing it.
Pricing: Different pricing options can be explored, but some type of cost-based methodology is usually considered the best practice. This
calls for significant resources from regulators to ensure correct implementation. It may also be prudent to explore an option pricing
approach that sets initial infrastructure prices low to encourage market entry, but allows prices to then rise over time. This may provide an
incentive for new market entrants to eventually build their own facilities as they build market share. In any case, pricing should allow
infrastructure owners to gain a reasonable rate of return, ensuring ongoing provision and maintenance. There is no point to mandating
below-cost provision.
Dominance or Significant Market Power (SMP): Policies to promote telecommunication development and competition require a
comprehensive framework for preventing and managing anti-competitive conduct. Regulators must prevent any abuse of market power or
dominance in cases where the infrastructure owner also competes downstream with other service providers in the same market. Such
policies should incorporate an implicit understanding that regulatory intervention is primarily needed only when an operator possesses SMP
and begins abusing that power. The standard best practices for intervention apply in infrastructure sharing regulation as they would in
pricing, interconnection and other areas of competition policy.
Enforcement: No policy can be effective without a sound enforcement framework, in which complaints can be brought and disputes
resolved with respect to sharing infrastructure. It is not possible to prescribe the exact enforcement measures that regulators should take;
each country has its own institutions and laws. The common thread, however, is a solid and effective mechanism for handling complaints,
with sufficient sanctions to discourage violations.
Incentive Creation: Markets respond well to commercial incentives, and policies should create financial incentives to share
infrastructure. In the functional separation context, BT’s allowable 10 per cent investment return on network assets is one example. Other
such incentives could include regulatory exemptions, subsidies, regulatory fee reductions, tax exemptions, or reduced local government
fees for rights-of-way or digging up city streets. Regulators could also consider awarding more spectrum to operators that share
infrastructure.
Role of Government: In promoting an open access model, governments need to take a firm policy position on their role in the sector. In
most developing countries, there has been a legacy of state ownership of incumbent operators. That often lives on in the current practice
of governments’ retaining some form of equity stake, even after liberalization policies have been implemented. Governments need to decide
if their role is to oversee and promote the sector or actively engage in it as market players. Evidence tends to warn against state
involvement in the sector as both a market player and policy maker.
Given the multiple ways to undertake infrastructure sharing, and how varying levels of market maturity and investment imperatives will
affect these decisions, it is difficult to provide a “best practice” template for implementation. Moreover, the following guidelines will need to
be adapted to different countries’ sector governance structures, some of which grant policy-making authority to ministries while others
give it to regulators. Nevertheless the following guidance is offered as a starting point.
National policy-makers need to:
■ Look at the market and enact enabling laws and regulations to facilitate the build-out of national infrastructure – including revised
licensing and interconnection policies to enable open access;
■ Co-ordinate with other government departments to ensure that the country’s non-telecommunication infrastructure (e.g., utility masts
and ducts) can be leveraged to facilitate telecommunications network deployment;
■ Design policies to boost and accelerate infrastructure investment;
■ Exhibit the required political will to follow through with clear, targeted, proportional regulation to achieve desired outcomes;
■ Design incentives for infrastructure investment in under-served and un-served areas (e.g., tax exemptions or rebates);
■ Consider a policy to separate the retail and wholesale functions of network infrastructure; and
■ Act as a clearinghouse for granting rights of way.
Local governments need to:
■ Assist operators to gain rights of way and access to ducts and poles;
■ Set up a clearinghouse for securing rights of way and other permits from multiple agencies or authorities;
■ Provide information, such as site surveys and geographic information system (GIS) data for public lands; and
■ Reduce operators’ costs for obtaining rights of way and other local permits for infrastructure deployment.
Regulators need to:
■ Embark upon consultation processes to (1) assess the market need for open access, (2) propose open access rules, and (3)
determine how to intervene to impose open access mandates in the most appropriate, directed and proportional manner;
■ Implement licensing/authorization frameworks to allow open access providers and create incentives for those who have spare
capacity on their networks to share that capacity;
■ Design regulatory interventions that are based on the technical reality of access at multiple levels of the network;
■ Create incentives to promote infrastructure sharing on commercial terms;
■ Improve transparency requirements for operators to publish relevant information for infrastructure sharing;
■ Decide whether to approve or require publication of reference sharing offers covering offerings such as collocation space and
connection services, power supply, air conditioning, maintenance access, etc.;
■ Establish where bottleneck facilities are and whether it is economically, technically or environmentally possible to duplicate such
facilities;
■ Where necessary, establish the cost methodology (i.e., cost, plus a fair rate of return) that will serve as the foundation for access
pricing;
■ Establish who is responsible for granting rights of way and assist operators in negotiating the complexities associated with dealing
with multiple agencies;
■ Establish sound monitoring and enforcement for implementing infrastructure sharing, including speedy dispute resolution among
operators;
■ Require the publication of reference interconnection offers (RIOs) or similar instruments by operators with significant market power
that specify the terms of access; sub-licensing if necessary; charges, billing, dispute resolution, etc.;
■ Use competitive bidding processes or auctions when authorizing municipal or backhaul providers;
■ Coordinate trenching and ducting operations among operators, and provide mechanisms for monitoring duct upgrading to ensure that
operators remove obsolete cabling to allow the introduction of third-party fibre;
■ Publish a list and identify critical infrastructure sites; and
■ Establish a dispute resolution mechanism.
Industry players need to:
■ Assess the business case for sharing, rather than duplicating infrastructure;
■ Move away from the assumption that excluding access to network elements is the only way to secure revenue;
■ Co-operate with regulatory/policy processes;
■ Improve transparency and publish on websites the details of existing and future facilities available for sharing; and
■ Coordinate trenching and ducting operations with other telecommunication operators and utilities.
6.5 Mobile Network Sharing

Note: This section is based in large part on a discussion paper entitled Mobile Network Sharing by Camila Borba Lefèvre. This discussion
paper was prepared for the 8th ITU Global Symposium for Regulators [1] in 2008.
This section explores mobile network sharing. Mobile network sharing offers a number of potential benefits to both operators and users. It
has particular relevance in developing countries where mobile operators have become or are fast becoming the dominant infrastructure
operators. Mobile networks also have a key role to play in rolling out services to remote and hard-to-service regions.
Mobile network sharing is also relevant in developed countries. Many initiatives aimed at providing free Internet access in urban areas use
mobile technology. Mobile network sharing also plays a key role in the structuring of triple-play and quadruple-play service packages.
This section outlines the policy benefits associated with mobile network sharing, the various forms of passive and active sharing, and
regulatory and policy issues associated with mobile network sharing.
Endnotes
6.5.1. Policy reasons for supporting mobile network sharing
This section sets out various policy reasons for supporting mobile network sharing. These policy reasons fall into three broad categories:
environmental and public health considerations; the facilitation of network deployment; and upgrading technology from 2G (GSM) to 3G (IMT
2000) and beyond. Each of these categories is discussed below.
Environmental and public health considerations
People generally view wireless communication masts and antennas as negative additions to the landscape. Local communities may object
to the construction of new sites because of the visual impact or environmental considerations. Also, residents may fear public exposure to
electromagnetic fields around masts and antennas. [1] Site sharing can limit such concerns and potential negative effects, since it limits the
number of sites while achieving the required coverage. [2] Another beneficial aspect of site sharing is the amount of energy that can be
saved when operators share electrical power, which is often in limited supply in developing countries. [3]
While sharing reduces the number of sites marking the landscape, it can also have adverse impacts. Because antennas generally have to
be separated from each other by a minimum distance in order to avoid interference, mast sharing usually requires taller (and more visually
disruptive) masts. Local planning authorities actually may prefer several small towers to one large one. More discrete (or disguised)
structures reduce visual intrusion, but cannot support more than one operator’s antenna.
Facilitating network deployment
Civil engineering costs can mount up when the number of building sites is relatively high in a mobile network roll-out. Moreover, operators
often run into practical difficulties in acquiring and developing adequate sites, obtaining the appropriate regulatory licences, and overcoming
public opposition to mobile towers.
Site sharing allows operators to reduce their capital and operating expenditures, bypass nettlesome planning and regulatory hassles, and
avoid potential environmental pitfalls. In short, site sharing can speed up network deployment and make it less expensive. Lower site-
development costs can pay dividends when they result in networks covering larger areas, increasing the likelihood of bringing wireless
services to sparsely populated rural areas – and at more affordable prices.
Upgrading technology from 2G to 3G
Infrastructure sharing can ease the transition from “second-generation” (2G) to “third-generation” (3G) mobile networks, by allowing
operators to collocate new 3G equipment on their existing towers and masts. This can help cut costs, even though 3G networks
commonly require significantly more sites. In the European Union, for example, 2G networks were deployed in the 900 megahertz (MHz)
spectrum band, while 3G licenses were assigned in the 1900-2100 MHz band. Because spectrum generally has a shorter range at higher
frequencies, 3G networks require more base stations (and therefore more sites) – a significant transition expense for 2G operators.
However, if those 2G operators can collocate 3G equipment on their existing 2G towers, they can enjoy significant savings as a result.
End Notes
[1] Current scientific evidence indicates that exposure to radiofrequency fields, such as those emitted by mobile phones and antennas, is
unlikely to have negative health effects. In response to health concerns raised by certain communities, the World Health Organization
(WHO) established a project to assess the scientific evidence of possible health effects of electromagnetic fields. See www.who.int/peh-
emf/en/index.html. The International Commission for Non-ionizing Radiation Protection (www.icnirp.de) has established guidelines for the
maximum level of radiofrequency levels in areas of public access from antennas and for users of mobile handsets.
[2] This principle is established in the European Union Directive [Framework directive], consideration 23: “Facility sharing can be of benefit
for town planning, public health or environmental reasons, and should be encouraged by national regulatory authorities on the basis of
voluntary agreements. In cases where undertakings are deprived of access to viable alternatives, compulsory facility or property sharing
may be appropriate. It covers inter alia: physical collocation and duct, building, mast, antenna or antenna system sharing. Compulsory
facility or property sharing should be imposed on undertakings only after full public consultation.”
[3] Today’s standard 3G equipment consumes about 4,000 KWh of Grey energy per year per node, which corresponds to 2.5 tons of CO2,
or the equivalent need of 120 trees per node to compensate for the environmental effect. In a developing country with no or little alternative
Green energy, network sharing can significantly reduce the environmental impact.
RELATED MATERIALS
Module 4, section 3.4.6, "Current and Emerging Form of Network Sharing"
6.5.2. When might mobile network sharing be appropriate?
Determining when mobile network sharing is appropriate requires consideration of both the anticipated benefits of such sharing and the
regulatory complexities associated with implementing the sharing arrangements.
Understanding the anticipated benefits of various forms of sharing helps to identify contexts where these forms of sharing will be
appropriate and most desirable. For example, national roaming arrangements are more important in a country that has remote or rural areas
that are un-served or under-served than in a smaller country with a largely urban population.
There are some benefits that apply to almost all forms of mobile network sharing. Network-sharing agreements generally benefit operators
and the general public from a cost perspective. They help operators avoid costs for building or upgrading redundant network sites and
allow them to gain additional revenue streams from leasing access. Operators also can achieve considerable savings in rent, maintenance
and transmission costs; these savings can be passed down to end-users through lower rates. Sharing arrangements may also achieve
economies of scale by combining operating and maintenance activities. Network sharing may also help operators to attain more efficient
coverage, since operators may choose to use only those sites that provide deeper and better coverage, decommissioning sites with poor
coverage possibilities. Operators can then reinvest those savings in upgrading their networks and providing better coverage and services
to end users.
Network-sharing agreements may also bring substantial environmental benefits, by reducing the number of sites and improving the
landscape. These types of benefits may be of particular relevance in contexts where tourism is a major industry in a country. In Caribbean
countries, for example, reducing visual disruption caused by mobile towers and antennae is important to maintaining the beautiful landscape
and beaches for which these countries are known.
The benefits of various forms of sharing must be weighed against the regulatory measures that will be necessary to facilitate these forms
of sharing. There are obstacles to overcome when dealing with network-sharing agreements. From an economic and practical point of
view, mobile network sharing is a complex process that requires substantial managerial resources. Some countries (often developing
economies) lack the regulatory resources and expertise necessary to address the complex issues that accompany some forms of mobile
network sharing such as Mobile Virtual Network Operators (MVNOs). Regulators need to analyze the potential benefits to be generated by
network sharing on a case-by-case basis, taking into account the specific characteristics of each market involved, the competition-related
concerns that may arise, and the relative difficulty associated with developing and implementing appropriate policies.
Generally speaking, network sharing is a useful tool for regulators and policy makers who want to encourage network deployment in un-
served or under-served areas. Several instruments can be used to promote network sharing. National roaming arrangements are probably
the most simple and effective arrangements. While roaming leads to a certain level of uniformity among operators’ offerings, this does not
necessarily restrict competition significantly. National regulatory authorities that have anti-competitive concerns may allow network sharing
for a limited period (for example, one or two years) in order to promote roll-out of initial phases of network deployment. After that,
operators could be required to provide coverage using their own networks.
Other types of arrangements, such as active infrastructure sharing, an open access model (allowing and promoting the entry of MVNOs)
and functional separation, may also work well to promote roll-out of wireless infrastructure and the advancement of competition. But these
types of arrangements may be difficult to monitor and regulate. Such measures require a strong regulator and an effective and efficient
judicial system, with appropriate enforcement powers.
When analyzing examples of network-sharing agreements around the world, regulators and policy makers should look at the way each
market has developed. For example, it is relevant to note that some network-sharing agreements in developed countries have presaged
later mergers between the companies involved. [1] In other cases, the companies involved in network sharing arrangements have not
merged with each other, but have seen consolidation take place in the market around them. [2] In the past few years, there has been a
large consolidation wave in the global telecommunication market across the globe. [3] Perhaps some of those operators could have
survived had they been allowed more freedom to share their infrastructure and to compete based on the quality of their service marketing
and delivery.
[1] This was the case with T-Mobile and Orange in the Netherlands and ATT and Cingular in the USA.
[2] After its network sharing arrangement with T-Mobile in the UK and in Germany, the company O2 was acquired by Telefonica.
[3] Worth mentioning are: the acquisition of Bell South by AT&T; the merger between Verizon en MCI; the acquisition of O2 by Telefonica of
Spain, the acquisition of Orange Netherlands by T-Mobile; the acquisition of a controlling interest in Telecom Italia by Telefonica.
6.5.3. Options for passive mobile network sharing
This section identifies the options available for mobile operators to share passive infrastructure elements of their wireless access
networks. The provider of the infrastructure can either be one of the operators or a separate entity set up to build and operate it, such as a
tower company. The passive infrastructure in a mobile network is composed mainly of:
■ Electrical or fibre optic cables;

■ Masts and pylons;
■ Physical space on the ground, towers, roof tops and other premises; and
■ Shelter and support cabinets, electrical power supply, air conditioning, alarm systems and other equipment.
A collection of passive network equipment in one structure for mobile
Passive Mobile Sharing:
telecommunications is generally called a “site.” Therefore, when one or more
Options Available in Site Sharing
operators agree to put their equipment on (or in) the same site, it is called “site
sharing” or “collocation.”
In site-sharing arrangements, operators might share space on the ground or on a

tower or rooftop. Depending on the location, operators could install antennas
directly on the structure (for example, a water tower or roof-top) or share a
mast. The next degree of cooperation would involve sharing support systems at
the site, such as power supply and air conditioning (often integrated in a site
support cabinet, or SSC). Telecommunication plant (antennas and transmission
equipment) is considered active infrastructure, which is discussed in
section 6.5.5. Operators often welcome site sharing as cost-effective, because
new sites can be costly, capital intensive, cumbersome to maintain and
environmentally risky.
Source: Telecom Regulatory Authority of India
(TRAI), Recommendations on Infrastructure
Sharing
6.5.4. Site Sharing Arrangements
There are several variations of site sharing arrangements. In terms of access, there are three basic types of site-sharing arrangements:
■ Unilateral – One operator agrees to provide access to its facilities to another operator;
■ Bilateral – Two operators agree to provide mutual access to facilities; or
■ Multilateral – Several operators agree on access terms.
Other variations include the number of sites involved in the arrangement: such agreements may pertain to just one site, or they could
provide a framework for access at multiple sites in a certain geographic region. Bilateral agreements for regional site sharing may be
particularly appealing for operators from an economic point of view. Sharing passive infrastructure in certain regions enables operators to
broaden service coverage over a larger geographical area, which is particularly attractive for operators subject to geographical coverage
obligations. Regional site sharing allows operators to save capital and operating expenditures and provides an alternative to roaming
arrangements.[1] Regulators should be careful, however, to head off collusion between operators, especially in highly concentrated
markets and where incumbent or dominant operators are negotiating bilateral sharing deals.
Standard terms and conditions in site sharing agreements

Objective of the agreement
Obligations of both parties
Term of the agreement

Applicable tariffs
Billing conditions
Service description
Implementation and coordination

Access to facilities and cooperation
Operations and maintenance
Subletting conditions (such as no subletting without the consent of the facility owner)
Term and termination
Penalties
Liability
Confidentiality
Representations and warranties
Amendments to agreement
Force majeure
Governing law and jurisdiction
Source : Camila Borba Lefèvre, Mobile Network Sharing, at: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html.
Most site-sharing agreements do not restrict competition between operators because operators generally retain independent control of
their respective networks and services. As a result, operators maintain different services and business plans, concentrating on their
different market niches. Full competition is assured where operators retain independent control over their radio planning and the freedom to
add sites, including non-shared sites. In that way, operators are free to increase their network capacity and coverage. Better coverage
and capacity can be a competitive advantage, allowing operators to distinguish themselves based on the quality, capacity and range of
their networks. It is therefore important that site-sharing agreements do not contain exclusivity clauses that would prohibit operators from
concluding similar deals with other parties.
Site sharing arrangements that fulfill these conditions are not likely to restrict competition among operators. [2] In fact, site-sharing
agreements may have a positive impact on competition, since the savings achieved may be passed on to consumers, increasing quality of
service and decreasing prices.
Finally, it is important to ensure that exchanges of information between site-sharing competitors are limited to information strictly necessary
for this purpose, such as technical information and location data for particular sites. Additional exchanges of confidential information should
be avoided in order to protect customers’ proprietary rights and privacy.
End Notes
[1] National roaming concerns a situation where the cooperating operators do not share any network elements as such but simply use
each other’s networks to provide services to their own customers. National roaming arrangements allow the roaming operator to rely
completely on the infrastructure of the operator providing national roaming, instead of building its own infrastructure in the roaming area.
According to certain competition authorities, national roaming agreements may restrict competition between the roaming operators.
[2] This was the opinion of the European Commission when it judged the site sharing arrangement for 3G mobile communications between
T-Mobile Deutschland and 02 Germany: Commission Decision of 16 July 2003, Case COMP/38.36.
6.5.5. Encouraging passive mobile sharing
In order to accelerate deployment of mobile networks to rural areas, policy makers and regulators may consider adopting regulatory
measures to promote site sharing. This section reviews measures that may be taken in order to encourage passive sharing in mobile
networks.
Mandatory sharing
In countries where mobile site sharing is mandatory, operators generally are required to allow third parties to share their facilities upon
request. The scope of mandatory site sharing may vary, depending on the needs of local markets and the policy objectives of the
government. For example, sharing obligations might only be imposed on operators that have significant market power (SMP) while site
sharing remains optional for non-dominant operators. The regulator might elect to require sharing of only certain infrastructure elements.
The scope of mandatory sharing may also vary based on the degree of independence operators have in negotiating the specific terms of
site sharing agreements.
For information about what regulators should consider when setting a policy for mandatory sharing, please see the Practice Note entitled
“Regulating Mandatory Site Sharing Agreements”. A link to this Practice Note is set out below.
Box 1: Site Sharing in New Zealand
Subject to limited exceptions, mobile site sharing is mandatory upon request in New Zealand, although access seekers and access
providers are free to set their own pricing arrangements for collocation. In 2007, an investigation conducted by the New Zealand
Commerce Commission (the Commission) found that although collocation agreements for mobile site sharing had been in place for many
years, collocation had occurred on less than 0.5% of available towers. The investigation further found that pricing for collocation
services was not the impediment to mobile collocation. Instead, the Commission considered that collocation had not occurred more
frequently because “incumbent operators had control over optimal co-location sites and incumbents had no or limited incentives to
support co-location (sic) by competing networks.” (Commerce Commission, Schedule 3 Investigation into Amending the Co-Location
Service on Cellular Mobile Transmission Sites, 14 December 2007, p. 9, para. 49.)
In response to these findings, the Commission undertook a process to determine the Standard Terms for collocation on mobile cellular
transmission sites. The Commission released its Determination on these Standard Terms (the “STD” or “Standard Terms Determination”)
in December, 2008. The Commission’s STD was aimed at enabling the efficient provision of mobile collocation services and at providing
access seekers and access providers with appropriate incentives to make efficient use of mobile network resources for the long-term
benefit of end-users. The Commission identified three aspects of the STD in particular that it considered will contribute to more rapid
collocation of mobile network transmission and reception equipment:
(1) the standard type site solution process;
(2) the ability for access seekers to make multi-site applications; and
(3) the Service Level capacity limit for each access provider of ten applications per access seeker per five day working period.
The Commission also stated that it would also monitor the implementation of the STD closely, given the limited progress that has been
made towards collocation. The Commission stated that it “will be carefully examining the Service Level performance reports, with
particular attention on the number of co-location (sic) Applications received and final approvals issued by Access Providers, as well
as Service Level defaults.” (Commerce Commission, Standard Terms Determination for the specified service Co-location on cellular
mobile transmission sites, Decision 661, December 11, 2008, p. xii, para. xxvii.)
Source: Commerce Commission, Standard Terms Determination for the specified service Co-location on cellular mobile transmission
sites, Decision 661, December 11, 2008.
Optional sharing
Policy objectives play an important role in deciding whether site sharing should be mandatory or optional. When the policy is geared toward
stimulating operators to invest in their own infrastructure, optional sharing may be applied.[1]
In many cases, operators may opt for site sharing voluntarily, in order to reduce costs. Regulators who wish to minimize market
interventions may take measures to stimulate site sharing without mandating it. Measures to promote site-sharing arrangements might
include: adopting model agreements; facilitating self-regulation; providing guidance on the types of sharing that is permitted; permitting the
sharing of government-owned facilities; and providing financial incentives for sharing. A Practice Note containing further information about
these options is linked below. This Practice Note is entitled “Measures to Promote Optional Passive Sharing”.
The role of local authorities
The ability to construct sites in certain areas may be limited by local land use or other regulatory restrictions. In most countries, local
authorities are involved in granting permission to install masts and towers for wireless communications. These city or township authorities
may take part in promoting site sharing, for example by requiring that new masts or towers be designed to accommodate more than one
operator. Local authorities may also require operators to place their equipment on existing masts, unless this is not possible for technical
reasons.
To promote site sharing successfully, local authorities must work closely with operators and their representatives. Disputes between
operators and local authorities can hinder network roll-outs and increase communities’ often-unnecessary fears of negative impacts from
tower sites. Associations of operators may be an important factor in establishing a dialogue between local communities and operators,
increasing awareness about the presence and the location of masts in their communities and any possible health or environmental effects
of those masts. Such associations may also develop, in cooperation with local communities, guidelines for building new sites or installing
new equipment at existing sites. The goal is to increase local participation, improve the availability of information and create legal certainty
for operators willing to roll out their networks.
It is also possible for national authorities to develop site-selection rules or guidelines to be followed by local authorities. These guidelines
could address environmental factors or distinguish between the conditions that apply in urban and rural, less-populated areas.
Encouraging the participation of infrastructure providers (tower companies)
Telecommunications operators that own towers, masts or other infrastructure may have incentives to prevent competitors from sharing
their sites. However, the outsourcing of tower operations may be an interesting option from a business perspective. Specialized tower
companies have every incentive to sell their services to as many telecommunications service providers as possible.
Tower companies own the towers and masts or, in some case, the whole site. They provide a variety of services to customers, such as:
radio and transmission planning; site acquisition; site construction and equipment installation; and/or site maintenance.
Outsourcing deals may create considerable financial value for operators and free them up to focus on their core wireless businesses.
Outsourcing of site infrastructure has been particularly successful in North America, where it is considered one of the key enablers of
effective mobile network roll-out.[2]
Although not as well known as in the United States, tower companies also are becoming more common in Europe.[3]
For information about how policy makers and regulators may promote the emergence of tower providers, please see the Practice Note
entitled “Tower Companies and the Promotion of Passive Sharing”.
Financial incentives
Sharing arrangements can often reduce costs and make wireless deployment more viable in many areas. In this regard, operators have an
incentive to pursue sharing arrangements. Governments can also implement various measures to create additional financial incentives.
Such measures could include tax and fee exemptions, the reduction of fees imposed by local authorities, and subsidies. The Practice Note
entitled “Encouraging Passive Sharing Using Financial Incentives” discusses these measures in greater detail. A link to this Practice Note is
set out below.
[1] Cf. Article 8 of the European Framework Directive, according to which one of the principle policy objectives of the European electronic
communications policy is to promote efficient investment in infrastructure.
[2] In the United States, the majority of mobile sites are owned by tower companies, and not by mobile operators. Among successful tower
companies in North America are: American Tower, Crown Castle and Spectra Site.
[3] The company Alticom B.V., which is a subsidiary of the French company TDF S.A., has recently entered the Dutch market in order to
operate a tower business for wireless transmission. TDF also operates a tower business in other European countries.
6.5.6. Active mobile network sharing
In addition to sharing passive infrastructure, operators may also share active elements of their wireless networks. Box 1 outlines the active
elements of a wireless network that can be shared. Operators may share those elements and keep using different parts of the spectrum
assigned to them. Although active infrastructure sharing is more complex, it is technically possible. Equipment manufacturers can supply
packages that have expressly been designed for active mobile sharing.
This section will describe several options available for active mobile sharing.
Box 1: Active elements of mobile networks Most of the examples used in this section focus on 3G network sharing, but the
that may be shared techniques could be applied to 2G mobile or broadband wireless networks, as
■ Antennas well. [1]
■ Antenna systems Active mobile sharing may not be permitted under the licensing regimes of some
■ Transmission systems countries. This is the case in India, for example, where the licensing regime for
■ Channel elements mobile telecommunications does not permit active sharing. Other regulatory
agencies may allow active sharing only with strict conditions, in the belief that
competing operators should utilize their own infrastructure independently.
However, there are indications that many authorities are reconsidering this as operators increasingly compete based on the price and
quality of their services and not on the features of their networks. The Telecommunications Regulatory Authority of India (TRAI) has
recommended a review of the existing licensing regime in India with respect to active infrastructure sharing.
This section will explore the different options available for operators that wish to share their 3G mobile networks. This information is also
relevant for sharing of networks based on other technologies, since their basic configuration is similar. Figure 1 provides a graphical
overview of the elements of a 3G mobile network that could be shared. These elements include:
■ The Node-B: This term refers to a base station placed next to an antenna. The Node-B contains equipment necessary to control the
transmission and reception of signals: power amplifiers, power supplies, air conditioning, support cabinets, alarm systems, the
transmission switch (in case of UMTS this is based on asynchronous transfer mode or “ATM” technology) and the TRX. Also
described as a transceiver, the TRX contains both a transmitter and a receiver and is responsible for sending and receiving signals at
the frequency assigned to the operator. The TRX is a very important device, because it enables communication with mobile handsets.
■ The Radio Network Controller (RNC): This piece of equipment controls the Node-B. One RNC is usually connected to several
Node-Bs (100 to 200). The RNC performs several important functions in a mobile access network, such as traffic and mobility
management. The RNC tracks where the subscribers of a mobile network are located and assigns them to the base station closest to
them. The RNC also controls the handoff of calls between cells.
Figure 1: Design of a Typical 3G Mobile Network

Source: Camila Borba Lefèvre, Mobile Network Sharing, at: www.itu.int/ITU-D/treg/Events/Seminars/GSR/GSR08/papers.html
■ The Core Network: This is the intelligent part of the network and includes the mobile switching centres (MSCs). In modern mobile
network architecture, the MSC is physically split into a mobile gateway (MG) and a mobile switching server (MSS). The MG switches
the traffic to and from the radio network and from external networks (the public switched telephone network and other mobile
operators). The MSS controls the traffic and customer services. In addition to the switching and control components, the core
network contains several databases, such as the subscriber data base or home location register (HLR), which identifies the
subscribers that are authorized to use the mobile network
Another element of the core network is the operations and maintenance centre (OMC). Part of the OMC controls the radio network
components, such as the RNCs and Node-Bs, and is responsible for traffic management on the network. In more intensive network-sharing
agreements, parties may have to share the OMC. This may raise competitive concerns, because operators would be able to access
information relating to competitors’ traffic and volume. Regulators may require operators to create an independent OMC to ensure that
information from the sharing parties remains separate. This would also allow independent network optimization, ensuring competitive
differentiation. Because of these regulatory issues, the OMC may be one of the elements not shared among operators.
[1] Many European mobile operators have contemplated active sharing of their 3G mobile networks. This was triggered, among other
reasons, by the high licence costs paid for 3G licences in Europe, the economic downturn that followed the 2000 3G-auctions and the
increasing doubt about the development of the UMTS-technology and the availability of adequate handsets. The need for a much larger
number of base stations for 3G, as compared with 2G, also made operators contemplate infrastructure sharing in order to reduce costs.
6.5.7. Active Mobile Network Sharing: Extended Site Sharing
Extended site sharing is when operators share not only the passive elements of a site, but also active equipment such as antennas,
combiners and transmission links. In extended site sharing arrangements, operators may also share the TRX (transmitter and receiver).
This requires parties to share the spectrum, as well. Spectrum sharing is technically possible, but it can raise regulatory challenges
because of rules dealing with spectrum optimization. Nevertheless, the Brazilian regulatory ANATEL has expressly allowed spectrum
sharing in case operators decide to share their networks in order to provide coverage in rural or remote areas (communities with less than
30,000 inhabitants).
The extended site-sharing option may increase capital and operating savings when compared with simple site sharing. The amount of
additional savings may be limited, however, because the additional costs of antennas and transmission equipment are relatively small.
While it is technically possible for operators using different frequencies to share an antenna, this option poses some technical challenges. It
may particularly difficult when radio optimization strategies are not aligned among the operators. While an antenna can be shared, it is
usually kept in a certain position, based on the operator’s own radio optimization strategy. The optimization strategy of the sharing operator
may require changing the antenna position – something that may be difficult or impossible when the antenna is shared.
Nevertheless, certain equipment manufacturers supply antennas that are adequate for antenna sharing.[1] This equipment includes Multi-
Operator Radio Access Network (MO-RAN) technology. Where MO-RANs are deployed they allow for two operators to share the same
base station and to share the radio network controller which directs the voice and data traffic back to the operator's own core network.
[1] Nokia, for example, is an equipment manufacturer and supplier that provides equipment intended for network sharing.
6.5.8 Active Mobile Network Sharing: Sharing the Radio Access Network
Rack sharing: In addition to the extended site–sharing option, operators can choose to install their active equipment in a shared cabinet or
rack (i.e., the housing frame encompassing the electronic and other hardware). In this rack-sharing option, other elements such as channel
elements, transmitter and receiver (TRXs) and power amplifiers remain physically separated, along with the transmission networks and
other elements of radio access, such as the Radio Network Controllers (RNCs). Power supply, air-conditioning, ancillary cabinet and alarm
installations can be shared.[1]
Depending on the actual situation, rack sharing may provide up to 5 per cent capital expenditure savings for an operator, per Node-B. If
battery backup is shared, this option may provide even further savings. Figure 1 graphically illustrates this option.
Figure 1: Rack Sharing

(i.e., sharing the physical frame housing electronics and other equipment and ancillary elements such as power supply, air-
conditioning, battery back-up, and alarm installation)
Full Radio Access Network (RAN) Sharing: In addition to extended site-sharing and rack sharing, operators may also share all the
elements of the Node-B. In case independent frequency control needs to be in place, the TRX and the power amplifier (PA) should remain
independent, allowing for radiation at each operator’s assigned frequency range. When the spectrum can be shared, operators may also
share the TRX and the PA.
Figure 2: Full RAN Sharing

In the case of full RAN sharing, regulatory authorities may require the shared elements to be functionally separated. In this scenario,
operators should retain independent control of all the parameters that determine the quality of the network, such as coverage, speed and
the handover parameters. The Dutch regulatory authority required functional separation when it assessed the proposed sharing
arrangement for 3G mobile services between operators Ben and Dutchtone.[2]
Functional separation implies that the communication between the RNC and the Node-B has to be under the independent control of one
operator, as far as that operator’s service is concerned. This communication may take place using the same cables and connections, but it
must be logically separated. In addition, operations, maintenance and network control should be separated. Those elements may be under
individual control by each operator or under joint control of an independent third party, which would operate the shared network on behalf
of the sharing parties. This independent third party could be a joint venture between the sharing parties or an independent company.
Figure3 illustrates a functionally separated RNC, which may be necessary due to regulatory requirements.
Figure 3: Separating the RNC Functions

[1] This situation may also be called “ancillary sharing”.
[2] See Decision of NMa of 11 October 2002, Case No. 2816/35.
6.5.9 Active Mobile Network Sharing: Core Network Sharing
Sharing of the core network is technically possible. However, core network sharing does raise some commercial concerns. The core
network performs several functions in areas essential to service performance, such as billing. The core network also contains a large
amount of confidential information on the operator’s business. These matters can complicate sharing the core network. There are,
however, other ways to share the core network, such as national roaming, or through a mobile virtual network operator (MVNO) structure.
In addition, with the emergence of so-called next-generation core networks, in which switching and the control/service functionalities are
physically separated, network sharing may move into the domain of core network switching while enabling service differentiation and
confidentiality.
National roaming
National roaming refers to an arrangement among operators to use each other’s networks to provide services in geographic areas where
they have no coverage. Such arrangements effectively multiply any one carrier’s ability to cover the entire country, without actually having
deploying infrastructure everywhere. There are several options for the geographic division of a country for the purpose of national
roaming. One common method is to assign different cities, provinces or regions to different operators. For more information on roaming
arrangements, please see the Practice Note, “National Roaming”. A link to this Practice Note is set out below.
Wholesale mobile access: Mobile virtual network operators
Broadly speaking, mobile virtual network operators (MVNOs) are entities that provide mobile services to end-user customers but do not
have their own radio spectrum; in some cases, MVNOs also do not have all the infrastructure necessary to provide mobile telephone
services. MVNOs offer services to their customers by reselling wholesale minutes that they have purchased from a mobile network
operator (MNO). Many MVNOs have backbone and back-office operations (including a billing and identification system) and only require use
of the mobile operator’s access (or “last mile”) network. These MVNOs thus avoid having to build out end-to-end mobile networks. One
form of MVNO, known as a “mobile service provider,” operates without any network facilities at all, simply buying and reselling minutes to
their end users. For more information on MVNOs, please see the Practice Note entitled “Mobile Virtual Network Operators”. A link to this
Practice Note is set out below.
6.5.10 Active Mobile Network Sharing: Backhaul Sharing
In certain rural or remote areas, backhaul networks can constitute bottleneck facilities. In areas where mobile traffic is low, the full
backhaul capacity may lie unused, making it ideal for sharing among different operators. Accordingly, where passive or active
infrastructure is shared on a tower or rooftop, operators may also share backhaul facilities. Backhaul sharing can be achieved either
through fibre cables or microwave links.
Regulatory or licensing conditions may preclude operators from sharing backhaul facilities, however, especially when spectrum is
employed (as with microwave links). This is often the case in rural or remote areas. For example, sharing of radio backhaul is not permitted
under the licensing conditions of mobile operators in India. When backhaul radio facilities cannot be shared, operators have to install
separate antennas on the towers, in addition to the antennas used to communicate with handsets. This increases the weight of the
antenna on the tower, requiring higher and heavier towers, increasing the cost of construction and adding to visual intrusion. Therefore, it
may be more practical to share fibre and limit the sharing of radio backhaul facilities to low-traffic regions. In addition, regulators and policy
makers intending to foster wireless broadband deployment may wish to encourage mobile operators to replace microwave links with fibre
links to add greater bandwidth.
6.5.11. Competition and Active Mobile Network Sharing
Active mobile sharing raises a number of competition-related concerns. The access that operators have to each other’s networks provides
them with access to confidential information about each other’s costs, operations, technology, and other key data. Sharing may therefore
provide opportunities for collusion on pricing, service packages, and network development.
Even without collusion, some forms of sharing such as national roaming may restrict competition by neutralizing key competition parameters
such as coverage, call quality and transmission rates. This neutralization occurs because roaming operators are restricted by the
coverage, network quality and transmission speeds available on the visited network, which are a function of the commercial choices made
by the visited operator. Regulators may be concerned at the resulting uniformity and lack of market differentiation. These types of concerns
were raised by the European Commission when it evaluated 3G network sharing arrangements between T-Mobile and O2 in Germany and
the UK. Although the European Commission restricted national roaming arrangements due to competition-related concerns, this decision
was subsequently overturned by the European Court of First Instance (CFI). For more details about the European Commission’s evaluation
of national roaming in the aforementioned cases, please see that attached Practice Note entitled “The European Commission’s Review of
National Roaming Agreements”.
Sharing may also hinder network roll-out since operators may find that it is cheaper to share infrastructure and provide services in the
same geographic areas as their competitors than to roll-out their network to under-served areas. Insofar as sharing increases collaboration
and reduces competition in the core and access levels of the network, sharing also reduces the incentive to innovate. Dynamic efficiency
may suffer as a result.
Box 1: Ofcom’s concerns regarding the impact of sharing on competition
“Network sharing could also have undesirable consequences for competition. For example, [mobile network operators] could
collaborate on network development and gain information about each other’s costs and plans, which may have a chilling effect on
competition in the retail market. Dynamic efficiency may also be lower with fewer networks able to provide high quality mobile
broadband services. End-to-end competition, i.e. at both the network and service level, could lead to greater innovation, which could
bring significant benefits for consumers. We note that the competition concerns would be amplified if the 900 MHz operators were
themselves to decide to share a single UMTS 900 network in response to the actions of their competitors. While it is difficult to quantify
the potential impact of these effects, Ofcom’s initial view is that there is a significant risk that both competitive intensity and innovation
in mobile broadband services would be weakened, with potentially serious impacts on consumer welfare.”
Source: Ofcom, Application of spectrum liberalization and trading to the mobile sector (20 September, 2007). This public
consultation document is available at: www.ofcom.org.uk/consult/condocs/liberalisation/liberalisation.pdf.
While the collaborative aspects of sharing pose concerns, sharing may also be used by operators, particularly those with significant
market power (SMP), to undermine their competitors. Refusals to grant access to network infrastructure, delays in responding to requests
for site sharing, poor service quality, and price gouging can sabotage initiatives to promote sharing where it would otherwise be
appropriate. Vertically integrated operators in particular have incentives to delay access and service requests, reduce service quality, and
over-charge operators who compete in the same downstream markets as they do.
With these concerns in mind, there are a number of ex ante measures that regulators may to avoid anti-competitive outcomes. These
measures include the following:
■ Impose geographic coverage requirements.

■ Set standards for quality of service indicators and time frames.
■ Require the publication of RIOs
■ Restrict exchange of confidential information.
■ Time limits: allow sharing in certain areas, for a certain amount of time. This allows for benefit of sharing for reducing costs so that
can help with roll-out, but phases it out as markets grow.
■ Functional separation.
Although sharing arrangements raise competition-related concerns, there may also be some competition-related benefits associated with
these arrangements. Network sharing agreements may help operators to offer services to more people in more geographic regions. The
operators can then compete based on brand, price and customer service. This applies in particular to rural and remote areas. One way to
balance the concerns about mobile network sharing with the benefits associated with sharing arrangements is to distinguish between
urban and rural areas when judging network-sharing agreements. Brazilian regulator ANATEL, for example, permitted mobile network
sharing in communities with less than 30,000 inhabitants, but not in other, larger communities, when it issued licences for the provision of
3G mobile services in 2008. Another option is to allow sharing for a period of time until operators have acquired a substantial customer
base in rural areas. Subsequently, the operators may be required to deploy their own networks and eliminate or reduce their reliance on
roaming.
Regulators need a thorough awareness of the competitive situation in the market when judging network-sharing agreements. Such
agreements should not affect important competition parameters, such as price and service packages. Cooperating operators should not be
allowed to exchange commercially sensitive information that may influence their future competitive behaviour. Where regulators impose
conditions or limitations (such as requiring part of the infrastructure to be functionally separate), they may wish to impose only those
obligations that are strictly necessary to preserve sustainable competition.
6.5.12 Regulatory and legal issues
Whether mobile sharing involves passive or active infrastructure, there are a number of key regulatory issues that must be addressed.
These issues are outlined below.
Licensing: Regulators and policy makers must assess the current licensing framework to ensure that it accommodates mobile network
sharing, whether in terms of active or passive infrastructure. In particular, regulators and policy makers should consider whether the terms
and conditions of existing licensees permit sharing and whether current categories of licences accommodate new types of industry
participants such as infrastructure providers (e.g., tower companies) and mobile virtual network operators. Other authorization-related
issues will also require considerations, such as fees and universal access obligations.
Mandatory or optional sharing: Regulators and policy makers must determine whether to mandate sharing or to permit it at the option
of the operator. A related issue is what types of sharing to permit and on what basis. A good example is spectrum sharing: although it is
technically possible to share spectrum, only the Brazilian regulator, ANATEL, has taken steps towards permitting it and then only in certain
limited contexts. By contrast, tower sharing is widely permissible around the world, and 3G network sharing is becoming more common. At
present, several countries including Jordan, Hong Kong China, and India permit MVNOs, though the terms under which MVNOs may operate
differ to quite a degree from jurisdiction to jurisdiction.
Standard terms and conditions: Regulators and policy makers may consider
adopting standard terms and conditions to govern sharing arrangements. Box 1: Commonly-recognized acceptable
Alternatively, they may opt to require dominant service providers to file grounds for refusing access and sharing
Reference Offers for sharing of essential facilities. One other option is to requests
stipulate that operators are free to negotiate their own terms and conditions for ■ Insufficient capacity (all existing capacity is
sharing on a commercial basis, and to require that operators enter into arbitration occupied or reserved).
or mediation if they cannot reach an agreement within a prescribed period of
■ Granting of access is technologically impossible.
time. The Bangladeshi Guidelines for Infrastructure Sharing, for example,
provide that parties must negotiate commercial agreements for infrastructure ■ The access or sharing request, if granted, would
sharing in good faith. However, if the parties are unable to come to an breach safety and reliability standards.
agreement, either party may ask the Bangladesh Telecommunication Regulatory
Commission (BTRC) for assistance. Any decision issued by the BTRC in such a case is final and binding. The Guidelines for Infrastructure
Sharing specifically state that disputes about tariffs or charges are to be resolved by the BTRC and that the BTRC's decision is final and
binding.
In most cases, it is preferable to allow operators to negotiate their own arrangements on a commercial basis. Regulators may safeguard
against anti-competitive behaviour in this context by setting some general principles for agreement and by establishing time deadlines for
completing agreements. General principles that regulators may consider adopting include:
■ Non-discrimination;
■ Provision of access and capacity on a “first-come, first-served” basis;
■ A requirement to return excess capacity and penalties for operators who order too much capacity;
■ Prohibitions on exclusivity arrangements;
■ Transparency; and
■ Acceptable grounds for refusing access and sharing requests.
Incentive regulation: Regulators and policy makers may consider implementing various measures designed to create incentives for
sharing generally and/or for complying with the terms and requirements surrounding sharing. One common approach is to require that all
operators keep a log of access requests, along with notes about the measures taken to respond to these requests. Operators may be
required to file these logs with the regulator on a regular basis or may simply be required to keep these records and to produce the logs
upon request. [1] Another measure that may be taken to enhance the efficiency of processing access requests is to allow operators to file
more than one access request with an operator at one time. In New Zealand, for example, the Commerce Commission approved a set of
standard terms for mobile collocation that allows access seekers to make up to ten requests to an access provider at a time.
Pricing: Regulators and policy makers must determine whether to set prices for sharing or to allow operators to negotiate prices, subject
to the general requirement that prices be fair and commercially reasonable. This decision typically turns on the infrastructure in question
and on the parties involved. Pricing controls are necessary where prices are unlikely to be fair or where pricing may be used to impose a
barrier to market entry. Thus, while it may be necessary to regulate the prices charged for access to essential facilities owned and
operated dominant service providers, it may not be necessary to set prices in other contexts where competition is healthier, such as tower
sharing. Where prices are controlled, the adoption of some form of cost-based methodology is generally considered the best practice.
Dominance or Significant Market Power (SMP): Policies to promote ICT sector development and competition require a comprehensive
framework for preventing and managing anti-competitive conduct. Regulators must prevent any abuse of dominance in cases where the
infrastructure owner also competes downstream with other service providers in the same market. Such policies should be premised on the
understanding that regulatory intervention is primarily needed only when an operator possesses Significant Market Power (SMP) and
begins abusing that power. The standard best practices for intervention apply in infrastructure sharing regulation as they would in pricing,
interconnection and other areas of competition policy.
Enforcement and dispute resolution: Whether governments choose a policy of mandatory or optional sharing, efficient enforcement
mechanisms and dispute resolution processes should be in place. It is not possible to prescribe the exact enforcement measures that
regulators should take; each country has its own institutions and laws. The common thread, however, is a solid and effective mechanism
for handling complaints and disputes, with sufficient sanctions to discourage violations. See the Practice Note entitled “Dispute Resolution
for Mobile Sharing” for more information about dispute resolution.
[1] See Booz Allen Hamilton Inc, "Telecom Infrastructure Sharing – Regulatory Enablers and Economic Benefits", November 2007, p. 8
available online at: www.boozallen.com/media/file/Telecom_Infrastructure_Sharing.pdf.
This section provides a set of best practices related to passive and active mobile infrastructure sharing.
■ Establish clear, objective and transparent policy goals involving network sharing.
■ Provide guidance on the types of sharing that will be allowed.
■ Establish clear guidelines for the conclusion of voluntary sharing agreements, including maximum time periods to conclude agreements
and to provide actual access.
■ Create efficient dispute settlement and judicial review mechanisms, including specialized dispute settlement bodies.
■ Allow and stimulate self-regulation.
■ Promote dialogue between authorities, community members and operators about the installation of infrastructure.
■ Allow network sharing, in particular site sharing and national roaming, in rural and remote areas.
■ Make a thorough and objective assessment of the competitive situation, including research on consumer preference and consumer
choice.
■ Only impose restrictions that are strictly necessary to promote competition and proportionate to the established policy goals.
■ Consider whether an open access model (such as the entry of MVNOs) or even functional separation would be viable, depending on
the market situation.
■ Consider providing subsidies for network sharing in remote or rural areas that are calculated to cover real costs and can be
distributed in a competitive fashion.
■ Monitor compliance with sharing requirements by requiring operators to provide regular reports and by creating an open dialogue with
operators.
■ Consider establishing websites publicizing the location of towers and other sites that may be suitable for infrastructure sharing.
6.6 Regulatory Issues Related to Sharing
Note: In addition to research done by Theresa Miedema, this section is based in part on two discussion papers prepared for the 8th ITU
Global Symposium for Regulators [1] 2008. The first paper, entitled Extending Open Access to National Fibre Backbones in Developing
Countries, was prepared by Dr Tracy Cohen and Russell Southwood and the second one, entitled Mobile Network Sharing, was prepared
by Camila Borba Lefèvre.
This section discusses various regulatory issues that are related to infrastructure sharing. Topics canvassed in this section include:
■ Selecting the appropriate approach to sharing;

■ National infrastructure sharing policies;
■ Licensing;
■ Pricing;
■ Competition and sharing; and
■ Regional initiatives.
6.6.1 Selecting an appropriate approach to sharing
There are a number of factors relevant to determining what type of sharing is most appropriate under different circumstances. Some of the
factors include market dynamics, policy objectives and concerns, and regulatory capacity. Each of these factors is discussed below.
Market dynamics
Different approaches are needed for different market dynamics. If multiple existing or potential players want to roll-out a network, then a
facilitating agency may succeed by offering passive infrastructure assets like rights of way and government land for sites. In developing
countries, however, the main challenge is often finding operators willing to go into under-served areas at all. The common view is that
network roll-outs do not make commercial sense in remote, thinly populated areas. Where nobody appears interested in entering such
markets, the government might need to take the primary risk by encouraging investment in a wider national network and then devising a fair
and efficient mechanism to share this resource with existing market players. The issue for governments is whether to duplicate elements
of existing networks or to “fill in” gaps in those networks. The latter might easily be seen as a suitable task for universal access funding,
but experience in some regions (Africa, for example) has shown that these funds often do not produce speedy network roll-outs.
simply do not allow for commercial evaluation to mature into investment. In such cases, the development of a national broadband strategy
may offer an opportunity to advance the move to full liberalization.
Box 1: Dominant form of infrastructure used to provide basic services and sharing arrangements to facilitate entry into market for basic
services
Fixed Mobile
■ Interconnection ■ Tower sharing
■ Local loop unbundling ■ Mobile site sharing
■ Collocation ■ Mobile Backhaul
■ Infrastructure sharing ■ National roaming
■ MVNO
Consideration should also be given to the type of infrastructure associated with potential sharing arrangements. Sharing, however limited in
its form and scope, is often necessary to introduce competition in the market in which the dominant infrastructure providers are involved.
Usually, the dominant infrastructure providers are engaged in offering fixed services. Stimulating competition in various markets for fixed
services may therefore involve interconnection and unbundling, including the unbundling of the local loop in more developed cases. In some
contexts, particularly in Africa, mobile service providers are fast becoming or have already become the dominant infrastructure providers.
Accordingly, various forms of mobile network sharing (e.g., tower sharing or even mobile virtual network operations) may be appropriate to
facilitate faster network deployment, in addition to measures such as interconnection requirements.
Primary policy concerns
It is prudent to have a clear sense of the policy objectives that sharing arrangements are meant to facilitate, as well as the policy concerns
that are most relevant in a given market context. Different forms of sharing serve different policy objectives. For example, tower sharing
may be an appropriate sharing arrangement if a key policy concern is rapid and efficient network deployment to rural or remote areas that
are under-served or un-served, whereas interconnection and local loop unbundling are aimed more at introducing competition to
established markets. Table 2 sets out what types of sharing arrangements are often appropriate in relation to particular policy objectives in
different market contexts.
Table 2: Policy objectives and sharing arrangements
Policy objective Market context Sharing Arrangements

Stimulate rapid and efficient network Un-served or under-served rural or -Tower sharing;
deployment remote areas
-Extended site sharing;
-Share rights of way;
-Mobile backhaul;
-Roaming.
If investment in infrastructure is a countervailing policy
concern, consider adding a sunset clause to period in which
sharing arrangements are permissible.
Introduce competition to domestic and Mature ICT sector with fairly -Interconnection;
international long distance markets and/or extensive existing infrastructure
-Collocation;
private line leasing markets
-Sharing of infrastructure such as duct access, trenches, and
rights of way;
-Infrastructure sharing (e.g., permit resale of capacity on
leased lines);
-International gateway sharing.
Introduce competition to domestic and Emerging ICT sector where -Interconnection;

international long distance markets and/or investment in infrastructure
-Collocation;
private line leasing markets 1 remains a critical issue
rights of way.
Introduce competition to local services Mature and converged ICT sector -Full local loop unbundling;
market
-Interconnection;
-Collocation and mobile site sharing;
rights of way;
-MVNOs.
Promote roll-out of FTTH or FTTB and Mature and converged ICT sector -Open access models;
general roll-out of broadband services
-Construction of national fibre core network;
- Interconnection;
-Collocation;
- Sharing of infrastructure such as duct access, trenches, and
rights of way;
-Backhaul.
Promote competition, with a particular Vertically integrated backbone -Interconnection;

concern for ensuring differentiation in service operator competing in
-Local loop unbundling;
levels and pricing downstream markets
-Collocation;
-Mandated access to bottleneck facilities.
Promote competition, with a particular Backbone operator is not active -Backhaul services;
concern for ensuring differentiation in service in downstream markets
-National roaming;
levels and pricing
-Core network infrastructure sharing/core network open
access models;
-Collocation.
Regulatory capacity
Sharing arrangements often raise complex competition-related issues. An important consideration related to selecting appropriate sharing
mechanisms for a particular market is therefore the capacity of the regulator to implement and monitor the sharing arrangements. At the
same time, while some regulatory issues associated with implementing forms of sharing are complex, there is ample international
experience to draw upon. Regulators may capitalize on this international experience and draw upon relevant “best practices” to facilitate
the implementation of various forms of sharing. The 2008 GSR Best Practices Guidelines on innovative sharing strategies is a useful
resource for regulators in this regard. This document is linked below.
Table 3 sets outs some considerations related to the regulatory capacity necessary to implement different types of sharing arrangements.
Table 3: Sharing and Regulatory Capacity
Sharing Arrangement Regulatory Capacity Considerations

Shared duct access, - Relatively light regulatory burden, and may be facilitated by statutory provisions, but may involve some
trenches, and rights of way monitoring of operators to prevent anti-competitive conduct in provisioning of services;
- Ample international experience from developed countries exists.
Collocation and site - Relatively light regulatory burden (often possible to allow operators to negotiate arrangements on a
sharing, including tower commercial basis), but may involve some monitoring of operators to prevent anti-competitive conduct in
sharing provisioning of services;
- Ample international experience.
Participation of - Relatively light regulatory burden, though authorisation framework is likely to be engaged;
infrastructure providers
- Growing international experience.
Interconnection - Some complex regulatory issues involved (e.g., pricing);

- Ample international experience in both developed and developing country contexts.
Local loop unbundling - Moderately high regulatory burden, involving complex issues, e.g., pricing and QoS-related matters;
- Considerable international experience in developed countries, but little experience in emerging markets.
National roaming - Moderately high regulatory burden involving several complex issues, e.g., pricing, QoS related issues,
concerns about uniformity of mobile services and pricing, and reduced incentives to invest in infrastructure;
- Growing international experience in developed countries, less so in developing countries.
MVNOs - High regulatory burden involving several complex issues, e.g., pricing, access to numbering resources,
interconnection rights, concerns about uniformity of mobile services and pricing, and reduced incentives to
invest in infrastructure;
- Early stages of international experience, with many lessons still to be learned.
Sharing of active - High regulatory burden involving several complex issues, e.g., pricing and risk premiums, balancing policy
infrastructure, e.g., lit fibre goals of stimulating competition and promoting investment in infrastructure, monitoring operators’ conduct to
prevent anti-competitive conduct, and ensuring fair access;
- Earlier stages of international experience, with lessons still to be learned.
Open access fibre core - Very high regulatory burden, involving numerous complex issues such as pricing, public funding (public-private
networks partnerships), functional, operational, or accounting separations, service level and QoS matters, as well as
broader national broadband strategies and Next-Generation Network strategies;
- Very early stages of international experience with many lessons yet to be learned.
6.6.2 National Infrastructure Sharing Policies
A growing number of countries are developing or have already developed national infrastructure sharing policies. These policies generally
set out the approach that will be taken to infrastructure sharing in the country, as well as the regulatory measures that will be adopted to
implement sharing. For example, a national infrastructure sharing policy may identify which network elements may be shared, the
conditions under which sharing should occur, various terms and conditions of sharing arrangements, including the basic approach to
pricing, and the regulatory measures, such as changes to the licensing regime, aimed at facilitating sharing. The Bangladeshi Guidelines for
Infrastructure Sharing, for instance, identify the primary policy reasons for mandating infrastructure sharing, identify the infrastructure that
may be shared, set out the types of sharing that are mandatory, outline a procedure for making sharing requests, set out certain
requirements necessary to facilitating sharing, and clarify what operators should do in the event of a dispute.
The development of a national infrastructure sharing policy allows operators, policy makers, regulators, and other stakeholders to engage
with issues related to sharing, network deployment, competition-related concerns and innovation-related concerns in a comprehensive
manner. The approach taken to sharing will depend in large on the state of ICT market development in the country. Since different market
dynamics require different approaches, sharing policies that are appropriate in an industrialized country may not be appropriate in a
developing country, and vice versa. In each case, a careful balance must be struck between the benefits associated with sharing, such as
greater efficiency, faster network deployment, reduced operating and capital expenditures, and lower end-user prices, with the potential
risks of sharing, such as anti-competitive conduct and disincentives to innovate.
Box 1: Developing a national infrastructure sharing policy
Before framing a national infrastructure sharing policy, officials should consider two important questions:
(1) whether the creation of shared infrastructure promotes competition and lowers prices for consumers, and
(2) whether such sharing arrangements encourage operators to make technical innovations that lower costs and improve services.
National infrastructure sharing policies also offer an opportunity for government to introduce wider measures designed to promote
cooperation between different sectors. Operators of national fibre backbone networks may fruitfully work with hydro companies, electric
companies, oil and gas pipeline operators, and railways to facilitate the laying of fibre, for example. Moreover, there is growing recognition
that utility companies such as hydro-electric, gas, water and sewer, and ICT service providers, should coordinate their efforts to construct
trenches and conduits. This type of collaboration not only lowers the costs for these companies, but it may also reduce public disruption
and aggravation associated with the construction of trenches and conduits on public roadways and other places.
6.6.3 Licensing
There are two issues relevant to licensing practices and implementing sharing. First, in order to facilitate sharing, it may be necessary to
revisit licensing terms and conditions to ensure that sharing is permitted. Second, a number of countries have established specific licences
for the provision of infrastructure such as towers. Both of these issues are discussed below.
Permitting sharing in the licensing regime
In countries where licences are required to provide network services, governments and regulators may have to review licensing terms and
conditions in order to ensure that operators and service providers are permitted to share infrastructure. Regulators should also review the
universal service obligations of licensees to ensure that licensees can share infrastructure as they work to rollout the network. The
Telecommunications Regulatory Authority (TRA) in Lebanon, for example, plans to encourage infrastructure sharing to facilitate the
fulfillment of licensees’ rollout obligations for future broadband access networks. Similarly, Lebanon’s draft mobile licences permit the
licensees to construct, maintain and operate mobile networks “whether alone or with other providers”.[1]
Similarly, in Canada, the agency responsible for licensing radiocommunication carriers, Industry Canada, imposed mandatory site sharing
requirements and mandatory roaming requirements on all radiocommunication carrier licensees. Industry Canada has also adopted a policy
that forbids licensees from entering into exclusive site sharing arrangements. For more information about these licence conditions, please
see the Practice Notes entitled “Canada – Mandatory Roaming Conditions for Radiocommunication Carrier Licensees” and “Canada –
Mandatory Antenna Tower and Site Sharing Conditions for Radiocommunication Carrier Licensees”. Links to these Practice Notes are set
out below.
Infrastructure licensing
In some countries, governments have created specific licences for the provision of infrastructure. Infrastructure licensees typically do not
provide network services. Instead, they provide network infrastructure such as towers and sites and manage these infrastructure
elements for operators. The licensing regimes of Nigeria, India, and Pakistan include infrastructure provider licences, for example.
In other countries, the provision of infrastructure or network access may be encompassed by the general authorization framework or the
unified/multiple service authorization regime. In Singapore, for example, Facilities-Based Operator (FBO) Licensees are authorised to deploy
and/or operate any form of telecommunications network, systems, or facilities that is used by any person to provide telecommunications
and/or broadcasting services to third parties. These third parties may include other licensed telecommunications operators, business
customers, or the general public.
Similarly, in Uganda, a Public Infrastructure Provider (PIP) Licence authorises licensees to establish, operate, and maintain infrastructure for
the provision of communications services to the public (if properly licensed to do so) and/or to offer infrastructure commercially for use by
Public Service Provider (PSP) Licensees. In order to provide services to the public, a PIP Licensee must also hold a PSP Licence. PSP
Licensees are authorized to provide voice and data services to the public. PSP Licensees must use the capacity or infrastructure of a PIP
Licensee. Alternatively, as indicated, an operator may hold both a PIP and a PSP Licence.
In some cases, more extensive legislative changes are necessary in order to implement a strategy for sharing and network rollout. In South
Africa, for example, the creation of a new state-operated company called “Infraco” to operate the national fibre network assets of two
state corporations required an Act of Parliament and an amendment to existing sector legislation.[2]
[1] TRA, Statement on Innovative Infrastructure Sharing and Open Access Strategies, at www.tra.gov.lb
[2] Infraco was been established through an Act of Parliament and by way of an amendment to the sector legislation with the stated aim of
empowering government to make strategic interventions on infrastructure investments whenever it deems it necessary. By doing so,
government intends to enable itself to initiate government intervention to ensure strategic ICT infrastructure investment and thereby more
easily be able to address the challenges of reducing the cost to communicate by providing infrastructure to other operators at wholesale
rates; improve on government service delivery and support the country’s economic policy.
6.6.4 Pricing
There is a general consensus that infrastructure sharing should be based on cost-oriented pricing and open access models. Countries
have differed, however, on the approach taken to establishing costs. The Jordanian regulator, the Telecommunications Regulatory
Authority, for example has recommended the adoption of incremental costing methodologies such as long-run incremental costs (LRIC). In
New Zealand, the Commerce Commission has approved the use of international benchmarking as a means of setting cost-oriented pricing
for unbundled bitstream access backhaul services. For an overview of the different approaches that various countries have taken to
pricing infrastructure sharing, see Table 1 below.
In most countries, pricing controls are only imposed on operators that exercise Significant Market Power. Non-dominant operators are
usually free to set their own prices for access to their network infrastructure, though they are often subject to a general requirement to
adopt fair and reasonable prices and to respond to requests for access in a timely manner. This asymmetric regulation of pricing on
infrastructure sharing reflects the underlying reason for adopting pricing controls: to prevent anti-competitive conduct such as the abuse of
dominance and vertical price squeeze practices.
Table 1: Select country examples of costing approaches to infrastructure sharing
European Union (Draft Mandated access to existing network Historical costs minus depreciation. Pricing should take into
recommendations) infrastructure that are not active (e.g., ducts, account the existing physical capacity, extent of depreciation of
civil engineering works, and ancillary services) existing facilities plus the operating costs of the efficient operator.*
*
European Union (Draft Mandated access to new network infrastructure Costs plus a project –specific risk premium to be included in the
recommendations) that are not active (e.g., ducts and civil costs of capital for the investment risk incurred by the operator.
engineering works) (Greenfield projects)*
The Capital Asset Pricing Model is identified as an adequate
instrument to calculate the return that is permitted on equity
capital to reflect the risk premium.*
Jordan Collocation, infrastructure sharing, bit stream Cost-based using Total Service Long Run Incremental Cost
unbundling, international gateway access “Plus” (TSLRIC+)
services
New Zealand Unbundled bitstream access backhaul Benchmarking against prices for similar services in comparable
provided by incumbent countries that use a forward-looking cost-based pricing model.
Pakistan Infrastructure sharing provided by operator No regulation of prices, subject to a general requirement that
without Significant Market Power prices be “fair and reasonable”
Singapore Connection Services for Submarine Cable Forward-Looking Economic Costs (FLEC) using Long Run Average
Landing Stations provided by dominant Incremental Cost (LRAIC) methodology
operators
Saudi Arabia Access services (e.g., local loop unbundling Regulator bases pricing on dominant service provider’s Long Run
and bit stream access) Incremental Costs (LRIC) and input/output data for Access Services
and other appropriate benchmarking data
* indicates a draft decision or recommendation that has not yet been formally adopted.
Source: Theresa Miedema, based on various regulators’ decisions, tariffs, and guidelines.
The deployment of new network infrastructure (Greenfield Projects) by a dominant operator requires particular consideration when setting
pricing controls for infrastructure sharing. This issue has become particularly salient with the introduction of Next-Generation Networks
(NGNs) and the roll-out of broadband fibre networks to homes and businesses (FTTH and FTTB). The pricing for access to new
infrastructure should include costs plus a risk-related premium. The risk-related premium provides a return on capital investment designed
to provide appropriate incentives for network investments. Prices should be set with a view to allowing operators to recoup their
investments and to continue their growth strategies. At the same time, prices should be low enough to facilitate market entry and to
stimulate competition. The regulator’s challenge is to find the balance between setting prices low enough to reduce barriers to market entry
and to maximise efficiency gains from sharing, but high enough to encourage investment in network infrastructure. See Box 1 for the
European Commission’s statement of this challenge with respect to the pricing of infrastructure sharing in the context of Next-Generation
Access networks (NGA networks).
Setting the appropriate rate
Box 1: Striking the balance between competition and investment in the pricing for of return on new
infrastructure sharing in the context of NGA networks infrastructure may be a
In its draft Recommendation on Regulated Access to Next-Generation Access Networks (NGA contentious issue. When the
networks), the European Commission made the following comment about the balance that must be UK incumbent operator, BT
struck when setting prices for infrastructure sharing: “The return that is allowed ex ante on equity agreed to implement
capital to finance NGA networks should strike a balance between providing adequate incentives for functional separation in its
companies to invest (implying a sufficiently high rate of return), while at the same time promoting network operations, part of
efficiency and sustainable competition and maximising consumer benefits (implying a rate of return that the arrangements with
is not excessive).” Ofcom, the U.K. regulator,
involved setting a return on
Source: Commission of the European Communities, Draft Commission Recommendation on regulated the new wholesale
access to Next Generation Access Networks (NGA), Annex I “Pricing principles for duct usage and the company’s capital
usage of other civil engineering works and other elements which are not active”. investments. In return for
agreeing to greater
Note that this document is a draft recommendation. At the time of publication, the final report had not functional separation in its
yet been released. operations through the
creation of BT Open Reach,
BT was permitted to have a
ten percent investment
return on Open Reach’s network assets. In January 2007, however, BT began arguing that it had not managed to achieve anything like this
rate of return. Ofcom noted that it was “aware of BT’s concerns with respect to revenue and profit level in the future.” As this example
illustrates, setting the terms for access to a dominant operator’s new network infrastructure will generally setting an appropriate rate of
return and may require re-visiting this rate of return.
As a final note, general pricing principles should be extended to setting the terms and conditions for the pricing of shared infrastructure.
For example, existing network elements should be unbundled as much as possible and priced separately so that access-seeking operators
are only required to pay for those elements that they actually require. Principles of non-discrimination should also be applicable. For more
information on pricing generally, see section 5 of this Module.
6.6.5 Competition and sharing
Competition-related issues are a major theme in discussions surrounding sharing initiatives. There are three main dimensions to the
relationship between sharing and competition:
■ First, the capacity of sharing to enhance competition;

■ Second, the risk that sharing will undermine competition; and
■ Third, anti-competitive conduct that can impede sharing.
The capacity of sharing to enhance competition
There are three ways that sharing promotes competition. First, by reducing the cost of network deployment, sharing mitigates one the
biggest impediments to market entry and makes it possible for more players to enter into the ICT sector. Second, sharing offers a means of
addressing barriers to competition, such as the control of bottleneck facilities by dominant operators. Finally, sharing has the potential to
create new types of players in the market who introduce new dynamics to traditional forms of offering services. For example, sharing has
given rise to a new type of core and access competitor, namely, infrastructure providers. Sharing has also introduced new types of
service providers who compete with more traditional service providers in various markets. An example is Mobile Virtual Network Operators
(MVNOs).
In general, it is preferable to take a “light touch” regulatory approach when considering measures that may be adopted to enhance
competition through sharing. Thus, there is a preference for measures that facilitate and encourage sharing, but that do not mandate
actions unless necessary (e.g., in the cases of operators with Significant Market Power or “SMP”). Some of the measures that regulators
and policy makers may take to promote the positive benefits of sharing for competition include:
■ Incentive-based regulations to encourage non-dominant operators to share infrastructure with each other;
■ Removing restrictions to sharing that may exist in the terms and conditions of authorizations;
■ Maintaining central registries of information about certain types of infrastructure that is capable of being shared (e.g., antennae
registries or tower logs) so that new entrants can better plan their business strategy and network deployment;
■ Working with industry groups to formulate recommended “best practices” for various forms of sharing;
■ Mandating sharing of essential facilities that are owned or operated by dominant service providers/operators with SMP; and
■ Ensuring that the authorisation framework accommodates new types of players like infrastructure players and MVNOs.
The risk that sharing will undermine competition
Some of the same factors that create the potential for sharing to enhance competition also carry a risk that sharing will undermine
competition. The ability of new entrants to enter the market without extensive network infrastructure of their own raises concerns about
the health of infrastructure-based competition in the sector. The concern is that there will be inadequate incentives for operators to invest
in infrastructure. Moreover, infrastructure owners, operators, and providers may lack sufficient incentives to continue to invest in new
infrastructure if they are required to share access to this infrastructure at rates that do not adequately compensate them for the risk
involved in deploying new infrastructure.
In addition to concerns about stagnation in infrastructure-based competition, there is a risk that sharing will occur in a manner that allows
operators to engage in anti-competitive conduct. For example, dominant operators may provide access to its bottleneck facilities to other
operators, but on terms and pricing that prejudice the access-seeking operators. Dominant operators may use sharing to engage in price
squeeze practices. Operators may also use sharing to engage in anti-competitive conduct such as collusion. Sharing raises issues about
the access that competitors may have to each other’s commercially-sensitive information, including, for example, plans for expansion,
operations data, and lists of customers.
While the concerns that sharing may undermine competition are legitimate, these concerns are not unique to sharing. Regulators and policy
makers have experience in addressing the types of issues that are raised by sharing. Thus, it is not necessary to “re-invent the wheel”
when considering how to manage the risk that sharing may undermine competition. Some of the measures that may be adopted to address
this risk include:
■ Ensure that operators that provide access to their infrastructure are permitted to earn a large enough return on their capital investment
in order to continue to give them incentives to build out infrastructure;
■ Require that operators who enter into sharing agreements file the agreements with the regulator;
■ Require operators sharing facilities to take measures to safeguard the confidentiality of their information;
■ Allow sharing subject to a “sunset clause” where appropriate: if there is a concern that infrastructure sharing will inhibit competition,
but there is a need to facilitate network deployment in the short-term, allow sharing (e.g., network roaming or tower sharing) for a
limited time period;
■ Monitor activity in the sector and enforce competition regulations; and
■ As a measure of last resort, require functional separation for vertically-integrated operators with SMP that provide access to
bottleneck facilities to operators who compete with them in downstream markets.
Anti-competitive conduct that can impede sharing
Where sharing is permitted, concerns may arise that operators will engage in various forms of anti-competitive behaviour to make it difficult
for other operators to access their facilities. Common forms of anti-competitive conduct that can impede sharing include:
■ Delays in responding to requests for access;

■ Refusals to provide access;
■ Refusals to provide information or delay in providing information about infrastructure available for sharing;
■ In the case of access to bottleneck facilities owned or operated by a vertically-integrated operator with SMP, price squeeze practices;
■ Providing access-seeking operators with services at inflated rates or with a poor quality of service; and
■ Demands for unreasonable access rates.
These types of conduct are not specific to the context of sharing and access to infrastructure. There are a variety of well-established
regulatory measures and tactics to address such conduct. Examples of regulatory approaches and measures that may be employed to
deter or to address anti-competitive conduct in the provision of sharing include:
■ Require and enforce the principle of non-discrimination in the provision of access services so that operators are not permitted to offer
better rates or levels of services to themselves and their affiliated companies than they do to unrelated operators;
■ Require transparency in the terms and conditions for access and sharing, particularly in the case of operators with SMP (operators
may be required, for example, to post a reference offer on their websites or to file such an offer with the regulator, or operators that
enter into a sharing or access agreement may be required to file a copy of the agreement with the regulator);
■ Adopt a “first come, first served” policy for allocating capacity or responding to sharing requests;
■ Establish penalties for operators that reserve capacity on networks (including their own) that far exceeds their demonstrated needs
(e.g., fine operators that reserve capacity and then fail to use a set threshold of this capacity within a certain period of time);
■ Establish base level service agreements and rates for sharing and access services;
■ Allow competitors to negotiate sharing arrangements on a commercial basis subject to the proviso that either party to a negotiation
may seek binding arbitration if the parties are unable to reach an agreement on the terms and conditions within a certain period of time;
■ Establish time deadlines for replying to requests for information or access;
■ Establish ex ante reasons that will be deemed legitimate and acceptable for refusing to provide access or sharing upon request, along
with guidance about what kind of evidence will satisfy the regulator about the legitimacy of a refusal;
■ Where necessary, mandate functional and/or accounting separations;
■ Require operators to maintain a log of all requests for sharing or access and require operators either to file annual reports on the log
or to maintain accurate records that may be examined or audited if a complaint is filed;
■ Streamline a process for lodging complaints with the regulator concern anti-competitive conduct; and
■ Adopt efficient dispute resolution procedures.
6.6.6 Regional Initiatives
While sharing is often viewed as a local or national issue, there is scope for regional sharing initiatives. The West African community, for
example, has adopted a regional regulatory approach to harmonizing the ICT sector that encourages infrastructure sharing. In January
2007, the Economic Community of West African States (ECOWAS) heads of state and governments adopted the Supplementary Acts that
cover ICT policy, the legal regime, interconnection, numbering, spectrum management, and universal access. The Act on Access and
Interconnection states in Article 10, point 2, that “National Regulatory Authorities shall encourage infrastructure sharing between
incumbents and new entrants concerning, in particular, posts, ducts and elevated points, to be made available mutually on a commercial
basis, in particular where there is limited access to such resources through natural or structural obstacles.”[1]
In addition to harmonizing regulatory approaches, various countries may work together to share access to major international gateways
such as submarine cables or satellite services. There are several regional initiatives in Africa at present involving shared submarine
cables, for example. Box 1 highlights some examples of regional sharing initiatives. In addition, the 2008 GSR discussion paper on
international gateway liberalization contains information about regional sharing initiatives. This discussion paper is available for free at:
www.itu.int/ITU-D/treg/publications/ITU_Trends2008_Chapter6.pdf.
Table 1: Regional Initiatives in Sharing Infrastructure
Initiative Number of Countries involved Interconnection points

South Atlantic 3/West Africa submarine cable 36 Cable landing points:
Portugal (Sesimbra & Chipiona)
Spain (Altavista)
Senegal (Dakar)
Cote d’Ivoire (Abidjan)
Ghana (Accra)
Benin (Cotonou)
Nigeria (Lagos)
Cameroon (Douala)
Gabon (Libreville)
Angola (Cacuaco)
South Africa (Melkbosstrand)
SEACOM 7 Cable landing points:

France (Marseilles)
Egypt (Sidi Kerir)
Backhaul through protected ring structure:
Egypt
Djibouti
Madagascar
Mozambique
Kenya
Tanzania
South Africa
Express fibre pair:
South Africa to Kenya
Tanzania to India
Kenya to France
East African Submarine Cable System (EASSy) 21 Cable landing points:

South Africa
Sudan
Madagascar
Somalia
Kenya
Tanzania
Comores
Islands
Djibouti
Backhaul:
South Africa
Tanzania
Uganda
Rwanda
Burundi
Sudan
Madagascar
Somalia
Kenya
Botswana
Somalia
Sudan
Mauritius
Zambia
Djibouti
Lesotho
Nigeria
Source: Theresa Miedema, based on information from the SAT-3/WASC/SAFE website (www.safe-sat3.co.za), the SEACOM website
(www.seacom.mu/intro.html), the EASSy website (www.eassy.org) , and the Fibre for Africa website
(www.fibreforafrica.net/main.shtml).
[1] See www.itu.int/ITU-D/treg/projects/itu-ec/Acts_E.zip

Competition and Pricing: Index
COMPETITION AND PRICE REGULATION HOME
1 OVERVIEW: PUTTING ICT REGULATION IN CONTEXT
2 COMPETITION POLICY AND THE ICT SECTOR



2.3.2 Regulation

2.4.2 Market Power
2.5 Common Forms of Anti-Competitive Conduct

2.5.4 Cross-Subsidization
2.5.6 Customer Lock-In
2.6 Remedies for Anti-Competitive Conduct

2.6.3 Remedies for Cross-Subsidization
2.6.5 Remedies for Customer Lock-In

3 REGULATING FOR INTERCONNECTION

3.2 Key Concepts

3.2.2 Unbundling

3.3.2 Long-Run Incremental Cost Modelling

3.4.3 Retentions for Fixed-to-Mobile Calls

3.5.3 Regulating State-Owned Operators
3.6 Cross-Border Interconnection

4 NEW PARADIGMS: VOICE OVER IP AND IXPS


4.2 About VoIP
4.2.1 Types of VoIP




5 REGULATING PRICES

5.3.3 Single- and Multiple-Service Firms

5.5.2 Determining Mark-Ups over TSLRIC

Practice Notes on Competition and Price Regulation
Reference Documents on Competition and Price Regulation

Reference Documents on Competition and Price
Regulation
Competition Policy
Competition and Regulation Issues in Telecommunications
Department of Justice and Federal Trade Commission, United States 1992 Horizontal Merger
Guidelines
European Commission Guidelines on Market Analysis
Implementation of the Non-Accounting Safeguards
Implementation of the Telecommunications Act
Jamaica Fair Competition Act, 1993
Jamaica Telecommunications Act, 2000
Malaysia Guideline on Substantial Lessening of Competition
New Zealand Commerce Commission, Mergers and Acquisitions Guidelines
US Federal Communications Commission, Telecommunications Carrier’s Use of Customer

Proprietary Network Information and Other Customer Information
US 1992 Horizontal Merger Guidelines
ITU/WTO Workshop on Telecom and ICT regulation
Interconnection
Anguilla Interconnection Decision, October 2005
Botswana Mini-Case Study
Botswana Ruling on Interconnection Dispute Between BTC and Mascom
COSITU Frequently Asked Questions
COSITU Model Description
Ex post Regulation of the Tariffs of PCCW-HKT under a New Fixed Carrier Licence
Jamaica Office of Utilities Regulation, Determination Notice on C&WJ RIO
UK Office of Communications, Wholesale Mobile Voice Call Termination: Statement
Unbundling Policy in the United States
US Unbundling: Interview with Professor Ingo Vogelsang
Wholesale Mobile Voice Call Termination: Statement
ITU/WTO Workshop on Telecom and ICT regulation
ITU West African Common Market Project: Harmonization of Policies Governing the ICT Market in the
UEMOA-ECOWAS Space, Interconnection
Voice Over IP and IXPs
Barbados: Voice over Internet Protocol Draft Policy
European Regulators Group, Common Statement for VoIP Regulatory Approaches
OFTA Regulation of Internet Protocol Telephony
OFTA Services-Based Operator Licence
The Treatment of VoIP under the EU Regulatory Framework
VoIP: Developments in the Market
Regulating Prices
Barbados: Decision on Standards of Service
Barbados: Price Cap Mechanism Decision
Barbados Unregulated Services Policy
Ex post Regulation of the Tariffs of PCCW-HKT under a New Fixed Carrier Licence
The Bahamas Telecommunications Company's Application to Reduce International Long Distance SEE ALSO:
Rates: Consultation Document
The Bahamas Telecommunications Company's Application to Reduce International Long Distance Practice Notes on Competition and
Rates: Statement of Results Price Regulation

Module+2.+Competition+and+Price+Regulation-ICT Regulation Toolkit

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Module+2.+Competition+and+Price+Regulation-ICT Regulation Toolkit

Uploaded by

Copyright:

Available Formats

Module 2.

Competition and Price Regulation

Downloaded Tuesday, April 06, 2010

Module 2. Competition and Price Regulation

Overview: Putting ICT Regulation in Context

Competition Policy and the ICT Sector

Regulating for Interconnection

New Paradigms: Voice Over IP and IXPs

Module 2. Competition and Price Regulation

Specifically, this section provides an overview of:

1.1 Key Developments in the ICT Sector

■ The shift from analog to digital technologies,

The Shift from Analog to Digital

Data Replaces Voice

■ Increased business data requirements,

Table 1: Comparison of Network Requirements for Conventional Voice and Data

Conventional voice service Data / information services

Packet Switching Replaces Circuit Switching

Broadband Replaces Narrowband Networks

Intelligence Migrates from the Core to the Edge

The Increasing Role of Wireless

Implications of these Trends

Table 2: Regulatory and operational implications of technological shifts

The Shift from

Market Structure and Level of Competition

1.2 Trends in ICT Regulation

This section summarizes some key trends in telecommunication reform, covering:

Liberalization and the Introduction of Competition

■ What we mean by competition, and why it is important

2.1 Forms of Competition

2.1.1 Perfect Competition

Perfect competition requires a number of conditions:

2.1.2 Effective Competition

2.1.3 Market Contestability

2.1.4 Sustainable Competition

This aim faces a number of challenges in the ICT sector.

■ The long term objective of ongoing, sustainable competition,

2.2 Why Focus on Competition?

Module 6, Section 4, Legal and Institutional Impacts of Regulation: Impact of Convergence

2.2.1 Benchmarking Competition by Sector

Worldwide Comparison by Sector

Figure 1: Competition by Sector Worldwide

Figure 2: Competition in the Local Services Sector

Domestic Long Distance

Figure 3: Competition in the Domestic Long Distance Sector

International Long Distance

Figure 4: Competition in the International Long Distance Sector

Figure 5: Competition in the Mobile Sector

Figure 6: Competition in the Internet Services Sector

Figure 7: Competition in the Leased Lines Sector

Module 2, section 2.2.2., Comparison Table: Competition by Sector and Region

2.2.2 Comparison Table: Competition by Sector and Region

Table 1: Competition by Sector and Region

Partial competition & Duopoly 11 11 12 15 4 8

Partial competition & Duopoly 5 3 3 7 3 3

Arab States (21)

Partial competition & Duopoly 4 4 4 7 5 4

Asia Pacific (38)

Partial competition & Duopoly 10 11 11 10 11 7

Europe and CIS (52)

Partial competition & Duopoly 5 3 6 15 0 6

Partial competition & Duopoly 35 32 37 54 23 28

Module 2, section 2.2.1., Benchmarking Competition by Sector