Global Broadband Battles: Why the U.S. and Europe Lag While Asia Leads 9781503625693

Citation preview

Global Broadband Battles

ADVANCE PRAISE FOR GLOBAL BROADBAND BATTLES Global Broadband Battles is a fascinating and detailed account of the reasons why Japan and Korea have risen to world leadership in broadband communications, while the United States and Europe have been lagging. Fransman’s book forcefully presents important policy lessons for companies and governments: competition should be fostered, the entry of new firms should be supported, and a regulatory regime allowing new competitors to be innovative should be encouraged. —Franco Malerba, Director, Centre for Research on Innovation and Internationalization, Bocconi University, Italy This book offers the reader a detailed explanation of the rise of South Korea to world leadership in the penetration of broadband, as well as the renaissance of Japan and the relative decline of the United States and Europe. It is compulsory reading for all those who wish to sort out these puzzles. —Chris Freeman, Professor Emeritus, University of Sussex, United Kingdom This is a remarkable collection of essays by leading international authorities in the field and should be a standard work of reference for anyone seeking to understand differences between countries in the present development and future prospects of vital communications technologies. The many contrasts between Asia, Europe, and the United States are telling and revealing in their focus on the interdependence between technology development, institutions, and organisation. As well as by telecommunications experts, this book should be read by anyone interested in the links between competition, innovation, and economic performance. —Stan Metcalfe, Jevons Professor, University of Manchester, United Kingdom

innovation and technology in the world economy

Editors mart in kenney University of California, Davis/Berkeley Round Table on the International Economy bruce kogut Wharton School, University of Pennsylvania

Other titles in the series The Triumph of Ethernet: Technological Communities and the Battle for the LAN Standard urs von burg Territories of Profit: Communications, Capitalist Development, and the Innovative Enterprises of G.F. Swift and Dell Computer gary fields Ivory Tower and Industrial Innovation: University––Industry Technology Transfer Before and After the Bayh–Doyle Act in the United States d av i d c . m o we r y, r i c h ard r. ne lson, bhaven n. samp at, and arvids a. ziedonis Locating Global Advantage: Industry Dynamics in the International Economy mart in kenney and r ichard flor ida, eds.

Contributors t akanor i ida inho chung manqiang liu johannes m. bauer jackie kraff t franz büllingen cr ist iano antonelli pier p aolo p atrucco sven lindmark p er björstedt

Global Broadband Battles Why the U.S. and Europe Lag While Asia Leads

edited by mart in fransman

Stanford Business Books An imprint of Stanford University Press Stanford, California 2006

Stanford University Press Stanford, California © 2006 by the Board of Trustees of the Leland Stanford Junior University. All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or in any information storage or retrieval system without the prior written permission of Stanford University Press. Printed in the United States of America on acid-free, archival-quality paper Library of Congress Cataloging-in-Publication Data Global broadband battles: why the U.S. and Europe lag while Asia leads / edited by Martin Fransman. p. c.m.—(Innovation and technology in the world economy) Includes index isbn 0-8047-5305-9 (cloth : alk. paper)—isbn 0-8047-5306-7 (pbk. : alk. paper) 1. Broadband communication equipment industry—United States. 2. Broadband communication equipment industry— Europe. 3. Broadband communication equipment industry— Asia. 4. Broadband communication systems—United States. 5. Broadband communication systems—Europe. 6. Broadband communication systems—Asia. 7. Telecommunication policy—United States. 8. Telecommunication policy—Europe. 9. Telecommunication policy—Asia. 10. Telecommunication—Technological innovations—United States. 11. Telecommunication—Technological innovations—Europe. 12. Telecommunication—Technological innovations—Asia. 13. Competition, International. i. Fransman, Martin. ii. Series. hd9696.b763u825 2006 384.3—dc22

2005033240

Original Printing 2006 Last figure below indicates year of this printing: 15 14 13 12 11 10 09 08 07 06 Typeset by G&S Typesetters, Inc. in 10/12.5 Electra Special discounts for bulk quantities of Stanford Business Books are available to corporations, professional associations, and other organizations. For details and discount information, contact the special sales department of Stanford University Press. Tel: (650) 736-1783, Fax: (650) 736-1784

This book is dedicated to Tammy and Dee

Contents

Tables and Figures Preface 1 Introduction Martin Fransman part one.

asia: japan, korea, and china

xi xiii 1 59

2 Broadband, Information Society, and the National System in Japan Takanori Ida

65

3 Broadband, the Information Society, and National Systems: The Korean Case Inho Chung

87

4 Broadband Access Development in China Manqiang Liu part two.

north america: the united states

5 Broadband in the United States Johannes M. Bauer part three.

europe: france, germany, italy, and sweden

6 Emergence and Growth of Broadband in the French Infocommunications System of Innovation Jackie Krafft

109 127 133 165 172

7 Development of the Broadband Market in Germany Franz Büllingen

195

8 Broadband in Italy: Timing in Inter-modal Rivalry Cristiano Antonelli and Pier Paolo Patrucco

219

Contents

x

9 The Swedish Broadband Market Sven Lindmark and Per Björstedt

240

Conclusion: The Importance of Institutions Martin Fransman

267

Contributors

271

Index

275

Tables and Figures

tables 1.1 Evolution of the Internet and mobile communications, 1995–2006

4

1.2 Some measures of national performance in broadband

7

1.3 World’s top 10 incumbents offering broadband access, ranked by speed, 2003

11

1.4 World’s top 11 new entrants offering broadband access, ranked by capacity, 2003

12

1.5 World’s top 10 incumbents offering unmetered broadband access, ranked by price, 2003

14

1.6 Percentage of broadband lines unbundled in Europe, March 2004

23

1.7 DSL market share of incumbent and competitors (based on local loop unbundling), 2003–2004

23

1.8 VoIP diffusion, February 2005

26

1.9 Asian broadband subscribers, early 2004

50

2.1 Area coverage of NTT East and West

68

2.2 Managerial status of NTT East and West

73

2.3 Trends in narrowband and broadband Internet users

74

2.4 ADSL market trends at NTT East and West

77

2.5 Changes in standard ADSL charges

77

2.6 FTTH market trends at NTT East and West

79

2.7 Changes in standard FTTH charges

80

3.1 Number of subscribers and market share of each service provider

89

Tables and Figures

xii

3.2 Monthly charges for high speed Internet services, October 2004

90

3.3 KII project budget

97

3.4 Budget for Informatization and Information Facilitation Fund

102

4.1 Growth of telephone subscribers and the Internet in China

111

5.1 Leading broadband suppliers as of September 2004

135

5.2 Technology mix of broadband supply

143

5.3 Selected broadband prices, October 2003 and February 2005

148

6.1 Concentration of infocommunications firms and employment in France, 2002

185

8.1 Diffusion of broadband access in Italy, first halves of 2003 and 2004

226

8.2 Penetration of broadband in Italy, percentage of population

226

8.3 Market shares in the Italian broadband market

227

8.4 The digital divide in broadband in Italy, 2003

230

9.1 Internet services market shares

244

9.2 Selection of IT initiatives by public authorities

248

9.3 Broadband market shares

250

9.4 Internet subscribers per access technology (thousands)

253

figures 1.1 Broadband penetration in OECD countries, 2005

9

4.1 ADSL subscribers in China

117

4.2 Internet users by type of access

121

4.3 Proportion of users by type of access

121

5.1 Usage patterns of narrowband and broadband subscribers

139

6.1 Broadband penetration rate in France and Europe, 2002–2004

175

6.2 Shared access and full unbundling in Europe, 2004, in millions of lines

177

6.3 Number of firms and net entry in the French telecommunications industry, 1997–2002

182

7.1 Development of DSL market shares, 2000 –2004

197

7.2 Operators and their market share of the broadband cable market

201

7.3 Groups of suppliers and consolidation processes in the PWLAN market

208

7.4 Forecast of the broadband Internet penetration rate in Germany until 2015

215

Preface

This book began on one of my regular visits to Japan. Prior to my visit I had become aware that in the area of broadband communications something unusual was happening in Japan. As had happened earlier in the field of mobile communications (where Japan pioneered the mobile Internet and where the quality of mobile phones far exceeded what was available elsewhere in the world), so too in the area of broadband services were being rolled out that were unavailable in other countries. Japanese broadband speeds were unparalleled and an intense price war was rapidly bringing broadband Internet access prices down to the lowest levels in the world. My discussions with Japanese friends made clearer the rapid changes that were occurring. Before even having time to adjust fully to the negative effects of the telecommunications (telecoms) bust (that began globally from March 2000) the Japanese telecoms industry had become embroiled in a new rollercoaster ride. NTT, the incumbent Japanese telecoms carrier, was under intense pressure. This pressure was so great that it was even beginning to affect the company’s financial performance. The pressure came from two sources. The first was from the Japanese ministry of communications that also acts as the telecoms regulator. The ministry was insisting that NTT open its network to its broadband competitors who would be given direct access to NTT’s customers. Furthermore, the ministry was imposing what was, comparatively speaking, possibly the lowest price in the world for access to NTT’s network. This facilitated the second source of pressure: entry by aggressive new competitors into the broadband market, able to serve NTT’s customers largely over NTT’s network. NTT was working hard to deal with these twin pressures. But the company’s response was not simply defensive. NTT retaliated by rapidly increasing its

Preface

xiv

broadband speed and sharply reducing its price (even though this bit into revenues already being hit hard by the fall in telephony revenues resulting from the substitution of mobile calls for fixed calls and, increasingly, the growing use of Internet telephony, a service in which Japan is leading the world). However, NTT went even further, technologically leapfrogging its competitors (both in Japan and abroad) by rapidly installing what is widely regarded as the Rolls Royce of broadband technologies, namely optical fibre to the home (FTTH).1 Even though FTTH was widely regarded in the U.S. and Europe at the time as being prohibitively expensive, NTT pressed on in its conviction that this was in the longer term the only way to go. True, in so doing NTT was also responding to the threat posed by unregulated electricity companies (unregulated in the area of telecoms) which had their own optical fibre networks which they were beginning to see as being also a commercial communications asset. But, as the most R&D-intensive of the world’s telecoms carriers, NTT had since the late-1990s espoused a VIP vision of the future of telecoms-visual, intelligent, and personal. And optical technologies generally were seen as being key in realising this vision. While trying, with the help of my Japanese friends, to understand these unfolding events in Japan, I became increasingly aware of very different events in the broadband area in other parts of the world. Like many other observers I had been amazed by the unparalleled Korean performance in broadband penetration (shown vividly in Figure 1.1 in Chapter 1). Increasingly, the experiences of Japan and Korea raised what I soon came to think of as the Japan/Korea puzzle: Why do these two Asian countries lead the world in broadband? They are not globally dominant in areas such as the Internet, computing, software, and microprocessors. Furthermore, since 1989 the Japanese economy has been in a state of recession and Korea, although growing rapidly, is not yet one of the richer countries. Just as fascinating a puzzle emerged from the U.S., which I began thinking of as the U.S. puzzle: Why is it that the U.S.––which gave birth to the Internet and leads the world in areas such as computing, software, and microprocessors––is lagging significantly behind in broadband Internet? In 2004 an FCC (Federal Communications Commission, the U.S. regulator) commissioner

1 Unlike the electromagnetic signals used by the conventional broadband technologies over copper cable, TV cable, and wireless communications systems, optical fibre uses light signals that give this technology the highest bandwidths, making it the superior technology.

Preface

complained that the U.S. was ranked eleventh in one of the key broadband league tables.2 Europe presented its own puzzle, the European puzzle. The European Commission had just worked hard and creatively to put in place a new European regulatory framework for electronic communications that was in the process of being transposed into member country institutions. There was a strong sense that Europe had finally got it right and had the best regulatory framework in the world. However, the comparative global data, such as it was, was beginning to reveal that European countries––particularly the largest including France, Germany, Italy, and the United Kingdom––were lagging behind not only Japan and Korea but also the U.S. These three puzzles, it seemed to me, were begging an explanation. The problem was that the bits of information and analysis that would have to become the jigsaw puzzle pieces making possible a clearer overall picture were scarce, partial, and fragmentary. In search of a more comprehensive global descriptive picture I visited both the OECD (Organisation for Economic Co-operation and Development) in Paris and the European Commission in Brussels. I also benefited from a comprehensive review of telecoms being undertaken by the new British regulator, OFCOM, which included a short appendix making international comparisons. These and other sources confirmed the global pecking order: Japan and Korea were clearly in the lead and well ahead while the U.S. followed with Europe even further behind (looking at the world’s largest economic powers). The main problem that remained was one of explanation. How were these puzzles to be explained? This was a significant problem since I was already all too aware that there were no available analyses, data, or measures that would easily yield a satisfactory answer. My own knowledge was fragmentary, dependent on my own direct experiences (mainly in Japan and Europe) and the relevant bits and pieces I had come across. It was in thinking about how to tackle my incomplete knowledge that I suddenly realised that, through my own network of contacts, I already knew (or knew of ) many of the people who had knowledge of what was happening in broadband in most of the key countries in the world. Perhaps they would be motivated to share their knowledge of their own country in return for a collective development of a global broadband picture. As things turned out this was indeed the case, and the motivation was sufficiently strong so that no fund-raising or financial inducements was needed. 2

See Figure 1.1 in Chapter 1.

xv

Preface

xvi

It is necessary to add a brief explanation regarding choice of countries to be included in this book. Japan and Korea, the world leaders in broadband, were obvious and indeed necessary candidates. The same is true of the U.S. as a result of its dominance of the Internet and the computing industry. In Europe I decided to include the four largest economies, large enough to make comparisons with Japan and the U.S. meaningful: France, Germany, Italy, and the UK (the latter included in the various introductions in this book). In addition, Sweden was chosen as a good representative (though with its own specific characteristics) of the smaller European economies that have performed well in broadband from a comparative international perspective. As a global leader in IT generally, Sweden seemed a sensible choice. Finally, I added China. Although clearly the odd man out, China’s size and rapid growth has already make it the world’s largest market in mobile communications and the same will, in the not too distant future, be true for broadband. This seemed to justify its inclusion. While the present book is essentially a collective knowledge-creating exercise, the attempts to answer the question posed in the subtitle of this book are presented largely (though not exclusively) in Chapter 1. It must be strongly emphasised, however, that although I have benefited greatly from formal and informal discussions regarding parts of this chapter with my co-authors, I alone bear responsibility for the analysis, information, and conclusions. It would be incorrect to infer that my co-authors necessarily share my views (although this is also to say that they do not necessarily disagree with some or all of these views). Apart from my co-authors, to whom my first acknowledgement of indebtedness must go, there are a number of others who must also be mentioned. In Japan I have benefited greatly from discussions with Mr Norio Wada, President and CEO of NTT, Mr Kiyoshi Fujita, President and CEO, InfoCom, Takahiro Ozawa and Teruaki Ohara also of InfoCom, and many of their staff. My friend and colleague, Professor Sadahiko Kano at Waseda University in Tokyo, has over the years taught me a good deal about Japan’s telecoms system and how it works in reality. Also in Japan, Professor Shumpei Kumon and Adam Peake of GLOCOM have long been sources of stimulating conversation (and indeed Adam was involved from the start in this project). At the OECD in Paris I learned from discussion and information supplied by Sam Paltridge and Dimitri Ypsilanti, whose comparative international data on broadband is widely known around the world and quoted. In Brussels the European Commission’s broadband specialists, Lucilla Sioli and Laura Pontiggia, gave me the benefit of both their reports and their wisdom. I also learned a great deal from senior managers at a number of telecoms companies

Preface

of whom special mention should be made of Alison Ritchie, BT’s head of broadband. I have long benefited from close interaction regarding European telecoms and more general issues with my friends and colleagues Cristiano Antonelli, Erik Bohlin, and Jackie Krafft. My understanding of the complexities of the U.S. broadband system was greatly improved by discussions with David Allen and Bill Melody. Those detecting the “ ‘national systems’ ” issues that are a key component of this book will not be surprised to have me acknowledge my intellectual debt to friends such as Bengt-Ake Lundval, Chris Freeman, Stan Metcalfe, and Dick Nelson. In my own institute, the Institute for Japanese-European Technology Studies, particular thanks are due to Lynne Dyer who, as always, was meticulous and diligent in undertaking the numerous tasks that need to be done in preparing a multi-authored study such as the present one. Her reliable presence will be greatly missed. Last, but by no means least, I must acknowledge the support of my family, Tammy, Jude, Karrie, and Jonathan, without whom the trials and tribulations of research and writing would be immeasurably more difficult. It is to Tammy and my mother, Dee, to whom this book, with gratitude, is dedicated. Edinburgh, April 2005

xvii

Global Broadband Battles

1 Introduction Martin Fransman

global broadband battles the broadband Internet Era is often portrayed in the media as a battle. According to this portrayal the United States and Europe battle with Japan and Korea, which have taken a decisive lead in the broadband field. Battles within countries are depicted as just as intense as battles among them. In the 2004 American presidential election, George Bush and John Kerry vied with each other in their promises to speed broadband diffusion in the U.S. Comparative national performance in broadband is seen as being almost as important as GDP (Gross Domestic Product) as an indicator, not only of economic wellbeing but also of national pride. The top advisers of presidents and prime ministers are known to keep a close eye on the broadband performance league tables. The combatants in the broadband battles are seen to be not only politicians and policymakers but also companies and their regulators. According to this view, telecommunications (telecoms) companies battle with cable TV companies, incumbents battle new entrants, and regulators fight with all of them in one field or another. Furthermore, the battles pitch those living in sparsely populated rural areas against their urban counterparts and the digital have-nots against the digital-haves. the purpose of this book This book provides a deeper understanding of the forces driving the Broadband Era in two ways: first, through an analysis, undertaken in this introduction, of the dynamic forces of change in the Broadband Era that situates this

martin fransman

2

era within the broader context of the evolution of the Internet; and, second, through a comparative analysis of the development of broadband infrastructure and services in nine key countries: in North America, the U.S.; in Asia, Japan, Korea, and China; and in Europe, France, Germany, Italy, Sweden and the UK.1 the three broadband puzzles The comparative analysis of these nine countries reveals three important puzzles, all of which are resolved in the second section of this chapter: The U.S. Puzzle: Why is it that the U.S.—which gave birth to the Internet and leads the world in areas such as computing, software, and microprocessors—is lagging significantly behind in broadband? In 2004 an FCC (Federal Communications Commission, the U.S. regulator) commissioner complained that the U.S. was ranked eleventh in one of the key broadband league tables.2 The Japan /Korea Puzzle: Why do these two Asian countries lead the world in broadband? They are not globally dominant in areas such as the Internet, computing, software, and microprocessors. Furthermore, since 1989 the Japanese economy has been in a state of recession, and Korea, although growing rapidly, is not yet one of the richer countries. The European Puzzle: Why is Europe lagging behind not only the leading Asian countries but also the U.S.? The European Commission has created a new regulatory framework for Europe that it hopes will unleash new forces of competition. National regulatory authorities in the leading countries of the European Union (EU) claim to be broadly satisfied with the performance of their national systems. Yet, as this book will show, Europe is a laggard in the Broadband Era. As these three puzzles imply, the broadband battles of the Broadband Era have thrown up some important questions that beg explanation. In order to deal with these questions it is necessary to understand more about broadband. what is broadband? Broadband may be thought of as access to the Internet provided at speeds significantly higher than those used by the traditional method, namely narrowband, or dial-up. Although definitions of broadband speed vary (and are tending to increase over time), currently most definitions refer to speeds in excess of 128 kilobits per second (kbps).3

Introduction

the broadband era: the latest stage in the evolution of the internet The Broadband Era represents the latest stage in the evolution of the Internet. Table 1.1 shows the evolution of the Internet and mobile communication from 1995 to 2006. Commercialisation of the Internet It was only in 1995 that Bill Gates, in an internal Microsoft memo, acknowledged the importance of the Internet and accepted that the Internet would drive Microsoft, rather than the other way round. From this time there was a massive global diffusion of the Internet, which in turn drove increasing demand for data capacity and services.4 The Internet radically transformed the telecoms industry, causing it to converge ever more closely with the computer industry, creating a new infocommunications—or ICT—industry.5 Mobile Communications From 1995 the Internet diffused rapidly around the world, however, from around 1997 it was mobile communications, rather than the Internet, that provided the main source of growth in revenue and profitability for the telecoms industry. Although the origins of cellular mobile telephony can be traced back to research done in the early post-war years in AT&T’s Bell Laboratories, by as late as the mid-1980s many were of the opinion that mobile communications would not become a mass market. Kurt Hellstrom, later to become President of Ericsson (the Swedish company that is one of the world’s largest suppliers of mobile equipment), recalled that his mobile division was regarded by others in the company as “something strange happening on the outskirts of Stockholm.” At that time the company was prioritising its digital switch for fixed-line communications. When AT&T in the mid-1980s asked the consultancy company McKinsey to predict the number of mobile phones that would exist in 2000, their answer was 900,000. In the event there were 40 million. From around 1997, first in the Scandinavian countries and Japan, and then in the other rich industrialised countries, mobile voice services began to diffuse rapidly, adding significantly to the revenue and profits of the mobile operators. However, by the early 2000s the first stage in the evolution of mobile communications—the stage based on voice telephony—was beginning to run out of steam in the rich countries as markets reached saturation levels.6

3

A. Phase One: 2000 – Broadband Access – Increasing speed/ bandwidth – Falling prices

– Rapid adoption of narrowband Internet – Rapid increase in demand for data-carrying capacity

B. Phase Two: 2004 – From Broadband Access to Broadband Services – New/improved broadband CASs (content, applications, and services; e.g., voice-over-Internet, Internet TV, and video)

From 2000: Broadband Internet

From 1994: Narrowband Internet

Internet

table 1.1

– Approaching saturation in rich countries by 2000 – Lagged but rapid diffusion in developing countries (e.g., China)

From 1997: Voice Telephony

Mobile

– Texting, e-mail – Mobile Internet (e.g., i-mode in Japan) – Mobile images (still pictures, 3G, video) – Mobile broadband (3G, WiFi, WiMax/ WLL [wireless local loop], satellite, etc.)

From 2000: Mobile Data

Evolution of the Internet and Mobile Communications, 1995–2006

Introduction

The search for a new source of growth initiated the second stage in the mobile communication industry’s evolution. Specifically, the industry’s corporate leaders began looking toward mobile data, and new services based on mobile data, as a new source of growth. Around the world “texting” became an important source of growth in mobile data traffic while in Japan NTT DoCoMo pioneered the mobile Internet in the form of the i-mode service.7 Later, again pioneered in Japan, camera phones provided a further source of new data traffic. However, by 2006 it was still not clear whether sufficient consumer demand would exist to turn the new mobile data services into a new source of rapid growth in the mobile communications industry. Phase One: Enter Broadband Access In the late-1990s the first commercial broadband services appeared in countries like the U.S. and Japan. However, only since 2002 has broadband access to the Internet begun diffusing rapidly in rich countries, where it is provided not only by telecoms companies but also by cable TV companies using cable modems. (The technologies that they used, as well as other alternative technologies, are discussed later in this chapter.) Why did broadband begin to diffuse rapidly in rich countries from around 2002? Although a more detailed answer is given to this question below, it is worth mentioning a few of the general determinants here. Incumbent telecoms operators in these countries began at this time to commit to broadband, partly in search of a new source of revenue, profits, and growth. As already noted, growth in their mobile subsidiaries began to slow as mobile voice markets began to saturate. Furthermore, the telecoms bust (that began in March 2000 when stock markets crashed) and the general slowing in economic activity soon took their toll. From being the darlings of stock exchanges and other financial markets, telecoms companies (together with the now infamous dot.coms) quickly became the devils. In the U.S., high-profile bankruptcies, compounded by massive fraud, further mired the telecoms waters. Although WorldCom’s insolvency, the largest in U.S. corporate history, was the most extreme, dubious and even outright illegal practices were found to be widespread in the telecoms industry. While the European telecoms industry was spared the same degree of illegality, the main European telecoms operator incumbents emerged from the telecoms bust with excessively high levels of indebtedness, largely the result of exuberant merger and acquisition activity during the preceding boom. For all these telecoms companies—and for the telecoms equipment companies that supplied them—the search for a new source of revenue and profit became a high

5

martin fransman

6

priority. Broadband soon seemed to offer the most promising way out of the doldrums. Cable TV companies also became interested in broadband at the same time. They, too, suffered from the stock market collapse and the economic downturn as well as pressures from consolidation. Furthermore, by broadbandenabling their cable TV networks (an expensive business) they saw an opportunity to increase their competitiveness while entering new markets. In addition to their traditional television, they could also offer high-speed Internet access as well as new services such as voice telephony (the so-called triple play bundle of services). A further factor encouraging the take-up of broadband was regulation that reduced entry barriers for companies wanting to compete with the incumbents, thus increasing their expectations of rates of return to be earned in what was coming rapidly to be seen as a set of new and attractive broadband-related markets. (The role of regulation is examined in more detail later in this chapter.) With all these companies motivated to contest the emerging broadband market, a new growth area quickly emerged. Phase Two: From Broadband Access to Broadband Services Things move quickly in the broadband market, as the experience of South Korea (henceforth referred to as Korea) vividly indicates. For a combination of reasons (analysed in detail in Chapter 3 and later in this chapter) Korea led the global field in broadband access penetration. However, by the middle of 2004 it was already becoming apparent that broadband access itself was losing steam as a significant driver of growth in revenue and profitability. On 4 May 2004, the Financial Times reported: South Korea’s top fixed-line phone operator, KT Corp, reported a 65 percent slump in quarterly profit on Tuesday, missing forecasts as landline customers switched to mobiles and broadband Internet growth slowed. [KT] supplies half of the country’s broadband connections, but the market is nearing saturation, with penetration rates already the world’s highest. However, analysts hope operating profits will improve this year as a cost-cutting programme bears fruit. [emphasis added]

As the first stage in the evolution of the fixed broadband market—based on broadband access—runs out of steam, as is already happening in Korea, two trends are becoming apparent. First, broadband speed/ bandwidth is increasing rapidly; the greater the degree of competition in the domestic market, the greater the increase in speed. Second, increasing attention is being

Introduction table 1.2 Some Measures of National Performance in Broadband Measure of Performance

Measure 1 Measure 2 Measure 3 Measure 4 Measure 5 Measure 6

Availability Penetration Capacity/speed Price Quality of access Goodness of fit with the needs of users

paid to the demand side, to the content, applications and services (including managed services) that it is hoped will drive growth in revenue and profitability and give a good return on investment in broadband infrastructure. measuring performance in broadband Five Performance Measures Not surprisingly, there are a number of possible measures of national performance in broadband. Some of these are shown in Table 1.2. Measure 1, availability, is the least demanding. In most countries most broadband connections are provided either by ADSL (a technology, discussed in more detail below, developed in the 1980s that allows broadband signals to be sent over the telephone copper cables that connect homes and small businesses), offered by telecoms operators, or by cable modems. Availability measures the proportion of homes (or of population) that has access to a broadband connection if they want it.8 However, availability does not refer to those who actually have broadband connections. The latter is provided by Measure 2, penetration, which refers to the percentage of homes (or population) that already has a broadband connection. But penetration per se says nothing about the capacity/speed that users receive or about the price that they pay. Accordingly, penetration figures hide a significant variance between countries regarding the latter two characteristics. It is therefore necessary to supplement Measures 1 and 2 with Measures 3 and 4. But this is easier in principle than in practice. The reason is that there are many players in the broadband market of any country. The players themselves may operate in different parts of the country and usually have a number of different broadband offerings. To make matters even more complicated, advertised speeds (usually measured in terms of downloaded bits per second) may not measure up to actual received speeds,

7

martin fransman

8

partly because the latter can depend on the user’s distance from the local exchange and the number of people using the system at the time. Little wonder, then, that at the time of writing there is only one serious international comparative study that examines speed and price.9 Even Measures 3 and 4 have shortcomings. One of the most important is that they say nothing about quality of service. For example, a very high speed connection at a very low price may be unreliable. The problem, however, is that internationally comparable measures of quality of service—referred to in Measure 5—do not exist. Measures 1 to 5 all refer to the supply of broadband connections. They reveal nothing about the use that is made of them. In other words, the demand/ user side is ignored. Clearly, this is a major problem with these purported measures of “performance.” At the end of the day, availability, penetration, speed, price, and even quality of access are all means rather than ends—what really matters is what users want to do with the bandwidth and connectivity that they have access to and how well they feel they are able to do what they want with it. Measure 6—goodness of fit with the needs of users, the holy grail of broadband—is intended to capture this important point. The problem, however, is that currently we are very far from any measures of Measure 6. Korea as Leader in Penetration The most widely accepted global comparison of broadband penetration is that provided by the OECD (Organisation for Economic Co-operation and Development). The latest data are shown in Figure 1.1. A number of conclusions emerge from Figure 1.1. The first is the outstanding performance in broadband penetration by Korea, a clear outlier. Korea is the only country close to reaching 25 subscribers per 100 inhabitants. Denmark and Canada come second and third, both with about 17 subscribers. The United States, as already noted, comes eleventh with about 11 subscribers. Japan comes eighth with around 13 subscribers. Sweden comes ninth with slightly fewer subscribers than Japan. The other European countries covered in this book come significantly behind those already mentioned in this paragraph. France comes fifteenth while the UK comes sixteenth, with about 8 subscribers each. Germany comes eighteenth and Italy is placed twentieth, with around 7 and 6 subscribers, respectively.

Greece Turkey Mexico Slovak Republic Czech Republic Poland Ireland Hungary New Zealand Australia Luxembourg Italy Portugal Germany Spain United Kingdom France OECD Austria Finland United States Norway Sweden Japan Belgium Switzerland Iceland Netherlands Canada Denmark Korea 0

5

10

15

20

Subscribers per 100 inhabitants DSL

Cable modem

Other platforms

figure 1.1 Broadband penetration in OECD countries, 2005 Source: OECD Communications Outlook, 2005.

25

martin fransman

10

Speed Incumbents’ Speed

In Table 1.3 the world’s top 10 incumbent telecoms operators are ranked according to their highest speed broadband offering.10 Several important points emerge from Table 1.3. The first is the superior performance of Japan. Not only is Japan and its incumbent, NTT East, the front-runner in terms of capacity, it is significantly ahead of the rest of the global pack. Measured in terms of download speed, NTT East’s measure was 100 Mbps (megabits per second) while that of the second incumbent—KT from Korea—was only 13 Mbps. The reason for this superior performance is also given in Table 1.3. While NTT East used FTTH (fibre to the home), KT used a DSL (digital subscriber line) local access connection. The broader issues relating to FTTH are discussed in more detail below. The second point is that Japan’s performance is also far superior in terms of absolute price, that is, the price of the high-capacity broadband connection, irrespective of the speed that is provided for this price. This is even more starkly clear when prices are measured, as they should be, in terms of purchasing power parity (which assumes that money has the same value in all countries, thus giving a better comparison). As Table 1.3 shows, the Japanese price is significantly below that of the next few countries. While NTT East’s price was $42.46, that for Korea’s KT was $67.61, and for the U.S.’s Verizon, $204.95. Price measured per 100 kbps improves Japan’s relative performance even more. Indeed, only two countries in the global top 10 had absolute prices slightly lower than Japan’s; Belgium’s (ranked seventh) was almost the same at $42.40, while New Zealand (ranked eighth) was $39.84. However, the latter was only for a speed of 2.1 Mpbs. Third, the performance of the U.S. is also worth noting. The U.S. is clearly behind front-runners Japan and Korea. The U.S. lag is particularly noticeable in terms of two comparisons. The first is between the speed offered by Japan compared to the one offered by the U.S.—100 Mbps versus 7.1 Mbps, respectively. The second is in terms of price between these two countries—$42.46 versus $204.95, respectively. The fourth point is that, significantly, none of the large European countries feature in the global top 10, namely France, Germany, Italy, and the UK. The reason for their poor relative performance will be discussed later in this chapter. Table 1.3 also shows the relatively good performance of several of the smaller European countries, including Denmark, Belgium, and Sweden. Reasons for the Swedish performance will be discussed later.

Introduction

11

table 1.3 World’s Top 10 Incumbents Offering Broadband Access, Ranked by Speed, 2003

Country

Company

Japan Korea United States Denmark

NTT East KT Verizon TDC Kable

Finland Canada Belgium New Zealand Sweden Poland

Elisa Bell Canada Belgacom Telecom NZ Telia Dialog

Access Type

FTTH DSL ADSL Cable Modem ADSL ADSL ADSL ADSL ADSL ADSL

Download Upload Speed Speed (Mbps) (Mbps)

Monthly Charge $ (U.S.)

$ (PPP)

Mbps per Month

100.0 13.0 7.1 4.1

100 N/A 0.8 0.3

52.45 42.01 204.95 125.84

42.46 67.61 204.95 100.22

Unl’td Unl’td Unl’td Unl’td

4.1 3.1 3.0 2.1 2.1 2.1

0.5 0.6 0.1 0.3 0.4 0.5

441.43 45.98 44.75 34.32 57.34 68.11

394.52 54.09 42.40 39.84 49.32 141.59

Unl’td Unl’td 10,000 500 10,000 Unl’td

Only includes the incumbent with the highest ranking capacity offer for each country. PPP ⫽ purchasing power parity. Unl’td ⫽ unlimited. source: Calculated by author from OECD, DSTI /ICCP/ TISP(2003)8, 19 November 2003, Annex Table 5, p. 79.

New Entrants’ Speed

The incumbent’s performance can give a false impression of what is happening in a country’s broadband market. The incumbent usually provides broadband services more extensively (in terms of product range and geographical coverage) than the new entrants. It is therefore necessary that the incumbent be included in analysing the broadband performance of a country. However, in many of the segments of the broadband market it is new entrants that often lead the way, both in terms of technologies as well as services (quality and price). Furthermore, an incumbent’s highest speed service may not be typical of its general offerings in the country. For this reason it is necessary also to analyse the broadband offerings of the new entrants. This is done in Table 1.4. Table 1.4 is different from Table 1.3. Rather than including only the bestperforming firm from each country, in Table 1.4 the top 11 new entrants, ranked in terms of the speed of their best offerings, are included regardless of their country. This ensures that weight is given to the number of country firms in the global top 11. On the basis of Table 1.4, it may be assumed that countries with the largest number of companies in the top 11 are the best performing, particularly if there is also a strong positive correlation between these countries and the best-performing countries in Table 1.3, which shows the bestperforming incumbents ranked in terms of capacity/speed.

martin fransman

12

table 1.4 World’s Top 11 New Entrants Offering Broadband Access, Ranked by Capacity, 2003

Country

Japan United Kingdom Japan Japan Sweden Korea Canada Korea Sweden Finland Italy

Company

USEN WRBB eAccess Yahoo! BB Bostream Hanaro Gulf Islands Thrunet Bredbands bolaget OulunPuhelin FastWeb

Access Type

FTTH Fixed Wireless ADSL ADSL VDSL DSL Fixed Wireless DSL Ethernet LAN ADSL Fibre/ LAN

Download Speed (Mbps)

Upload Speed (Mbps)

100.0 5.4 40.0 26.0 26.0 20.0 11.0

Monthly Charge $ (U.S.)

$ (PPP)

Mbps per Month

100.0

35.38 15.94

28.84 15.10

Unl’td 1000

1.0 1.0 26.0

38.46 25.19 48.04 41.25 23.00

31.35 20.54 41.33 66.38 27.06

Unl’td Unl’td Unl’td Unl’td 1000

10.0 10.0

31.74 38.63

51.08 33.23

Unl’td Unl’td

10.0 10.0

45.16 79.37

40.36 87.81

Unl’td Unl’td

source: Calculated by author from OECD, DSTI /ICCP/ TISP(2003)8, 19 November 2003, Annex Table 1, p. 56. Includes only the highest capacity offering for each new entrant. PPP ⫽ purchasing power parity. Unl’td ⫽ unlimited.

Again, it is clear from Table 1.4 that Japan is the best-performing country in terms of speed. First, in this table three of the eleven new entrant companies are Japanese, including the first, third, and fourth. Second, the first two of these companies—USEN and eAccess—have download speeds that are significantly higher than the other companies in the global top 11. Sweden has two companies included, in fifth and ninth position, while Korea also has two, in sixth and eighth position. The U.S. is not included at all in the list, while Canada has one company, in seventh position. It is also worth noting that the UK company is something of an anomaly (as is the Canadian company), being a very small company with a very small wireless network offering service in a very limited part of the country. For the other countries the new entrants featured are amongst the main competitors to the incumbents. One of the reasons for Japan’s superior performance in terms of capacity/speed is the relatively substantial deployment of optical fibre to the home (FTTH), which is shown in Tables 1.3 and 1.4. (Although Sweden is probably the second country in terms of FTTH diffusion, this technology has not influenced Swedish performance in Tables 1.3 and 1.4, although it is possible that the Ethernet LAN referred to in Table 1.4 for the new entrant

Introduction

Bredbandsbolaget uses FTTH.) A more detailed discussion of FTTH is undertaken later in this chapter. Tables 1.3 and 1.4 are based primarily on a ranking according to speed. Price is then given for these services. But a rather different picture emerges if providers are ranked according to price and speed is then added to these rankings. Price Table 1.5 shows the results of ranking providers according to price first and then speed. The following observations may be made about Table 1.5. The first point is that when it comes to the price of relatively basic broadband access in November 2003, Japan and Korea do not perform nearly as well as they do when rankings are in terms of speed (Table 1.3). Indeed, as shown in the second last column, where purchasing power parity (PPP) prices are given, Japan comes seventh out of the top 10 and Korea comes eighth. The U.S., on the other hand, is second with the same download speed as Japan’s. Denmark and Sweden are third and fourth, and Canada, fifth. Data published by the Japanese ministry of communications confirms that it is only in the 10 –100 Mbps and the 2–10 Mbps categories that a Japanese company (Yahoo! BB) has the lowest price of the world’s top 5 leaders. In the other categories lower than 2 Mbps Japanese companies are not amongst the top 5, although Korean companies are.11 However, a rather different picture emerges when market prices (rather than purchasing power parity prices) are used, which is the basis for the ordinal ranking of countries in Table 1.5. Then Korea becomes the country with the lowest price and Japan comes tenth. The U.S. is third with a price very little different from the second country, Canada. We do know that price competition has increased significantly in Japan since the end of 2003 and that a severe price war is currently underway. In several other countries there are also price wars, France and the UK being examples. Indeed, the general global trend is for substantial broadband price deflation. Tables 1.3, 1.4, and 1.5, as well as the differences between market prices and purchasing power parity prices, make the dangers clear of unqualified mediatype pronouncements about “the world leaders in broadband.” The figures are subject to interpretation and, depending on the precise definition of “leadership,” it is possible to put different spins on the data in order to arrive at different conclusions (something that governments, unfortunately, are apt to do).

13

martin fransman

14

table 1.5 World’s Top 10 Incumbents Offering Unmetered Broadband Access, Ranked by Price, 2003 Monthly Charge

Country

Korea Canada United States France Turkey Switzerland Denmark Sweden Poland Japan

Company

Access Type

Download Upload Speed Speed (kbps) (kbps)

$ (U.S.)

$ (PPP) (Rank)

Mbps per Month

KT Bell Canada SBC

DSL ADSL

1500 1024

320

25.21 26.81

40.57(8) 31.53(5)

Unl’td Unl’td

ADSL

1500

128

26.95

26.95(2)

Unl’td

France Telecom/ Wanadoo TurkTelecom Swisscom TDCKable TV Telia TPSA NTT East

ADSL

128

33.96

31.83(6)

Unl’td

ADSL ADSL Cable Modem ADSL ADSL ADSL

128 300 256

32 50 64

34.88 35.47 35.85

70.48(9) 25.85(1) 28.55(3)

Unl’td Unl’td Unl’td

250 512 1500

64 128 512

36.09 38.09 40.25

31.05(4) 79.18(10) 32.81(7)

Unl’td Unl’td Unl’td

source: Calculated by author from OECD, DSTI /ICCP/ TISP(2003)8, 19 November 2003, Annex Table 6, p. 87. Only includes the incumbent with the highest ranking price offer (i.e., the lowest price) for each country. PPP ⫽ purchasing power parity. Unl’td ⫽ unlimited.

Japan as World Leader in Both Speed and Price There is little dispute that Korea is the world leader in terms of broadband penetration (as is shown clearly in Figure 1.1). Similarly, as suggested by Tables 1.3 and 1.4, there is little doubt that Japan leads the world in broadband speed. However, it is the case that Japan also leads in terms of price. Although this was not apparent from the global study done by the OECD in November 2003 (shown in Table 1.5), more recent research has confirmed Japan’s leadership in price. The research in question was undertaken by the British regulator OFCOM as part of a wide-ranging strategic review of the country’s telecommunications published in November 2004. According to this research, which focuses on “entry level” broadband products, “Japan emerges as the cheapest country [in the world] for both residential and business users.” Accordingly, OFCOM concludes, “This analysis confirms Japan’s acknowledged position as a broadband price benchmark.” Regarding speed, the study concludes that the “basic speeds available in Europe are significantly below those in Japan and Korea, and to a lesser extent, the U.S..”12

Introduction

The main reason for Japan’s superior performance in both speed and price is the greater intensity of competition in this country’s broadband market. This is discussed in more detail later in this chapter. explaining the puzzles 1: the drivers of the broadband era Driver 1: Technical Change Transforming the Knowledge Base of the Industry One of the most important drivers of the Broadband Era is new technologies. As Figure 1.1 shows, the two most important technologies currently used to provide broadband access are DSL (digital subscriber line) and cable modems. DSL

ADSL (asynchronous digital subscriber line) is a technology that allows broadband data to be sent over the traditional copper telephone lines that connect most homes and small businesses. It has become the incumbent telecoms operators’ broadband technology because it is these companies that own and control the copper telephone lines (also known as the local loop, the loop consisting of two copper wires that connect the customer’s premises with the local telephone network). It is called asynchronous because the download speed is faster than the upload speed. ADSL was first developed in the late 1980s by Joseph W. Lechleider, formerly of Bellcore Laboratories. AT&T was divested in 1984, separating the telephone company’s long-distance division together with its research laboratories (Bell Laboratories) from its regional divisions (the regional operating companies, or Baby Bells as they were called). The former became the new AT&T, while the latter were divided into several separate regional companies (the best known of which now are Verizon and SBC). Parts of Bell Laboratories were split off to become Bellcore Laboratories, the purpose of which was to do research and development work for the Baby Bells. For the Baby Bells and their researchers, an obvious question was whether they would be able to use their local access networks (their copper local loop) to provide services such as TV and video to customers over their telephone lines. In this way the Baby Bells would be able to enter new markets, competing with satellite and cable TV companies. It was Lechleider who first demonstrated, in a mathematical paper, that this was possible. This made technically feasible a new service called video

15

martin fransman

16

streaming, which in turn facilitated video-on-demand (VoD). Further important research, implementing ADSL, was done by John Cioffi of Stanford University.13 Essentially, ADSL involves splitting the telephone line, allowing one channel to provide voice communications and a second, operating at higher frequencies, to deliver high-speed data (thus allowing broadband Internet access). This, in turn, facilitates two of the most important characteristics of broadband communications; always-on functionality, and the ability of the user to use the Internet while receiving phone calls (capabilities that did not exist with narrowband, or dial-up, Internet access). It was these characteristics, together with a third—namely, the faster speed of broadband—that initially drove demand for broadband, rather than broadband-specific content, applications, or services. (The demand for broadband is considered in more detail later in this chapter.) Unfortunately for the Baby Bells, however, video-on-demand did not take off in the 1990s. Given the state of the technology, the cost of VoD, and the availability of other alternatives (such as video rental shops), consumer demand failed to materialise. In this sense ADSL was at the time a failed technology; although it was technically feasible, it failed to find a market. But fortunately, and also unexpectedly, a new demand was to emerge that ADSL was well-placed to serve, namely Internet access. As noted earlier, from 1995 massdemand for Internet access emerged around the world. Narrowband dial-up access made use of the copper telephone lines to provide Internet access. With ADSL it became possible to provide broadband Internet access over the same lines. Furthermore, incumbent telecoms operators who owned and controlled these lines, as we have already noted, had a particularly strong incentive to exploit this technology.14 With the high-speed data channel that it provided, ADSL also allowed providers to offer flat-rate charges (i.e., one price, irrespective of usage) while ensuring that the telecoms network was able to survive the massive increase in traffic that flat-rate charges led to. The existence of flat-rate tariffs was a major factor behind the dramatic surge in Internet adoption that occurred around the world in the latter 1990s.15 Cable

Co-axial cables, used for cable TV, have also been used, together with cable modems, to provide two-way broadband communications. However, broadband-enabling traditional cable TV networks require significant investment, and this accounts for the fact that (a) broadband via cable modems has only

Introduction

emerged gradually, (b) in most countries ADSL accounts for a significantly larger proportion of broadband subscribers than cable (see Figure 1.1), and (c) why in some countries such as Germany (see Chapter 7), despite widespread availability of cable TV, broadband-via-cable has been very limited. Optical Fibre

Optical fibre is a technology developed in the 1970s.16 In this technology light (created by lasers) is used to send signals through fine glass fibres. Light travels much faster than the electrons that are used in copper cables, and light is not affected by electromagnetic interference, which can influence electrons. Furthermore, different colours of light (i.e., different light frequencies) can be used to send many signals simultaneously through the same optical fibre, thus increasing significantly the carrying capacity of that fibre.17 As a result of these advantages, optical fibre can provide extremely high bandwidth/speed. It is therefore the superior technology in terms of speed. However, optical fibre in the local access network suffers from two disadvantages. The first is that it is expensive to install, requiring in most countries that underground conduits are prepared. The second is that optical fibre is a fixed facility, lacking the greater flexibility and ubiquity of the wireless technologies that are discussed below. Fibre to the home (FTTH) refers to optical fibre connections that go all the way to the subscriber’s home or business. Fibre to the curb (FTTC) only goes to the curb, usually relying on the copper telephone line to complete the connection (perhaps using VDSL). Fibre to the node (FTTN) only goes to the node in the local network. (FTTX is the generic term, covering all these kinds of fibre connections.) As the superior technology, optical fibre has been widely installed in the trunk /core/ backbone network (i.e., the centralised part of the network with a high level of traffic aggregation). However, as a result of the expense, in most countries FTTX has not yet been widely deployed. As will be seen later, Japan (and to a lesser extent, Sweden) are significant exceptions to this generalisation. Indeed, in Japan in 2005 more than 8% of broadband connections were provided over FTTH. The Japanese incumbent, NTT, has provided FTTH to more than 80% of local exchanges, and FTTH is available in all municipal cities. The availability of FTTH in Japan is a significant contributor to the superior speeds achieved in Japan (shown in Tables 1.3 and 1.4). In the U.S. major regulatory debate has taken place over the incentives that are necessary to encourage investment by regional incumbent telecoms operators in FTTH (see Chapter 5).

17

martin fransman

18

Wireless

In 1896, 20 years after Alexander Graham Bell invented the telephone, Guglielmo Marconi developed the radio transmitter, which allowed him to transmit long-wave radio signals across the Atlantic in 1901. Although an old technology in widespread use, by the 1970s many researchers felt that as a way of transmitting communications signals, radio had serious drawbacks. Bell Laboratories went so far as to cut back its research on radio. While radio signals are subject to interference, optical fibre with its higher carrying capacity and speed and being less prone to disturbance was seen as the transmissions technology of the future. The development of cellular mobile communications in the 1970s, however, and its widespread adoption in the late-1990s, changed these perceptions.18 The technologists, it became apparent, had focused too narrowly on the techniques of transmission and had ignored the users. Users, it turned out, were willing to put up with interference and, on occasion, lack of signals in return for mobility. Cellular mobile soon provided not only mobility for the user but also a way for the user to be located. These advantages fuelled the rapid global diffusion of mobile communications from the late-1990s. By the early years of the new millennium, advanced second-generation (2G) and new third-generation (3G) cellular mobile systems were capable of offering broadband speeds.19 Particularly in Japan (through NTT DoCoMo’s i-mode service) mobile Internet access became very popular. Another set of wireless broadband technologies is wireless local area networks (LANs), popularly known as WiFi (Wireless Fidelity). This technology allows an over-the-air connection to be made between a radio-enabled client (e.g., laptop or phone) and a base station connected to the local access network (or another client). First discussed in 1990, this technology was developed by the Institute of Electrical and Electronic Engineering’s (IEEE’s) 802 committee, winning final approval in June 1997.20 Since the user can be mobile (within the footprint of a WiFi base station), some analysts believe that WiFi will become a serious competitor to cellular mobile communications. Most incumbent telecoms operators have developed WiFi hotspots (e.g., in airports, train and subway stations, and coffee houses) and have integrated them into their service offerings. A further wireless technology is WiMax, a fixed wireless access technology that covers a far wider area than WiFi. Some believe that WiMax will be able to provide wireless mobile coverage over entire city centres, also becoming a serious competitor to cellular mobile networks. WiMax is covered by the 802.16 standard.

Introduction Other Broadband Technologies

Other broadband technologies include satellite (particularly suited to some large, sparsely populated areas), powerline (broadband connectivity offered through ordinary electricity cables connecting homes and businesses to the electricity grid), and laser access (using low-powered laser beams). Driver 2: Competition between Technologies It is necessary to distinguish a further driver of the Broadband Era, competition between technologies. The point here, simply, is that technologies are not developed in isolation. Rather they interact with the developers of each competing technology paying particular attention to the performance parameters (e.g., speed) on the basis of which they compete. Competition between technologies, therefore, becomes an important driver; the prize for developers being substantial markets and profits if “their” technology is capable of delivering superior performances. For example, voice-over-the-Internet (VoIP) is competing with regular wireline and mobile telephony; Internet TV (IPTV) is competing with terrestrial analogue and digital and satellite TV; and all the access technologies mentioned in the last subsection are competing with one another. It is necessary to distinguish four types of competition: between technologies; between networks; between services; and between companies. The last of these is considered further in the following subsection. Driver 3: Competition between Companies Regulation

Left entirely to markets, there would be limited competition in the telecoms industry (barring competition stemming from radical, Schumpeterian technical change). There are two reasons for this. First, very high fixed costs (e.g., the investment costs of establishing networks) lead to economies of scale (the larger the firm, the greater its output and the lower its costs), which over time will result in either monopoly or oligopoly. Second, a further force pushing in the same direction is network externalities (where the benefit of membership of a network increases with the number of members of that network). For example, the greater the number of people on a phone network the greater the benefit for each member of the network because of the larger number of people who can be reached. Exactly the same reasoning applies to other cases, such as Windows or e-Bay users. The presence of substantial economies of

19

martin fransman

20

scale and network economies is likely to mean that markets will soon become monopolistic or oligopolistic. However, regulators, representing the interests of consumers and society as a whole (and playing their role in national political processes), have not been willing to leave outcomes entirely to market forces. Specifically, they have tended to intervene where there have been bottlenecks dominated by firms with significant market power. Examples in the broadband area are regulators’ measures regarding infrastructure competition and local loop unbundling (discussed in more detail below). In a number of countries, the UK and Korea being examples, regulators at the beginning of the Broadband Era hoped that infrastructural competition would emerge and speed broadband diffusion. In this way it was hoped that bottlenecks would be eliminated.21 In the UK, for instance, from 1991 to 1997 “regulation aimed principally to promote infrastructure competition, particularly in access.”22 However, in this hope regulators were often frustrated. As part of a fundamental Strategic Review of Telecommunications undertaken in 2004 (the most far-reaching since 1991), the regulator OFCOM concluded: “We believe that UK telecoms regulation has yet to overcome the problems of enduring economic bottlenecks and lack of equality of access to these bottlenecks.”23 Why has there been only limited infrastructural competition? OFCOM identifies two of the reasons: In many markets it has proved to be very hard for . . . competitors to overcome incumbency advantages. Often this is due to the cost economics of building competing infrastructure [such as economies of scale]. In some cases it may also be due to other features of the market, such as barriers to customers switching suppliers in numbers sufficient to make competitive provision economic.24

The global telecoms bust that occurred from mid-2000 made matters worse by sharply decreasing the funding that financial markets were willing to channel to non-incumbent telecoms network operators. In the face of such frustrations, regulators increasingly turned to a more modified form of infrastructural competition known as local loop unbundling. Local Loop Unbundling

One of the most important influences on the evolution of broadband access has been the unbundling of the local loop. Local loop unbundling (LLU) is a regulatory concept. The idea is to give new entrant telecoms operators access to the local network of the incumbent operator. LLU is a process whereby the incumbent’s exchange lines are physically disconnected from the incumbent’s network and connected to the new entrant’s network. This allows the new

Introduction

entrant, by installing its own equipment in the incumbent’s local exchange, to use the incumbent’s local loop to provide services directly to customers. A major benefit of LLU is that it facilitates a degree of facilities-based competition since it gives a new entrant an important degree of control over the local access network, allowing it to make its own differentiating innovations while at the same time having direct access to the final customer. This has given new entrants an incentive to compete with incumbents by differentiating their service to the final customer through innovation. However, it is also possible that countervailing costs have been incurred as the incumbent’s incentive to invest in upgrading the local loop has been diminished as a result of low-priced access25 to this network being given to its competitors by regulation. The key issue here is whether the regulated access price is sufficiently high to give the incumbent a reasonable rate of return on its investments in the local loop, while at the same time being sufficiently low to give new entrants a sufficient incentive to engage in innovative competition with the incumbent. As the last sentence makes clear, a delicate balancing act is needed to ensure sufficient incentives on both sides. As we will see later, this is a key issue in the case of expensive new local loops, such as fibre to the home. It is worth distinguishing local loop unbundling from two other regimes that deal with access to the final customer. The first regime involves the sale of wholesale products by the incumbent. Under this regime the incumbent retains its sole control over the local loop, but is required by the regulator to make one or more wholesale products available at regulated prices to its competitors who will compete with it in the retail market. This has been the dominant form of broadband access in countries like the UK. The second regime involves new entrants having full control of their own separate local networks, whether using the same technologies as the incumbent (intra-platform competition) or other alternative technologies (inter-platform competition). This regime gives the new entrant full control of its own facilities, without having to enter into transactions with the incumbent and regulator. This may be thought of as full facilities-based competition. Important questions arise regarding which of the three alternatives—unbundling, wholesale access, or full facilities-based competition—is the most efficient from a social cost-benefit point of view (i.e., for the country as a whole). While these questions will not be exhaustively examined here, two relevant points will be made. The first point is that while full facilities-based competition may be the most beneficial, it is also the most costly. From a cost-benefit point of view, therefore, it may not be the most efficient alternative. (One reason why it is likely to be the most beneficial is that it encourages competition, not only between companies

21

martin fransman

22

but also between different networks and technologies, thus leading to the improvement of these networks and technologies over time.) Second, by giving the new entrant full control over its own local access network, allowing it to compete through innovation, local loop unbundling may provide for more vigorous competition than the wholesale option. There is evidence (e.g., from the UK) that the wholesale alternative may be the least efficient of the three alternatives. The main problem is that while it does provide for strong competition in the retail broadband market, and while this can lead to significantly lower retail prices, it provides, as we have seen, little incentive for investment and innovation in networks and technologies, particularly where the incumbent retains de facto monopoly control over the wholesale market. Unbundling the Local Loop: A Cross-Country Comparison

Unbundling in Europe. Table 1.6 provides information on the proportion of broadband lines that are unbundled in the selected European countries covered in this book. In France and Germany the proportion—just over 11%— is similar. In the UK, however, the proportion is only 0.5%. As pointed out earlier, this reflects the fact that until recently most DSL broadband in the UK is provided on the basis of wholesale products provided by the incumbent, BT. In Sweden, the proportion is similar to that in France and Germany, and the same is roughly true for Italy. Unbundling Globally. For Germany and the UK, Table 1.7 confirms the data provided in Table 1.6. But it also allows a comparison to be made with the two Asian countries—Japan and Korea—and the U.S. Clearly, among the five countries included in Table 1.7, Japan and the UK are the two outlying cases. The relatively greater proportion of unbundled lines in Japan (operated by non-incumbents) is striking. Similarly, the minute number of unbundled lines in the UK also stands out. Also noteworthy is the similarity between the U.S. and Korea according to this measure. In both, unbundled lines controlled by competitors to the incumbent amount to about 15% of total lines while the incumbents control the remaining 85%. Germany, although significantly less than the U.S. and Korea, is far closer to them than to either Japan or the UK. Unbundling and Cross-Country Performance. What does this data on unbundling tell us about cross-national broadband performance? One possible hypothesis is that the greater the DSL market share (based on unbundling) held by competitors to the incumbent, the better that country’s performance. However, the evidence does not clearly support this hypothesis.

Introduction

23

table 1.6 Percentage of Broadband Lines Unbundled in Europe, March 2004 Country

Sweden France Germany Italy Britain

Lines Unbundled (%)

11.7 11.4 11.3 10.9 0.5

source: The Economist.

table 1.7 DSL Market Share of Incumbent and Competitors (Based on Local Loop Unbundling), 2003–2004 Country (Incumbent)

Germany (Deutsche Telekom) Korea (KT) Japan (NTT East & West) United Kingdom (BT) United States (Baby Bells)

Incumbent’s Share (%)

Competitors’ Share (%)

91

9

82 36

18 64

99 85

1 15

source: Author’s sources.

Take, for example, broadband penetration, shown in Figure 1.1. In terms of penetration, the countries included in this book fall into three groups. Korea is alone in the first group with the highest penetration by far. The second group includes Japan, Sweden, and the U.S. The European countries—France, the UK, and Germany—fall into the third (worst-performing) group. Relating unbundling (shown in Tables 1.6 and 1.7) to penetration performance (shown in Figure 1.1) raises two problems. First, although Korea and the U.S. have about the same proportion of lines unbundled (Table 1.7), Korea performs significantly better in terms of penetration. Second, although Japan has by far the largest proportion of unbundled lines (Table 1.7), it, too, is significantly outperformed by Korea on this measure.

martin fransman

24

Unbundling can also be related to performance in terms of speed (Tables 1.3 and 1.4) and price (Table 1.5). To summarise this data, Japan dominates performance in terms of speed, with Korea coming second. In terms of price, however, Korea dominates, with Japan coming last in the top 10 (although, as was pointed out earlier, since this data was collected there has been a broadband price war in Japan that has given Japan the world’s lowest prices). The problem, however, is that the Korean performance in speed and price cannot be explained by unbundling alone. As Table 1.7 shows, Korea and the U.S. have similar proportional unbundling (about 15%), but Korea significantly outperforms the U.S. in terms of speed (Tables 1.3 and 1.4) although not in terms of price (Table 1.5).26 Furthermore, while Japan has a substantially higher proportion of unbundled lines in the hands of nonincumbent competitors, it did not at the date reflected in this table outperform Korea in terms of price (Table 1.5), although it did in terms of speed (Tables 1.3 and 1.4). Conclusion. Regarding the link between unbundling and performance, therefore, we may conclude that definitive evidence that greater proportional unbundling leads to better overall performance does not exist. This is not to say, however, that unbundling has no significant effect on performance. Rather, unbundling by itself is not a definitive determinant of performance. Other factors are also at work. The point is that in providing cross-country explanations of broadband performance a number of factors need to be taken into account of which unbundling may be one. Disruptive Competitors

As will be seen in more detail later in this chapter, the activities of “disruptive competitors” are an important determinant of global broadband performance. Disruptive competition may be defined as existing when competitors to the incumbent have been so aggressive with their pricing that they do not cover their costs and end up making short-run losses. Their hope is in this way to gain market share and possibly force the exit of some of their competitors. In the longer term they hope they will be able to price more profitably. Later we will return to disruptive competitors in explaining the three performance puzzles.

Introduction

Driver 4: The Co-evolution of Demand Demand is always a key determinant of the evolution of any industry. In the case of broadband, however, demand appears to be playing different roles in the countries examined in this book. In most of the countries there appear to be few broadband-specific services that are driving adoption of broadband services. Certainly, there are no broadband-specific “killer applications” driving broadband adoption. For most users in the mass market (residential subscribers and small businesses; i.e., excluding large businesses), e-mail remains the most important service, as was the case with narrowband dial-up. In general, most users spend most of their time using their broadband Internet connections to do much the same as they did with their narrowband (see individual chapters). The three most important characteristics influencing broadband adoption are (1) always-on functionality, (2) the separation of voice and data channels (which allows the user to use the phone while also using the Internet), and (3) the higher speed of a broadband connection. In terms of services, peer-to-peer services, specifically music downloads, have been significantly enhanced by broadband, and video-intensive applications and services will become increasingly important. However, these services have not yet become the main drivers of broadband adoption. From the demand perspective, Korea and Japan appear to be somewhat different from the other countries. In Korea (see Chapter 3), online gaming was an important application that helped to kick-start the country’s unequalled penetration. In Japan, to a greater extent and earlier than in other countries, voice-over-the-Internet (VoIP) has played an important role. In China, as shown in Chapter 4, attempts have been made to co-ordinate efforts to create broadband-specific content, applications, and services. It is hoped that this will encourage greater broadband uptake. Voice-over-the-Internet (VoIP)

VoIP was first offered commercially in 1995 when an Israeli start-up, Vocaltec, produced the software that would allow PC users to speak to each other over the Internet (the requisite software could be e-mailed to another would-be user). Although reception was originally poor (voice packets would fail to arrive quickly enough), the technology gradually improved, and today with a broadband connection VoIP is competitive with conventional telephony. In Japan, VoIP was offered free to broadband subscribers with the same operator, making this service a killer of traditional voice revenues as much as a broadband killer application, that is, VoIP has become a disruptive service,

25

martin fransman

26

table 1.8 VoIP Diffusion, February 2005 Country

France Germany Japan United Kingdom United States

Number of VoIP Users

220,000 110,000 4,900,000 50,000 1,000,000

source: European Commission, quoted in The Guardian, 12 February 2005.

posing a serious threat to telecoms operators who still receive a substantial proportion of their revenue and profits from voice services. Take-up of VoIP in several countries in early 2005 is shown in Table 1.8. Japan’s lead in this area is striking. Internet TV, Internet Video, Triple Play, and Home Gateways

Perhaps the most significant potential of broadband, one still to be realised, is that it allows the information society to include the visual society. Increasing broadband speeds allow the development and consumption of videorich contents, applications, and services. Putting visuality together with one of the most important characteristics of the Internet, namely interactivity, creates an important potential that constitutes the still-unrealised promise of broadband. One of the trends in the broadband market has been for providers to bundle several services in order to offer consumers triple play, that is, voice telephone, broadband Internet access, and TV/video. In the U.S., for example, both telecoms operators and cable TV companies are competing on the basis of triple play, and this is one of the drivers of the adoption of FTTH by regional incumbents such as Verizon and SBC (see Chapter 5). In France, broadband competition between the incumbent France Telecom and new competitors like Iliad/ Free and Neuf is leading to the evolution of a home gateway that will create another important potential, namely the fully networked home and small business. Home gateway—essentially a set-top box that links to an ADSL line—will integrate the three main information and communication platforms in the home: the phone, the PC, and the TV. In addition, it will also integrate other networked appliances (e.g., games consoles, audiovisual equipment, heating and security systems, and even refrigerators and microwave ovens). This will provide a new battleground for those companies contesting the home and small business market. This terrain has long been the battlefield for

Introduction

consumer electronics companies—originally in the area of television, audio, video, and games—and computer hardware and software companies focused mainly on the PC. The advent of broadband will allow telecoms companies (operators and equipment suppliers) to join the fray using their distinctive competencies in reliable networking and the provision of connectivity. Driver 5: Government Leaving aside regulation (already discussed), government—both national and local—can also play an important role influencing broadband diffusion. However, there appear to be some significant differences between countries regarding government’s actual role. Generally, all the governments covered in this book have put a high priority on improving their broadband infrastructure and services. And all have taken measures aimed at broadband-enabled improvements, for example in education, health, and e-government (i.e., the improvement of government functioning through the use of electronic communications). However, the most noticeable differences between countries seem to be the willingness of national governments to play a direct role in financing broadband infrastructural investment. At the local level, however (i.e., regional and municipal), there appears to be a more uniform willingness to intervene to improve broadband infrastructure and access. The role of government will be analysed in more detail in the section “Explaining the Puzzles 2: Country Differences,” which deals with the three puzzles identified at the outset of this chapter. Driver 6: Geography, Demography, and the Digital Divide Geography and demography are major drivers of broadband diffusion. The reasons are obvious. In urban areas, and areas of concentrated population within urban areas (such as apartment blocks), the costs of broadband installation per household are lower. The concentration of population, particularly in Seoul, is an important part of the explanation of Korea’s high penetration (see Chapter 3). But the same logic applies to all countries (as shown in the other country chapters). Furthermore, within the population it is younger adults who are earlier adopters, as with the Internet generally. In rural, sparsely populated areas, conversely, broadband is more costly and the degree of competition between broadband suppliers far lower. Partly as a result, technologies—such as wireless local loop and satellite—that are relatively more efficient within this context, tend to be more prevalent. These

27

martin fransman

28

issues raise important questions about the so-called digital divide (i.e., the division between the digital haves and have-nots), social inclusion, and the universal service obligations of broadband operators. explaining the puzzles 2: country differences The Japan/Korea Puzzle Why do these two Asian countries lead the world in broadband? Their industries are not globally dominant in areas such as the Internet, computing, software, and microprocessors. Furthermore, since 1989 the Japanese economy has been in a state of recession and Korea, although growing rapidly, is not yet one of the richest countries. Strong, Innovative Incumbents

A necessary condition for the Japanese and Korean success in broadband has been the existence of strong and innovative incumbents. Their countrywide networks have provided ADSL-based broadband for the majority of broadband subscribers in their countries. Furthermore, together with facilitating regulation (analysed in more detail below), their networks have also been used by new entrants competing with them. NTT, the Japanese incumbent, is a good case in point.27 As shown in Table 1.3 earlier, NTT leads the world’s incumbents in terms of speed and, more recently, price. As a result of NTT’s networks, subscribers in Japan’s major cities are routinely offered broadband speeds of 100 Mbps over optical fibre connections while their U.S. and European counterparts rarely get more than 5 Mbps over copper. Part of the reason for this impressive performance lies in NTT’s R&D capability. Working closely with leading equipment vendors (both Japanese and non-Japanese), NTT has consistently given a significantly higher priority to R&D than its Western counterparts. This is evident in NTT’s relative R&D intensity. While NTT allocates about 3.5% of its sales to R&D, the equivalent figure for BT is about 2%, and it is even lower for AT&T. Since the late-1990s NTT has been committed through its research-led activities to a VIP (visual, intelligent, and personal) vision of the future of telecoms, a vision that will finally be realised in the Broadband Era. Behind the rapid increase in broadband speeds in Japan lie the activities of NTT’s R&D engineers working handin-hand with the company’s equipment vendors. A similar story is true of KT, the Korean incumbent. Although lacking NTT’s size and R&D capability, KT has also been necessary for Korea’s broadband performance.

Introduction

However, a further part of the Japanese and Korean stories has been the intense competitive pressure that has been exerted on these strong and innovative incumbents by regulation-induced disruptive competitors. Riding on the back of the incumbents’ networks—a ride made possible by facilitating regulation—these competitors have forced the incumbents to respond. And the incumbents have done so by improving availability, penetration, speed, and price—precisely the indicators of broadband performance. However, as pointed out in the final section in this chapter, questions remain regarding whether the incumbents have been pushed too far in the regulator’s eagerness to improve the broadband performance of their countries. Disruptive Competitors

One of the main theses of this book is that a major reason for the superior broadband performance of Japan and Korea is the activities of disruptive competitors (defined earlier in this chapter). The activities of these disruptive competitors have resulted in a relatively greater intensity of competition in the broadband market in Japan and Korea than in the other countries examined here. It is acknowledged that this thesis may well be controversial. It certainly requires further research in order to refute or support it. In this section more detail will be provided on the main disruptive competitors in Japan and Korea, Yahoo! BB and Hanaro, respectively. Yahoo! BB. Yahoo! BB has been the main disruptive competitor in the broadband market in Japan, competing vigorously against the incumbent, NTT. Other competitors, also aggressive, include companies such as KDDI (the former incumbent international telecoms operator), eAccess, and cable TV and electricity companies. In 2005 NTT and Yahoo! BB each controlled approximately one-third of the total Japanese broadband market. Yahoo! BB’s founder, Masayoshi Son, is Japanese-born and of Korean descent. He made his first million dollars before graduating from the University of California, Berkeley. In 1981 at the age of 23 he established SoftBank, a company that began by distributing PC software in Japan largely from Microsoft. With the massive growth of the Internet from the mid-1990s, Son developed his company’s strategy around it. In January 1996 Son established the Yahoo Japan Corporation, a jointly owned venture between SoftBank and Yahoo! of the U.S. The initial aim of Yahoo Japan was to provide an Internet information search service. In June 2001 the start of broadband services provided by Yahoo! BB was announced, a joint venture between SoftBank and Yahoo Japan. In April 2002 Yahoo! BB began BB Phone, a broadband phone service based on VoIP.

29

martin fransman

30

SoftBank was one of the major beneficiaries of the dot.com boom in Japan. However, with the dot.com bust in 2000 the company was also one of its main casualties. Nevertheless, Son has retained his belief in the money-making potential of Internet-related businesses and with the coming of the Broadband Era he has articulated a strategy that makes broadband the common threat tying his many Internet-related companies and activities together. From the start Son assumed an aggressive role and quickly became regarded as something of a maverick. However, it was Japan’s government that created the main opening for Son’s broadband activities. In July 2000 the Ministry of Communications (MIC) ordered the incumbent NTT to open all its exchanges for the co-location of equipment owned by competitors. In this way local loop unbundling (LLU) became an important influence shaping the evolution of the Japanese broadband market. Furthermore, MIC enforced very low prices for LLU access by competitors to NTT, arguably even below NTT’s costs (see below for further details). These government measures considerably eased the entry of Yahoo! BB into the broadband market. However, justifying the appellation “disruptive,” Son, assisted by low LLU charges, aggressively priced his broadband access services. At one stage free modems were given to new subscribers, signed up outside subway stations.28 In addition, Son was innovative, aggressively marketing the VoIP broadband phone service that provided free calls to other Yahoo! BB subscribers. The intention behind these moves was to rapidly increase market share and drive weaker competitors out of the market in the hope that in time more profitable pricing could be established. More recently Son has made some significant moves to own and expand his own telecoms infrastructure. In 2004 he purchased the fixed network of Japan Telecom.29 And in the same year he bought Cable and Wireless’ IDC network.30 Son believes that these assets will put him in a better position to tackle NTT. However, some analysts have their doubts about their value. With a broadband market share roughly the same as the incumbent, NTT, Son has succeeded in his market share objective. However, the cost of this disruptive approach is evident in the continuing losses that Yahoo! BB has made since its inception, although the losses have been declining. SoftBank as a whole has also continued to make losses, largely as a result of Son’s broadband ventures, though they, too, are declining. Some have remained unconvinced by this strategy. In August 2004 the editorial Lex column in the Financial Times contained the following comment about SoftBank: SoftBank has more than a whisper of the devil-may-care attitude about it—it is cavalier about borrowing and spending, as witnessed by its $7bn of interest-bearing debts

Introduction and four years in the red. It appears to have more attachment to grandiose visions than predictable earnings. Investors are left to draw consolation from the 43 per cent stake in Yahoo Japan, on which SoftBank is sitting on an unrealised $15.5bn gain.”31

Faced with such disruptive competition, the incumbent, NTT, has been forced to respond by significantly lowering broadband access prices and substantially increasing speeds, in part by an accelerated programme of fibre to the home (FTTH).32 It is this process, also impacting on the other broadband competitors, that in large part lies behind Japan’s superior relative global performance in price and speed. However, questions need to be addressed in determining whether this superior performance in the short term will also result in superiority in the medium to long term. This important issue is tackled in the final section in this chapter. Hanaro. On 5 March 1997 the Korean Ministry of Information and Communications (MIC) announced application procedures for selecting a new local telecoms service provider to challenge the monopolist incumbent, KT.33 (As in Japan, the ministry also acts as regulator of the telecoms industry.) KT was established as an independent company in 1982, although the government kept full ownership until 1993. In May 2002 KT was fully divested. Seven Korean conglomerates (chaebol in Korean) agreed to create a company to compete with KT. On 13 June 1997 the Ministry selected this company, Hanaro Telecom, as Korea’s second local telecom service provider. YunSik Shin was chosen as its CEO. In April 1999 Hanaro launched its broadband Internet and local phone service. However, Hanaro was not the first to offer broadband in Korea. That honour is held by another new entrant, Thrunet, which began commercial broadband services using cable modems in July 1998. Also in April 1999, the same year that Hanaro began, the incumbent, KT, started its ADSL-based broadband service. KT was slow to enter the ADSL broadband market for fear of cannibalising its ISDN (integrated services digital network) service. Significantly, infrastructural competition launched the broadband market in Korea. Both Thrunet and Hanaro at first depended on cable TV networks for their broadband services. It was only in 2002 that MIC introduced local loop unbundling (LLU). According to Inho Chung (see Chapter 3), “It is not clear whether or not lateness of the introduction of LLU . . . was the intended result of government policy for facilities-based competition.” He suggests that LLU may have been introduced in 2002 to prevent excessive investment in broadband infrastructure and to assist the diffusion of broadband in sparsely populated rural areas where there is less competition.

31

martin fransman

32

Significantly, as Chung points out, MIC also intervened as an act of omission rather than commission. While in some areas of telecoms MIC imposed a floor on KT’s pricing, protecting the new entrants temporarily against predatory pricing from the incumbent KT and giving them time to establish themselves before facing the harsh winds of full competition (i.e., a kind of infant industry protection), this did not happen in broadband. And KT responded, positively embracing ADSL, after seeing the new entrants make rapid progress in the market. In turn, Hanaro and Thrunet responded aggressively with lower prices, stimulating the market even further. Chung suggests there was a further demand-side factor that entered into the calculations. This was the slow growth of the ISDN market for high-speed Internet access. Both incumbent and new entrants would have had an incentive to establish low prices to compete with ISDN and to see the market grow. At any rate, however the process was caused, the new entrants had little option but to propagate low broadband prices. As things turned out, this was to have seriously negative consequences for both Hanaro and Thrunet. Things came to a head in 2003. The causes were both intense price competition as well as a saturation of broadband demand (the other side of the high Korean penetration coin). In this year Hanaro suffered a significant deterioration in its performance. The company’s financial position was so serious that it had to consider the possibility of seeking foreign resources, even though this would compromise its independence. This dilemma led to divisions in the company. On 3 July 2003 the Board of Directors voted down plans to attract foreign capital. But the financial pressure proved to be too great. On 9 September 2003 the company signed a contract with an American consortium for the sum of $1.1 billion. The consortium was led by Newbridge Capital and the largest American insurance company, American International Group (AIG). They acquired a 40% controlling stake in Hanaro in exchange for cash and loan guarantees. Korea’s LG Group, formerly the largest shareholder in Hanaro, holding 18% of it shares, had opposed the deal and formed its own alliance with the Carlyle Group in an unsuccessful attempt to counter the American bid. (LG is the owner of Dacom, the third Korean fixed telecoms supplier.) Thrunet, the first provider of broadband in Korea, fared little better. In early 2003 the company went bankrupt. In February 2005 it was bought out of receivership by Hanaro for $460 million, which triumphed against Dacom, who had also put in a bid. Hanaro enjoyed a landmark year in 2004. For the first time since its founding in 1999, the company earned a net profit. In early 2005 Hanaro (before its acquisition of Thrunet) controlled 24% of Korea’s broadband market compared to KT’s 51%.

Introduction

The process of intense competition outlined in this subsection is one of the main determinants behind Korea’s impressive relative global performance in broadband. However, as in the Japanese case, whether this good short-term performance can be turned into equally good medium- to long-term performance requires further discussion. This is provided in the final section in this chapter. Strong Regulation

There are indications that telecoms regulators in Japan and possibly Korea have tended to take a stronger line with their incumbents than in Western countries. Unfortunately, however, there are no straightforward indicators of the strength of regulation34 with the result that we have no option but to rely on indirect evidence. One instance of such evidence is the rare occasion when regulators compare themselves to other regulators. A case in point is the British regulator OFCOM, which compared Japan and the UK in undertaking its “strategic review of telecommunications” in 2004. With reference to the early period of Japanese telecoms liberalisation in the mid-1980s, OFCOM concludes that “In the early stages after privatisation [the Japanese ministry of communications] adopted what can be considered a more interventionist approach to regulation of NTT when compared to the UK, with direct setting of prices and returns for key products.”35 The battles between the Japanese ministry of communications and NTT, from the beginning when NTT became a private company (though with a majority government shareholding) in 1984, are legendary. These battles included several attempts made by the ministry to break NTT up, efforts that were thwarted after a long and complicated struggle involving many other important players in the Japanese political process.36 In the end a compromise was reached. NTT was transformed into a holding company (including the research and development laboratories) that controlled the group’s five major businesses—NTT East, NTT West (the regional fixed-line companies), NTT DoCoMo (the mobile company), NTT Communications (the long-distance and international company), and NTT Data (the information services company)—and its many subsidiaries. With the coming of the Broadband Era, OFCOM provides its own reconstruction of what happened in Japan: Despite the restructuring NTT remained unenthusiastic about introducing DSL to its customers and opened only a small number of exchanges for co-location with heavy limitations on equipment space and the number of lines available for un-

33

martin fransman

34

bundling. However, MIC [the ministry of communications] was swift to intervene and in July 2000 ordered NTT to open up all exchanges for co-location and to lift limitations on rack space. The effects were quickly felt as SMEs [including firms such as Yahoo! BB and eAccess] took advantage of cheaper prices.37

Furthermore, the ministry insisted on imposing a very low unbundling access charge on NTT (the price that NTT is allowed to charge its competitors for access to its local loop in order to reach its customers). OFCOM explains how the ministry justified such a low cost: Prices for LLU [local loop unbundling] in Japan are now the lowest in the world. OFCOM understands from discussions with MIC that it believes that NTT had already amortised the cost of the “last mile” of copper assets which is being paid for by the consumer in basic line rental and call charges. LLU costs in Japan are therefore only the new net costs of the bypass of the PSTN [public switched telephony network] switch. . . .38

How does this tough, interventionist approach in Japan compare with that adopted by the regulator in the UK? OFCOM does not deal with this question explicitly. However, an answer can be gleaned by looking elsewhere in its report. One section of OFCOM’s strategic review is titled “Achieving equality of access.” This refers to access to the incumbent’s network provided by BT Wholesale. Access is given both to BT Retail as well as its competitors. The issue that OFCOM addresses both here and elsewhere in its strategic review is the concern that BT has failed to give equal access to its competitors. One of OFCOM’s conclusions on this issue is worth quoting in full: Competition has delivered very substantial benefits to consumers in the last twenty years; for example, in terms of much lower prices and enhanced choice. But the clear consensus of the responses to Phase 1 [of OFCOM’s strategic review] was that even though substantial effort has been focused on it over the last twenty years, the problem of lack of equality of access has yet to be resolved. For example, [Cable and Wireless] argued that: In the world of broadband, BT was allowed to create an LLU [local loop unbundling] product which was prohibitively expensive, not industrialised and not fit-for-purpose, which meant that it was entirely unsuitable for mass-market takeup. The result is that there is currently virtually no competition in broadband based on LLU. We believe that similar stories could be told about carrier pre-selection, wholesale line rental, partial private circuits, and indirect access in their early days.39

As shown in Table 1.6, Britain was alone amongst the large European countries in having a negligible amount of unbundling (although OFCOM has since given high priority to LLU). OFCOM’s conclusion begs the question

Introduction

regarding why in this respect regulation has been so weak in the UK. In fact the question is raised, not about OFCOM itself, but about its predecessor regulator, Oftel, which regulated British telecoms through most of this period. It was only in 2004 that OFCOM took over as a “super-regulator,” assuming Oftel’s responsibilities in addition to other regulatory duties in areas that include spectrum allocation and broadcasting. More light on this issue is thrown by the head of Oftel who had responsibility for the regulation of broadband (and the rest of telecoms) in Britain. In a remarkably frank interview with the Financial Times, David Edmonds, Oftel’s Director General, admitted that he was himself responsible for some serious mistakes. In his own words: “If I knew then what I know now I would have handled local loop unbundling differently. I should have realised earlier that BT was playing a long game. My hope was that the industry would work it out for itself.”40 According to the Financial Times, “he said the lesson from these mistakes was to be more ‘directive.’” The contrast between the Japanese and British approaches, evident from these quotations, is substantial. But the evidence provided in this book suggests that some of the other major Western countries have also taken a weaker approach in regulating their incumbents compared to Japan. The U.S. chapter, for example, shows that the FCC changed tack significantly in 2003 when it made important changes in the unbundling obligations imposed on the regional incumbents, easing the requirements and therefore adding to the costs imposed on competitors. The German chapter shows that the incumbent, Deutsche Telekom, commanded 88% of the country’s broadband market, far higher than in any other country included in this study. This implies that in Germany, too, regulation was far weaker than in Japan. In concluding this subsection, two further points must be made. The first, noted in the earlier general discussion of local loop unbundling, is that it is incorrect to equate toughness of regulation with increased net benefit to society. The regulator needs to strike a balance. For example, if unbundling regulation is too tough the incumbent may not have a sufficient incentive to invest in upgrading the local loop. On the other hand, if it is not tough enough competitors may suffer and there may not be sufficient intensity of competition. Debate continues in the U.S. and other countries regarding whether the right balance has been struck. The second point is that the present discussion raises the very difficult issue of the political process of regulation and its determinants, an area that is very much understudied (most analytical studies being devoted to what regulators should do, rather than to what they are doing and why). Part of the problem in

35

martin fransman

36

studying this process is that many of the significant things that happen do so behind closed doors, despite the avowed obligations of Western regulators to be transparent. But some cross-country institutional differences are apparent. In Japan and Korea, for example, the regulatory function is performed by the ministry of communications that not only has responsibility for making general communications policy but also has a significant shareholding in the incumbent (though the latter is in the process of reducing). In Western countries, on the other hand, the approach has been to institute a regulatory body with specialist regulatory obligations and with an arm’slength relationship to the ministries with responsibility for communications (e.g., the FCC, OFCOM, ART, and RegTP in the U.S., UK, France, and Germany, respectively). Since the regulatory function in Japan and Korea is directly embedded in the ministerial bureaucratic and attendant political process, the resulting tendency is for telecoms regulation and broader ministry communications policies to be more closely aligned than is usual in Western countries. A further consequence is that regulation tends to be less transparent than in the West. However, the politics of regulation in the American FCC (discussed in Chapter 5), the role played by party politicians in the U.S. (in contrast to the usual European case where regulation is more the function of non-political party functionaries), and the present discussion of the British case serve as a reminder of the political intricacies that are also part and parcel of Western regulatory processes. Direct Government Intervention

Particularly in Korea (see Chapter 3), government has played a direct, wellplanned, and decisive role in enabling the development of broadband infrastructure. To a large extent this has been through the provision of funding, including subsidies of various kinds. But it has also involved other forms of government enablement (e.g., the provision of building certification, which has encouraged broadband provision). Unfortunately, cross-country data on government funding for broadband (infrastructure and services) are not available (partly because of the complex forms and programmes of government intervention and the different levels of government involved). However, the strategic review of the British regulator OFCOM offers some insights into differences regarding the role of government that it perceived. In the case of Korea, OFCOM concludes that: There is no doubt that the pro-active approach of the Korean government in promoting investment in broadband infrastructure has been a key reason for the stag-

Introduction gering success of Korea in terms of broadband penetration levels. What is less clear is whether such levels of intervention would be practical [possible?—MF] in the UK.”41

Regarding Japan, OFCOM’s conclusions on the direct role of government (excepting its regulatory role) are more nuanced. In the case of optical fibre, for example, OFCOM concludes that the direct effect of the tax incentives and subsidies given since 1995 “remain unclear.” Regarding the early roll-out of fibre to the home, it concludes that “Tax and fiscal incentives may have played a small part but regulatory policy appears to be a more significant factor.”42 Sweden is the only Western country included in this book where government has played a comparable role to that in Korea. As the Swedish chapter shows, government in Sweden—at the national, regional, and municipal levels—has played a crucial role in facilitating the development of broadband infrastructure in the country. Demand

Although Japan and Korea, like their Western counterparts, have not had broadband-specific killer applications that have driven broadband diffusion, demand for broadband-related services seems to have played a more significant role in these Asian countries than in their Western counterparts. A specific example is voice-over-the-Internet (VoIP), which, as we saw in the earlier discussion of NTT and Yahoo! BB, has been an important driver of broadband take-up in Japan. The Korean chapter shows that interactive computer games, education, and the development of a significant number of Korean-language web sites have all made an important contribution to broadband penetration. By contrast, the chapters on Western countries seem to suggest that broadband-specific services have been somewhat less important in these countries and that to a large extent Western broadband users spend most of their time doing much the same with their broadband Internet connections that they did with narrowband. However, more research is needed on this intriguing observation. Geography and Demography

There is little doubt that geography and demography are also explanatory factors accounting for the superior performance of Japan and Korea. The general points were made in the earlier discussion of this issue (see Driver 6) and will not be repeated here.

37

martin fransman

38

The U.S. Puzzle Why is it that the U.S.—which gave birth to the Internet and leads the world in areas such as computing, software, and microprocessors—is lagging significantly behind in broadband? In 2004 an FCC commissioner complained that the U.S. was ranked eleventh in one of the key broadband league tables (see Figure 1.1). Intensity of Competition

In explaining the U.S. puzzle, one hypothesis that will be discussed in this subsection is that the intensity of competition has tended to be lower in the U.S. than in Japan and Korea. It must be stressed, however, that this hypothesis needs further research in order to refute or support it. Unfortunately, there is no straightforward operational measure of intensity of competition that can be readily applied across countries. This means that it will be necessary to rely on indirect evidence, with the danger that there may be ambiguity and therefore disagreement regarding the inferences that are to be drawn from the evidence. There are three sets of factors that might be taken to support the above hypothesis. The first is that in the U.S. the former regional Bell companies (e.g., Verizon, SBC, Bell South) seem to have been relatively successful (using both judicial and political processes) in limiting attempts made, notably in the 1996 Telecommunications Act, to increase the intensity of competition facing them. Evidence supporting this contention comes in the regulatory changes (and the debates accompanying them) made in 2003 (and documented in the U.S. chapter). These changes reversed some of the earlier regulatory measures requiring these companies to offer unbundled access to their competitors at long-run incremental cost (LRIC). It seems clear (see Chapter 5) that these reverses have had the effect of reducing the intensity of competition facing these regional companies that are the main providers of DSL-broadband in the U.S. However, as also noted in Chapter 5, the same measures may also have increased the incentive to invest—and the ultimate amount invested—in broadband. A particularly important instance is FTTX, where Verizon and SBC have recently announced large investment programmes. Second, the U.S. does have more broadband competition between telecoms operators (using mainly ADSL) and cable TV companies (using cable modems) than almost all other countries (see Figure 1.1). Furthermore, this competition is having significant effects. Again, a case in point is the investment in FTTX just referred to (although the U.S. chapter also notes that these

Introduction

investments have been scaled back with the reduction in regulatory pressure since 2003). However, this competition is oligopolistic (competition amongst a few dominant companies) with all the limits on intensity that this form of competition implies. Both sets of competitors—telecoms operators on the one hand and cable TV companies on the other—are dominant incumbents in their own traditional markets, telephony and cable broadcasting, respectively. With the coming of convergent networks, technologies, and services, they are beginning to compete increasingly in the newly converged markets. However, the competition is still that between giants, newly testing each other’s territories, rather than the kind of disruptive competition that, as was shown in the previous section, emerged in Japan and Korea. Thus far there seems to be little indication in the profitability of these two groups of U.S. antagonists that truly intense competition exists. As an aside, in the U.S., it is worth noting, a distinction needs to be drawn between the long-distance telecoms companies—notably AT&T, MCI / WorldCom, and Sprint—and the former regional Bell companies (like Verizon and SBC). The former have been subjected to intense competition in long-distance and international markets. Partly as a result of this competition these companies have suffered significantly: AT&T is being taken over by SBC; MCI / WorldCom went infamously bankrupt and at the time of writing there is a battle between Verizon and Qwest to acquire this company. The former regional Bell companies, as just noted, have been able to avoid the same intensity of competition. Third, in Japan and Korea, by contrast, there is evidence that the incumbents have been subjected to intense competition. There are two indicators of this intensity that were highlighted. The first is the substantial losses that have been made by Yahoo! BB and Hanaro. These losses were documented in the last section. The second indicator is market share. This indicator needs to be taken together with the first. This reveals that the market share of the competitors was won, not through normal oligopolistic competition, but through what has been termed in this book “disruptive competition” where short-run losses are made in the interests of longer term market share and profitability. U.S. market share data for the DSL broadband market—which accounts for 37.4% of the total broadband market, 57.6% coming from cable—is provided in the U.S. chapter. The following conclusions, with a bearing on the intensity of competition, emerge from this exhibit and the related discussion in the U.S. chapter: 1. The five largest cable and telecom operator companies held 63.4% of the U.S. broadband access market. All of these companies may be regarded as incum-

39

martin fransman

40

bents. [The companies are, in order of market share, Comcast (20.1%), SBC (14.4%), Time Warner (11.4%), Verizon (10.0%) and Cox (7.5%).] 2. The top 10 broadband providers (all cable and telecom operators) held 83% of the market. 3. The top 18 providers (all cable and telecom operators) held 95% of the market. 4. Amongst the top-ranked companies only three new entrants were founded since the mid-1990s. [These were Covad (1.6%), Bright House Networks (2.1%) and Mediacom (1.1%).]

In Japan, by comparison, the incumbent, NTT, holds about 33% of the broadband access market while its disruptive competitor, Yahoo! BB, also has about 33%. In Korea, the incumbent, KT, has 51% of the market, while the disruptive competitor it faces has (after its acquisition of Thrunet) between 30% and 40% of the market. These new entrant companies were founded in 2001 and 1997, respectively. The argument according to the intensity-of-competition hypothesis, based on the evidence presented in this subsection, is that the superior broadband performance in Japan and Korea relative to that of the U.S. is to be explained by a higher intensity of competition. As a result of this more intense competition in Japan and Korea, both their incumbents and the rivals that compete with them have responded by improving availability, penetration, speed, and price to a greater extent than have their U.S. counterparts. However, to repeat, further research is needed to put this hypothetical conclusion on a firmer footing. Regulation

Read this subsection in conjunction with the subsection titled “Strong Regulation” in the Japan/Korea Puzzle section. There are indications that broadband regulation in the U.S., regarding both access to the existing network of the incumbents and the regulation of new optical fibre networks, is relatively weaker compared to Japan and Korea. This emerges from the Triennial Review of the unbundling regulations relating to incumbent telecoms operators that the FCC announced it would conduct in 2001 (adopted in February 2003 and released in August 2003). In the U.S., “much regulatory focus since 1996 has been devoted to unbundling elements of the [former regional Bell companies’] networks to create conditions for intra-platform competition,”43 even though the then Chairman of the FCC, Michael Powell, made it repeatedly clear that his favoured overall solution was facilities-based competition. Following the Triennial Review, a majority of the FCC’s commissioners agreed that in order to encourage investment in new optical fibre access networks the commission would

Introduction

“forbear”—that is, refrain from regulating—for a period of five years. In making this decision they explicitly assumed that there would continue to be interplatform competition from other networks and technologies. This regulatory absence of regulation, it was hoped, would provide sufficient incentive for the incumbents to invest in FTTX. Consensus within the commission broke down over the question of whether existing rules regarding unbundling and line-sharing44 imposed on the incumbents should also be removed (see Chapter 5). The majority view was that these rules should be steadily withdrawn. This view was opposed by Chairman Powell, however, who argued that the existing line-sharing rules encouraged facilities-based competition. In his view, removal of the rules would reduce the intensity of competition and increase prices.45 This brief story of changing regulation in the U.S. regarding access by competitors to the new and existing networks of incumbents may be compared with the stories of similar regulation in Japan and Korea discussed in the last section. The comparison indicates fairly clearly that a far stronger line was taken with the incumbents by the Asian regulators than by their American counterparts (and by Oftel in the UK as discussed above). But does this provide sufficient evidence to warrant a conclusion that broadband regulation has been “stronger” in Japan and Korea than in the U.S.? (As suggested in the last section, a strong regulator may be defined as one that is able to formulate and enforce regulations that it judges are in the public interest, even when they conflict seriously with the interests of the dominant companies negatively affected.) Ideology, vested interests, and politics being what they are, it is likely that at least two responses will be constructed in response to this question. The first response, in line with the views of Chairman Powell, is that broadband regulation has indeed been stronger in Japan and Korea and that the 2003 FCC decision represented a reduction in the intensity of competition in the U.S. Further evidence supporting this response is that, up to the time of writing, the Japanese regulator has insisted on applying the same unbundling rules on access to the incumbent’s new optical fibre access network as apply to the existing copper network (discussed above), although this approach is being contested. The second response might be that although the unbundling rules in the U.S. have been relaxed, intensity of competition (at least equal to that in Japan and Korea) will be ensured, first, by the growing strength of competition between telecoms operators and cable TV companies and, second, by competition from future new entrants using alternative technologies and networks (outlined above). This response is inconsistent with the hypothesis previously put forward suggesting that disruptive competition in Japan and Korea has re-

41

martin fransman

42

sulted in a greater intensity of competition than oligopolistic competition between telecoms and cable companies in the U.S.. (For the foreseeable future, serious across-the-board competition is unlikely to come from competitors to the incumbents using new technologies as alternatives to DSL /copper, cable modems, and optical fibre. This conclusion may be deduced from Figure 1.1.) On the basis of the evidence presented in this chapter, however, it is my view that regulation has indeed been stronger in Japan and Korea than in the U.S. (and Europe—see below). However, it is accepted that this is a contestable conclusion. But even if the conclusion is correct, it still remains for further research to explain the complex politics of regulation and the vested interests that lie behind these politics. As the British regulator OFCOM noted in its strategic review of telecommunications, the U.S. regulatory structure is “complex and legalistic.”46 A deeper understanding of this phenomenon is still needed. New Technologies and New Entrants

As noted in the “Intensity of Competition” subsection, only three new entrants into the U.S. broadband market were founded since the mid-1990s. These were Covad, Bright House Networks, and Mediacom, and together they held only 4.8% of the market. This contrasts with the situation in Japan and Korea where the main competitors to the incumbents—Yahoo! BB and Hanaro, respectively—each held about one-third of the market. This prompts brief reflection on some of the differences between the broadband market in the U.S. and other parts of the American information and communications sector. Entrepreneurial new entrants, and ways of financing them, have been key distinguishing features of the U.S. system of innovation, providing a significant part of the explanation of the U.S. global lead in areas such as computing, software, microprocessors (and other fields such as biotechnology).47 Famous examples include Microsoft and Intel, which successfully challenged the incumbent IBM in computing, software, and semiconductors, as well as the numerous new entrants that subsequently emerged to challenge these two challengers. Frequently, the new entrants made their challenge on the back of new technologies.48 The contrast with the U.S. broadband market, dominated by incumbent telecoms operators and cable TV companies, on the basis of two “incumbents’ technologies”—that is, DSL and cable modems—is striking. Why is the U.S. broadband market so different in this respect? Factors explaining the paucity of new entrants in the U.S. broadband market include high fixed costs and scale economies, the dominance of the telecoms regional operators (reinforced by the telecoms bust, which weakened new entrants com-

Introduction

paratively more than the incumbents), the strength of cable TV companies (in the market and politically), and pro-incumbent regulation. These related factors are probably sufficient to explain the relatively small number of new entrants in the U.S. broadband market, compared both to other U.S. information and communication markets as well as to the Japanese and Korean broadband markets. Furthermore, thus far new entrants and potential new entrants have not been helped very much by the availability of competitive new technologies. Essentially, as already noted, DSL and cable modems are incumbents’ technologies. However, alternative technologies—particularly those coming from the computer industry, an area of comparative U.S. global strength—have not yet, as all the chapters in this book show, had a significant impact on the broadband market. These technologies include the IEEE technologies such as WiFi (802.11X), WiMax (802.16), and Mobile-Fi (802.20) discussed earlier. In some cases the incumbents have been able to leverage their existing complementary assets in order to dominate new technologies. A case in point is the telecoms incumbent operators that have come to dominate WiFi networks, not only in the U.S. but all over the world. WiFi networks have simply become an extension of their existing networks, providing an extension to their existing services and enabling these incumbents to reap economies of scope. One possible future trajectory for the broadband market would give these alternative technologies a far larger share of the market while creating new openings for the kinds of entrepreneurial new enterprises that have powered U.S. global leadership in other information and communications markets. While such an occurrence is likely to be a long way in the future—if it occurs at all—it may allow the U.S. to play to one of its most important strengths. The European Puzzle Why is Europe lagging behind, not only the leading Asian countries but also the U.S.? The European Commission has created a new regulatory framework for Europe, which it claims will unleash new forces of competition. National regulatory authorities in the leading countries of the European Union claim to be broadly satisfied with the performance of their national systems. Yet, as this book shows, Europe is a laggard in the Broadband Era. Read this section in conjunction with the corresponding sections on the Japan/Korea and U.S. puzzles.

43

martin fransman

44

Intensity of Competition

In explaining the European puzzle, one hypothesis that will be discussed in this subsection is that the intensity of competition in the European countries included in this book has tended to be lower than in either the U.S. or Japan and Korea. It must be stressed, however, as with the other two puzzles, that this hypothesis needs further research in order to refute it or support it. DSL versus Cable Modem. As noted under the U.S. puzzle section, the most important source of broadband competition in the US has come from rivalry between telecoms operators (using DSL) and cable TV companies. In the European countries included here, by contrast, with the partial exception of the UK, this source of competition has been relatively weak in contrast to the U.S. Evidence to support this conclusion comes from Figure 1.1. As this figure shows, in Italy and Germany the share of broadband penetration accounted for by cable modems is negligible. It is only slightly higher in France. Sweden has a higher ratio of cable modems to DSL than these three countries. It is only in the UK that the ratio is substantial, though DSL still outweighs cable modems while in the U.S. the opposite is the case. The country chapters provide more detailed accounts of why cable has not mounted a more significant challenge to DSL in Europe. For example, in Germany, although there is a relatively high penetration of cable TV, for the complicated reasons examined in the German chapter there has been little attempt to turn the cable TV infrastructure into a broadband platform to compete with Deutsche Telekom’s substantial domination of the broadband market (astoundingly, this incumbent still has 88% of the total market). The French chapter shows that France Telecom’s continuing interests in cable have helped to create a perverse incentive discriminating against DSL– cable modem inter-platform competition. In Sweden it is only since 2003 that DSL– cable modem competition has increased, mainly the result of the incumbent Telia’s cable subsidiary (Com hem) being spun-off as a regulatory pre-requisite for the merger between Sweden’s Telia and Finland’s Sonera. Although in the UK, Figure 1.1 suggests that cable–DSL competition is strongest (of the European countries included here), several factors have in practice constrained somewhat this form of competition. The most significant of these has been the financial weakness of the two main cable competitors to the telecoms incumbent, BT. Both these companies—ntl and Telewest—were severely affected by the bursting of the telecoms bubble after 2000, and their financial problems limited their ability to compete strongly in the broadband Internet access market.

Introduction

For these reasons it may be concluded that although DSL– cable modem competition exists to a varying extent in these European countries, it does not play as important a role as in the U.S. Disruptive Competitors. It is also clear that although relatively recently companies that can legitimately be called disruptive competitors (in the sense defined above) have emerged in several of the European countries, their significance as a source of intense competition is not nearly as great as in the cases of Japan and Korea (discussed under the Japan/Korea puzzle section). This hypothesis is supported by evidence on market share. At the one extreme lies Germany, where, as already noted, the incumbent Deutsche Telekom has an astounding 88% share of the broadband market. One of the main competitors to Deutsche Telekom is QSC.49 However, as the German chapter shows, this competitor only has about 3% of the small company segment of the broadband market and around 1% of the large company market. It is noted in this chapter that “the lack of competition in the DSL market is one of the most crucial points in the current regulation debate [in Germany].” In the UK much is sometimes made of the fact that the telecoms incumbent, BT, has only around one-quarter of the retail broadband market, a very low share compared to most other countries. It is suggested that this has resulted in vigorous competition that, in turn, has led to relatively low broadband prices in the UK.50 However, this fact is seldom coupled with the other fact that until 2005 BT controlled almost all of the DSL wholesale market. (The way it has worked is that BT Wholesale, which as a player with significant market power is regulated, has supplied its wholesale DSL product both to BT Retail as well as to other retail broadband Internet access providers. It is precisely the worry that BT Wholesale has not provided “equivalent” access to its sibling, BT Retail, and to its other competitors that was one of the main motivations of the substantial strategic review undertaken by the regulator OFCOM.) In the UK wholesale DSL market, therefore, there has been little competition to BT. Excluding the cable companies ntl and Telewest, there has been no substantial competition to BT from the kind of disruptive competitors that have played such an important role in Japan and Korea.51 (As was shown earlier, unbundling has been virtually non-existent in the UK.) In France, as the French chapter shows, a crucial change took place from 2002. Having been inspired by the Japanese and Korean examples, the regulator ART introduced an aggressive programme of unbundling. By lowering entry barriers into the broadband market in this way, new entrants were

45

martin fransman

46

encouraged to enter. This has considerably increased the competitive pressure brought to bear on the incumbent, France Telecom. While in 2005 France Telecom retained 57% of the broadband market, new entrants had made a significant impact. They included Iliad (brand name, Free) with a 17% market share, Telecom Italia (brand name, Alice) with 8%, and Neuf and Cegetel with 5% each. Iliad/ Free makes an interesting comparison to Yahoo! BB and Hanaro. iliad/free. Iliad began as a provider of content in France’s Minitel era. (Minitel was France’s proprietary networked system to provide information and content to French homes and offices. It was later displaced by the Internet.) In 1999 Iliad created an Internet Service Provider (ISP) called Free, which was granted telecoms licences from the French government allowing it to create a network and offer telecoms services to the public. Iliad decided that it was necessary to own its own network since, in the group’s words, this “would allow [us] to control the technical aspects and pricing of [our] services for the routing of both data (Internet) and voice (telephony).” In 2000 the group sold its Minitel business. Indicating both confidence in the group, as well as perhaps reflecting the exuberance of the time, Goldman Sachs invested €15 million in the group in 2000. In the same year Free’s network began to be constructed and interconnections made to the France Telecom network. Significantly, in 2001 Iliad began doing R&D in order to develop Freebox, a combined modem and DSLAM (digital subscriber line access multiplexer), which would in time allow it to control its own network with direct access to final customers and therefore to offer innovative services that would compete effectively with France Telecom and its Internet subsidiary, Wanadoo, over ADSL lines. In the same year the regulator ART introduced local loop unbundling while imposing some of the lowest LLU prices in Europe. In 2002, Free began using unbundled lines in Paris and in the same year launched its broadband service. By March 2003 Free became the second largest provider of ADSL in France and the same year it rolled out its broadband services to the majority of residents in France. In August 2003 it launched its ADSL fixed-telephony service using its Freebox modem. Free was also the first company to provide TV over ADSL, using triple play—broadband Internet access, telephony, and TV/video—to compete with France Telecom. Iliad’s strategy is “to pursue its policy of winning new ADSL broadband customers by combining a competitive pricing policy with a focus on the quality

Introduction

of its offerings (such as continuous increases in bandwidth and innovative ADSL-based telephony and audio-visual content) and its technical and customer support services.” By competitive pricing policy, Iliad has meant a price war with France Telecom, a war that has significantly reduced prices and therefore operator revenues. However, strictly speaking it would be incorrect to call Free a disruptive competitor in the same boat as Yahoo! BB in Japan or Hanaro in Korea. The reason is that, unlike these Asian new entrants, Free has been making profit and distributing a dividend since 2001. In 2003 it made a profit of €34 million on revenues of €293 million.52 With a market share of 17%, Iliad is still significantly behind Yahoo! BB and Hanaro who, as noted earlier, have about one-third each of the local broadband market. Significantly, both the regulator ART and Iliad have been strongly influenced by the Japanese and Korean experiences in shaping their broadband policies and strategies. Iliad has consciously based its business model on that of Yahoo! BB, with the notable exception of the extent of its price cuts (Yahoo! BB has been making losses while Free, as just noted, has been making profits). Changes such as these in France from 2002 have helped turn the country into one of the fastest-growing in Europe in the broadband area. Other European countries, including Britain and Germany, have been influenced by the French example. In Sweden aggressive competitors have also played an important role, helping to make Sweden one of the best-performing countries in broadband, not only in Europe but also in the world (as Figure 1.1 and Tables 1.3, 1.4, and 1.5 show). In terms of market share (and apart from Com hem, Telia’s spun-off former cable subsidiary discussed earlier) two companies provide the strongest challenge to the incumbent, TeliaSonera: Bredbandsbolaget and Bostream. Bredbandsbolaget—also known as B2—started as a development project within a Swedish Internet consultancy company, Framfab, in 1998. In 1999 it began operations and in 2000 it received loan-finance from Cisco for equipment. Its investors include Investor (the largest industrial holding company in the Nordic region whose other holdings include Astra Zeneca and Ericsson), the Carlyle Group (a global investment company from Washington, DC), and Access Industries (a global private equity investment firm). B2 has a strategic partnership with the National Swedish Rail Administration. Frequently making losses, B2, like its counterparts Yahoo! BB in Japan and Hanaro in Korea, may be regarded as a disruptive competitor.

47

martin fransman

48

Bostream was established as a broadband provider in the financial year 1998 –99. By 2004 it had become Sweden’s largest DSL broadband provider after TeliaSonera with its own access platform based on local loop unbundling. In June 2003 Bostream launched its own ADSL and VDSL services (in Table 1.4 Bostream came fifth in the world in terms of capacity/speed with its VDSL service, while Bredbandsbolaget came ninth with an Ethernet LAN-based service). Unlike Bredbandsbolaget, Bostream has prided itself that it was able to achieve strong growth in revenue and profits “without the need for any external financing or access to the debt markets.”53 Nevertheless, in 2004 it was Bredbandsbolaget that acquired Bostream, creating an even stronger competitor to TeliaSonera. The combined company had 23% of the broadband market. In Italy, eBiscom/ FastWeb has come to play a similar role, competing with Telecom Italia, the incumbent, as shown in the Italian chapter. In 2004 FastWeb controlled 10% of the Italian broadband market. Conclusion To conclude, this section shows that in some of the European countries included in this book new telecoms operator entrants, largely using DSL, have in recent years come to offer significant competition to the incumbents. To some extent Germany and Britain have been exceptions to this generalisation. However, to return to the issue of Europe’s relative lag in broadband compared to Japan/Korea and the U.S., the information examined in this section has also suggested that the intensity of competition emanating from these competitors has been relatively weaker in Europe than in either Japan/Korea or the U.S. In the case of the Europe–Japan/Korea comparison of the role played by disruptive competitors there are two noticeable differences. The first is the significantly higher market share in Japan and Korea of the disruptive competitors compared to the share of their counterparts in Europe. Second, the extent of the financial losses made by the Japanese and Korean disruptive competitors suggests that they have chosen (or been forced) to compete more intensely on price than their European counterparts. However, in the European–U.S. comparison, the source of the relatively greater intensity of competition in the U.S. comes from cable-based competitors to the telecoms operator incumbents, as shown in the previous section.

Introduction Regulation

There is evidence that with the relative weakness of DSL– cable modem competition in Europe, unbundling has become an important determinant of performance. In the section on the Japan/Korea puzzle it was shown that unbundling played an important role in Japan and, more recently, Korea. This is being increasingly realised by regulators as the following quotation from OFCOM, the British regulator, shows: Local loop unbundling has had a big impact in Japan and France and a large amount of growth in these two countries can be attributed to broadband services based on shared local loop products. These products tend to be significantly cheaper than services using incumbent wholesale products. [emphasis added]54

Several points may be made, however, about unbundling in Europe. The first is that unbundling began significantly earlier in Japan than in Europe (though in the U.S. unbundling started earlier than in Japan). In Japan ADSL services via unbundled lines were introduced in 1999. In France, one of the first European countries to take unbundling seriously, the emphasis on unbundling only began from 2002. In countries like Britain and Germany unbundled lines providing broadband access are insignificant. Second, Japanese regulators appear to have been far tougher in terms of establishing the regulated price that the incumbent is entitled to charge competitors for access to its unbundled lines. As OFCOM notes, “Prices for [local loop unbundling] in Japan are now the lowest in the world.”55 Third, this suggests, as noted in the subsection “Strong Regulation” under the Japan/Korea puzzle section, that regulation has been stronger (as defined earlier) in Japan and Korea56 than in Europe. However, how this is to be explained in terms of a more detailed understanding of the politics of regulation in both sets of countries is an intriguing question that also requires further research. china Of the countries included in this book China is clearly the odd man out since it is the only developing country. However, as in many other areas, in broadband China is the exception that merits close attention. The reason, as shown in Table 1.9, is that although broadband penetration is still relatively low in China, in absolute terms the country already is, or soon will be, the largest broadband adopter in the world (as it is in mobile communications). Several points of a comparative nature are worth highlighting in this introductory chapter (see the China chapter for further details).

49

martin fransman

50

table 1.9 Asian broadband subscribers, early 2004

Country

DSL Subscribers (thousands)

Total Broadband Subscribers (thousands)

China Japan Korea

10,950 10,270 7,920

11,390 13,640 11,920

source: Information supplied to author from InfoCom.

Although Internet access began in China only in 1994, broadband access started in 2000. As Table 1.9 shows, broadband access is dominated by DSLproviding telecoms operators, the most important being China Telecom and China Network Communications (CNC). Cable companies, for reasons Manqiang Liu explains in the China chapter, have not been significant competitors in the Chinese broadband market. Similarly, new entrants—the largest of which is Great Wall Broadband57—have also not been substantial competitors. This leads Liu to conclude that “broadband access competition in China is insufficient.” Two demand-side factors in China are worth mentioning. The first is the role of the SARS outbreak in stimulating demand for broadband access, a useful facility for Chinese householders and firms whose face-to-face contact was limited by fear of infection. The second is the attention being co-operatively paid by stakeholders to increasing the supply of broadband-related content, applications, and services aimed at encouraging broadband demand. Clearly, the Chinese government plays an extremely important role in shaping the evolution of the broadband market. As in Japan and Korea, it is the ministry (Ministry of Information Industries, MII) that is also the telecoms regulator. Furthermore, the government is the main owner of the most important broadband providers, telecoms operators as well as cable providers.58 Government has also influenced the broadband market through several crucial direct interventions, such as when it restricted in June 2001 broadband access to 13 cities (including the main ones, such as Beijing, Shanghai, and Guangzhou). As in most of the countries examined in this book it was in 2003 that broadband demand began to take off in China. Unlike these countries, however, the size of the Chinese population and economy imply that China will shortly become one of the most important countries in the world in the broadband field.

Introduction

broadband performance in the short run and economic performance in the long run? Up to this point readers may have received the impression that it has been assumed that broadband performance in the short run has a positive effect on a country’s economic performance in the long run. This, it may have been assumed, is why the outcomes of the global Broadband Battles have been taken so seriously. However, there is little evidence to support this assumption. Indeed, one of the most remarkable features of the Broadband Battle is the paucity of evidence supporting the assumption—implicit in many of the policies and actions taken in the broadband field—that broadband, at least as measured in the commonly used measures of performance discussed in this chapter, is “good” for the economy. The absence of evidence exists as much in Asia as it does in North America and Europe. One case in point is the Strategic Review of Telecommunications undertaken by the British regulator OFCOM. In a 126-page review that devotes a significant amount of attention to broadband and next-generation networks, only two pages are allocated to “UK productivity and competitiveness.” Most of this section reports on the views of respondents to the earlier Phase 1 of the review. In terms of credible evidence, all the Review says is that “there is some academic evidence that take-up of information and communications technologies (ICT) is an important contributor to economic growth.”59 However, the Review goes straight on to acknowledge that “it is not yet possible to quantify this contribution accurately in Western European countries” (ibid). Although in a footnote on page 38 the Review refers to an academic study done in 2001—“which looked at the impact of telecoms infrastructure on economic growth in 21 OECD countries between 1970 and 1990 and identified a strong causal link”—it offers very little evidence regarding the impact of broadband adoption on productivity, growth, and competitiveness. Indeed, only two bits of evidence are provided. The first is in a consultant’s study from November 2003 that concludes that “due to the growth in the number of broadband connections, annual UK GDP could be up to £21.9 billion higher than it would otherwise have been by 2015” (p. 37). This study was commissioned by the Broadband Industry Group, a group consisting largely of broadband suppliers. The second bit of evidence comes from a survey of the British Chambers of Commerce, which “showed that 84 percent of [small and medium-sized enterprises] identified some benefits from broadband adoption, 46.4 percent thought that they had benefited from improved productivity, 45.3 percent from reduced costs and 13.4 percent from increased sales” (p. 38).

51

martin fransman

52

But this evidence, such as it is, does not warrant the conclusion that countries that perform better according to the broadband performance benchmarks—such as availability, penetration, speed, and price—are also likely as a result to record better productivity, growth, and competitiveness. It is therefore reasonable to conclude that the “Broadband Promise” at this stage rests more on uncorroborated belief, hope, and self-interest than on hard evidence (which, of course, is not to say that broadband does not have a significant, positive impact on these output variables). Several observations are worth making in this connection. The first is that in terms of the impact of broadband on productivity, growth, and competitiveness, it is necessary to distinguish between firms and households. Most of the discussion, statistics, and research focus on households. For example, this is the main determinant in statistics such as broadband availability and penetration. Relatively little research has been done on how firms, large and small, are adopting broadband—using technologies like DSL, cable modems, and FTTX—and the uses to which they are putting them. As far as households are concerned, it was shown earlier that thus far, there generally seems to be little that they do with a broadband connection that they did not do with a narrowband dial-up connection. Until now, therefore, it seems that the adoption of broadband by households has had minimal impact on productivity, growth, and competitiveness. However, this is likely to change over time—and change increasingly rapidly—with the development of broadband-related video-rich content, applications, and services. Second, the question arises as to whether a lag in the adoption of broadband by a country is likely to have a negative impact on its relative productivity, growth, and competitiveness. In short, to refer to the subtitle of this book, does it matter whether the U.S. and Europe lag behind Japan and Korea? While more research is needed to give a firmer answer to this question, it is worth noting, as mentioned in the U.S. chapter, that there is little evidence to suggest that the U.S. suffered economically as a result of its earlier significant lag behind Japan and Europe in mobile (cellular) communications, an area in which U.S. performance has now caught up. If the same were to be true in the broadband field then a similar lag would not matter. Third, there is also the possibility that “good” performance in broadband according to the measures may have a cost in terms of a trade-off in other areas. For example, we have seen that in Korea both the main new entrant competitors to the incumbent KT—Hanaro and Thrunet—suffered serious financial difficulty as a result of their disruptive competition. Consolidation has been the outcome, with Hanaro buying Thrunet, possibly along with a reduction in the intensity of competition. The main competitor to Japan’s

Introduction

incumbent NTT—Yahoo! BB—has thus far failed to make profit, and there are some signs that NTT has been put under serious pressure (more so than its Western counterparts) as a result of the disruptive competition. Thus, while intense competition in Japan and Korea has undoubtedly had a bracing and in some respects positive effect, it may also have produced some negative economic outcomes. Having said all this, however, it remains the case that the shift to broadband communications—a shift that will continue as speeds get ever faster—does represent a qualitative and radical change in the communications infrastructure of modern economies. This change will have significant effects, economically as well as socially and politically. Most importantly, perhaps, the move to broadband is allowing the information society to include the visual society, a transformation with significant and fundamental ramifications. It remains the case, however, that at the present time the absence of convincing evidence of the benefits of broadband—which, after all, provides the rationale for much of the exercise—is a remarkable characteristic of the transition to the broadband society. Paradoxically, so far we have the grin, but not the cat. It remains to be seen when the cat will join the grin.

notes 1. The analysis of the UK is integrated into the present chapter, while each of the other countries referred to here is analysed in a distinct chapter. 2. This table is reproduced in this chapter as Figure 1.1. 3. According to a key OECD publication that makes global broadband comparisons, “While there is, in fact, no standard definition of the speed necessary for a connection to be considered ‘broadband,’ there is widespread agreement that this should be faster than basic rate ISDN [integrated services digital network] which operates at 128 kbps. In this document most attention is paid to offers higher than 256 kbps. . . .” OECD (2003), p. 4. According to a publication from the European Commission, “Definitions of broadband vary widely. According to ITU [International Telecommunication Union, a United Nations body] Recommendation I.113, broadband means transmission capacity that is faster than primary rate ISDN. The FCC considers broadband to mean speeds in excess of 200 kbps, and the OECD a service with downstream capacity of at least 256 kbps.” Commission of the European Communities (2004), p. 4. 4. A detailed history of the evolution of the Internet and its symbiotic relationship with the telecommunications (telecoms) industry is found in Fransman (2002). 5. For an understanding of how the infocommunications industry works, see the Layer Model developed in Fransman (2002) and at http://www.TelecomVisions.com. 6. Nevertheless, in rapidly growing developing countries such as China, fast growth continued so that by 2005 Chinese mobile operators were able to add 5 million new mobile subscribers each month.

53

martin fransman

54

7. For a detailed analysis of the development of i-mode in Japan and an explanation regarding why it succeeded while its European counterpart failed, see Fransman (2002). 8. In fact, “availability” can be more vague than one might imagine. For example, availability statistics may be based on local telephone exchanges that have been ADSLenabled, ignoring the fact that connections are still required to be made both in the exchange itself as well as within the home for a working broadband connection to be established. This requires additional effort and expenditure than simply enabling the exchange. 9. See OECD (2003). 10. Table 1.3 is based on an OECD study published in November 2003. At the time of writing, this report remains the latest international comparison of broadband speed. However, it should be noted that the figures are changing rapidly and some of them are now out of date. 11. However, the information, from a Japanese ministry of communications web site, does not give a date for this data. Judging from the other data presented, it is likely that the figures are for the end of 2003. 12. OFCOM (2004a), pp. 5– 6. 13. Cioffi (1997), pp. 395ff. 14. This was not an unmixed incentive. Prior to the advent of ADSL telecoms companies had developed another high-speed access technology (though not as fast as ADSL) called ISDN (integrated services digital network). To begin with, many incumbent telecoms operators were slow to introduce ADSL for fear of cannibalising their ISDN services. However, actual and potential competition soon removed this reluctance. 15. The generic term xDSL refers to several different kinds of DSL, one of which is VDSL (very high data rate digital subscriber line) that allows for speeds that are high enough to compete with optical fibre local access connections, although VDSL operates only over short distances and is affected by the number of users. 16. For a detailed history of the development of optical fibre, see Fransman (1995). 17. This is a technology called D-WDM (dense wavelength division multiplexing). Multiplexing is a process that allows many users to share access to a communications transport facility. 18. The first proposal to use cellular systems in the field of mobile communications in order to make most efficient use of limited spectrum was put forward in 1947 and discussed in Bell Laboratories. The first publication on cellular radio was by Shulte and Cornell of Bell Labs in 1960. 19. These were offered by two competing standards, both based on code division multiple access (CDMA), a method of spreading spectrum transmission for digital wireless personal communications networks that allows a large number of users simultaneously to access a single radio frequency band without interference. The two standards are WCDMA, used largely in Europe and Japan, and CDMA 2000 (and later versions), used largely in the U.S. 20. This has resulted in a number of WiFi standards, including 802.11b, 802.11a, and 802.11g.

Introduction 21. According to the UK telecoms regulator OFCOM, “By an enduring bottleneck we mean a part of the network where not only does one operator have significant market power (SMP), but where effective, infrastructure-based competition is unlikely to emerge in the medium term” (OFCOM, 2004a, p. 52). 22. OFCOM (2004a), p. 52. 23. OFCOM (2004a), p. 52. 24. OFCOM (2004a), pp. 52–53. 25. A key regulatory issue has revolved around how best to calculate this price. Many economists have proposed that the price should be based on the long-run incremental cost (i.e., the additional cost in the long-run of producing an extra unit of output, assuming some output has already been produced). However, the matter has been contentious since some incumbents have claimed that this price does not give them a sufficient rate of return on their investment in the local loop, thus jeopardising such future investment. 26. In terms of purchasing power parity (Table 1.4), the U.S. is significantly cheaper than Korea. 27. NTT is analysed in detail in Fransman (1995) and Fransman (2002). 28. In presenting its cost and loss figures, Softbank still separates the effect of high customer acquisition costs, showing how much better the figure would be without them. 29. Japan Telecom was one of three companies allowed to compete with NTT when competition was first introduced into the Japanese telecoms services market in the mid-1980s. The other two were DDI and Teleway Japan. In the early 2000s Japan Telecom was acquired by Vodafone, which wanted its J Phone mobile subsidiary. The fixed network part of Japan Telecom was sold, first to a U.S. private equity company, and then to Softbank. 30. IDC was one of two companies allowed in the mid-1980s to compete with KDDI, NTT’s counterpart that was previously the monopoly incumbent offering international services (NTT provided only domestic services). IDC was majority-owned by the British company Cable and Wireless, and a Japanese trading company. In the late-1990s Cable and Wireless fought a rather bitter battle with NTT to acquire the whole of IDC. 31. Financial Times, 30 August 2004. 32. It is worth emphasising, however, that NTT’s commitment to FTTH is not solely a competitive response to Yahoo! BB. In the late-1990s NTT articulated a strategy of VIP (visual, intelligent, and personal) and committed itself to massive investment in FTTH. This strategy was also strongly supported by the ministry of communications. Strong competition in urban areas from Yahoo! BB and, it is important to add, electricity companies with their own unregulated optical fibre networks, has speeded up NTT’s investments in FTTH. 33. This section draws on Chapter 3, in addition to providing further information. 34. It is no easy matter even to define such “strength,” let alone to develop an operational measure. Ultimately, a strong regulator is one that is able to formulate and enforce regulations that it judges are in the public interest, even when they conflict seriously with the interests of the dominant companies negatively affected. But to measure this ability, and compare it across countries, is difficult.

55

martin fransman

56

35. OFCOM (2004b), Annex O, p. 12. 36. For a detailed study of these battles, see Fransman (1995). 37. OFCOM (2004b), Annex O, p. 13. 38. OFCOM (2004b), Annex O, p. 13. 39. OFCOM (2004a), p. 65. Carrier pre-selection refers to a customer’s ability to select the services of an alternative carrier without having to use an additional prefix number or use a dialler box. 40. Financial Times, 11 August 2003. 41. OFCOM (2004b), Annex O, p. 10. 42. OFCOM (2004b), Annex O, pp. 14 –15. 43. OFCOM (2004b), Annex O, p. 17. 44. Four FCC Commissioners were in favour of keeping the line-sharing rules, although for different reasons. Democratic Commissioners Copps and Adelstein supported them as a tool to keep markets open; Commissioners Powell and Abernathy supported line-sharing on the grounds that it would strengthen the incentive of the incumbent local exchange carriers to shift investment from DSL to optical fibre (where sharing obligations were eliminated), which was seen as the real alternative technology to cable. However, for political reasons they could not agree on keeping the rules in the Triennial Review Order and were also not able to reintroduce them later. I am grateful to Johannes Bauer, author of the U.S. chapter, for pointing this out to me. 45. This interpretation of views within the FCC draws on OFCOM (2004b), Annex O, p. 17. 46. OFCOM (2004b), Annex O, p. 17. 47. See, for example, Nelson (1993) and Mowery and Nelson (1999). 48. For example, DEC challenged IBM on the basis of the minicomputer; Microsoft and Intel challenged both of them on the back of the PC in operating systems and microprocessors, respectively. 49. Cologne-based QSC is the main ADSL competitor to the German incumbent. In April 2000 the company went public and the same year it completed the construction of its DSL network, which covered the 40 largest cities in Germany. According to the company’s publicity material, QSC “developed from a pure infrastructure vendor to a service provider and further to a solutions provider for broadband applications.” At the present time, the company provides nearly complete coverage for more than 60 of Germany’s largest cities. 50. Although Table 1.5 provides little comfort for this argument, the suggestion is that this exhibit refers to the situation at the end of 2003 while more recent data reveals that relative UK broadband prices have improved significantly since then. 51. Recent competitors to BT include companies like Easynet, Bulldog (just acquired by Cable and Wireless), and Video Networks. But they have an insignificant share of the broadband DSL market. 52. The Economist, 21 August 2004, p. 61. 53. Company’s web site. 54. OFCOM (2004b), p. 37. The reasons why unbundling has produced superior performance to wholesale broadband products were given earlier in the subsection dealing with unbundling of the local loop. 55. OFCOM (2004b), Annex O, p. 13.

Introduction 56. As shown in the section on Korea and in Chapter 3, Thrunet and Hanaro began their broadband services using cable networks. The Korean authorities started by encouraging full facilities-based (inter-platform) competition and only later moved to unbundling, partly in order to avoid a build-up of overcapacity. 57. Great Wall Broadband is owned mainly by the Great Wall Computer Group, and CITIC (China International Trust and Investment Corporation) also has a significant stake in the company. 58. China’s four main telecoms carriers—China Telecom, CNC, China Unicom, and China Mobile—are listed companies with most shares owned by the Chinese government. Two government departments manage these companies. SASAC (Stateowned Assets Supervision and Administration Commission) manages the assets of these enterprises and appoints their main leaders. MII (Ministry of Information Industries) is responsible for making telecoms industrial policy. Broadcasting and television, including the cable companies, are managed by SARFT (State Administration of Radio, Film, and Television). Some have argued that this separation of the governance of telecoms and broadcasting has impeded the process of convergence between the two in China. I am grateful to Manqiang Liu, author of the China chapter, for providing the information in this and the previous footnote. 59. OFCOM (2004a), p. 37.

references Cioffi, John. 1997. ADSL: The beginning of the information superhighway. Annual Review of Communications, International Engineering Consortium. Commission of the European Communities. 2004. Communication from the Commission to the Council, the European Parliament, the European Economic and Social Committee, and the Committee of the Regions: Connecting Europe at High Speed: National Broadband Strategies. COM (2004) 369 final, Brussels, 12 May. Fransman, Martin. 1995. Japan’s computer and communications industry: The evolution of industrial giants and global competitiveness. Oxford: Oxford University Press. ———. 2002. Telecoms in the Internet age: From boom to bust to . . . ? Oxford: Oxford University Press. Mowery, David C., and Richard R. Nelson, eds. 1999. Sources of industrial leadership: Studies of seven industries. Cambridge: Cambridge University Press. Nelson, Richard R., ed. 1993. National innovation systems: A comparative analysis. Oxford: Oxford University Press. Office of Communications (OFCOM). 2004a. Strategic review of telecommunications, Phase 2 consultation document. London: OFCOM, 18 November. ———. 2004b. The communications market 2004. London: OFCOM. Organisation for Economic Co-operation and Development (OECD). 2003. Benchmarking broadband prices in the OECD. DSTI /ICCP/ TISP(2003)8, 19 November, Paris.

57

part one. asia: japan, korea, and china

part i examines the development of broadband communications in three Asian countries. The first of these countries, Japan, is the world leader in terms of broadband speed and price and also leads in diffusion of the superior fixed broadband technology, namely fibre to the home (FTTH). The second country, Korea, is the global leader by far in broadband penetration. The significant lead of these two countries was documented in Chapter 1. They are included in Part I of this book because their lead allows us to use them as a benchmark against which the other countries can be compared. A third Asian country, China, is also included. Although China is not a world leader in broadband, in this field as in so many others in telecommunications and elsewhere, both the rapidity and scale of Chinese achievements justify its inclusion as a special case worthy of close attention. In Chapter 2 Takanori Ida makes a significant contribution to our understanding of Japanese broadband performance. In so doing he documents what is perhaps the most remarkable of Japan’s achievements in the broadband area, namely the rapid diffusion of FTTH. The significance of this achievement becomes apparent when account is taken, as noted in Chapter 1, of the substantial superiority of optical fibre as a means of transporting broadband communications. The picture that Ida implicitly paints is indeed remarkable and insufficiently appreciated. This picture is one of Internet users in cities such as Tokyo and Osaka routinely receiving speeds of 100 megabits per second while their counterparts in New York, London, and Paris usually get below 10 megabits. As Ida points out, currently (at the time of writing) 12 percent of Japanese broadband users get their fast Internet access over FTTH. The incumbent telecommunications operator, NTT, aims to provide FTTH to

martin fransman

60

30 million subscribers by 2010, half of all Japanese households with telephones. Furthermore, FTTH is also available in medium-sized towns, although Ida argues that more must be done to extend the availability of FTTH in smaller provincial towns in order to overcome the new “fibre divide” and reap the externalities that fibre provides in areas such as health, education, and social services. Japan’s achievement with FTTH is all the more remarkable when it is understood that in the rest of the world the conventional wisdom is that fibre to the home is generally too costly, and the associated benefits in comparison to alternative technologies such as ADSL too small, to justify large-scale adoption. As shown elsewhere in this book, in the U.S. incumbent operators such as Verizon and SBC have succeeded in getting the FCC (their regulator) to “forbear,” that is, abstain from regulating access to their fibre local access networks, on the grounds that this is necessary to create a sufficient incentive for their construction of FTTH networks. Both these companies have begun implementing FTTH programmes worth several billions of dollars. My own interviews with leading European operators, however, revealed that none of them has plans for large-scale adoption of FTTH in the foreseeable future. In Korea, too, as shown in Chapter 3, FTTH is currently considered to be too expensive. So how has Japan been able to do what most other countries find financially unfeasible? This question is all the more intriguing when it is added that far from “forbearance” the Japanese incumbent, NTT, which provides most of the FTTH in Japan, is required by its ministry-regulator to provide low-cost access to its fibre networks to its competitors, thus significantly reducing the company’s rate of return from its huge investments in this area. As noted in the last paragraph, in the U.S. the incumbents were able successfully to argue that forbearance was a necessary condition for their investments in FTTH. Ida’s story in answer to this important question asked in the previous paragraph is embedded in an analysis of Japanese political economy. Pointing to the long-standing conflict between NTT and the ministry of communications (that also acts as telecoms regulator),1 and the attempts of the latter to break NTT up into several independent companies, Ida argues that partly as a defensive move NTT undertook to develop FTTH within Japan as a whole. The implicit argument was that only a company with the resources and capabilities of the integrated NTT could deliver on this promise. Ida, however, acknowledges that NTT’s commitment to FTTH was also the result of the company’s R&D-inspired VIP vision (i.e., visual, intelligent, and personal communications), a point also made in Chapter 1. Taking this argument further, Ida suggests that the failure of the ministry to break up NTT (the result of a vigorous defence by NTT) had unforeseen

Asia: Japan, Korea, and China

consequences that contributed crucially to the later success of FTTH in Japan. More specifically, in order to finance its costly FTTH programme NTT was able to draw on the significant contribution made to company resources by its mobile subsidiary, NTT DoCoMo. Had NTT been broken up, this source of finance would not have been available. However, this is only part of the Japanese FTTH success story. The availability of resources does not by itself explain why these resources were invested in FTTH. The other part is intense competition, a determinant of the success of Japan and Korea emphasised in Chapter 1. However, in this case the competition comes not from other telecoms companies but from diversifying companies in a neighbouring utility industry, namely electricity companies. According to Ida, a major reason for the rapid diffusion of FTTH in Japan is competition emanating from subsidiaries of the major electricity companies. These companies have two crucial advantages that have motivated them to enter the Internet access market (with the potential to offer triple play services).2 They already have their own optical fibre networks and, furthermore, they are not subject to telecoms regulations as is NTT. More specifically, they are not required to give access to competitors who may like to use their networks. These companies constitute a considerable potential threat to NTT, which has speeded its FTTH programme partly in an attempt to counter the threat. Until recently this competition story has been largely confined to the Kansai area (i.e., western Japan with Osaka as its major city). A corporate scandal in the main electricity company in the Kanto area (dominated by Tokyo) slowed this company’s entry into FTTH, although it is now moving in the same direction as its Kansai counterpart. As this brief account of Japan’s leadership in FTTH makes clear, intense competition is a key part of the explanation although political pressure from the ministry-regulator and historical path-dependencies (such as the failure to break NTT up) also played a role. As shown in Chapter 1, the intensity of competition in Japan was also a major driver of this country’s performance in ADSL-based broadband. In Chapter 3, Inho Chung paints a very different picture of the other Asian leader of the global broadband industry. As noted in Chapter 1, Korea has a remarkable lead in broadband penetration, so much so that penetration in this country has already reached saturation point. As in Japan, where NTT and the disruptive competitors who use its network are at the heart of the broadband story, so in Korea the incumbent, KT, and its competitors who also use its network are major drivers of broadband. Indeed, it was shown in Chapter 1 that intensity of competition together with strong regulation are the main features that these two Asian countries have in common.

61

martin fransman

62

However, in Korea a dominating figure in the broadband painting is that of government. In Japan, by contrast, government (apart from its regulatory role) is only a minor player. In Korea in telecommunications, as in many other parts of the economy, it seems to be accepted that government has a legitimate and productive role to play in supplementing market forces. Chung, accordingly, concludes that in Korea “government seems to have played a very important role” in encouraging broadband performance. It has done so in numerous ways that include the provision of soft finance for the purpose of improving the broadband infrastructure. For example, in order to encourage the developing of broadband networks in the home the government has invested $14 billion up to 2007. The hands-on approach of the Korean government is also evident in its use of other non-financial policy tools, such as its systems of certification for the provision of broadband in buildings and apartments. But as Chung notes, demographic and geographic factors have also been important. For example, more than 50 percent of the Korean population live in apartment blocks making it relatively easy and cheap (per household apartment) to provide broadband connections (a fact that many Western studies have emphasised, sometimes one suspects in an attempt to attribute Korea’s remarkable penetration rate to special circumstances). Clearly these factors, while constituting important initial conditions, are not sufficient to explain Korean performance. We still have to understand the combination of factors on the supply side that led to these and other households being offered broadband connections at attractive prices, and on the demand side the circumstances that resulted in households deciding to allocate part of their budgets to broadband as opposed to the other alternatives. It is here that factors such as intense competition (documented in Chapter 1), which also affected the demand side through the high-profile advertising of the rivals, enter the causal chain. Chung also emphasises the importance of demand-side factors such as the prevalence of Korean-language portals and web sites and Korean Internet films, music, and online games. Another significant contrast with Japan is the role played by local loop unbundling (LLU)3 in Korea. LLU was introduced in Korea only in 2002. This had the effect of encouraging competitors to the incumbent operator, KT, to develop their own broadband infrastructure, rather than relying on KT’s. Both Thrunet and Hanaro initially depended heavily on broadbandenabled cable networks, which they used as a platform to attack KT. This enhanced inter-modal competition between ADSL and cable in Korea. Chung suggests that the introduction of LLU was motivated partly by concern at excessive investment in broadband infrastructure and partly by the attempt to expand broadband availability in rural areas where ADSL is often the only

Asia: Japan, Korea, and China

option. However, here, too, government, according to Chung, played an important role, in this case by not intervening and leaving developments to market competition. By not limiting downward movements in KT’s broadband prices, the government did not give its competitors the luxury of highly profitable prices. Instead, they had no option but to challenge KT on the basis of low prices. This benefited consumers and contributed to a rapid diffusion of broadband, although, as we saw in Chapter 1, it also led eventually to financial difficulties in both Hanaro and Thrunet. A further significant difference between Japan and Korea is in the area of FTTH. As Chung shows, unlike in Japan, FTTH has not yet taken off in Korea primarily as a result of the relatively high cost of this option. Why Korean electricity companies are not playing the same role in FTTH as their counterparts in Japan remains an intriguing question that still needs to be researched, not only in Korea but also in the U.S. and the European countries. China’s picture, painted by Manqiang Liu in Chapter 4, not surprisingly in view of this country’s particular circumstances and history, is very different from the cases of Japan and Korea. Here, too, the figure of government looms large, much larger than in Korea and indeed so large that it dominates the whole of the painting. Not only does government in China make policy for and regulate the telecommunications industry, it also owns and directly or indirectly (through major senior personnel appointments) controls the main competing telecoms companies. But as Liu documents, it is a government that is strongly committed to the development of broadband. This pre-existing commitment was considerably reinforced during the SARS outbreak in 2003 when large numbers of the urban population, in public and private sectors, were forced by their desire to minimise face-to-face contact to depend on the Internet for communications. Under these circumstances the benefits of broadband were clearly demonstrated. The fallacy of thinking of China as an undifferentiated developing country is demonstrated by the statistics that reveal that of Beijing’s population 50 percent are connected to the Internet of which 64 percent have broadband connections. There will be an additional boost to statistics such as these when China hosts the Olympic Games in 2008, an event that the government is intent on using as a springboard for the further development of both fixed and mobile broadband communications. A particularly interesting aspect of the development of broadband in China, as Liu documents, is the cooperative efforts being made by many complementary players to develop broadband-specific content, applications, and services. Realising that until now speed and always-on functionality remain the principal differentiating characteristics between broadband and narrowband, and that people by and large do much the same with their

63

martin fransman

64

broadband connections as they did with narrowband—a situation that is mirrored in all countries, developed and less-developed—attempts are being made in China to encourage the co-evolution on the demand side of content, applications, and services. In this area lessons may well emerge in China with relevance for other countries. However, Liu is also frank about some of the shortcomings of broadband in China. One of the most important of these is the limited role played by competition. For reasons that Liu analyses, cable companies are not significant competitors to the ADSL-based incumbents. Furthermore, although there are some private sector new entrants competing with the incumbents— notably Great Wall Broadband, a joint venture between the computer company group Great Wall and China International Trust and Investment Corporation (CITIC) established in 2000—they have not performed particularly well and some, including Great Wall, face serious financial difficulties. The extent of government control over the whole area of broadband must also be included as a serious shortcoming. But despite these limitations, China being China with its rapid rate of growth in output, exports, and income, and broadband being a key enabling infrastructure, it is reasonable to predict that important broadband outcomes may be expected in the future from this country. As this introduction to Part I has amply demonstrated, drawing on the Asian chapters in this book, the dynamism of broadband performance in these three countries—established statistically and analytically in Chapter 1— is due to a combination of very different circumstances in each country. In the cases of Japan and Korea these circumstances have produced worldleading performances in broadband. It is against these countries, accordingly, that the other countries examined in this book are benchmarked. Martin Fransman

notes 1. For further details on this conflict see Fransman (1995). 2. For a definition and discussion of triple play services see Chapter 1. 3. See Chapter 1 for a more detailed discussion of local loop unbundling.

reference Fransman, Martin. 1995. Japan’s computer and communications industry: The evolution of industrial giants and global competitiveness. Oxford: Oxford University Press.

2 broadband, information society, and the national system in japan Takanori Ida

a “broadband big bang” has swept through various parts of Asia including Japan and South Korea, where the popularity of broadband services has grown explosively. But why did this happen? It is indeed a puzzle, as Martin Fransman has said. In the last decade of the 1990s, Japan suffered from the collapse of the bubble economy (1985–1991) and the prolonged period of recession that followed it. The Nikkei Stock Average—which had once topped $352 (PPP $227)—fell to less than $100 and negative economic growth was also recorded.1 Finance businesses found themselves in difficulties with bad loans, manufacturing industries were plagued by de-industrialization, and structural reforms of the Japanese economy were never completed. But in the midst of this economic turmoil, the telecommunications industry began a sudden spurt of growth in 2000, especially in the ADSL and FTTH markets, and is now leading the world in terms of price and service standards. Professor Fransman attributes this boom to a combination of vicious competition and strong regulation by the Japanese government. Based on his discussions, I will argue that these factors originate from damaging interactions between NTT and newcomers to the telecommunications market, and from the policy of deregulation that developed due to the fact that NTT was not broken up. Japan’s broadband market first started to take off in 2001 when SoftBank BB joined the ADSL market with aggressively low prices. ADSL users greatly appreciated the always-on connectivity and low pricing offered by this service. However, they were not so happy with the communication speeds because the speed of ADSL is highly asymmetric (the download speed is much faster than the upload speed) and because the effective transfer rates drop dramatically

takanori ida

66

with increasing distance from the local exchange. It is therefore likely that ADSL users will eventually migrate to FTTH out of necessity. The cost of FTTH services was originally much higher than ADSL. But in 2002, KOpticom (a subsidiary of Kansai Electric Power Co. Inc.) broke into the FTTH market with their own optical fiber network, thereby initiating competition in terms of the installed equipment base. This led to a rapid reduction in the price of FTTH, almost eliminating the price differential with ADSL. SoftBank BB was able to participate in this price war because NTT had been strongly regulated into providing line sharing at very low prices. Also, while NTT was duty-bound to allow other carriers to use its optical fiber, the electric power companies were able to participate without this obligation. This bias against NTT benefited its competitors. But why was NTT able to financially endure these excessively harsh regulations? To put it simply, it is because NTT is a comprehensive communications provider that includes both fixed-line and mobile phone services. In particular, its profits from the rapidly growing mobile phone sector allow it to accommodate deficits in the fixedline phone sector in an implicit way. Abundant resources of capital are important for making the transition from old public exchange phone networks to new optical IP networks. The NTT group had the money to achieve this. However, the broadband big bang is causing a new problem, which is being called the “digital divide.” Optical IP networks now offer complete coverage of urban districts but have made little progress in the provinces. Furthermore, in provincial regions there are more households with single elderly occupants and IT literacy levels are very low. Since there is no demand for broadband in the provinces, it has been argued that there is no point installing optical fiber in these regions. Personally I am opposed to this view. Broadband is not just a faster means of accessing the Internet, it is a means of providing public services such as HDTV broadcasts, remote medical checkups and consulting, remote education and remote administration. The people most likely to benefit from these public services are probably the less “IT-savvy” members of regional communities where the transport situation is not so good. Therefore, since the introduction of broadband services has so far relied on competition between NTT and newcomers in urban areas, the government will eventually have to lead the way in closing the digital divide in the provinces. These are the main points I will be making in this chapter, which is structured as follows: section 2 discusses the reasons behind the success of the broadband big bang in Japan; section 3 describes the broadband services available in Japan, especially ADSL and FTTH; section 4 discusses the digital divide and measures for overcoming it; and conclusions are presented in section 5.

Broadband, Information Society, and the National System in Japan

the reasons for japan’s success In the 1990s, Japan’s telecommunication industry was said to be lagging 10 years behind Europe and the U.S. But the broadband big bang that started in Japan in the following decade attracted a great deal of attention from overseas. This turnaround in the fortunes of Japan’s telecommunications industry probably looked like a “puzzle” to foreign analysts. According to an analysis by Ida (2005), the factors behind the success of broadband in Japan are (1) unintentional benefits of Japan’s telecommunications policy; (2) intentional benefits of Japan’s telecommunications policy; (3) the appearance of newcomers with strong capabilities and desires; and (4) reluctant but concerted efforts by NTT’s management. These factors are discussed in detail below. Unintentional Benefits of Japan’s Telecommunications Policy Unintentional benefits occurred because NTT was not broken apart like AT&T. The fact that NTT was left intact was advantageous when it came to installing a nationwide optical fiber network. The events relating to the proposed breakup of NTT are summarized below. In 1982 the Second Administrative Examining Committee proposed breaking up the NTT public corporation; in 1985 the NTT public corporation was privatized, with the breakup of NTT to be reconsidered within 5 years; in 1990 the Telecommunications Council issued another report on the breakup of NTT; in 1991 it was decided that the proposed NTT breakup was to be reconsidered after a further 5 years according to “government measures”; and finally, in 1996, the Telecommunications Council published its final report on the breakup of NTT (Ministry of Internal Affairs and Communications).2 In December 1996, the Ministry of Posts and Telecommunications and NTT finally reached the following agreement3: (1) NTT was to be reconfigured under a holding company into separate regional companies for the east and west of Japan and a long-distance company; and (2) the system of subsidiary companies must allow NTT to make inroads into the international communications arena. This compromise agreement is said to have been influenced by the then Prime Minister Ryutaro Hashimoto who wanted Japan to have a strong flag carrier. These 15 years of discussion over NTT’s breakup had a beneficial effect on its investment in optical fiber. In 1990, to resist the breakup proposals put forward by the Ministry of Posts and Telecommunications, NTT unveiled its plans for contributing to society with a VI&P (Visual, Intelligent, and Personal

67

takanori ida

68

table 2.1 Area Coverage of NTT East and West

Urban areas Provincial towns Other Total

1999 NTT East / West (%)

2000 NTT East / West (%)

2001 NTT East / West (%)

2002 NTT East / West (%)

2003 NTT East / West (%)

2004 (est.) NTT East / West (%)

60 / 50 36 / 25 19 / 10 41 / 29

65 / 56 43 / 37 29 / 17 48 / 37

84 / 68 62 / 43 41 / 36 66 / 50

82 / 84 78 / 66 49 / 50 76 / 67

95 / 92 86 / 86 55 / 63 81 / 80

98 / 93 88 / 89 57 / 64 83 / 81

sources: NTT East and West, http://www.ntt-east.co.jp/index_e.html, http://www.ntt-west.co.jp/index_e .html.

Communication Service) concept based on B-(Broadband-) ISDN, and announced that it would finish the installation of a nationwide optical fiber network by 2015. NTT’s announcement of the VI&P concept is said to have preempted U.S. vice president Al Gore’s proposal of an “information superhighway” in 1993. Once the U.S. began to stir, the Japanese government hurriedly established a “bureau for the promotion of an advanced information and telecommunications society” in 1995 and stated that their fundamental policy was to establish a nationwide optical fiber network by 2010. In the middle of the arguments over the breakup of NTT, it was essential that the first proposals for the installation of Japan’s optical fiber network should come from NTT rather than the Ministry of Posts and Telecommunications. Since the 1990s, NTT East Japan and NTT West Japan have continued to invest over $200 billion (PPP $157 billion) in optical fiber installation. This flow remained unaffected by the regrouping of NTT. The area coverage of NTT East / West is shown in Table 2.1. At present, NTT’s installation of nationwide optical fiber facilities is getting close to achieving its target. Japan’s optical fiber facilities are a side-effect of the contention surrounding NTT’s breakup and the fact that NTT was not broken up contributed to the ongoing investment in optical fiber. Intentional Benefits of Japan’s Telecommunications Policy An intentional benefit was the government’s policy of thoroughly deregulating NTT’s regional communications networks. The excessively harsh and imbalanced regulations imposed on NTT helped to promote the activities of many newcomers and, as a result, NTT’s infrastructure ended up being used efficiently by the whole society. The process behind this deregulation policy is summarized below.

Broadband, Information Society, and the National System in Japan

In 1987 an additional charge scheme was introduced whereby long-distance NCCs paid NTT charges for local calls with no interconnection charge system in place; in 1994 an interconnection charge system was introduced by the communication carriers, whereby public network interconnection charges were calculated based on historical expenses; in 1999 dry copper unbundling was introduced and line sharing started; in 2000 public network interconnection charges were determined based on long-run incremental costs (LRIC) and dark fiber unbundling was introduced; and in 2001 regional IP network unbundling was introduced. Two major consequences of Japan’s broadband big bang were the appearance of line sharing in 1999 and dark fiber in 2000. Thanks to the policy of thoroughly open access to NTT’s regional communications networks and optical fiber, it became possible for newcomers to offer ADSL and FTTH services without having their own infrastructure. The degree of openness of Japan’s regional communication networks in 2001 compared with foreign networks is summarized as follows: 1. Japan: LRIC introduction (good); metal unbundling (good); optical fiber unbundling (good). 2. U.S.: LRIC introduction (fair); metal unbundling (fair), optical fiber unbundling (fair). 3. UK: LRIC introduction (good); metal unbundling (good); optical fiber unbundling (bad). 4. France: LRIC introduction (bad); metal unbundling (good); optical fiber unbundling (bad). 5. Germany: LRIC introduction (bad); metal unbundling (good); optical fiber unbundling (good).

Although the Ministry of Posts and Telecommunications was unable to bring about a breakup of NTT, it thoroughly supported NTT’s local loop unbundling (LLU) and as a result the regional communications networks became more open in Japan than in any other country in the world. It should also be noted that low charges were set for line sharing and dark fiber. NTT calculated that the initial line sharing cost would be $3.7 (PPP $2.9) but this was eventually brought down to $1.7 (PPP $1.3). NTT also claimed that the dark fiber cost of subscriber systems would be $185 (PPP $146) but this was brought down to $73 (PPP $57) during trial service provision and to $48 (PPP $38) during full-scale provision. This LLU policy of low charges forced NTT to slash its cost prices, allowing a wide range of newcomers onto the scene. NTT was able to withstand these harsh competition-promoting policies thanks to the rich financial reserves of the intact NTT group (especially NTT DoCoMo).

69

takanori ida

70

The Appearance of Newcomers with Strong Capabilities and Desires To understand how the broadband big bang occurred in Japan, it is important to consider the arrival of newcomers with strong capabilities and desires. In particular, the presence of SoftBank BB in the ADSL market and K-Opticom (a subsidiary of Kansai Electric Power Co. Inc.) in the FTTH market cannot be overlooked as these companies came to dominate the market in terms of price. In fact, NTT East and West were only just able to keep up with their low-price strategies. ADSL services first became available in Japan in 1999 but they made very little progress because of their high cost and restricted service areas. ADSL did not take off in Japan until 2001 when SoftBank BB joined the ADSL market with a monthly charge, including ISP fees, of $21 (PPP $17), which is half what NTT were charging. SoftBank Japan is a PC software company which was founded in 1981 by Masayoshi Son. The history of this company is summarized below. In 1981 Masayoshi Son established SoftBank Japan; in 1996 SoftBank invested as a primary stockholder in Yahoo! Inc.; in 1998 the stock was listed in the First Section of the Tokyo Stock Exchange and SoftBank introduced a holding company system; in 2001 “Yahoo! BB” (8M) ADSL services were launched; in 2002 “Yahoo! BB 12M” was launched and the number of subscribers topped 1 million; in 2003 “Yahoo! BB 26M” and the IP phone services were launched and the number of subscribers topped 3 million; in 2004 “Yahoo! BB 45M” was launched and SoftBank began field trials of 3G mobile phone services while the number of its ADSL subscribers topped 4 million. By forging links with other companies such as Microsoft, Sony, and Tokyo Electric Power, SoftBank has started various forms of communication and broadcasting businesses but so far without success. Of the services started by SoftBank, only the ADSL services have come to anything. However, even the ADSL business is unsuccessful as far as SoftBank is concerned. Although SoftBank originally aimed to turn a profit after reaching 1 million subscribers, it was still $810 million (PPP $640 million) in the red in 2003 despite having passed the 3 million subscribers mark. It has been predicted that losses in the broadband sector will decrease after 2004 but it seems very likely that the ADSL market will sooner or later be absorbed by the FTTH market. So although SoftBank may have fired up the ADSL market, it has not yet succeeded in this business. Meanwhile, NTT East and West have had no choice but to follow SoftBank’s lead in price reductions, resulting in Japan having the cheapest and fastest broadband connections in the world. In other

Broadband, Information Society, and the National System in Japan

words, the winners of the price war in Japan’s ADSL market are consumers, not businesses. SoftBank will probably move from ADSL services into FTTH services at some point but they are unlikely to have a smooth ride along the way. But without SoftBank’s dynamic influence, Japan’s broadband big bang would probably never have happened. If SoftBank was responsible for getting the ADSL market moving, then the prime mover behind the FTTH market is probably K-Opticom.4 K-Opticom is a subsidiary of Kansai Electric Power Co. Inc.—a brand name to rival NTT West—and is aiming to compete on the basis of its own optical fiber installations, which are completely independent of NTT’s optical fiber. The history of K-Opticom is briefly summarized below. In 1988 K-Opticom started as a communications equipment service for the Kansai region. In 2000 it acquired a type 1 telecommunication carrier’s license and in 2002 it launched the “eo” FTTH services. In 2004 K-Opticom slashed the cost of eo services and also launched optical IP phone services. K-Opticom inherited an electrical communication network from Kansai Electric Power and it also had no problems finding ducts and telegraph poles to use for the installation of optical fiber. Due to this advantage, K-Opticom has achieved a coverage of 70% of households and is now able to provide FTTH (100M) connections for a charge of just $58 (PPP $46), including ISP charges.5 Apart from NTT, only electric power companies are able to compete on the basis of their own installed equipment. Power companies also have greater freedom because, unlike NTT East and West, they are not duty-bound to provide a universal service or to let anyone else use their optical fiber facilities. In 2004, K-Opticom cut its prices by 20% and began providing optical IP phone services with quality guaranteed to be on a par with fixed-line phone services. So it appears that NTT’s ultimate rivals are electric power companies. Reluctant but Concerted Efforts by NTT’s Management After NTT was privatized it became embroiled in a dispute with the Ministry of Posts and Telecommunications over the breakup plans and it has trodden a thorny path ever since. The changes that took place in NTT between its privatization in 1985 and the regrouping in 1999 are quite surprising, for example: (1) the number of employees fell from 313,600 (in 1985) to 138,200 (in 1999); (2) the number of branch offices fell from 70 to 11 and the number of branch stores from 1,700 to 47; (3) the number of telephone subscribers increased from 45,300,000 to 58,470,0006; and (4) long-distance call charges (for a 3-minute daytime call) decreased from $3.7 (PPP $2.9) to $0.8 (PPP $0.6).

71

takanori ida

72

As a result of shedding 10,000 employees per year, the number of employees has decreased by 175,400, which is more than half the number of employees in 1985.7 NTT has also substantially rationalized its branch offices and branch stores. Meanwhile, the number of telephone subscribers has increased while call charges have plummeted. In the middle of this desperate management rationalization, NTT aimed to soften the calls by the Ministry of Posts and Telecommunications for a breakup of the company and to respond to criticisms from the long-distance NCCs. In either case, the management efforts made by NTT after privatization have to be highly regarded. But when NTT was regrouped in 1999 the structure of the industry changed from fixed-line voice services to mobile and IP services and the financial situation of NTT East and West rapidly deteriorated due to the appearance of competing policies such as the introduction of MyLine carrier selection services and long-run incremental costing. It became clear that NTT East and West would not survive this competition unless they underwent further drastic management rationalization. Therefore in 2002 NTT East and West implemented the following structural reforms: (1) a total of 100,000 positions, including 60,000 from NTT East and West, were transferred to outsourcing companies and their salaries were cut by 15–30%, while a further 16,400 jobs were lost through voluntary retirement, resulting in personnel costs being cut by a total of $1 billion; and (2) cuts were also made in various other forms of expenditure such as investment in facilities. As a result, overall costs were reduced by $1.5 billion. Assuming the management rationalizations carried out at NTT before the regrouping were the result of external pressure the structural reforms they made after regrouping were performed under their own volition as a matter of survival. The managerial status of NTT East and West after the regrouping is shown in Table 2.2. After the NTT regrouping the turnover of NTT East and West steadily decreased. The fall in revenue from voice communication services was particularly severe. Due to the structural reform of NTT East and West in 2002 the management performance made a slight recovery. However, it is doubtful whether this recovery will continue because, with the growing popularity of broadband and increasing use of IP phone services, subscribers are canceling their fixed-line contracts so that NTT East and West do not receive any call charge revenues at all. That is why the NTT holding company published a mid-term management strategy in November 2004 to announce their plans to move away from phones into broadband: (1) to promote the growth of high-speed broadband services and aim for 30 million broad-

Broadband, Information Society, and the National System in Japan

73

table 2.2 Managerial Status of NTT East and West

NTT East Japan Turnover

2000 $100m (PPP $100m)

2001 $100m (PPP $100m)

259 (204)

238 (188) (⫺7.7) 0.4 (0.3) 0.7 (0.5)

218 (171) (⫺8.6) 4.5 (3.5) 5.9 (4.6)

210 (165) (⫺3.3) 8.0 (6.3) 9.1 (7.1)

223 (175) (⫺8.8) ⫺15.4 (⫺12.1) ⫺15.8 (⫺12.4)

205 (161) (⫺8.0) 3.9 (3.1) 4.2 (3.3)

201 (158) (⫺2.2) 8.0 (6.3) 8.4 (6.6)

Operating profit Ordinary profit NTT West Japan Turnover

3.1 (2.5) 1.3 (1.0)

Operating profit Ordinary profit

⫺9.3 (⫺7.3) ⫺9.8 (⫺7.7)

244 (192)

2002 $100m (PPP $100m)

2003 $100m (PPP $100m)

sources: NTT East and West, http://www.ntt-east.co.jp/index_e.html, http://www.ntt-west.co.jp/index_e .html. Percentage growth shown in parentheses. PPP ⫽ purchasing power parity.

band service users by 2010; and (2) by 2010, to draw up concrete plans for switching the whole country from the fixed-line phone network to an optical IP network. Hitherto, NTT’s structural reforms have mainly involved cutting back on personnel expenses. The structural reforms that NTT will face in the future involve dismantling the public exchange telephone network and installing a nationwide optical IP network. Moreover, NTT East and West will remain duty-bound to provide a universal phone service and open up their optical fiber resources. This means NTT has a heavy social burden to carry. No matter how many newcomers appear on the scene it is probably only NTT that is capable of providing broadband services on a nationwide scale. Ieyasu Tokugawa (1543–1616), the first shogun of the Edo Bakufu government, is reported to have once said, “Peasants should be taxed not to death, but almost to death.” It looks like the same could also be said of NTT. the current state of the broadband market The growth of broadband Internet access in Japan has been remarkably quick. This section describes the current situation of broadband in Japan. Every month the Ministry of Internal Affairs and Communications issues a report on the changing numbers of people using Internet service providers. From the end of 1999 to the end of 2004 the trends in narrowband Internet users and broadband Internet users were as shown in Table 2.3.

takanori ida

74

table 2.3 Trends in Narrowband and Broadband Internet Users

Narrowband Broadband ADSL FTTH CATV Internet

December 1999 (million)

December 2000 (million)

December 2001 (million)

December 2002 (million)

December 2003 (million)

December 2004 (million)

10.59 0.15 — — —

15.81 0.63 0.01 — 0.62

19.75 2.82 1.52 — 1.30

21.19 7.81 5.65 0.21 1.95

19.19 13.64 10.27 0.89 2.48

18.62 18.63 13.33 2.43 2.87

source: Ministry of Internal Affairs and Communications, http://www.soumu.go.jp/english/index.html.

Narrowband Internet users grew steadily until 2002 but then started falling again from 2003. On the other hand, the number of broadband Internet users remained very low until 2001, but the number of ADSL users grew explosively in 2002 with the number of broadband Internet users eventually overtaking the number of narrowband Internet users in August 2004. By the end of 2004 about 70% of households were connected to some form of Internet service, comprising 35% narrowband users and 35% broadband users. A breakdown of the broadband Internet statistics is shown below. Every month the Ministry of Internal Affairs and Communications issues a report on the changing numbers of people using broadband Internet service providers. Since the start of the broadband boom in 2002 the numbers of broadband (ADSL, FTTH, and CATV Internet) users have varied as shown in Table 2.3. First of all, it can be seen that Japan’s broadband market is mainly growing in the ADSL sector. In 2003 the number of ADSL users passed the 10 million mark, making Japan the world’s second largest ADSL user behind South Korea. These figures also show that FTTH is making solid progress, with the number of users exceeding 1 million in 2004. Japan is the first country in the world where the number of FTTH users has passed this milestone. As for CATV Internet users, although they now number more than 2 million, the rate of growth in this sector is not very large. Japan’s broadband usage is thus not dominated by CATV Internet, unlike the U.S. With regard to the standard charges for these services, it could be said that Japan’s broadband is the fastest and cheapest in the world. An ITU publication entitled “Internet Report: the Birth of Broadband” compared the Internet access speeds and charges of various countries to evaluate the charge per 100 kbps bandwidth. The results show that Japan ($0.09) is far faster and

Broadband, Information Society, and the National System in Japan

cheaper than the U.S. ($3.53) or UK ($6.37) although the report does not take the Japanese FTTH service into consideration.8 Next, I will discuss the demand substitution properties of broadband with reference to Ida and Kuroda (2005). This survey was implemented using a web questionnaire from the viewpoint of costs and efficiency and obtained responses from a sample of 1,013 users in environments where all five of the following types of ISP service were available: (1) fixed-line telephone dial-up (DU); (2) always-on connected ISDN; (3) ADSL; (4) CATV Internet; and (5) FTTH. After eliminating leaked entries and outliers from the results, a total of 799 valid samples were obtained. The selection ratios were as follows: DU 2%; ISDN 5%; ADSL 67%; FTTH 8%; and CATV 18%. Closer examination of the data shows that the average monthly expenditures (the sum of Internet communication charges and Internet provider charges) were as follows: DU $37 (PPP $29); ISDN $48 (PPP $38); ADSL $40 (PPP $32); FTTH $55 (PPP $43); and CATV $48 (PPP $38). The average nominal speeds of these services were DU 52 kbps, ISDN 65 kbps, ADSL 10 Mbps, FTTH 82 Mbps, and CATV 11 Mbps. Broadband and narrowband differ widely in terms of price and speed, and even within broadband there are large differences between ADSL, CATV, and FTTH. The reasons given for using ISP services were, in decreasing order: (1) searching the web, 44.7%; (2) email, 30.7%; (3) online shopping, 5.5%; and (4) chat rooms and message boards, 4.2% (multiple responses were allowed in this question). Ida and Kuroda (2005) calculated the “elasticity of demand” for ADSL subscribers, that is, the ratio of the rate of change of monthly charges to the rate of change of subscriber access demand. As a result, they obtained a value of 0.8, which shows that ADSL is less elastic than FTTH (3.2) or CATV broadband (2.5). However, since the ADSL market is very large-scale, they calculated the elasticity separately for low-speed ADSL (1–1.5M), medium-speed ADSL (8 –12M), and high-speed ADSL (24 – 40M) and obtained the following results: low-speed ADSL, 10.6; medium-speed ADSL, 2.6; and high-speed ADSL, 9.1. To sum up, there is very little elasticity of demand in the market for ADSL, which is the current market leader, but when this elasticity is calculated separately for low-speed, medium-speed, and high-speed ADSL it can be seen that there is much more elasticity in the demand especially for lowspeed and high-speed ADSL. This indicates that there is a significant influx and outflow of subscribers at the extremities of the ADSL market. The general status of Japan’s broadband market has been described above, but in the following subsection the ADSL and FTTH markets are described separately.

75

takanori ida

76

The Current State of the ADSL Market At present (September 2004), 72% of Japan’s broadband users are on ADSL connections. Let us start by examining the ADSL-market trends in line subscriber numbers at NTT East and West and at other communications carriers (see Table 2.4). The number of ADSL subscriber lines is increasing every year and had reached 26% of all households by September 2004. These figures prove that ADSL has been the driving force behind the growth of broadband services in Japan. The market share of NTT East and West has been decreasing from 41% to 37% but has recently stabilized. This reflects the fact that NTT has faced fierce competition in urban areas but still has a large share of the market in the provinces. In fact, the recent expansion of the ADSL market is based on increased usage of ADSL in provincial regions. However, a few clouds have recently appeared on the horizon of the ADSL market. First of all, there are still large differences between different parts of Japan in terms of the percentage of households with ADSL connections. As of September 2004, Tokyo has the greatest coverage (35%), while Kagoshima prefecture has the lowest (15%). This is due to reduced usage in remote areas, because ADSL cannot be provided far from the exchange. Second, there has been a fall-off in the rate of increase in the number of ADSL contracts. From 2002 to 2004 the rate of increase dropped from 28% to 12% and then to 6%. This probably means that the ADSL market is becoming saturated, and sooner or later the number of ADSL subscribers is likely to start falling as users switch to FTTH services. Next, let us look at the supply structure of the ADSL market. As of September 2004, the numbers of ADSL contracts from different providers break down as follows: NTT East and West, 4.74 million subscribers (38% market share); SoftBank BB, 4.49 million subscribers (35% market share); eAccess, 1.78 million subscribers (14% market share); and ACCA Networks, 1.25 million subscribers (10% market share). Between them, the two leaders NTT East and West (38%) and SoftBank BB (35%) lead the ADSL market with a combined market share of 73%. For the top four companies, the combined market share rises to 97%, showing that the ADSL market is dominated by a small oligopoly. However, the price competition between these top four players is very fierce and prices have come down a long way. This has been going on since SoftBank BB suddenly came in with a price of just $21 (PPP $17) in 2001, forcing NTT and the other carriers to cut their own prices substantially. This is illustrated numerically in Table 2.5, which shows the changes in standard ADSL charges (including ISP charges).

Broadband, Information Society, and the National System in Japan

77

table 2.4 ADSL Market Trends at NTT East and West September 2002 No. of subscribers (million)

September 2003 No. of subscribers (million)

September 2004 No. of subscribers (million)

0.94 0.79 2.50 4.22

1.91 1.52 5.79 9.23

2.62 2.12 8.06 12.80

NTT East Japan NTT West Japan Others Total

source: Ministry of Internal Affairs and Communications, http://www.soumu.go .jp/english/index.html.

table 2.5 Changes in Standard ADSL Charges

NTT East (1.5M) NTT East (8M) NTT East (12M) NTT East (24M) SoftBank BB (8M) SoftBank BB (12M) SoftBank BB (26M)

March 2001 $ (PPP $)

September 2001 $ (PPP $)

September 2002 $ (PPP $)

September 2003 $ (PPP $)

54 (43)

52 (41)

43 (34) 44 (35)

21 (17)

21 (17) 24 (19)

41 (32) 41 (32) 42 (33) 42 (33) 21 (17) 24 (19) 25 (20)

source: Ministry of Internal Affairs and Communications, http://www.soumu.go.jp/english/index.html. PPP ⫽ purchasing power parity.

These figures show how much SoftBank BB’s involvement has broken down the prices. As a result of these low prices, SoftBank BB has managed to gain a market share comparable to NTT East and West in just two years. But although SoftBank BB originally aimed to break even after gaining 1 million subscribers, it is still in the red even with 4 million subscribers. SoftBank BB’s aggressive pricing policy and billboard advertising campaigns may have been successful at gaining subscribers, but they have not been able to translate these gains into increased profits. There is still a very large difference between the charges levied by NTT East and West and by SoftBank BB. But it is interesting to note that in spite of this difference they both have more or less equal market shares. This is due to their different business models. The business model of NTT East and West is called “vertically separated.” NTT only sells ADSL circuit services, leaving users free to choose their own ISP. As a result, users have a great deal of freedom over their choice of circuit service and ISP but they have to pay charges both to NTT and to their ISP. Accordingly, it is suitable for users who want to

takanori ida

78

continue using the same ISP even though they end up paying higher charges. On the other hand, SoftBank BB has a “vertically integrated” business model, which means that they provide the ISP functions as well as the ADSL lines. SoftBank BB also ties these services in with its “BB phone” IP phone services. As a result, users have no freedom in their choice of circuit service and ISP but they only have to pay charges to SoftBank BB. This approach combines the convenience of a “one-stop” service with low charges. Finally, let us look at the demand structure of the ADSL market. According to a report by the Ministry of Internal Affairs and Communications entitled “Review of competition in the telecommunications area” (2004), the main reasons given for choosing ADSL are (1) always-on connectivity (65%), (2) fixed pricing (40%), (3) bandwidth (30%), and (4) low charges (29%) (multiple responses allowed). Compared with conventional dial-up connections, it can be seen that ADSL is rated more highly for its always-on connectivity and fixed prices. However, since ADSL provides less bandwidth than FTTH, users will probably start migrating from ADSL to FTTH at some point. Among users that considered choosing ADSL but decided against it, the reasons given were (1) insufficient bandwidth (24%), (2) awkward to set up (21%), and (3) too expensive (17%), so evidently dissatisfaction with the bandwidth offered by ADSL is the largest factor (multiple choices allowed). The results of asking current ADSL users what kind of circuit they would like to be using one year later were (1) ADSL (37%), (2) FTTH (34%), and (3) CATV Internet (7%). To sum up, the users who have enjoyed always-on connections and fixed pricing are likely to develop higher standards and become dissatisfied with ADSL leading them to switch over to FTTH instead. The Current State of the FTTH Market At present (September 2004), 12% of Japan’s broadband users are on FTTH connections. First, let us look at how the number of line contracts has changed at NTT East and West and at the other carriers as shown in Table 2.6. The number of FTTH line contracts is still small, although it is increasing every year, and at present (September 2004) it is only used in 4% of households. However, from 2005 substantial growth is expected in the number of users ditching ADSL in favor of FTTH, and the coverage of FTTH is likely to grow sharply in the future. The market share of NTT East and West is still high and is tending to increase (37% to 58% to 60%). Of the other carriers besides NTT East and West, the electric power companies are putting up a strong fight and currently occupy 13% of the market. K-Opticom (a subsidiary of Kansai Electric Power Co. Inc.) is doing particularly well. However, the

Broadband, Information Society, and the National System in Japan

79

table 2.6 FTTH Market Trends at NTT East and West

NTT East Japan NTT West Japan Others Total

September 2002 No. of users (thousand)

September 2003 No. of users (thousand)

September 2004 No. of users (thousand)

50 20 120 190

250 250 360 860

630 580 820 2,030

source: Ministry of Internal Affairs and Communications, http://www.soumu.go.jp/ english/index.html.

FTTH market differs from the ADSL market in that it is so far only making serious progress in urban areas. It has only reached more than 5% of households in just six administrative regions of Japan, such as Tokyo and Osaka. On a finer level, the FTTH market can be divided into housing complex consumers and detached house consumers. As of September 2004, the change in the market share of NTT East and West in each of these sectors is as follows. For housing complexes: (1) NTT East and West, 270,000 subscribers (32% market share); (2) USEN, 170,000 subscribers (20% market share); (3) power companies, 80,000 subscribers (9% market share); (4) Fiber-Bit, 70,000 subscribers (8% market share); and (5) others, 280,000 subscribers (31% market share). For detached houses: (1) NTT East and West, 940,000 subscribers (81% market share); (2) power companies, 200,000 subscribers (17% market share); (3) USEN, 20,000 subscribers (2% market share); and (4) others, 60,000 subscribers (1% market share). The market share of NTT East and West clearly differs between the FTTH market for housing complexes and the FTTH market for detached houses. Surprisingly, NTT East and West have an even lower share in the housing complex FTTH market than in the ADSL market. Meanwhile, in the detached house FTTH market, the market share of NTT East and West is far higher at over 80%. However, the electric power company carriers are also putting up a strong fight in the FTTH market for detached houses in urban areas and the FTTH market seems to be highly variable between housing complexes and detached houses, and between urban regions and provincial regions. Next, let us look at how the charging levels have changed. Each company continued to reduce its prices after 2004, shown in Table 2.7, so essentially the price differential between FTTH and ADSL continued to decrease. In particular, K-Opticom and KDDI have started to provide optical IP phone ser-

takanori ida

80

table 2.7 Changes in Standard FTTH Charges

NTT East Japan (housing complexes) USEN (housing complexes) NTT East Japan (detached houses) USEN (detached houses)

September 2002 $ (PPP $)

September 2003 $ (PPP $)

35 (28) 36 (28) 54 (42) 48 (38)

32 (26) 31 (24) 42 (33) 44 (35)

source: Ministry of Internal Affairs and Communications, http://www.soumu.go.jp/ english/index.html. ISP charges are not included in the figures for NTT East Japan but are included in the figures for USEN. PPP ⫽ purchasing power parity.

vices with quality guaranteed to match that of fixed telephone lines, so by subscribing to FTTH users can get rid of their fixed-line telephones. Accordingly, cases are arising where FTTH works out cheaper than the combination of ADSL and a fixed phone line. FTTH business models can be divided into three types. NTT East and West employ a vertically separated model in which it lets users choose ISP freely, K-Opticom employs a vertically integrated model in which it has their own optical fiber, and KDDI employs a vertically integrated model in which it rents their optical fiber. In 2003 KDDI brought out a “triple play” option combining high-speed Internet access with optical IP phone and video delivery services, and in 2004 K-Opticom also brought out a triple play package and reduced their charges. Faced with this sort of fierce competition from other carriers, NTT East and West also brought out an optical IP phone service from 2005. For NTT East and West providing optical IP phone services means canceling fixed-line contracts, so this means they are swallowing a bitter pill that will not necessarily lead to increased profits. Furthermore, since SoftBank BB seems likely to move from the ADSL market into the FTTH market at some point in the future, it is likely that fresh price cuts will occur. Finally, let us look at the demand structure of the FTTH market. According to a report by the Ministry of Internal Affairs and Communications entitled “Review of competition in the telecommunications area” (2004), the main reasons given for choosing FTTH are (1) very large bandwidth (72%), (2) always-on connectivity (35%), (3) fixed pricing (20%), and (4) home circumstances (20%) (multiple responses allowed). This shows that the advantages of FTTH basically boil down to its high speed. On the other hand, among users that considered choosing FTTH but decided against it, the reasons given were (1) charges are too high (63%), (2) set-up costs are too high (39%), and (3) awkward to set up (32%), so obviously cost is the main deterrent (multiple

Broadband, Information Society, and the National System in Japan

responses allowed). However, the price differential between FTTH and ADSL is rapidly closing, and since it offers the possibility of triple-play contracts it is able to command much higher prices. Accordingly, although FTTH has so far been given a wide berth due to its high charges, it will gain more subscribers in the future as prices fall. However, this still only applies to urban regions, which are the only places where FTTH has so far been made available. No FTTH services have yet been started up in provincial regions so it will be necessary to start by providing FTTH services in these areas. This issue is dealt with in more detail in the next section. the broadband digital divide There are currently two forms of digital divide in broadband services. The first is the digital divide between urban and provincial regions. Although FTTH is now starting to become available in cities and towns with a population of more than 50,000 (October 2004) there is hardly any provision for FTTH starting in towns or villages with smaller populations. For ADSL the critical mass is a population of 5,000.9 The second is the digital divide that occurs even in the same community. As of April 2004, there are many regions where FTTH is still unavailable but, even where it is available, it can only be provided to a small fraction of households. In the future, as FTTH becomes available to a wider range of communities, the regional digital divide will probably become as important an issue as it is in ADSL.10 With regard to the issue of switching from older networks to newer ones, the digital divide can be rephrased as the following two issues: (1) How can metal-based telephone networks be disposed of quickly and systematically? (2) How can an optical IP infrastructure be deployed quickly and systematically? The first issue is the problem of how to avoid causing a digital divide with old universal telephone services. The second issue is the problem of how to prevent a digital divide in new broadband services. The first issue (disposing of the old infrastructure) will probably have to be addressed before the second issue (installing new infrastructure). Both involve time lags and will probably require the implementation of connection policies involving the use of wireless or cable technology or the like. But eventually an official program for addressing the second issue will have to be clearly laid out in order to gain public approval for plans to resolve the first issue. The first and second issues are therefore interrelated. The second issue cannot be resolved without resolving the first, and the first cannot be resolved without having plans in place for resolving the second.

81

takanori ida

82

Market mechanisms do not assess the true value of services that have external effects. In other words, the installation of an optical network would create social benefits and cost savings in the field of public services such as health, welfare, education, and government, but since there is an intrinsic limit to how far these fields are implicated in market pricing mechanisms it is difficult to incorporate this social value into the pricing structure of optical IP networks. Is It Possible for FTTH Networks to Extend into Provincial Regions? Since an optical IP infrastructure is very costly it is often said that it is not worthwhile extending the infrastructure into provincial regions with a small demand density. Let me explain why I think this opinion is incorrect. There is no denying the fact that installing an optical IP infrastructure is a massive undertaking. According to rough calculations it could cost something like $3,000 (PPP $2,800) per household depending on the initial conditions. Naturally the cost would be higher in provincial regions. However, in economics it is important to compare opportunity costs. Although several carriers have started providing optical IP services in urban regions, maintaining the public exchange telephone networks only in provincial regions will in itself lead to higher costs. I do not think the public are likely to accept a situation where cheap optical IP services are available in urban regions while the provinces have to pay higher charges just to keep their existing telephone services. To put it another way, it took a whole century for the telephone network to be completed. Installing a nationwide telephone network was much harder than installing a nationwide optical IP network is likely to be. That is because NTT Co. (more precisely, NTT Public Co.) had to build the telephone network from scratch, including the infrastructure such as exchanges, ducts, telephone tunnels, and telegraph poles. An optical IP network, on the other hand, could be installed more cheaply because it is possible to utilize this hardware legacy of the telephone network. In fact, over the past decade, Japan has made steady progress in converting networks to optical technology as far as the parts just beyond the subscriber lines and is now close to achieving the important target of implementing a nationwide FTTH network. Once an optical IP network has been set up over the whole country it will cost less to maintain the optical IP network due to differences in network technology between public exchange networks and IP networks. It is also likely that the concept of a message area (MA) limited to a telephone exchange unit will also cease to be relevant. Furthermore, in an optical IP world, people would be more likely to use high-level services such as video messaging than low-level services such as ordinary voice calls, thereby

Broadband, Information Society, and the National System in Japan

facilitating the introduction of new ways of working such as video conferencing and telecommuting and leading to savings in terms of travel time and commuting costs. As I have mentioned already, an optical IP network can build directly on the existing telephone network infrastructure such as switching offices, ducts, telephone tunnels, and telegraph poles while the exchange equipment and metal lines can be disposed of so in the long run it can be achieved economically. The rest depends on how the initial cost burden can be alleviated by concentrating the sparse demand in provincial regions. Is There a Demand for FTTH in Provincial Regions? It is often said that there is no demand for FTTH services in provincial regions where many elderly people live, so there is no need to provide these services. I think this view is wrong because it is based on the assumption that FTTH is just an extension of existing Internet browsing services. If the service is only going to be used with a PC then ADSL is quite adequate instead of FTTH. And anyway very few elderly households in the provinces actually own a PC. Instead, optical fiber connections are needed to provide television services. A throughput of several tens of Mbps is needed to transmit and receive highquality digital video, but this can easily be provided when the transmission medium is an optical fiber. It also eliminates the need for building radio masts. Furthermore, the proportion of elderly households in Japan is set to increase sharply in the future. In the provinces, the inferior transportation situation makes it difficult for people to visit their local hospitals or council offices. By making effective use of optical IP services to connect homes directly to public utilities people would be able to avail themselves of just about any service from the comfort of their own home. Subscribers that require FTTH services for purposes such as online gaming and streaming delivery are likely to be few in number. It would be nice if elderly subscribers in the provinces could be allowed to spend their time more productively by having easy access to all kinds of services—e.g., health, welfare, care, education, and administration—in their own homes through a TV connected to a broadband network, even if they are not initially aware of how they might benefit. The Final Stretch Toward a Nationwide FTTH Network NTT has declared its target of implementing an FTTH service environment with 30 million users by 2010. There are currently 60 million households with subscriber telephone contracts, so the figure of 30 million represents half of these. This is a commendable move because NTT now has a concrete target to aim for and is striving to connect with the next stage of development.

83

takanori ida

84

With regard to charges, the differential between ADSL and FTTH has almost disappeared. To use ADSL subscribers must pay a monthly phone subscription charge of over $14 (PPP $11) and an ADSL usage charge of about $20 (PPP $16). On top of that they also have to pay for their phone calls. On the other hand, FTTH users have to pay a FTTH usage charge of about $40 (PPP $31). But since they can use optical IP phone services they are able to dispense with a subscriber phone line and hence it is often the case that FTTH works out to be the cheaper of the two. Accordingly, in terms of cost there are no financial obstacles to prevent current ADSL users from migrating to FTTH. And what about services? This is a more important issue. So far, nobody is providing “killer content” that fully utilizes the superior upstream and downstream bandwidth of 100 Mbps offered by FTTH. That makes FTTH nothing more than an efficient replacement for existing services such as video phone services, which can be carried by IDSN networks, and multi-channel broadcasts, which can also be carried by cable or satellite systems. Many people are frantically searching for services with a bit more potential for profit. I do not think that optical IP services themselves need to be profitable. Although the market prices are determined by reflecting the utility to users and the costs to industry, they do not provide a suitable reflection of various “external effects.” The social value and market value of an optical IP infrastructure are only separated by positive external effects. So what sort of external effects can be expected to result from the completion of a nationwide optical IP infrastructure? In terms of public policy on issues such as health, welfare, education, and administration, the value of an optical IP infrastructure, market pricing structures are of limited use. The public services that depend on the availability of an optical IP infrastructure may not be all that highly profitable but they ought to be of great benefit to society. An optical IP infrastructure should be compared not with telephone networks but with major infrastructure elements such as airports, freeways, and high-speed rail links whose costeffect ratio is questionable in terms of social need. conclusion In this chapter I have described the current state of Japan’s broadband services. First, I separated the reasons for Japan’s “broadband big bang” into (1) unintentional benefits and (2) intentional benefits of Japan’s telecommunications policy, (3) the involvement of a wide range of newcomers and (4) NTT’s self-preserving efforts. Next, I analyzed Japan’s broadband services, citing the latest figures for the ADSL and FTTH markets. Finally, I discussed the digital divide issue and some measures for overcoming it.

Broadband, Information Society, and the National System in Japan

Providing a public infrastructure requires public vision. The government should therefore pledge itself to providing a nationwide optical fiber network by a particular year (e.g., 2015). It will also be necessary for NTT to respond to this pledge. Due to the distractions caused by the breakup issue between NTT and the Ministry of Posts and Telecommunications, Japan’s telecommunications industry is still failing to engage in basic discussions about the future. Now they have finally been released from the ghosts of the past. The Ministry of Internal Affairs and Communications and NTT should eliminate future uncertainties by presenting a clear vision of the construction of a nationwide optical fiber network. What is needed now is confidence in the future of the telecommunications industry. acknowledgments I would like to thank Professor Martin Fransman (the editor of this book) for his helpful advice on writing this chapter and Adam Peake for providing me with useful materials to work from. I would also like to thank Hideyuki Ohashi and all the staff at the Ministry of Internal Affairs and Communications who worked with me on a joint study to assess the effective competition in Japan’s telecommunications. Most of the data presented here was taken from this effective competition review.

notes 1. In this chapter an exchange rate of $1 ⫽ ¥108 and an OECD PPP (purchasing power parity) rate of $1 ⫽ ¥137 are used based on the 2004 statistics published by the OECD. 2. The council’s report called for a regrouping of NTT along the following specific lines: (1) NTT to be regrouped into a long-distance carrier and two separate regional carriers for east and west Japan; (2) the long-distance carrier should be completely privatized and should admit competition in international communication, CATV, content provision, and regional communication services; (3) the regional carriers should be administered as semigovernmental corporations for the time being but should admit competition from other carriers between regions. 3. In January 2001, the Ministry of Posts and Telecommunications was regrouped together with the Management and Coordination Agency and the Home Affairs Ministry to form a new Ministry of Internal Affairs and Communications. In this chapter I will use the titles applicable at each point in time. 4. It should be noted that Tokyo Electric Power Company (TEPCO), the largest electric power company in Japan, started late in providing an FTTH service compared with Kansai Electric Power. The reason for this is that the companywide cover-up of

85

takanori ida

86

the troubles of atomic power generators was revealed in 2002. The president of TEPCO took responsibility for it and resigned, but TEPCO was obliged to refrain from developing new businesses on a large scale for some years. However, TEPCO has aimed at a recovery in the FTTH market since 2005. Accordingly, the competition in the FTTH market will become fierce even in East Japan, including the Tokyo metropolitan area. 5. They also achieved an operating profit in 2003 when they purchased Osaka Media Port. 6. The number of subscriber telephones peaked at 61.46 million in 1997 and has been declining since then. 7. However, 72,600 of these positions were outsourced so the number of acute job losses was 100,000. 8. The detailed figures are (1) Japan $0.09; (2) Korea $0.25; (3) Belgium $1.15; (4) Hong Kong $1.27; (5) Singapore $2.21; (6) Macao $2.56; (7) New Zealand $2.71; (8) Canada $3.25; (9) Holland $3.36; (10) U.S.A. $3.53; (11) Israel $3.98; (12) Germany $4.42; (13) Denmark $5.81; (14) Austria $5.89 (source: “Internet Reports: Birth of Broadband,” ITU, 2004). 9. The details are given as follows: (1) 50,000 or more (478 cases)—FTTH 93.9%, ADSL 100%; (2) 10,000 through 50,000 (1186 cases)—FTTH 33.8%, ADSL 99.2%; (3) 5,000 through 10,000 (783 cases)—FTTH 10.6%, ADSL 89.9%; (4) 1,000 or less (676 cases)—FTTH 2.1%, ADSL 48.8% (note: the numbers represent the percentage of communities where the service is at least partially available; source: Ministry of Internal Affairs and Communications, http://www.soumu.go.jp/s-news/2005/ 050201_1.html). 10. The details are given as follows: (1) available to 100% of households—FTTH 4.6%, ADSL 51.3%; (2) available to 80 –100% of households—FTTH 9.6%, ADSL 20.6%; (3) available to 50 – 80% of households—FTTH 6.8%, ADSL 9.3%; (4) available to 0 –50% of households—FTTH 9.2%, ADSL 5.1%; (5) available to 0% of households—FTTH 69.8%, ADSL 13.7% (note: the numbers represent the percentage of communities where the service is at least partially available; source: Ministry of Internal Affairs and Communications; http://www.soumu.go.jp/s-news/2005/ 050201_1.html).

references Ida, T. 2005. The broadband market in Japan. In Japanese telecommunications market and policy in transition. Edited by R. Taplin and M. Wakui. Routledge. Ida, T., and T. Kuroda. 2005. Discrete choice analysis of demand for broadband in Japan. Journal of Regulatory Economics 29.1: 5–21. International Telecommunication Union (ITU). 2004. Internet report: The birth of broadband. Available online at http://www.itu.int /. Ministry of Internal Affairs and Communications. 2004. The review of competition in the telecommunications area. Available online at http://www.soumu.go.jp/ english/index.html.

3 Broadband, the Information Society, and National Systems The Korean Case Inho Chung

great changes have taken place in the telecommunications and information services industry throughout the 1990s and one of them was the explosion in the supply of multimedia services integrating voice, data, and graphics. In the centre of that change the Internet stands out as a dominant driving force. Because the Internet makes highly diverse and in-depth information available in a way not possible through other existing media, it has become a very important objective for governments to provide the general population with access to this media. Governments have to create the right environment and fulfil many conditions in order to achieve this goal, but the construction of broadband infrastructure and the availability of broadband services in regard to both access and content are the most fundamental of these requirements. Governments are able to choose a variety of approaches to achieve their goals. They could, for instance, allow the market to solve problems by itself, or they could enact a more active intervention policy to promote market demand. Also, governments have to strike a balance in the trade-off between fair competition and market development. In determining which approach to take, governments have to consider not only their telecommunications industry but also their cultural, social, economic, and demographic conditions. Korea is the number one country in the world in the broadband services market. According to statistics published by the OECD (Organisation for Economic Co-operation and Development), the broadband penetration rate in Korea is more than two times that of any other advanced country. This phenomenon is somewhat surprising because Korea had never been considered a leader in telecommunications and information technologies, although it was one of the countries with the fastest growing telecommunications and IT sectors.

inho chung

88

To understand this interesting phenomenon and to gain insight from the Korean case, it will be useful to study the Korean broadband services market, including not only the industry itself but also the relevant political, regulatory, and socioeconomic aspects. Although it is still too early to contend that the Korean case is an example that other countries should follow, it does provide other countries with a very interesting case study. In this chapter, I will cover the current state of the broadband market in Korea and explain what factors contributed to the success of broadband penetration. After that, I will discuss how this market will develop in the future and which emerging or potential services will lead this market. the market situation The Structure of the Broadband Market In July 1998 the commercial broadband service in Korea was launched by Thrunet, a new service provider using cable modem, followed by Hanaro Telecom in April 1999, a new local telephone company. In the same month, Korea Telecom (KT), the largest telephone operator, began to provide broadband services using ADSL technology. Although KT was the last starter in this market, it caught up with the forerunners easily because of its pre-existing national network, strong marketing capability, high brand recognition, and abundant financial resources. The Korean broadband market is well balanced in terms of technologies and business structure. Of the current 11 million broadband users, more than 58% subscribe to DSL; cable modem is second with around 4 million subscribers comprising 34% of all broadband users; and LAN is third with an 8% market share. KT, the number one carrier in Korea, is the most dominant service provider in this market, providing broadband services by xDSL technologies and Ethernet. Hanaro Telecom, the number two service provider, mainly depends on cable modem technology with a small proportion of xDSL technologies, while Thrunet, the number three service provider, relies totally on cable modem technology. KT accounts for about 51% of the total market, followed by Hanaro Telecom with almost 24% and Thrunet with just over 11% market share (see Table 3.1). There are also scores of small service providers including cable TV companies in this market. Although KT has a total market share of over 50%, it competes rigorously with other service providers in urban areas, especially those areas with large numbers of apartment buildings. Other competing service providers concentrate their efforts on profitable urban sectors, and ignore large parts of rural

Broadband, the Information Society, and National Systems

89

table 3.1 Number of Subscribers and Market Share of Each Service Provider At 31 December 2001 At 31 December 2002 At 31 December 2003

KT Hanaro Thrunet Cable TV Others

Number of Subscribers

Market Share (%)

Number of Subscribers

3,858,194 2,059,960 1,302,638 — 584,723

49.4 26.4 16.7 — 7.5

4,922,395 2,872,351 1,301,620 367,135 941,985

Market Share Number of (%) Subscribers

47.3 27.6 12.5 3.5 9.1

5,589,058 2,725,563 1,293,364 619,103 951,411

At 30 June 2004

Market Share (%)

Number of Subscribers

Market Share (%)

50.0 24.4 11.6 5.5 8.5

5,900,038 2,775,638 1,289,057 758,603 894,489

50.8 23.9 11.1 6.5 7.7

source: Ministry of Information and Communications, Korea, www.mic.go.kr.

areas where there is not much demand and a lack of economy of density. According to statistics collected by KT, by the end of 2003 its market share was 43.7% in Seoul and its vicinity where approximately 40% of the population lives; this was remarkably lower than its market share in other areas, which stood at 57.2%. This phenomenon is even more striking in respect of the cable TV companies; their market share in Seoul and its vicinity was 10.1%, while their average market share in other areas was no more than 3.8% in the same period. This result shows that the competitive service providers actually used a cream-skimming strategy throughout the take-off period. Price Competition The price level of broadband services in Korea is one of the lowest in the world. This is one indicator of the performance of Korean broadband service providers and is one of the success factors in the market at the same time. The high population density, especially in urban areas, enables service providers to lower their costs, but they set their prices at a level that was insufficient to recover their costs, at least in the short run (see Table 3.2). It is clear that these service providers have been disruptive competitors in the sense that their prices did not cover the short-run marginal costs. As a result of this pricing strategy both Hanaro and Thrunet face serious financial difficulties; Hanaro has had to sell large shares of its equity to foreign financial institutions, and Thrunet is now under the management of the judicial system and awaiting new ownership.1 It is not clear, however, whether these low prices for broadband services were intentional or the result of the competition with KT, which is able to provide such services at low prices without worrying too much about its costs.

inho chung

90

table 3.2 Monthly Charges for High Speed Internet Services, October 2004

VDSL ADSL Cable

KT Hanaro KT Hanaro Hanaro Thrunet

Monthly Charge ($)

Monthly Modem Rental Fee ($)

26.2–52.4(38.9–77.8) 41.0–48.9(60.9–72.6) 26.2–34.9(38.9–51.8) 33.1(49.2) 29.6(44.0) 24.4–29.6(36.3–44.0)

2.6(3.8) 4.3(6.4) 2.6(3.8) 4.3(6.4) 4.3(6.4) 7.8(11.6)

Maximum Speed

2–20 Mbps 2–8 Mbps

Number of Service Types

4 2 2 1 1 3

sources: KT, www.kt.co.kr; Hanaro, www.hanaro.com; Thrunet, www.thrunet.com. PPP in dollars shown in parentheses.

One theory suggesting that providers intentionally adopted such a disruptive pricing strategy from the beginning is based on the fact that earlier versions of broadband services such as ISDN did not become popular in the market mainly due to their high price levels. Accordingly, service providers probably determined that, to some extent, a low price level was necessary to promote demand in the emerging, promising market. The hands-off policy of the Korean government could be one of the reasons why they pursued such an aggressive policy. Broadband belonged to the advanced services, distinguished from the basic services such as telephone services according to the legal classification system of the Korean telecommunications business law, and therefore did not require much intervention or regulation by the government. Service providers may have believed that they could not hope to hide under the umbrella of protective pricing provided by the government for markets such as international and domestic long-distance telephone markets in their early stages of competition. In those markets the government forced the incumbent, KT, to keep its prices higher by predetermined amounts than those of new entrants. This policy was adopted to give new entrants some time to build their market share and financial resources before they were exposed to intense competition from the incumbent. Without such protection, they might have had no choice but to get a large customer base with lower prices earlier than their competitors. The stringent price competition with KT is also one of the factors that forced the new service providers, such as Hanaro Telecom and Thrunet, to adopt such a disruptive pricing strategy. They incurred enormous initial sunk costs because they had to construct their own networks from scratch or lease networks from cable TV companies, but they could not set their prices at an

Broadband, the Information Society, and National Systems

adequate level to cover their costs because of the competitive pressure from other providers, especially KT. Considering that the speed of penetration in this market would not have been so fast if the competition had been normal, in the sense that service operators had set their prices higher than marginal costs, their disruptive strategy turned out to be one of the important factors for the success of the Korean broadband services market. On the other hand, the competitors’ disruptive pricing strategy has had a great impact on the behaviour and strategy of KT, the current leading service provider. The very aggressive promotion of Thrunet, followed by Hanaro Telecom, was probably one of the market forces that led KT to choose xDSL as its main solution for high speed Internet. Before that, there had been a very heated and prolonged debate inside KT on the choice of technology to provide broadband services, especially between ISDN and xDSL. Many engineers preferred ISDN to xDSL because of ISDN’s advantage in speed, flexibility, and controllability. However, the popularity and fast penetration of the new entrants’ services in the market forced KT to choose ADSL as it could neither provide ISDN services in many of its service areas nor provide this service at prices comparable to an ADSL service. ADSL was an easy solution for KT because it could install nationwide ADSL networks using the existing telephone networks in a very short period of time. Success Factors The performance of the Korean broadband services market cannot be explained satisfactorily by any one factor. For example, many consumers were well educated and were accustomed to the Internet through dial-up services, the prices of which were partly flat and cheap, and they were ready to use broadband services when they became available. This, however, fails to explain why broadband services have been slow to catch on in the U.S., a country with similar customer conditions. The following factors have probably jointly contributed to the rapid penetration of broadband services in Korea. The “Hurry, Hurry” Mentality

It is interesting to note that the general psychology and characteristics of Koreans and their social environment played a significant role in facilitating the rapid spread of broadband services. In general, they show a tendency to accomplish their work as soon as possible and they place a high value not only on how well it was done but also on how fast it was done. This characteristic

91

inho chung

92

seems to play an implicit but very important role in the fast penetration of broadband services. Koreans are dissatisfied with connections slower than those of their neighbour and have an inexorable demand for faster connection speeds. Government Policy

As the Korean government did not impose any strict regulation on the broadband market the carriers did not have to adhere to any conditions on entering this market or in pricing their services. This policy encouraged many carriers to enter and to compete rigorously with each other. This vigorous competition reduced prices to a low level, which was helpful in obtaining a critical mass of customers for sustainable market growth in a relatively short period of time. The three major players, KT, Hanaro Telecom, and Thrunet, competed not only on prices but also in many other marketing activities, so speeding up the diffusion of services. Geographical Advantages

In Korea more than 50% of households live in apartment complexes, and population density is very high, which enables the carriers to realise economy of density and lower their costs for building and operating networks. It is also important that more than 93% of households are within 4 km of central offices because the quality of broadband services using DSL technologies critically depends upon the distance from central offices. In fact, the quality of DSL services deteriorates rapidly if the distance from central offices is over 4 km. Changing Markets Subscribers to broadband services have increased dramatically, although the growth rate has slowed considerably in recent years. The number of subscribers had reached 11.3 million by September 2003, covering 74% of households. However, the net increment of subscribers decreased after reaching its peak in 2001. As the market matures and new customers become difficult to find, competition between service operators to lure customers away from each other will become even harder. Recently, aggressive selling by the cable TV companies of broadband services bundled with cable TV services is the most outstanding trend in the Korean broadband market, and this approach seems very threatening to other operators. This is an attractive alternative for cable TV subscribers, who can cut their costs and get one-stop service by just switching their broadband service

Broadband, the Information Society, and National Systems

provider to the cable TV company to whose broadcasting services they currently subscribe. The cable TV companies are rapidly becoming a new competitive force in this market with a small but fast growing market share (6.8% by August 2004). They provide broadband services in order to create a new source of income as their traditional cable TV markets mature and do not generate much revenue due to low prices. They are very disruptive players because they often provide their services at less than half the prices of the major players. However, some in the industry argue that in the end they will only be minor competitors because the quality of their services deteriorates drastically as the traffic on their network grows. In providing broadband services they mainly use their own cable TV networks with some conduits leased from KT at very cheap prices that were set before they entered the broadband market. As they become more threatening as competitors, KT is now trying to raise the prices for conduit leases because it is one of the reasons why cable TV companies are able to lower theirs. The Effects of Broadband Services on Economic Activities The effects of broadband Internet services can be observed in many different sectors. The most remarkable phenomenon is that a high proportion of offline activities are being shifted into online activities, and this is most remarkable in the financial sector. For example, the number of online banking users has dramatically increased from 0.12 million in 1999 to 22 million in 2003, and more than 30% of all transactions are being carried out online. Online stock trading also showed explosive growth from $10 billion (PPP $14 billion) in 1999 to $2,827 billion (PPP $4,198 billion) in 2003, and currently more than 50% of all transactions are carried out online. In addition, in the business sector Internet transactions are showing an outstanding growth trend with the value of online transactions increasing from $51 billion (PPP $75 billion) in 2000 to $208 billion (PPP $309 billion) in 2003, and 17% of all business sector transactions are now carried out online. According to a survey in 2002, 18% of companies with more than 50 employees were involved in e-commerce. The Various Types of Broadband Internet Service The wide availability of broadband access facilitates the development and use of Internet applications and services, but at the same time the reverse is also true. In the initial stages of broadband services, Internet sites and services were

93

inho chung

94

not well developed in Korea, so users frequently had to access the websites of other countries, especially the U.S., to get information and use services. However, as the number of Internet users increased enormously with the spread of broadband services, the number of Korean websites grew very fast and diverse services and applications in Korean became available. The existence of these rich and varied applications in the Korean Internet market is thought to have facilitated the penetration of broadband services by attracting even more users. Web Portals

In the Korean Internet services market the web portal is one of the most important applications that contributed to the explosive growth of Internet usage through broadband networks. As is currently observed in other countries, portals provide very diverse services to Internet users, such as web searches, e-mail, shopping, dictionaries, music downloads, and so on, to expand their own customer bases. Some companies are now charging for premium services to avoid too much dependence on Internet advertising as their primary source of revenue. Daum, the number one portal in Korea, attained its current position by providing well-developed and convenient community services for its members, while Naver, the second largest portal in Korea, is well-known for its powerful exploring services, especially for Korean-specific information and knowledge in various fields. Yahoo is only the fourth-ranked portal, which is surprising considering its dominant position in the U.S. and many other countries, followed by Cyworld operated by SK Telecom, the dominant mobile telephone carrier in Korea, which is particularly popular among young people due to its mini-homepage services. The rigorous competition among the portals works as a driving force to expand the Internet user base and the area of usage by developing many new and attractive applications. Internet Music

Music in various formats, such as mp3 and wav, is downloadable from the Internet free of charge, which makes this a very important type of content on the Internet, although not in commerce. The wide availability of free music webcasting and download services contributed to the rapid penetration of broadband services but, at the same time, it also contributed greatly to the shrinking of the music industry. The strong resistance from the offline music industry and the subsequent court ruling in its favour are expected to change these services to charged premium services. The decision to charge for music downloads and web-casting, however, could be an important turning point in

Broadband, the Information Society, and National Systems

the growth of commercial music services throughout the Internet, although at least a temporary downturn in the use of Internet music seems inevitable. Currently, the most popular music service in Korea, which is also commercially successful, is a bell sound download from the Internet to mobile phones for use as a ring tone. The various bell sounds available attract many customers, mainly young people, and had a market share of 77% in commercial music services on the Internet in 2002.2 Internet Films

The film industry in Korea has experienced great success and explosive growth in both the number of viewers and films. Some Korean films have received awards at important international film festivals, and many have become hits in the market with a domestic film market share of more than 50%. The popularity of films makes them a very important application in Internet services in Korea. Currently, after being shown in theatres, many films are available on the Internet before video or DVD release, while some are released on the Internet at the same time as they are in theatres. This phenomenon shows the status of the Internet as a new channel of distribution for films although this market is still in the early stage of development. There are three types of distribution through the Internet. First, web portals provide film services as one of their charged services and are trying to develop this as one of their main revenue sources. Second, there are some independent Internet film sites that buy the screening rights from many film studios. Third, there are partnerships between Internet film companies and independent film studios that make films solely for the purpose of Internet viewing. Online Games

Online games account for one-third of the domestic games market in Korea. This share is known to be much higher than the average for other countries and the early penetration of broadband services could be the explanation. In the early stages of broadband penetration, the availability of sites for PC use (called PC bangs in Korean or Internet cafes in English), principally for sending e-mail and searching the Internet, enabled many customers, mainly young people, to access the Internet through a broadband network without having to have their own PCs or having to subscribe to broadband services. Although the number playing online games at home has been increasing steadily due to the wide availability of high speed Internet at home, PC bangs still play a signifi-

95

inho chung

96

cant role in facilitating online games by providing ubiquitous access as well as many charged games at flat rates. The availability of high speed Internet led to the development and release of many domestic online games, with some of them becoming extremely popular and attracting many young players, making this a very important application in the Internet services market. Internet Newspapers

Internet newspapers are a very popular service and seem to be replacing parts of the offline newspaper market. Major newspaper companies also operate online sites and provide differentiated services to Internet users. There are newspapers that provide news services only through the Internet, and they are becoming increasingly popular by providing very specialized and up-to-date information on topics such as stock markets and information technology markets. They are, however, also experiencing severe competition from web portals that provide real-time news. Another problem is that most Internet newspapers provide their services free of charge to the public. This makes the Internet newspaper business less attractive as a commercial application, but many more online newspapers are starting to charge subscription fees or usage fees for some specialised and premium services. This trend is expected to make the Internet newspaper a promising application in the future. Online Education

Online education is one of the most important applications in the Korean broadband market. The main programme provided online is for supplementary lessons for schoolchildren and for certification and language tests. In particular, a new programme for the university entrance test provided by public education broadcasting has been assessed as having greatly contributed to the use of online education. The very intense competition to enter top level universities in Korea leads high school students to a heavy dependence on private education and, while elite teachers are in great demand, they are available only to those students who can afford the very expensive tuition costs. The government views this as a social problem of inequality and has tried to solve it by creating programmes of lessons by famous private lecturers broadcast through the Internet and TV broadcasting. Recently, online education has extended to the market for business training in corporate and adult education. In 2003 online education had 28% of the total corporate education market.

Broadband, the Information Society, and National Systems

97

table 3.3 KII Project Budget

KII-G (government) KII-P (private) KII-T (KOREN) R&D costs Total

First Stage (1995–1997) ($m)

Second Stage (1998 –2000) ($m)

Third Stage (2001–2005) ($m)

151.(224) 1,731.(2,570) 21.(32) 698.(1,037) 2,601.(3,863)

233.(346) 6,081.(9,031) 26.(39) 274.(407) 6,614.(9,823)

363.(540) 15,153.(22,503) 113.(167) 1,190.(1,767) 16,819.(24,977)

Total ($m)

747.(1,110) 22,965.(34,104) 160.(238) 2,162.(3,210) 26,034.(38,663)

source: Ministry of Information and Communications, Korea, www.mic.go.kr. PPP in million dollars shown in parentheses.

the policy and role of the government The National Plan for Information Infrastructure In the development of the Korean broadband services market the government has played a very important role. It has not only helped facility-based competition to take place in this market by adopting a non-intervention or hands-off policy, but it has also established comprehensive plans for constructing an infrastructure for broadband services and has initiated several programmes to facilitate the use of the Internet through broadband networks (see Table 3.3). KII-G (Korean Information Infrastructure-Government)

The government launched the Korean Information Infrastructure (KII) Plan in 1995 with public financial resources and laid the foundation for high speed Internet services. It divided the spanning periods into three stages, each of which had a sub-goal: in the first stage, 1995–1997, the government set the goal of constructing ATM testing networks connecting about 80 major areas, and in the second stage, 1998 –2000, it had a target of completing the ATM switched network connecting all the major cities and a large number of small and medium-sized towns. The accumulated amount spent on the KII-G project until 2000, the end of the second stage, was $384 million and is expected to reach $747 million by the end of 2005. As a result of this plan, 272 ATM switches were installed during the second phase and large cities were interconnected at a speed of 2.5 to 5 Gbps, while small and medium-sized cities were interconnected at a speed of 155 to 622 Mbps. Government organisations, public research centres, and major libraries were connected with each other through this government-owned

inho chung

98

network, and elementary and secondary schools were able, with special discount programmes, to access high speed Internet services using this network from 2000. In the third stage, the government set a more ambitious goal of completing a super-speed information network covering the whole country that anyone would be able to access at any time, anywhere, and with any terminal. The government also plans to keep upgrading the established network facilities. According to this plan, by 2005, 84% of households will have access to high speed Internet services and mobile phone users will be able to use Internet services at 2 Mbps. In addition, government agencies and public firms are expected to use services at up to 622 Mbps by that year. KII-P (Korean Information Infrastructure-Private)

The Korean government also encouraged carriers to invest in constructing broadband access networks in parallel with the KII-G project, which aimed to construct backbone networks. The carriers are expected to spend $22,965 million on this project by the end of 2005. The government did not set any regulations on entry and pricing in this market, nor did they impose any obligation on the incumbent network holders such as telephone carriers and cable TV companies to provide unbundled network elements. This laissez faire policy led to fierce facility-based competition among carriers and, as a result, investment in the access part of broadband networks has been greatly facilitated. For example, more than 60% of commercial and public buildings with more than six floors are connected to the public network by optical fibre (fibre to the office) and many apartment complexes have access to fibre optic networks, or at least to the communications offices inside those complexes (fibre to the curb). KII-T (KOREN)

The government established KOREN (Korean Advanced Research Network) interconnecting research institutes in six major cities for the purpose of network testing. The research institutes are able to carry out their research and experiments jointly using this test bed without worrying about network investment costs. This network is also linked to other countries in Asia to allow joint research. In 1998 KOREN was linked to Japan and Singapore to form the Asia Pacific Information Infrastructure (APII) test bed.3

Broadband, the Information Society, and National Systems

Other Programmes National Education Programme on Information

In 2000 the Korean government also established the “Plan of Providing Information Technology Education to 10 Million People,” which had the objective of providing Information Technology education to housewives, soldiers, the disabled, and the elderly, either free of charge or on a discounted basis. This project has been praised for having contributed to the expansion of the consumer base of the Internet and information services. The programme mainly consisted of providing participants with basic lessons on PCs, including some popular and practical applications, and how to use the Internet through public and private institutions. The government provided financial assistance to the private education institutes that participated in this programme. Certification for Broadband Buildings and Apartments

The government introduced a very unique system of “certification for broadband buildings and apartments” in May 1999. Under this system, the government issues certificates to buildings and apartments if they have the infrastructure for high speed Internet that satisfies an a priori set of criteria. The government awards such buildings and apartments with emblems representing their capability to provide in-house Internet and automation services. This system was evaluated to have been successful in promoting in-house informatization by encouraging many new buildings and apartments to be constructed with such facilities. Internet PC Programmes

In 1999, the government began to offer to the general public, 40% of whom live in rural areas, PCs with basic features at low prices. These PCs were made by firms who applied for the programme and were selected by the government. The low prices and various payment options, such as paying on an instalment basis, were popular with consumers and this helped many people become familiar with PCs as well as preparing them to use Internet services. Also, some consider that this programme contributed to the decrease in the general prices of PCs by putting competitive pressure on the PC market in Korea. IT Education in Schools

Another policy aimed at improving people’s ability to use PCs was the “School Informatization Program,” in which the government provided elementary and

99

inho chung

100

secondary schools with high speed Internet access lines. About 10,000 schools benefited from this programme and were linked to the high speed Internet network by December 2000. The government also gave certificates to students who passed examinations on the use of PCs and the Internet, and encourages universities and colleges to give some credits to students with such certificates. Local Loop Unbundling (LLU) The Ministry of Information and Communication (MIC), the Korean government department in charge of telecommunications, has not only made important decisions on telecommunications policy affecting the development of the broadband market but has also become involved with the regulation of carriers. There is another regulatory agency, the Korean Communications Commission (KCC), which has a remit for dealing with issues concerning fair competition environments and consumer protection of telecommunications services as well as with arbitrating disputes among telecommunications services carriers and between users and carriers. In reality, however, MIC has made the major decisions in domestic telecommunications policy and regulation, including local loop unbundling (LLU), and the role of the KCC has been very limited. MIC did not impose obligations on the incumbent telecommunications operators to provide unbundled elements of their local loop to their competitors until 2002. As a result, this non-intervention policy led new entrants in the broadband services market to construct their own network facilities. It is not clear whether the belated introduction of LLU into local telecommunications services markets was the intended result of government policy for facility-based competition. Indeed, Hanaro Telecom, the new and secondranked local telephone company in Korea, was actually not very active in developing their share in the local telephone market but instead put more emphasis on developing the newly emerging high speed Internet market. Hanaro mainly depended upon cable TV companies’ networks in providing their broadband services. This strategy was not irrelevant because Hanaro mainly targeted office buildings and apartment complexes in urban areas where cable TV networks were already installed. For this reason, it was not vital for the company to lease the local network elements of KT that were more appropriate for ADSL services. In 2002 the government changed its policy and decided to introduce LLU. Such a policy appeared to be quite rigorous from the point of view of the incumbent telecommunications operator, KT, because the government not only forced KT to provide all kinds of LLU to competitors, including full line

Broadband, the Information Society, and National Systems

sharing, bit stream sharing, and high speed Internet lines, but also to do so at prices arguably lower than costs. There have been serious debates between the government and operators on the level and method of calculating proper prices for using network elements based upon a long-run incremental cost model. However, it was evident that the government had changed its policy to discourage the facilities construction competition among operators as that might result in having too much capacity. Also, it might have been the hope or expectation of the government that such a policy would intensify competition in high speed Internet services in rural areas where cable TV networks are not so widespread. However, at the very least, the lack of LLU may have been helpful in encouraging double mode competition (i.e., competition between ADSL and cable modem) because Hanaro Telecom might have chosen ADSL technology if LLU was strongly enforced by the government. Judicial Institutions With the exception of the regulatory agency, the role of judicial institutions in regard to the activities of competitors seems not to be as significant in Korea as it is in the U.S., where the courts make important decisions that greatly impact on the industry. This phenomenon is understandable because KT was a government-owned public company4 and therefore could not take a strong position against government policy. Also, it was more profitable for new operators, such as Hanaro Telecom and Thrunet, to win more favourable policies from the government by persuasion or lobbying than to appeal to the courts. Fund for Information Technology The Korean government operates a unique financial funding programme called the “Information Facilitation Fund” (see Table 3.4). It was established in 1996 by a special law and consists of contributions from the government and telecommunications operators as well as income from investments and interest. The fund is regulated by a law relating to the construction and utilization of high speed information networks, the business of providing information to the public, R&D on IT, the creation and diffusion of standards in IT, and the education of people employed in the IT industry. The government selects the companies involved from applicants that satisfy certain criteria and gives them a loan at very low interest rates over a very long period of time. Many countries operate similar kinds of funds for information facilitation, but these generally come with some limit on their coverage and length of

101

inho chung

102

table 3.4 Budget for Informatization and Information Facilitation Fund

Budget for Information Facilitation Percentage of General Budget IFF Total

1996 ($m)

1997 ($m)

1998 ($m)

1999 ($m)

2000 ($m)

2001 ($m)

436 (648)

2002 ($m)

2003 ($m)

619 (919)

742 (1,102)

1,078 (1,601)

1,187 (1,762)

1,313 (1,949)

1,407 (2,090)

1,459 (2,167)

0.79

0.99

1.05

1.40

1.43

1.42

1.47

1.50

313 (465) 749 (1,113)

484 (718) 1,102 (1,637)

676 (1,003) 1,418 (2,106)

742 (1,102) 1,821 (2,704)

653 (970) 1,840 (2,732)

857 (1,272) 2,169 (3,222)

1,039 (1,543) 2,446 (3,633)

934 (1,388) 2,393 (3,554)

source: Ministry of Planning and Budget, Korea, www.mpb.go.kr. PPP in million dollars shown in parentheses.

time. However, this fund is unique in its wide coverage and flexibility, and it enables the government to provide loans with only slight restrictions to firms in the IT industry. The fund is managed and operated by a public institute, the Institute of Information Technology Assessment, and approximately $6,494 million (PPP $9,645 million) of the total $8,984 million (PPP $13,342 million) in the fund had been loaned to IT companies on a cumulative basis by the end of 2003 (www.iita.re.kr). In particular, in 1999 the government provided loans of $77 million (PPP $114 million) to companies for investment on access networks required for broadband services, which was helpful in allowing new business to flourish. Similar loans have been provided for broadband services since then, but with additional conditions for investment in less densely populated areas. new development and future perspectives Changes in the Pricing System As the broadband Internet access market matures, the revenue growth rate is beginning to stagnate. Market revenue reached $3.1 billion (PPP $4.7 billion) in 2003, but the growth rate fell from 177% in 2001 to 15% in 2003. Also, the number of subscribers had reached 11,618,000 by June 2004 with a penetration rate of 77.9%, which seems to be close to the saturation point. However, telecommunications carriers have tried to expand this market by changing the price structure from a fixed-rate system to a usage-based system. This move is facing strong resistance not only from heavy Internet users but also from below average

Broadband, the Information Society, and National Systems

users because they have some vague fear that this change might result in unexpected drastic increases in Internet prices. Because of this fear it is not easy to see such a change in the pricing system in the near future, but service providers will not give up trying to make these changes under the impending financial difficulties and the burden of traffic becoming increasingly heavier over time. New Emerging Technologies and Services The advancement of high speed Internet services is enabling service operators to develop many potentially promising technologies and services. They are expected to be a new growth engine for service operators who are suffering as a result of the present sluggish market. Their efforts to develop new technologies and services are supported by the government, which also expects to find a new growth engine for the national economy in the IT industry. VDSL and LAN

VDSL is a DSL technology and can reach a speed of 52 Mbps, at least theoretically. This technology allows customers to access the Internet at a speed of 13 Mbps, which is much faster than ADSL but has a shorter service range of 1.5 km away from central offices compared to 6 km for ADSL. Fortunately, however, more than half the population live in communal buildings such as apartment complexes, which usually lie within the service range. This technology has been available since August 2002 and has rapidly replaced ADSL, partly due to KT’s aggressive strategy aimed at getting customers back in urban areas where it had suffered from poor quality relative to their competitors’ services. The availability of VDSL services at prices not much higher than those of ADSL made this service popular, with 46.3% of new subscribers selecting this service over other alternatives in November 2002. Another new technology that is mainly installed in apartment complexes is the LAN system, which makes it possible to access the Internet and make a telephone call with a single connection. This technology is also important strategically to the suppliers because of its low installation and operating costs and easy upgrade for multimedia services that might be popular in future home networking services. Wireless LAN: Nespot

From the perspective of wired telecommunications companies, the next stage of development in high speed Internet services is to make the access part of

103

inho chung

104

the Internet wireless using wireless LAN technology. The Korean government allowed them to provide commercial services using 2.4GHz that had been allocated for public use under the condition that they would not cause disturbances such as spectrum interference. Telecommunications companies are currently approaching the market from two directions; one of their targets is hotspots such as airline terminals, hotels, department stores, and so on, while the other is businesses and households that have more than one PC. It is clear that some consumer groups welcome this alternative for the freedom it gives from the complexities and inconveniences of wiring but, at least in the beginning stages, this service does not yet appear to be successful in the market because of its poor quality and lack of appropriate applications. KT is selling such a service by the name of “Nespot” with a bundling service for mobile voice by the name of “Nespot Swing.” Portable Internet: Wibro

Another technology wired telecommunications companies hope to use to increase the wireless Internet is a portable Internet service called Wibro (Wireless Broadband). This technology is still not available commercially as it is waiting for the allocation of 2.3GHz among service operators. The limitations of wireless LANs in coverage and mobility motivated some service providers to look for other technologies, and they found such possibilities in Wibro using the IEEE 902.11g standard. This technology is said to provide wireless services to users outdoors, even while driving at speeds of up to 60 km per hour. However, this wider coverage and higher mobility demands a large initial investment, which makes using this technology risky for businesses. Also, it is expected to compete with the next generation mobile service, HSDPA (High Speed Data Packet Access), which is an improvement on WCDMA (Wideband Code Division Multiple Access) when it is combined with voice. This service uses 2.3GHz, the band of spectrum that was originally allocated for Wireless Local Loop, the technology designed to provide broadband services to customers in islands and mountain areas via satellite, but which was discontinued due to its low usage and reallocated for the use of the portable Internet in 2002. FTTH

It is evident that the current ADSL / VDSL technologies, which use existing copper lines, have limits in both speed and capacity. With this restriction the already matured market for high speed Internet access services has forced

Broadband, the Information Society, and National Systems

operators to consider replacing copper lines with optical fibre to create new markets for applications requiring large bandwidth. However, the prohibitive cost makes it impractical to install optical fibre to every household, known as FTTH (Fibre to the Home). Telecommunications operators, including KT, have found a more practical solution, at least for the time being, in FTTC (Fibre to the Curb). According to KT’s FTTC plan, fibre deployment in the access network is expected to grow to 120,000 km by the end of 2007 from 50,000 km at the end of 2003. In the FTTC scheme, operators introduced the concept of a “cell” that is very similar to the existing cell model in the cellular mobile network. Areas around the central office are divided into cells where the cell location and size depends on factors such as current and future customer density. There are many problems, both from the supply and demand sides, in making profitable business out of the FTTH network. The supply side problem is mainly related to costs. For example, experts in the industry contend that the deployment cost of FTTH has to be lower than 150% of the current DSL deployment cost to make any economic sense, but reducing the cost to this level will not be easy considering the current DSL deployment cost. Also, they are sceptical about the existence of services or applications in the near future that will make the FTTH infrastructure profitable within a reasonable period of time. In November 2003, the Korean government announced the BcN (Broadband convergent Network) Plan. It set the goal of transforming currently separated and narrowband networks into integrated broadband networks on which all forms of services can be provided by 2010. To achieve this ambitious goal, the BcN Plan divides the objectives into three stages: in the first stage pioneer networks will be constructed by the end of 2005, on which various research and tests will be carried out; in the second stage, to the end of 2007, networks covering major parts of the country will be constructed and service operators will begin to provide commercial services; and, finally, in the third stage, it sets the goal of completing nationwide networks and providing services to 20 million subscribers at 50 to 100 Mbps. The Korean government selected three consortia, led by KT, SK Telecom, and Dacom, respectively, as the test operators for the broadband convergent network in September 2004. Starting in July 2005, these companies plan to provide trial services, such as Internet services via digital TV over the pioneer broadband networks and to provide trial users with home gateways and set-top boxes free of charge for a certain period of time. These companies will mainly concentrate their investment on upgrading transmission facilities and terminals and enlarging their capacities to avoid spending too much on optical fibre.

105

inho chung

106

However, some telecommunications companies, especially KT, have announced plans to increase the number of optical fibre lines, with priority given to certified apartments of the highest degree, that is, those in which each household is connected to the Internet through more than four optical lines and at least one LAN cable. For such buildings, construction companies do not have to install copper cable for cable TV and, for this reason, they are encouraged to install optical fibre in newly constructed apartments to obtain such certification. This system is expected to greatly facilitate optical fibre installation and to help service providers to transmit their multimedia services, such as home networks and IPTV services, on this upgraded network. Home Networks

The Home Network is one of the promising services that can be provided on a broadband network. It allows users to control electronic devices in their home, such as TVs, refrigerators, and air conditioners, from outside. In 2003, the Korean government announced the “Home Networking Plan” in which 61% of all households would be provided with such services by 2007. Currently some carriers, including KT, are providing trial services in Seoul and plan to extend service areas into all the major cities depending upon the result of these trials and the market situation. VoIP

The construction of the broadband network and the penetration of high speed Internet services enabled consumers to get easy access to VoIP (Voice over Internet Protocol)services. VoIP has succeeded in attracting a large consumer group in international telephone services but has not been so successful in domestic markets. This result may be explained by the fact that international services users are willing to endure the inconvenience and complexity of making a PC-to-PC call to avoid the high prices of existing services but for domestic services they prefer PSTN (Public Switched Telephone Network) services that are of better quality and more convenient than VoIP. The phone-to-phone service is not popular with users, partly because of the lack of numbers assigned to their phones and partly because of the already lower prices of existing services. The Korean government announced that it would assign numbers to VoIP telephones if their quality satisfies an a priori set of criteria. This policy is expected to facilitate the penetration of VoIP, especially among business customers who are more sensitive to prices.

Broadband, the Information Society, and National Systems

conclusion This chapter explains the short history and current situation of the broadband services market in Korea. It seems evident that Korea is in the forefront, not only in the penetration of current broadband services such as ADSL and cable modem but also in developing future and next generation services. This brilliant performance could be attributed to market, geographic, and demographic factors to some extent. It is important, however, to note that the facility-based competition combined with non-interventionist policy and various incentive programmes greatly contributed to the earlier establishment of the market compared with other countries. The broadband infrastructure and diverse services available on it are expected to be a great resource for the approaching new age of integrating services between wired and wireless, between voice and data, and between telecommunications and broadcasting. For this reason many governments are making great efforts to increase their resources more rapidly than other nations and to find the best policy for their circumstances. The Korean experience should be a very helpful case to study for this purpose, although it is still an ongoing case.

notes 1. The consortium led by New Bridge Capital, the well-known international investment bank, obtained the status of dominant shareholder in Hanaro Telecom, defeating the consortium led by LG Group, the owner of Dacom, in October 2003. Also, at the end of 2004, there were three candidates, Hanaro Telecom, Dacom, and a subsidiary of City-Corp Bank, applying for the new ownership of Thrunet. 2. The market size of Change Bell Services was $60 million (PPP $89 million) followed by music streaming services of $16 million (PPP $23 million) in the commercial internet music market in 2002 (source: Korea IT Industry Promotion Agency, http://www.kipa.or.kr). 3. APII is a co-operative body that was established to promote international cooperation by linking Asia Pacific countries via high speed communications networks. 4. KT was a 100% government-owned public firm until the government began to sell its share to the private sector from 1993. KT was completely privatized in 2002.

references Helanski, Howard A. 2002. Competition and regulation in broadband communications. In Broadband: Should We Regulate High-Speed Internet Access? Edited by

107

inho chung

108

Robert W. Crandall and James H. Alleman. Washington, DC: AEI-Brookings Joint Center for Regulatory Studies. Jackson, Charles L. 2002. Wired high-speed access. In Broadband: Should We Regulate High-Speed Internet Access? Edited by Robert W. Crandall and James H. Alleman. Washington, DC: AEI-Brookings Joint Center for Regulatory Studies. Kelly, Tim. 2001. Government strategies to encourage broadband. Paper presented at the ITU Regulatory Implications of Broadband Workshop, in Geneva, Switzerland. http://www.itu.int /osg/spu/ni/ broadband/workshop/presentations/ kelly.pdf. Korea Information Strategy Development Institute. 2004. IT industry outlook of Korea. Available online at http://www.kisdi.re.kr. Lee, Nae-Chan. 2002. Broadband Internet Service: Korea’s Experience. Seoul: KISDI. Ministry of Information and Communication. 2002. Broadband Internet in Korea. Available online at http://www.mic.go.kr. ———. 2003. White Paper: Internet Korea. Available online at http://www.mic.go.kr. ———. 2003. Informatization White Paper. Available online at http://www.mic.go.kr. ———. 2004. IT 839 Strategy. Available online at http://www.mic.go.kr. OECD. 2003. Developments in local loop unbundling. Working Party on Information and Telecommunications Services Policies, DSTI /ICCP/ TISP(2002)5/ FINAL. ———. 2004. The development of broadband access in rural and remote areas. Working Party on Information and Telecommunications Services Policies, DSTI / ICCP/ TISP(2003)7/ FINAL.

4 Broadband Access Development in China Manqiang Liu

china is the largest developing country in the world, and over the last decade both the telecommunications sector and Internet in China have taken off and have been increasing dramatically.1 China became the number one country in terms of the telephone subscriber base in 2003, and the penetration rate for Internet users has reached 6.7%. At the same time broadband development is increasing at an ever accelerating pace. This chapter will discuss the process of broadband development, broadband supply and demand, and its likely progress in the coming years in China. the basic environment of broadband development in china Before our discussion about broadband, it is necessary to keep in mind the following basic facts about China that have a bearing on the development of broadband in that country. • China is the largest developing country in the world, and the population increases by about 10 million every year. A baby boy, China’s 1.3 billionth citizen, was born on 6 January 2005 in a Beijing hospital. In 30 years’ time, China will reach its population peak of 1.6 billion, to be followed by a gradual decrease. This means that for many products and services China is a huge real or potential market. • The Chinese economy began to take off in the early 1980s from a poor base and has been growing rapidly at an annual rate of over 9% for more than 20 years.

manqiang liu

110 •

•

•

•

•

It is predicted that it will continue to grow at the rate of 7– 8% for the next decade and longer. As the economy grows, the income level of both the urban and rural population is increasing gradually. In 2003, the per capita disposable income of urban households was 8,472 RMB yuan ($1,0232 or PPP $4,6183), compared to 1,510 RMB yuan ($316) in 1990. At the same time, households’ expenditure on telecommunications services has also steadily increased. In 2000, the per capita annual telecom expenditure was about $28 (PPP $131) or 3% of per capita annual total expenditure and it is predicted that the rate will have risen to 5– 6% in 2020. This provides the demand basis for telecommunications and broadband development.4 China is a vast country with imbalances in the economic and social development of different regions. Generally, the eastern part of China has developed quite fast, while central and western China is comparatively poor and has developed more slowly. This obviously has implications for both the telecom sector and Internet development. It is usually the case that telecom services and Internet connections are launched first in eastern China and big and central cities, and then gradually spread to out-of-the-way regions and smaller cities a few years later. China’s telecommunications sector began to take off in the early 1990s from a low base and, since then, its annual growth rate has been over 20% or about three times GDP (Gross Domestic Product) for the last decade. By June 2004, there were 295.5 million fixed-line telephone subscribers and the penetration rate was 22.9%. The number of mobile subscribers had reached 305.3 million and the penetration rate was 23.6%. Both fixed-line and mobile subscribers are increasing by over 5 million each month, and China has become the number one country in the world in terms of its telephone subscriber base (see Table 4.1). It is predicted that by 2020 the number of fixed-line telephone subscribers will be 600 million, while mobile telephone subscribers will number 730 million.5 Since the initial launch of cable television networks in China about 40 years ago, the number of cable television subscribers had risen to 120 million by 2003, about one-third of the total around the world.6 In 1994 China connected to the Internet. During the last decade, in common with the rest of the world, the Internet has been developing dramatically in China. By June 2004, there were 87 million Internet users; computer hosts had 36.3 million; and the total bandwidth of leased international connections was 53,941 M. These figures were, respectively, 140, 121, and 2,894 times higher than those of October 1997 (see Table 4.1). The penetration rate of Internet users had risen to 6.7% by the end of June 2004.

The fast growth of China’s economy, telecommunications, and Internet laid a solid foundation for the development of broadband.

Broadband Access Development in China

111

table 4.1 Growth of Telephone Subscribers and the Internet in China 1990 1995

1997

1998

1999

2000

2001

2002

2003

2004

Telephone subscribers Fixed telephone 6.9 40.7 70.3 87.4 108.7 144.8 180.4 214.4 263.3 295.5 (million) Mobile telephone 0.02 3.6 13.2 25.0 43.3 85.3 144.8 206.6 268.7 305.3 (million) Fixed and mobile 6.9 44.3 83.5 112.4 152.0 230.1 325.2 421.0 532.0 600.8 (million) The Internet Computer hosts 299 747 3,500 8,920 12,540 20,830 30,890 36,300 (thousand) Internet users 620 2,100 8,900 22,500 33,700 59,100 79,500 87,000 (thousand) Domain names 4,066 18,396 48,695 122,099 127,319 179,544 340,040 382,216 registered under CN Number of “WWW” 1,500 5,300 15,153 265,405 277,100 371,600 595,550 626,600 websites Total bandwidth 19 143 351 2,799 7,598 9,380 27,216 53,941 of leased international connections (M) sources: Ministry of Information Industry, China; “Statistical Survey Report on the Internet Development in China, 1997–2002,” China Internet Network Information Center.

the broadband market structure in china The Evolution of China’s Telecommunications Market Under China’s economic planning system there was only one telecom carrier, China Telecom, until reform of the telecommunications market began in 1994. During the last decade, this sector experienced several major changes: China Unicom was set up in 1994; China Telecom was split into four companies in 1998; China Netcom and China Tietong were founded in 2000; and, in 2001, China Telecom was split in two, one part becoming the present China Telecom, while the second part and China Netcom were merged to become China Network Communications Group Corporation (CNC). Following these changes there are, at present, six carriers providing basic telecommunications services; China Telecom, CNC, China Mobile, China Unicom, China Tietong, and China Satcom. China Telecom and CNC provide fixed-line telecom services; China Mobile provides GSM mobile services; China Unicom has both fixed-line and mobile (GSM and CDMA) licences, but its fixed-line service is very small;

manqiang liu

112

China Tietong provides fixed-line telecom services mainly in areas near railway lines; and China Satcom runs satellite telecom services. Provisions for Domestic Telecommunications Networks In the “Telecommunication Services Classification Catalogue” (amended in February 2003) issued by the Ministry of Information Industry (MII), the section on “Domestic Communication Infrastructure Services” covers the building and operating activities of the backbone network and the metropolitan area network. Currently five companies are permitted to run these services; China Telecom, CNC, China Unicom, China Mobile, and China Tietong. According to the Classification Catalogue, broadband access services belong to the network access service under the terms of basic telecommunications services and only companies holding the relevant licences are entitled to operate them. In 2000, broadband access construction started in China and an upsurge soon appeared around the country. Besides the basic telecommunications carriers, many new entrants, including cable companies, new broadband access companies, ISPs, and real estate agents, entered this field. However, the MII was of the opinion that this had resulted in some chaotic situations in the broadband access market. In an attempt to resolve the problems, the MII published the “Framework Opinions about Opening Broadband Customer Premise Network (CPN) Market” in June 2001. Under the Framework Opinions, the MII opened up the broadband access market in 13 cities, including Beijing, Shanghai, and Guangzhou, and provided that broadband access services were to be operated as telecom added value services. In all other cities, companies without basic telecommunications service licences were prohibited from operating these services and, as a result, many new entrants were forced to exit the market. In addition, China Broadcast Network Co. Ltd. (CBN), led by the State Administration of Radio, Film, and Television (SARFT), owns a fibre backbone network transmitting broadcast and television signals around the country. Local cable television companies also own metropolitan area networks and HFC access networks in their cities. The Major Players in China’s Broadband Market At present the major players in China’s broadband market include China Telecom, CNC, local cable television companies, and some new entrants such as, for example, Great Wall Broadband Network Service Co. Ltd. (GWB).

Broadband Access Development in China China Telecom and CNC

China Telecom was the leading fixed-line telephone carrier until 2001. In that year, as mentioned above, China Telecom was split into two parts. One part became the present China Telecom Corporation Limited (China Telecom). The second part and China Netcom were merged to form the present CNC. China Telecom and CNC hold 70% and 30%, respectively, of the national trunk-line transmission network assets owned by the former China Telecom and all local networks in their own regions. CNC runs fixed-line telecom services in 10 provinces, autonomous regions, and municipalities in North China. These are Beijing Municipality, Tianjin Municipality, Hebei Province, Shanxi Province, Inner Mongolia Autonomous Region, Liaoning Province, Jilin Province, Heilongjiang Province, Henan, and Shandong Province. China Telecom is the leading provider of wire-line telecommunications services in the 21 provinces, autonomous regions, and municipalities in South China. Its service regions consist of Shanghai Municipality, Guangdong Province, Jiangsu Province, Zhejiang Province, Anhui Province, Fujian Province, Jiangxi Province, Guangxi Zhuang Autonomous Region, Chongqing Municipality, Sichuan Province, Hubei Province, Hunan Province, Yunnan Province, Hainan Province, Guizhou Province, Qinghai Province, Gansu Province, Shaanxi Province, Xinjiang Uygur Autonomous Region, Ningxia Hui Autonomous Region, and Xizhang Autonomous Region. The two companies are allowed to run fixed-line services in their counterpart’s area and they have set up some branch companies but, so far, few inroads have been made on either side. CNC and China Telecom mainly provide ISDN, ADSL, and Ethernet access to their customers. As far as the other two fixed-line network carriers, China Unicom and China Tietong, are concerned, their market shares are very small, respectively, 0.1% and 0.15% in 2002 and, on the whole, they are not strong competitors. Cable Television Companies

The story of China’s cable television system is very different from that of the telecommunications sector. Cable television networks are widespread and grew gradually from their initial development about 40 years ago. There is no single unified cable television network around the country. The cable television stations and companies are owned and managed, respectively, by province, city, and county governments, and operate under the relevant regula-

113

manqiang liu

114

tions and guidance of state and local broadcast and television administrations. However, these cable companies are monopolistic in their local cable markets. They have the largest cable television networks around the world. As the only pilot city to converge three networks, Shanghai Cable Network Corp. began to reconstruct its HFC (Hybrid Fibre-Coaxial Cable) networks to two-way transmitting networks in 1999 and launched a broadband access service in 2000. Thereafter, cable companies in some large and medium-sized cities began to provide a similar service. Great Wall Broadband Network Service Co. Ltd. (GWB)

In 2000, several new entrants appeared in China’s broadband access market including Great Wall Broadband Network Service Co., Juyou Network, and Fangzheng Broadband. GWB is the largest and can be considered as representative of all these companies. GWB was established by Great Wall Group and China International Trust and Investment Corporation (CITIC) in 2000. Its headquarters is in Beijing and it has branches in more than 30 cities. These new entrants mainly provide Ethernet access services. broadband access development in china Broadband access development in China can be divided into three phases: the start-up phase (2000 –2001); the adjustment phase (2002); and the rapid growth phase (2003–present). The Start-up Phase: 2000 –2001 As mentioned above, in the late 1990s and early 2000s, Internet use developed rapidly and most customers used dial-up connections during that time. Inspired by the dramatic growth of Internet use and with excessive expectations of broadband development, a variety of companies entered the broadband access market. These companies included telecommunications carriers, local cable TV companies, new broadband companies (for example, GWB), and even city infrastructure providers and real estate agents. This was regarded as a new field in which all the competitors would start from scratch. The main focus of competition was on establishing as wide a user-base as possible in local communities and some called this a “broadband enclosure movement.” The telecommunications carriers launched ISDN and, later, ADSL services for telephone subscribers and Ethernet access to local communities. Meanwhile, the new broadband companies mainly used Ethernet technology,

Broadband Access Development in China

which was considered more advanced at that time. In order to begin the construction of Ethernet networks, these broadband companies and telecommunications carriers signed agreements with real estate agents and local communities. Cable TV companies had about 100 million cable subscribers around the country at that time but they had to rebuild their HFC networks to provide Internet connections. Meanwhile, they attempted to develop broadband content, for instance, VOD (video-on-demand) music programmes, video news, and so on. These companies, especially the new broadband companies, invested a large amount of money in broadband network construction. But these huge investments did not lead to an enthusiastic response from prospective users and the expected returns did not materialize as the number of subscribers grew very slowly. The following reasons are the most likely cause of this outcome. The first is high tariffs. The tariffs generally consisted of three parts; installation charge, modem cost, and monthly/hourly fee. Income level was one of the main factors that broadband players considered when determining tariffs. The per capita disposable income of urban households was about $828.5 (PPP $3,965) in all cities in 2001 with imbalances in different size cities. In especially large cities it was $1,105 (PPP $5,288), while in country towns it was only $680 (PPP $3,254).7 Generally speaking, the tariff in larger cities and relatively rich regions was higher than in smaller and relatively poor ones. For example, the installation charge was about $181 (PPP $867) in Beijing, while it was only $30.2 (PPP $144.5) in Chongqing. Monthly fees ranged between $12.1–$18.1 (PPP $57.8 –$86.7) and modem costs were between about $121 (PPP $578) to over $242 (PPP $1,156). When compared to the lower income level, on the whole broadband tariffs were very high in 2001 while the dial-up fee was relatively cheap. Second, there was not enough broadband content at that time to encourage users to adopt broadband. For most users, who mainly read the news, sent e-mails, and dealt with a few pictures, narrowband speed was acceptable. As a result, many new entrants faced financial difficulties. Meanwhile, the MII believed that too many new entrants had resulted in chaos and began to restrict the broadband access market in the summer of 2001. As mentioned above, the MII prescribed that new entrants could apply for broadband access licences only in 13 large and medium-sized cities. In all other cities, those companies without basic telecommunications service licences were forced to stop broadband network construction and services. These companies suffered heavy losses, but it provided the telecommunications carriers with a good opportunity to develop ADSL on a large scale later on.

115

manqiang liu

116

The Adjustment Phase: 2002 The year 2002 was something of an interlude. After the bustle of 2001, broadband players moved into a quieter period. They began to adjust their business strategies and pay more attention to broadband applications, services, and cooperation with content providers. The network construction pace of the cable companies, GWB, and other new entrants slowed down. Meanwhile, China Telecom and CNC began to improve their ADSL technology and networks, decrease prices, and steadily increase the number of ADSL subscribers. By the end of 2002, CNC had 800,000 ADSL subscribers while China Telecom had 2.5 million. The Rapid Growth Phase: 2003–Present Despite the bursting of the network economy bubble around the world, the Internet’s progress did not falter in China. The number of Internet users continued to grow rapidly while on-line applications and content increased gradually. Since the start of 2003, broadband access, especially ADSL, has also steadily grown. In China in 2003 the overriding event was the initial outbreak and spread of SARS. During the SARS epidemic people in the affected regions interrupted their regular routines and stayed at home most of the time to avoid infection. Consequently, they mainly got information and worked by electronic means, that is, via telephone, television, and the Internet. School and college students had their lessons transmitted over cable TV and the Internet, and people carried on business over the telephone and Internet. An increasing amount of broadband content appeared, such as video programmes, pictures, and voice, and users no longer found narrowband speed acceptable. As a result, the demand for Internet and broadband access greatly increased. Telecommunications carriers finally had the opportunity to develop fast broadband access and began a large-scale expansion of ADSL subscribers. During the summer of 2003, for example, Beijing Communication (the Beijing branch of CNC) started a series of promotional activities. The installation charge and the modem cost were cancelled and the hourly fee decreased by 50%. The monthly fee (no time limit) was reduced to $14.5 (PPP $65.8). Consumers had a positive response and the number of ADSL subscribers increased rapidly. By the end of 2003, CNC had 3 million ADSL subscribers. And the momentum is continuing even now. The growth in subscriber numbers was so fast that at times the company had to temporarily halt some of their promotions in order to install and adjust their equipment and networks. CNC anticipated increasing its subscriber base to 8 million by the end of 2004.

Broadband Access Development in China

117

20,000

Thousand

15,000 10,000 5,000 0

2002

2003 China Netcom

2004 plan China Telecom

figure 4.1 ADSL subscribers in China

The same story also applied to China Telecom in South China. By the end of 2003, China Telecom’s ADSL subscribers had increased to 7.5 million. In May 2004, China Telecom announced that they had 10 million ADSL subscribers and they aimed to add another 5 million by the end of 2004, giving a total of 15 million. Figure 4.1 shows ADSL users’ development in China. China Tietong, another fixed-line telecommunications carrier, also became a broadband competitor. By September 2004, it had over 10 million fixed-line telephone subscribers and 780,000 ADSL subscribers.8 However, China Tietong is a small competitor compared to China Telecom and CNC. Facing the threat from ADSL, the reaction of cable TV companies was almost non-existent. For a long time they took hardly any action, losing a lot of subscribers to the telecommunications companies as a result. In fact, the cable TV companies seemed to be indifferent to their weak position in the broadband access market. They are monopolistic players in their field and cable TV transmission is their main service and main source of revenue. In fact, in spring 2003, Beijing Gehua Cable received approval from the Beijing government to increase the monthly fee for cable television by 50%, which was a valuable windfall for them. Because the rebuilding costs of HFC were higher than for ADSL they were at a disadvantage in regard to cost competitiveness. Furthermore, they have no strong incentive to become more enterprising. For a long time cable television had been a monopolistic and non-profit sector, and these companies did not understand and did not want to enter new markets. However, as noted above, in summer 2003, Beijing Communication had to temporarily stop their promotional activities in order to construct networks and

manqiang liu

118

add new equipment. Taking advantage of this, Beijing Gehua Cable finally reduced its broadband monthly fee from $18.1 to $14.5 (PPP $82.2 to $65.8), with a yearly fee of just $120.8 (PPP $548). It also set up a separate company to expand its broadband service. However, as ADSL had already occupied most of the market, cable companies were only able to make limited progress. At present the new broadband companies are still in a difficult situation. Great Wall Broadband invested about $301.9 million in network construction and only has about 200,000 subscribers. It is now in a serious financial plight and is searching for new ways to go while attempting to raise more funds. The state of other new broadband companies is similar to GWB’s; indeed, some may be much worse. As a whole, broadband access competition in China is insufficient. In the ADSL market, China Telecom and CNC are the leading players in their own regions. China Tietong is beginning to provide an ADSL service but its share is very small. In the cable access market, local cable companies are monopolistic players. The Ethernet access market is competitive but the new entrants are very weak and are limited to the 13 open cities. This means that, currently, competition is mainly between different access methods from which, if available, users can make their own choices. the development of broadband content in china As early as 2002, people began to appreciate the crucial role in the industry of broadband content. It was realised that the broadband industry is a chain which, in order to succeed, needs co-operation between all its individual parts. Eventually, after several years of gestation, China’s broadband industry made significant progress in 2004 when many players, including telecommunications carriers, radio and television stations, programme producers, Internet games developers, and record companies, began to come together in order to build broadband content platforms. Telecommunications Carriers’ Broadband Alliances and Broadband Portals Fixed-line telecommunications carriers are well aware that with the fixed-line telephone market reaching saturation point, broadband services will provide them with new opportunities. So while continuing to develop broadband access they are, at the same time, attempting to increase their broadband revenue through broadband portal websites. However, in order to achieve this, they must find suitable partners.

Broadband Access Development in China

In February 2004, CNC launched its broadband portal website, Tiantian Zaixian (“Every day online”), and subsequently was granted the licence for audio/video programme transmission from the State Administration of Broadcast, Film, and Television (SARFT). Currently the audio and video content on the website includes news, TV shows, sports, games, and so on. On 17 May 2004, CNC announced the establishment of its broadband industry alliance, which comprises 17 firms and includes telecommunications and IT companies as well as content providers.9 Meanwhile, China Telecom had launched its own broadband portal, Chinavnet. On 19 October 2004, China Telecom set up the Xingkong Alliance with about 30 companies and more than 260 content providers have joined.10 IP Television The advantage that the broadcast and television media have is rich content resources and licences for playing audio/video programmes. Supported by telecommunications carriers, they are beginning to set up IP television channels and to provide more audio/video programmes on their websites. In June 2004, supported by China Telecom and CNC, CCTV (China Central Television Station) launched its IPTV website and there are now three branch stations, Beijing, Shanghai, and Jiangsu. The monthly fee is $3.6 (PPP $16.4) in Beijing and subscribers can watch most CCTV programmes on line. CCTV plans to open branch stations in other provinces and cities in the future. On 24 December 2004, Beijing People’s Radio Station, with the support of CNC, launched Beijing IPTV, providing audio/video programmes. This is the first IPTV operated by the broadcast media in China and covers Beijing, Tianjin, Hebei, and Heilongjiang.11 A few days later, on 28 December 2004, Shanghai Oriental IPTV started and now more of the traditional media are preparing to enter the IPTV sector. Internet Games One of the most popular broadband applications is Internet games, especially for young people. Almost 40% of Internet users play online games. According to an iResearch report, in the Chinese market in 2004, 52% of online games were from Korea, 14% from Hong Kong and Taiwan, and 27% from China’s mainland. The government has paid great attention to Internet games development. While enforcing the management of the market, the General Administration

119

manqiang liu

120

of Press and Publication and the MII have also produced policies to support Internet games development. Currently there are about 300 Internet games companies in China; of these, 150 are games developers, producing 200 kinds of Chinese games.12 Instant Messaging According to a recent survey, about 93% of Internet users use or have used instant messaging. Of these, 82.6% use Tencent QQ, a Chinese IM software, and 4.4% use MSN. In addition, 70% use the IM voice/video Function.13 Undoubtedly, many more users will spend much more time on instant messaging and this, in turn, will promote broadband access development. In spite of all the progress made so far, browsing websites, e-mail, and chatting are still the main applications for most users. On the whole, broadband content development is just at the starting stage in China as broadband industrial chains are still in the process of forming and gradually being integrated. the demand for broadband in china The Growth of Broadband Users According to a survey of users by the China Internet Network Information Center (CNNIC), the number of Internet users had reached 87 million with an Internet penetration rate of approximately 6.7%. Of these 87 million users, 31.1 million used broadband access.14 Figure 4.2 shows the growth of users by type of access. In fact, before 2001, there were only two types of access, leased line and dial-up, and most subscribers used the latter. In 2002, ISDN and ADSL were introduced and grew rapidly. By June 2004, although dial-up remained the main access method, the number of leased line and broadband users had exceeded the number of dial-up users. At the same time, there were 36.3 million computer hosts. Of these, 8.8 million computers had broadband access (ADSL and cable modem) while the number of leased line computers was 6.5 million. While the proportion of dial-up access has started to decrease, computer hosts using broadband access have a higher growth rate of over 60%, and the broadband user growth rate has reached almost 80% over the past two years. Clearly, broadband will become the main access method in the next few years, particularly in large and medium-sized cities. Figure 4.3 shows that ADSL is the most popular access method, with 70.8% of users, while those using either Ethernet or cable modem access amount to 24.5% and 7.8%, respectively.

Broadband Access Development in China

121

60,000

Thousand

50,000 40,000 30,000 20,000 10,000 0

97.1

98.0

99.0

Leased lines

00.0

01.0

Dial-up

ISDN

02.0

03.0

04.6

Broadband

figure 4.2 Internet users by type of access

ADSL Ethernet Cable modem Other 0

20

40

60

80

Percent

figure 4.3 Proportion of users by type of access

Internet access by enterprises shows the same trend. In 2003 the Center for Informatization Study (CIS), Chinese Academy of Social Sciences, carried out a survey which found an Internet penetration rate of 96.2% for Chinese enterprises. Compared to a similar survey in 2001 by the SETC (State Economy and Trade Commission), this was an increase of 7.2%. According to the CIS study the type of access used by 838 enterprises was as follows: 47.6% used ADSL; 20.1% used leased lines; 10.4% used dial-up; 7.9% used cable modems; and 3.9% used ISDN. Again compared with the SETC survey, ADSL access had increased by 45.2% while dial-up access had decreased by 35.2%, meaning that ADSL had overtaken dial-up to become the main access method. According to several surveys, most Internet users feel narrowband access is too slow while many broadband users are also dissatisfied with its access speed. In addition they want better access services and more broadband content.

manqiang liu

122

A survey by CNNIC of the factors households considered when choosing broadband providers showed that due to the dearth of supply and competition in many cities and communities there were no alternatives for 22.5% of households. About 11% of users wanted better technology, faster access, and quality services and, although broadband tariffs have decreased considerably, cost is still a concern for many users. Of those who do not use broadband now, over 50% will, or are likely to, install broadband within the next year. Only 11% of non-broadband users are not interested in moving to broadband. Considering the huge number of potential users and its present rapid growth, it is very likely that there will be a very significant increase in broadband demand over the next few years. Broadband Applications The majority of Internet users in China are young people. According to the 2003 CNNIC report 66.9% of broadband users were under 30 years old; 30.4% were aged between 31–50; and only 2.7% were over 50. The monthly income of 77.7% of users was less than $241.5 (PPP $1096). Another survey by iResearch Inc. in 2003 looked at the main online activities of both broadband and narrowband users. This survey revealed, firstly, that many online activities took place on narrowband; for example, e-mail, browsing the news, online chatting, and sending short messages. However, more bandwidth was necessary to better download software, listen to music, watch movies, and get online education programmes. Secondly, the main online activities and priorities of both broadband and narrowband users were broadly similar, which would perhaps indicate that there was insufficient content for broadband users. Finally, the Internet usage rate of broadband users was clearly higher than that for narrowband users. As far as business users are concerned, e-commerce has developed steadily over the past few years. According to the 2003 CIS survey, 31% of businesses had already constructed e-commerce websites; 9.4% were still in the process of construction; and 11.9% were planning to do so. Currently, approximately 90% of large enterprises do business over the Internet. conclusions and future prospects Accelerated Growth and Competition in Fixed-line Broadband China’s telecommunications sector will continue to grow rapidly. By November 2004, fixed-line telephone subscribers had increased by more than 50 mil-

Broadband Access Development in China

lion over the previous year, reaching a total of 313 million. Mobile subscribers also increased by nearly 60 million to 330 million. China’s Internet will continue to grow rapidly while the broadband access market has entered a period of accelerated expansion with a growth rate of over 60%, and this trend will last for a further 2–3 years. Meanwhile, although insufficient, broadband competition, especially between access methods, will continue and become keener. ADSL has become one of the most important services for both CNC and China Telecom. CNC aims to raise its broadband revenue from the current 3– 4% to 20%. In order to achieve this target, they are improving the access quality and bandwidth of ADSL as well as developing more broadband content. Also they are now starting to introduce an ADSL service of 1M/2M bandwidth in some cities, with experiments underway to increase the bandwidth to 2– 8M in the coming months. At present about 50% of Beijing’s households are connected to the Internet and, of these, 63.8% have broadband access. Beijing Communication predicts that Beijing’s entire population will use broadband before 2008. Meanwhile, despite its smaller size, China Tietong also competes in the ADSL market in certain areas of some cities. Although some local cable companies have begun promotional activities and offer slightly lower prices than ADSL, the growth of cable broadband subscribers is still very slow. For example, Shanghai is one of the leading cities in the development of cable access but has only around 200,000 users, about onefifth of the ASDL total. This may be because they have placed greater priority on the development of digital television. The reasons for this are that SARFT plans to provide digital television programmes during the Beijing Olympics in 2008 and to stop broadcasting the analogue TV signal by 2015. Another new entrant is FibrLINK, a company owned by State Power Telecommunication Companies. It has researched Power Line Communication technology for several years and is carrying out experiments in several cities. By August 2004, there were about 13,000 PLC subscribers in 297 communities. The monthly fee is $6 –$9.6 (PPP $27.4 –$43.8), which is a little less than for ASDL or cable access. However, PLC entered this market too late and its future is unclear.15 Despite being in a difficult period, GWB is trying to move ahead. They are attempting to raise new funds and hope to get licences in more cities besides the 13 with an open market. Nonetheless, all these players are aware that they have a potentially strong rival in wireless broadband.

123

manqiang liu

124

Wireless Broadband Access In 2003, China Mobile and China Unicom launched their wireless access services, GPRS and CDMA1X, respectively. The average access speed of GPRS is 30 – 40KB per second and, after several price cuts, the local monthly fee is $24.2 (PPP $109.6) without time limit. The average access speed of CDMA1X is 70 – 80KB per second with a local monthly fee of $36 (PPP $163.3) without time limit. With higher speed, China Unicom’s access service is more competitive than China Mobile’s and it now has about 260,000 subscribers. In a strict sense, these two wireless access services are not broadband. However, 3G (third-generation) wireless broadband access is currently in an experimental and testing phase in China. So far, the MII has not set an explicit timetable for determining 3G standards and issuing 3G licences, although Wang Xudong, Minister of the MII, recently announced that his department would put forward proposals for the country’s 3G strategy “at the proper time” in 2005. However, China Telecom and CNC are both anticipating getting wireless communication licences eventually and began some preliminary work when, in 2002, they started installing WLAN in some airports, large hotels, and conference centres.16 The Progress of FTTH Although ADSL technology has improved gradually and its speed is increasing, FTTH is still considered to be more advanced. Because it was more expensive, FTTH development in China was very slow for a long time but now the cost is decreasing and the situation is changing. The price of the optic transceiver has come down from more than $120.8 (PPP $548) to approximately $24.2 (PPP $109.6) and optic fibre is much cheaper than copper wire. Consequently, this has created the conditions for further FTTH development. Some experiments are being undertaken in Wuhan and Chengdu. Experts predict that by 2006, or even earlier, an FTTH service will be available in Beijing and other large cities, and that eventually FTTH will spread to mediumsized and small cities. Beijing Olympics 2008 Beijing will host the Olympic Games in 2008 and the Beijing Olympic Organization Commission has announced that the theme will be “Green Olympics,

Broadband Access Development in China

People’s Olympics, and Hi-Tech Olympics.” Advanced communication networks are one of the principal components of the “Hi-Tech Olympics.” Most of the telecommunications carriers will be involved in the creation of the “Digital Olympics,” the aim of which is “. . . securely, conveniently, timely and efficiently, to provide abundant, multilingual, individualized and affordable communication services to anyone, at any related locations and at any time.” In order to achieve this, Beijing will reconstruct and update its networks and equipment, providing broadband networks, broadband access, and broadband services. Six large communication projects will be undertaken and total investment will be over $12.1 billion. Similar projects will be planned and implemented in the other cities and provinces where Olympic events will be held. The Beijing Olympics will provide a strong impetus to speed up broadband development not only in those areas directly involved but also throughout the whole country.

notes 1. This chapter discusses broadband development only in China’s mainland and does not include Taiwan Province, Hong Kong Special Administrative Region, and Macao Special Administrative Region. 2. The nominal exchange rate of RMB to U.S.$: 8.28 since the mid-1990s; 4.78 in 1990 (source: www.stats.gov.cn). 3. The conversion factor of PPP dollar to RMB yuan: 1.775 in 2000; 1.729 in 2001; 1.780 in 2002; 1.823 in 2003; calculated on World Development Report, 2000, 2001, 2002, 2003, World Bank. 4. Li, Xuenong. 2004. Analysis on telecom consumption level in all-round well-off society. People’s Post & Telecom News, December 2. 5. Li (2004). 6. Xu, Zhuang. 2004. Study on countermeasure of cable TV industrialization in China. Available online at http://media.people.com.cn/GB/40724/40726/3072679 .html. 7. China Statistical Yearbook. 2002. National Bureau of Statistics, China. 8. China Tietong finished two business aims ahead of time. People’s Post & Telecom News, October 21, 2004. 9. CNC set up broadband alliance. Beijing Times, May 18, 2004. 10. China Telecom Xingkong alliance builds Internet industrial chain. People’s Post & Telecom News, October 20, 2004. 11. First IPTV started. Sina Tech, December 20, 2004. 12. Look back on China’s Internet games market. China Youth News, January 7, 2005.

125

manqiang liu

126

13. 70% of users pay attention to IM audio/video function. Communication Information News, February 11, 2004. 14. CNNIC: (1) percentages do not add to 100% as Internet users who adopt multiple accessing methods are recounted; (2) leased line users refer to beneficiaries of LAN, which connects to the Ethernet; (3) broadband users refer to beneficiaries of xDSL, cable modem, etc. 15. Users of PLC over 10,000. IT Manager World, November 9, 2004. 16. Wireless broadband enclosure started stealthily. International Finance News, May 23, 2003.

part two. north america: the united states

in his chapter on the U.S. contained in Part II, Johannes Bauer analyses the essential details of the broadband story in the United States. The canvas that Bauer paints is dominated by two groups of players: a small group of cable companies (including Comcast, Time Warner, and Cox) and several incumbent local telecoms operators (such as SBC, Verizon, and Bell South). Together the cable companies provide 58% of broadband connections in the U.S. while the telecoms operators account for 37%, although the latter are now growing at a more rapid rate than the former. Bauer makes clear that the principal dynamic of the U.S. broadband market is driven by competition between these two groups. As shown in Figure 1.1 in Chapter 1, the dominance of cable companies in the provision of broadband distinguishes the U.S. from almost all the other countries. However, Bauer’s painting also shows that these players are embedded in a set of institutions that help shape their visions, strategies, activities, and interactions and by so doing influence the outcomes, including comparative international performance, of the U.S. broadband industry. Amongst these institutions Bauer pays particular attention to those that affect the regulatory framework within which the players operate. The institutions include not only the Federal Communications Commission (FCC), the U.S. telecoms regulator, but also the broader political processes that influence the composition of the FCC, its politically appointed commissioners, and accordingly its decisions. In addition, account is taken of all the legal institutions in the United States—at federal, state, and local levels—that circumscribe the decisions and activities of the FCC and create strategic alternatives for players seeking to further their own interests. Unfortunately, the “political economy of regulation” has received far too little attention from scholars, not only in the U.S. but in all countries, partly as a result of the dominant approach in economics that pays significantly more attention to what regulators should do (based on concepts of optimal outcomes) than to what they actually do and why they do it.

martin fransman

128

As Bauer notes, there are significant differences between the networks of these two groups as a result of their different institutional embeddedness. For more than 100 years, the U.S. telephone network has been operated as a common carrier network, an open communications platform. He suggests that this openness was one of the reasons that the dial-up (narrowband) Internet took off so soon and so rapidly in the U.S. By contrast, however, cable television systems had historically been operated as private contract carriers, relatively closed networks fully controlled by their cable company owners. With the advent of broadband the issue arose as to whether broadband access provided by cable modems should be regarded as a common carriage activity, whether DSL provided by telecoms operators should be freed from common carriage obligations, or whether these two regulatory models should continue to co-exist. In March 2002, the FCC declared broadband access provided via cable modem as an interstate information service under FCC (and not local or state) jurisdiction and as such not subject to regulation. However, as noted by Bauer and in Chapter 1, a different approach was taken with the local telecoms operators (known as incumbent local exchange carriers, or ILECs). They were required to unbundle1 elements of their local access networks. Moreover, they had to do so at forward looking incremental cost prices (using the Total Element Long-Run Incremental Cost method, TELRIC). In effect, since the regulator concluded that much of the cost of the copper local loop had already been recovered through earlier charges, this meant that new entrant competitors (the competitive local exchange carriers, CLECs) wanting to compete by accessing the networks of the ILECs could do so at very low or even zero cost. This lowered barriers to entry and potentially increased the intensity of competition facing the ILECs. But, while an increase in competitive pressure was a desirable outcome, the other side of the unbundling coin contained a negative effect: it reduced the incentive of the ILECs to invest in new networks. As Bauer points out, this distortion of investment incentives was in part blamed for the slower rollout of DSL and optical fiber compared to cable modem service. Constrained by the effects of regulatory and legal institutions, the ILECs proceeded to try and change the offending institutional requirements. They challenged the line sharing and other unbundling provisions before regulatory agencies and in the courts. By 2003 they succeeded in their attempts, as detailed by Bauer and in Chapter 1 where the essential details of the complex politico-legal battle that took place are recounted. As Bauer concludes, “Overall . . . unbundling obligations in the [U.S.] broadband markets have been essentially eliminated” (p. 28).

North America: The United States

This significantly changed the rules of the ballgame. While the legal framework circumscribing the activities of competing cable companies and telecoms operators was made more compatible, the balance of power and hence reward between ILECs and CLECs was significantly altered. As Bauer observes, the short-term effect of the new rules is likely to increase the cost of inputs for the CLECs and broadband Internet service providers, thus reducing their incentive to invest. However, two different effects on the ILECs need to be distinguished. First is the increase in their incentive to invest resulting from their greater appropriation of returns from their networks (since the competing CLECs have reduced and more costly access to their networks). This is a positive effect from the point of view of the U.S. broadband system as a whole. But the second effect is negative. As Bauer notes, with reduced pressure coming from competitors using their copper local loops, the ILECs may also have less incentive to move to the superior technology, fibre to the home (FTTH). But, as Bauer also points out, there is another dimension to the problem that needs to be taken into account. By making the regulatory framework affecting telecoms operators more like that influencing the cable companies, the authorities have increased the “closed-ness” of the broadband system. Some have argued that this may negatively affect complementary innovation in content, applications, and services. In the area of FTTH, as shown in Chapter 1, the ILECs were able to persuade the FCC to forbear, that is refrain from regulating. This absence of regulation in the field of optical fibre access networks in the U.S. constitutes one of the more glaring differences between this country and Japan. As shown in the Part I introduction on Asia, in Japan the ministry-regulator has insisted that the optical fibre access networks of the incumbent, NTT, must be open to competitors at low interconnection prices. However, in response NTT has argued not only that this reduces its incentive to invest in such optical fibre networks but also its ability to finance these investments to which the company has already committed. Since optical fibre is the superior technology and is partly responsible for Japan’s superior performance in broadband speed, it is important to ask what effect this regulatory framework for fibre is having in the U.S. So far, according to Bauer, the results are not encouraging: “Prior to the promulgation of the new rules, the [ILECs] had promised to develop fiber-to-the-home (FTTH) if relieved from unbundling and line sharing rules. Once these rules were in place, Bell South and SBC backed off their initial promises and are now proceeding with plans to develop fiber only to the neighborhood or

129

martin fransman

130

to the curb [FTTC]” (pp. 16 –17). However, Bauer does caution that it is still too early to make longer term predictions. As this account of U.S. broadband makes clear, the main competitive game is being played by a small group of incumbent telecoms operators and cable companies. In other words, it is inter-modal competition between DSL and cable that is the dominant form of competition in this country. Intramodal competition between DSL-based incumbents and their competitors who are able to access their local networks at low prices—so important in some of the Asian and European countries included in this book—is far less significant in the U.S. This has important consequences for the intensity of competition in the broadband market in the U.S. Competition between telecoms operators and cable companies is real and will bring important benefits over time, including the rapid deployment of so-called triple play services (voice, Internet access, and TV/video). But crucially, this is oligopolistic competition where neither side has an interest in rocking the boat through disruptive activities. A foretaste of such oligopolistic competition, as Bauer notes, became apparent when telecoms companies initially began to deploy DSL services more aggressively. Bauer observes that the “cable service providers responded with increased download speeds but not lower prices” (emphasis added, p. 7). The contrast with the intense process of competition that resulted in both significant increases in speed and rapidly falling prices in Japan and Korea, as shown in Chapter 1 and the Japanese and Korean chapters, is striking indeed. Bauer accordingly concludes that “perhaps as a consequence of inter-modal rivalry, price competition was less important than service competition, leading to a lower intensity of competition in the broadband access markets [in the U.S.] than in other countries” (emphasis added, p. 35). If oligopolistic inter-modal competition between telecoms operators and cable companies is of its essence muted, in the U.S. system compensating increases in the intensity of competition have not been produced by intramodal competition. It is this conclusion that Bauer comes to when he observes that “the recent elimination of mandatory line sharing will increase wholesale prices for access to high-speed loops and reduce competition from new entrants” (emphasis added, p. 37). Perhaps the best indicator of this lower intensity of competition is the market shares of incumbents and new entrants, respectively, in the U.S. on the one hand and Japan and Korea on the other. A further indicator is the profitability of the two groups (more specifically, the losses that have been made by new entrants such as Yahoo! BB in Japan and Hanaro and Thrunet in Korea). These indicators were discussed in detail in Chapter 1.

North America: The United States

The overall conclusion seems to be, therefore, that for the reasons analysed the intensity of competition in the U.S. is significantly lower than in Japan and Korea, the global broadband leaders, a conclusion emphasised in Chapter 1. This leads to Bauer’s final conclusion that “it is questionable whether [the U.S.] will be able to fully eliminate the gap . . . [with Japan and Korea] anytime soon” (p. 37). In concluding this introduction to Part II, it is worth returning to the U.S. puzzle highlighted in Chapter 1: Why is it that the U.S.—which gave birth to the Internet and leads the world in areas such as computing, software, and microprocessors—is lagging significantly behind in broadband? The contrast between the process of competition in the U.S. broadband and computer sectors is striking. Although a detailed discussion of this important issue is not possible here, several pertinent observations may be made. The U.S. broadband market, as documented in Part II, is dominated by two groups of incumbents: telecoms operators and cable companies. Aggressive new entrants, not to say disruptive competitors (as defined in Chapter 1), have been far less significant than in Japan or Korea. Furthermore, new technologies, championed either by incumbents fighting to stay ahead or by new entrants battling to get their foot in the door, have also thus far played a minimal role. This is notably the case with FTTH, which is now beginning to be introduced in the U.S. but at a slow pace compared to Japan. In Japan, in striking contrast, the aggressive new entrant in the FTTH market, riding on the back of new optical fibre technology, is in fact the incumbent in its own (electricity) industry. But the threat that it poses to the incumbent, NTT, is sufficient to force the latter to rapidly deploy FTTH.2 As noted in Chapter 1, so far other new technologies—such as wireless (including the 802.11 WiFi family as well as WiMax), satellite, and laser-based technologies—have played a limited role in delivering broadband access, providing limited opportunities for new entrants, although this may change in the future. The history of the dynamic U.S. computer industry, which has also powered the software and semiconductor industries,3 is very different from this country’s broadband industry. IBM’s rise to effective dominance in mainframe computers in the 1960s and 1970s was largely the result of its R&D-led innovation together with related organisational and marketing innovations. However, this dominance was effectively challenged, first by competitors such as DEC riding on the back of minicomputer innovations, and then by the personal computer challenge mounted by competitors such as Intel and Microsoft using the microprocessor. Entrepreneurial new entrants exploiting new technologies and the financial markets supporting them were a crucial part of the global success story that followed. The contrast with the U.S.

131

martin fransman

132

broadband industry, dominated by two groups of incumbents engaging in oligopolistic competition using their incumbents’ technologies (DSL and cable modems), is substantial. Martin Fransman

notes 1. See Chapter 1 for a description and discussion of local loop unbundling. 2. It is worth noting that competition in the Japanese FTTH market between the incumbent telecoms operator, NTT, and the incumbent electricity companies is not an example of oligopolistic competition analogous to that occurring between telecoms operators and cable companies in the U.S. The reason, as shown in the Japan chapter, is that currently NTT is required to provide access to its optical fibre network at low prices to aggressive new entrants such as Yahoo! BB, which is expected to make increasing use of this opportunity. This means that NTT’s FTTH-based offerings will have to meet strong competitive challenges from these new entrants, a challenge that will also have repercussions for the electricity companies that have entered the broadband FTTH market. A relatively cosy oligopolistically competitive relationship such as that existing in the U.S. broadband market is therefore precluded in Japan by the challenges of these aggressive new entrants. 3. For a recent account see Mowery and Nelson (1999).

reference Mowery, David C., and Richard R. Nelson, eds. 1999. Sources of industrial leadership: Studies of seven industries. Cambridge: Cambridge University Press.

5 Broadband in the United States Johannes M. Bauer

an observer of the recent U.S. broadband debate would find almost unanimous consensus that ubiquitous broadband access and the applications and services it enables are critical for economic growth, social welfare, and a vibrant civic society. There is much less agreement as to whether the country is on the best path toward this goal. Key stakeholders on one end of the broad range of positions, including the Federal Communications Commission (FCC) and important broadband service providers, are confident that advanced communications platforms are deployed at a satisfactory pace and at reasonable prices (FCC 2004a). On the other end are pundits expressing concern about the country’s slow rate of broadband diffusion. They point to OECD data indicating that the U.S. has lost ground compared to other industrialized nations, slipping from rank 6 in 2002 to rank 11 in 2005 (OECD 2003, 2005) and the cost this may entail for the economy.1 Not surprisingly, these different perceptions have also generated widely divergent policy recommendations (Crandall and Alleman 2002; CSTB 2002; Ferguson 2004), somewhat bifurcated between the federal and state levels. The FCC, on its own initiative and in response to directions by the courts, has created an ever-more market-based framework to the consternation of new entrants and consumer groups. At the state and municipal levels trust that market forces will deliver broadband access in a timely fashion is often weaker. Consequently, numerous complementary initiatives are underway to accelerate broadband deployment, including measures of demand aggregation, subsidization of broadband investment, and investment in publicly owned infrastructure. Given the dynamics of technology and of broadband markets, it is too early to assess whether the optimistic or the more pessimistic view better reflects the

johannes m. bauer

134

outlook of broadband in the U.S. Rather, this chapter explores the organization of the American broadband markets and assesses its likely future development. The following section presents background data on the structure of the broadband market, the nature and intensity of competition, and the overall performance of the sector in more detail. Sections three and four address the demand and supply conditions of the industry. The fifth section reviews selected performance indicators and the sixth section is dedicated to broadband policy. Section seven integrates the evidence from a systems-of-innovation perspective, followed by conclusions and an outlook. the anatomy of the u.s. broadband market As of June 2004, 485 firms offered broadband services in the U.S., up from 105 in 1999 (FCC 2004b, table 6). Two hundred and ninety-eight suppliers provided DSL service, 129 cable modem service, and 281 firms offered other platforms, such as wireless broadband, with several firms present in more than one platform (FCC 2004b, Table 6). However, the main players in the broadband access market are well-established cable and telephone companies. As of September 2004, according to data by the Leichtman Research Group, the five largest cable and telephone companies served 63.4% of subscribers. Taken together, the top 10 service providers captured 83%, and the top 20 supplied 95% of the broadband access market (see Table 5.1). The remaining 5% is served by a competitive fringe of more than 450 service providers. Only three companies founded since the mid-1990s—Covad, Bright House Networks, and Mediacom—rank among the top 20 service providers. Covad entered the market after the passage of the Telecommunications Act of 19962 (the “Act”), which was designed to facilitate competition in the local loop, then the only remaining monopolistic domain. Taking advantage of the unbundling provisions of the Act, Covad started to roll out DSL service in 1997. Like other new entrants, the company initially benefited from the high stock market evaluations of competitive entrants. However, the resulting broadband glut and recent regulatory rulings (discussed in more detail below) that made it more difficult and expensive to get access to local loops negatively affected this industry segment. It remains to be seen whether privately negotiated partnerships, such as the one announced between Covad and Verizon in December 2004, will suffice to overcome these structural difficulties. A second non-traditional player is Bright House Networks. In addition to business data services via its advanced fiber optical network, the company offers integrated home entertainment and communication services, including broadband Internet access, 250 channels of television, Video on Demand (VoD), and High-

Broadband in the United States

135

table 5.1 Leading Broadband Suppliers as of September 2004 Cable

DSL

Provider

Subscribers

Market Share (%)

Comcast Time Warner Cox Charter Adelphia Cablevision Bright House Networks Mediacom Insight RCN Cable One Top Cable Total U.S. broadbanda

6,554,000 3,716,000 2,430,555 1,819,900 1,253,407 1,259,024 700,000

20.1 11.4 7.5 5.6 3.8 3.9 2.1

350,000 311,500 215,000 165,600 18,774,986 32,580,000

1.1 1.0 0.7 0.7 57.6 100.0

Provider

SBC Verizon Bell South Qwest Covad Sprint ALLTEL Cincinnati Bell Century Tel Top DSL Total U.S. broadbanda

Subscribers

Market Share (%)

4,679,000 3,253,000 1,872,000 956,000 524,900 432,000 216,885

14.4 10.0 5.7 2.9 1.6 1.3 0.7

123,000 120,869

0.4 0.4

12,177,654 32,580,000

37.4 100.0

source: Leichtman Research Group Inc., Research Notes 4Q 2004, p. 6, available online at http://www .leichtmanresearch.com/research/notes12_2004.pdf, last visited 22 January 2005. aEstimated based on FCC 2004 data.

Definition television in several locations throughout five states. Mediacom, another relatively new player, was founded in 1995 and focused on serving non-metropolitan markets with basic cable, premium cable services, and high-speed Internet access. A large number of small broadband service providers have entered the market at a local or regional level. No detailed figures as to individual market shares are available, but their aggregate national market share is at most 5% (see Table 5.1). Many new firms are rolling out wireless broadband but their joint share of the market is presently less than 2%. National figures obfuscate the differentiated spatial nature of the broadband market. None of the suppliers in Table 5.1 has a truly national footprint. Fragmentation tends to be higher in the cable broadband industry, which is populated by 8,875 systems, most of them, due to the prohibitive costs of system overbuilds, the only franchise in a location. Fragmentation is somewhat counteracted by the integrative effect of Multiple System Operators (MSOs), which own many systems. In 2004, the top 10 MSOs owned systems serving approximately 80% of all cable subscribers.3 The market for DSL services is populated by more new entrants, not least because telephone companies until recently were subject to mandatory line sharing provisions.

johannes m. bauer

136

Some rudimentary data is available on the number of competitors by geographic location. The FCC collects subscriber data by postal (ZIP) code. According to this information, in June 2004, only 5.7% of ZIP codes did not report any high-speed4 lines in service, 13.8% reported one service provider, and 80.5% reported two or more service providers (FCC 2004b, Table 12). Whereas 11.8% of ZIP codes reported ten or more suppliers, the median number of broadband service providers per ZIP code was three. These figures imply that more than 95% of the population lives in ZIP codes with at least one highspeed provider. As broadband is not offered evenly throughout ZIP codes, the data probably overestimates the actual availability of broadband. Competition in the wholesale broadband access market has likely been dampened by the recent policy changes regarding line sharing, which will be discussed in more detail later in this chapter. On the other hand, the backbone market is flush with capacity and, while moderately concentrated, is presently highly competitive (Pelkovits and Cerf 2003). Given the relatively high cost of entering as a facilities-based service provider, broadband competition can be modeled as a two-stage decision. In a first stage, the decision to invest in basic broadband capability is made. During this stage, suppliers are competing in a race against each other for first mover advantages and presence in a location (Faulhaber and Hogendorn 2000; Woroch 2002). Once the infrastructure is in place (and the expenses sunk), competition will shift to other parameters, such as enhanced coverage in a location, price, and service conditions. The intensity of competition during the first stage will depend on the number and financial health of potential service providers whereas competition during the second phase will be influenced more strongly by the actual number of firms that have entered the market during stage one. The patterns of initial market entry and of the subsequent competitive rivalry will also depend on local demand and supply side conditions as well as the specific legal and policy rules governing the market. Large segments of a broadband network are shared infrastructure. Whether or not multiple service providers will be able to co-exist in the market depends on whether demand is sufficient to recover the fixed costs of building the network (Bauer, Kim and Wildman 2005). The unique nature of the benefits from broadband may further complicate market coordination. Some of the benefits of broadband have the character of club goods; they accrue to a limited group of users, but within the group are public goods. For example, bringing broadband to a business park may be such a club good. In this case, the challenge to organize the required number of users to justify the investment will need to be overcome. Other possible benefits, such as lower crime rates due to better access to

Broadband in the United States

education, may be pure public goods, creating value for one or multiple communities or even an entire region at large. In this latter case, private willingness to pay may not truly reflect the total benefits of broadband. Decentralized market decisions may therefore not yield socially optimal outcomes. These economic features challenge the assumption that unfettered market forces alone will deliver the desirable level of broadband access and content. Where cable and DSL service providers compete head-to-head, empirical evidence indicates that diffusion rates are higher (Burnstein and Aron 2003). However, competition in some, especially rural, areas is relatively sparse, contributing to spatial variations in access (Grubesic and Murray 2004). These findings are supported by studies reporting a cross-price elasticity of demand between DSL and cable larger than one, indicating the strong substitutive relation between these services (Rappoport, Kridel et al. 2003). A broad range of price and service packages is offered and service providers offer introductory discounts to attract customers to their networks. There is evidence that price competition is presently relatively weak but that providers compete by differentiating their service conditions. Some degree of price competition is introduced by wireless broadband providers. Overall, the market seems to exhibit a classical pattern of oligopolistic competition, carried out in dimensions other than price. broadband demand Although they share some similarities, there are also important differences between the residential and business markets for broadband access, which justify a separate treatment. More advanced applications, especially in business environments, require higher bandwidth. For these users, broadband is a precondition to pursuing innovative business models. However, for many other services, such as e-mail or information search, which are widespread among residential users, broadband is not a necessary condition but rather a higherquality substitute for dial-up service. Forman, Goldfarb and Greenstein (2003) in their study of business use of the Internet, distinguish between “participation” (including e-mail, browsing, and passive document sharing) and “enhancement” (uses of the Internet to change business processes and to implement new services). Users and user groups seeking basic participation in online activities will likely be able to satisfy their needs with dial-up or slow broadband services whereas more advanced users will migrate to higher bandwidth. Furthermore, advanced applications and services also require higher complementary investments in hardware and skills, possibly increasing the barriers to their use.

137

johannes m. bauer

138

Household Demand Details regarding the services driving residential broadband demand can be gleaned from the October 2003 census data and the regular surveys by the Pew Internet and American Life Project (“Pew”). Impatience is a main motive to switch to broadband: 36% of respondents in the most recent Pew survey stated that their dial-up connection had been too slow; 21% mentioned that they wanted to download files faster (Horrigan 2004). Ten percent gave job related tasks as a reason to switch. Only 3% answered that they switched because the price had fallen to an acceptable level. These findings are, to a large degree, reflected in the census data, which show a similar overall pattern of uses for narrowband and broadband users, but higher levels of use for those with a faster connection (see Figure 5.1). The largest use of the Internet continues to be e-mail, followed by search for product and service information and the acquisition of news. More than half of the Internet users aged 15 and over purchased products and services online; 41.6% looked for information on health services or practices; 38.1% played games; more than the 35.7% who searched for government information. If these averages are broken down by the type of connection available (dial-up, broadband, no home access), it becomes visible that broadband users engage in all activities more intensely. The difference between home broadband users and dial-up users is particularly high for e-commerce (15.1 percentage points) and home banking (14.9 percentage points). Not surprisingly, it is also much higher for radio streaming and the watching of TV or movies online (13.6 percentage points). There does not seem to be one particular main driver of broadband demand. Before the business associations of the content industries, in particular the Recording Industry Association of America (RIAA) and the Motion Picture Association of America (MPAA) started their intense legal campaign against music and video downloads in 2003, peer-to-peer (P2P) file sharing was a very fast, if not the fastest, growing use of broadband. P2P traffic constituted the bulk of the traffic on many high-speed university networks, leading several universities to adopt policies limiting the use of their networks to file sharing deemed in compliance with pertinent copyright legislation. According to surveys conducted by Pew, the number of Internet users who are downloading music dropped from 29% in the spring of 2003 to 18% in the winter of 2004. Some of these users shifted their demand to paid online music services, such as Musicmatch.com, Roxio Inc., or Apple’s iTunes.5 At present, voice over IP (VoIP) is seen as a potential “killer application.” However, in a survey conducted in February 2004, only 27% of online

Broadband in the United States

139 Search for a job Search for government information Search for health information Get news, weather, or sports information Search for product or service information Bank online Trade stocks, bonds, or mutual funds Take a course online Purchase products or services Listening to the radio, viewing TV or movies Playing games E-mail, IM

0 No Internet at home

10 20 30 40 50 60 70 80 90 100

Dial-up at home

Broadband at home

figure 5.1 Usage patterns of narrowband and broadband subscribers Source: NTIA 2004, p. 9.

Americans indicated that they had heard of VoIP, and a mere 3% of Internet users stated that they were considering acquiring it at home.6 These findings suggest that presently voice service alone—despite the many more features offered by VoIP—may not justify switching to broadband. Another future driver is seen in Internet-based interactive entertainment services. Several of the

johannes m. bauer

140

leading incumbent local exchange carriers are adopting aggressive strategies to deploy fiber, with plans to deliver an integrated bundle of services to homes, including voice, Internet access, and entertainment. Internet TV (IP TV) perhaps bundled with other services may have a promising future. Early studies concluded that its economics are unfavorable compared to over-the-air broadcasting, cable television, or direct broadcasting via satellite (DBS) (Owen 1999).7 This assessment was based on the assumption that Internet TV would be configured more or less like traditional television. The rapid advance of sophisticated and cheap digital recording and storage devices, such as the TiVo, digital recorders, and massive data storage centers that can host thousands of movies, promises alternative architectures to deliver entertainment. Consequently, the initial skeptical position has given way to a more optimistic assessment. The main DSL service providers are all in the process of introducing IP TV, as are mobile service providers. The increasing exchange of digital photos, integrated with mobile multimedia messaging, may gradually grow demand for broadband. Moreover, emerging concepts of web services and grid computing will further push demand for broadband. In all, it looks most likely that the migration to broadband will be gradual and based on a mix of applications rather than one vanguard service or application. In February 2004, 40% of dial-up users indicated that they would like to have broadband, but a surprising 58% answered that they did not want broadband (Horrigan 2004). Dial-up users who indicated their desire to switch to broadband were willing to pay $9.40 extra per month. However, 37% of dialup users were not willing to pay any more for broadband. On average, dial-up users were willing to pay $6.20 per month more for broadband. With an average dial-up price of $23, the current level of broadband prices ($39 for DSL, $41 for cable) therefore seems above the trigger point for many dial-up users to switch. Given these responses, the practice to offer introductory packages at prices that are just below these switching points may explain the continued migration toward broadband. Two additional obstacles to further broadband adoption are its unavailability in many locations, especially in rural areas, and the reluctance of large segments of the American population to go online in the first place. In the Pew survey, 15% of all dial-up users stated that broadband was not available in their area (26% did not know whether it was available). In urban areas, only 11% of dial-up users thought broadband was not available, whereas in rural areas 22% stated it was unavailable (Horrigan 2004). Thus, adopting measures to support broadband deployment to rural areas would likely have a positive effect on overall broadband adoption. Despite the continued growth, there is also evidence that the vision of ubiquitous broadband adoption may be only slow in

Broadband in the United States

the making, if not fully unrealistic. Since the fall of 2001, when the share of adult Americans online reached 60%, it has hovered around that mark, with 59% of American adults online in December 2004.8 Many of those who are offline do not intend to go online in the near future. There is a generational effect in play, and younger people are more likely to be online. For example, census data indicate that in 2003, 70.6% of the age group 18 –24 was online, whereas only 44% of people 50 and older used the Internet (NTIA 2004, pp. A1–A2). More recent disaggregated data by Pew show that by December 2004, 72% of 18 –29 year olds but only 24% of the group aged 65+ were online.9 Thus, if younger generations retain their online use, it can be expected that the number of Internet users will increase further, albeit at a slow pace. Furthermore, the speed of migration will also depend on the continued availability of offline substitutes for activities that moved online or are mediated online. As the Internet becomes more tightly woven into society’s fabric and the number of services such as e-government, tele-education, and tele-health increases, additional incentives to move online may be generated. This effect will be counteracted by increasing concerns about privacy and security. The presently available evidence seems to indicate that the saturation rate for broadband will be significantly below full penetration. Business Demand Very little systematic evidence is available about the market for business users. FCC statistics pool small businesses with residential customers. Of the 32.5 million high-speed lines in operation in 2004, 2.4 million served medium and large businesses (FCC 2004b, Table 1). Large businesses have historically operated their own networks and were relatively independent of public networks. Prices for dedicated access have fallen during the past years. With prices for fractional T-1 as low as $500 per month, more businesses may opt for dedicated access rather than DSL. However, in many areas T-1 is not available and incumbents are slow to deploy it. Because historically cable systems served only few small business customers, DSL enjoys a slim advantage in this market. InStat / MDR forecasts that business DSL lines will increase from 1.8 million in 2003 to 4.3 million in 2008 (USTA 2004, p. 33), a slight increase from about 20% to 21% of the total number of DSL lines. The business broadband access market has a different structure and dynamic than the market for private households. Internet access has become a necessary input for many businesses. Forman, Goldfarb, and Greenstein (2003), using a census of business establishments for the year 2000, found that 88.6% of business establishments participate in the Internet. However, only

141

johannes m. bauer

142

12.6% used the Internet to enhance their competitiveness, with another 10.6% experimenting with advanced Internet applications. Although these figures have likely increased in the meantime, the pattern illustrates several important issues. Advanced business applications often require a significant complementary investment in equipment, training, and business re-engineering. The estimated complementary private investment is at least as high as the total investment in public networks (Forman, Goldfarb, and Greenstein 2003). Moreover, the effects of general-purpose technologies such as the Internet materialize only slowly and through trial and error. Contrary to claims that the Internet has had dramatic impacts in only a few locations such as Silicon Valley or New York City, Forman, Goldfarb, and Greenstein (2003) show that it is largely the existing structure of business activity that affects the demand for business Internet access and its use. Leaders are found in sectors such as banking, insurance, data processing, publishing, and broadcasting. Laggards are sectors such as general merchandising, waste management, and laundry services (Greenstein 2004). The authors also found differences in both participation and enhancement rates between metropolitan areas, smaller cities, and more rural areas. However, in their view these differences do not necessarily reflect a digital divide. Rather, they might reflect the underlying economics of thin labor markets in rural areas and missing synergies and knowledge spillovers outside of larger agglomerations. broadband supply As would be expected from the economic characteristics of broadband networks, service availability far exceeds actual subscriptions. According to industry data, in 2004, 91% of the households passed by cable were able to receive cable modem service (NCTA 2005, p. 6). In cable modem availability, the U.S. leads other nations head-to-head with South Korea, the Netherlands, and Canada (OECD 2004c). The U.S. fares less well with regard to DSL availability. Although the growth rate of DSL has outpaced cable modem service for the past years, by 2003, the latest year for which detailed data is available, only 75% of all U.S. access lines were equipped for DSL, placing the nation number 15 among the industrialized countries (OECD 2004c). In 2004, 57.3% of all high-speed access lines were provided by cable systems, 35.1% via DSL, 2% via fiber, and 1.3% were based on wireless technologies including satellite and fixed wireless (see Table 5.2). Wireless broadband access is more widely utilized in rural areas (9% versus 5% in the rest of the nation). Cable modem service is more widely used in rural than in urban areas (53% versus 48%) whereas the opposite is true for DSL (35% in rural areas versus 41% in

Broadband in the United States

143

table 5.2 Technology Mix of Broadband Supply Number of access lines

ADSL Other wireline Coaxial cable Fiber Satellite, fixed wireless Total

June 2000

June 2001

June 2002

June 2003

June 2004

CAGR (%)

951,583 758,594 2,284,491 307,151 65,615

2,693,834 1,088,066 5,184,141 455,593 194,707

5,101,493 1,186,680 9,172,895 520,884 276,067

7,675,114 1,215,713 13,684,225 575,613 309,006

11,398,199 1,407,121 18,592,636 638,812 421,690

86.0 16.7 68.9 20.1 59.2

4,367,434

9,616,341

16,258,019

23,459,671

32,458,458

65.1

In %

ADSL Other wireline Coaxial cable Fiber Satellite, fixed wireless Total

21.79 17.37 52.31 7.03 1.50

28.01 11.31 53.91 4.74 2.02

31.38 7.30 56.42 3.20 1.70

32.72 5.18 58.33 2.45 1.32

35.12 4.34 57.28 1.97 1.30

N/A N/A N/A N/A N/A

100.00

100.00

100.00

100.00

100.00

N/A

source: FCC 2004b, table 1, own calculations. CAGR ⫽ Compound annual growth rate.

the rest of the nation) (Horrigan 2005). Overall, between 2002 and 2004, broadband access grew from 6% to 15% for the rural population; from 15% to 30% for the suburban population; and from 7% to 29% for the urban population (Horrigan 2005). Cable modems had gained an early lead over DSL and have since retained their position as the dominant platform. Telephone companies argue that this lead of cable is the outcome of a regulatory framework that—until its abandonment in 2003—disadvantaged them compared to cable. While this effect was likely in play, it was probably aggravated by the political battle of the telephone companies to get relief from regulation. Telephone companies may initially also have underestimated the growth potential of broadband and the agility of cable systems to pursue this market. Since June 2000, the share of DSL has expanded considerably, although mostly at the expense of “other” wireline technologies, including access provided by the telephone companies that was functionally equivalent to DSL. Thus, overall, the relative share of telephone companies and cable service providers in the overall number of high-speed access lines has been nearly stable (see Table 5.2). At this point, neither fiber nor wireless broadband play a major role in the overall mix of broadband access technologies, although both platforms are expected to expand fast. The roll out of fiber has likely been affected by the regulatory strategies pursued by the Regional Bell Operating Companies

johannes m. bauer

144

(RBOCs, the initially seven and now four holding companies that emerged from the break-up of AT&T in 1984), who sought relief from regulatory rules forcing them to make their network services available to competitors (see section 6 below for a more detailed discussion). Only after the contours of new, less restrictive rules were taking shape, have they initiated a more aggressive roll out of fiber networks. However, in the process, their plans changed. Prior to the promulgation of the new rules, the RBOCs had promised to develop fiber-to-the-home (FTTH) if relieved from unbundling and line sharing rules. Once these rules were in place, Bell South and SBC backed off the initial promises and are now proceeding with plans to develop fiber only to the neighborhood or to the curb. Nevertheless, these plans may be the logical next step in the upgrade to fiber-to-the-home as the solution provides access speeds sufficient to serve consumer needs in the near future. In 2003, shortly after the FCC changed its rules, Bell South, SBC Communications, and Verizon, three former RBOCs, had adopted common standards for the deployment of fiber-to-the-premises (FTTP).10 In 2004, Verizon announced its intention to deploy FTTP widely in its local networks, with a goal of passing 1 million homes within its eight-state territory by the end of the year. Download speeds offered are 5, 15, and 30 mbps. Verizon recently announced agreements with Starz Entertainment and other content providers, giving it access to more than 100 television channels. Bell South, after a favorable declaratory ruling from the FCC clarifying that hybrid networks (combining fiber with older coaxial or even twisted pair wire technology) were also covered by the new rules, scaled back its initial promises to only deploy fiber-to-the-curb (FTTC), about 500 feet from homes. SBC announced in November 2004 that it would invest $4 –$6 billion in its “Project Lightspeed,” in many cases terminating fiber even further from the premises. Only in new developments will SBC continue to deploy fiber-to-the-premises, as it has done since 2001. In existing developments, the project will bring fiber to neighborhood nodes (FTTN) but not to individual homes. The company plans to serve 18 million households by the end of 2007.11 In FTTN areas, the subscribers will be served with VDSL, permitting speeds of up to 20 –25 mbps, sufficient to supply video, voice, and high-speed Internet access. SBC plans to start offering video services using its own network in the fourth quarter of 2005, using Microsoft’s IPTV platform for video-on-demand (VoD). In many urban markets the company already offers a first version of a “triple play” service packet, including voice, Internet access, and television, supplied by DBS provider Echo Star. However, the company does not offer fully bundled pricing and it is unclear how the relation with satellite providers will evolve once the terrestrial networks are capable of carrying entertainment programming.

Broadband in the United States

Wireless broadband access is a very dynamic field ranging from third generation (3G) wireless platforms to WiFi (802.11) and predecessors of WiMax (802.16) and Mobile-Fi (802.20). The U.S. was relatively slow to develop a framework for licensed 3G services. Unlike most other countries, the U.S. model of flexible licensing permitted the use of 2G spectrum for 2.5G and 3G services allowing service providers a gradual migration to higher-bandwidth services. However, service providers were limited by the relatively narrow allocation of 140 MHz of spectrum (compared to more than 200 MHz in most of Europe). In 2004, the FCC allocated an additional 90 MHz of spectrum to advanced mobile services, which may be used to deploy 3G services.12 However, because the bands are currently occupied by government users, the spectrum will not be auctioned until June 2006 to allow clearing of the bands.13 Presently, most wireless carriers have deployed 2.5G capability in their networks and are experimenting in some markets with higher data rates (FCC 2003a). In June 2004, Sprint announced that it would deploy CDMA-EVDO upgrades to major metropolitan areas during 2005, allowing data speeds of up to 2.4 mbps.14 In July 2004, AT&T Wireless (now merged with Cingular) in partnership with NTT DoCoMo announced the availability of 3G service in Detroit, Phoenix, San Francisco, and Seattle. Several carriers have introduced multi-media services on their wireless networks. For example, Sprint offers a 13-channel radio service for a fee, and Verizon has introduced V-Cast, a television-like service. Despite these initiatives, mobile Internet access in the U.S. lags the global leaders in Southeast Asia and Europe. A flurry of activity exists in unlicensed services, in which many entrepreneurs, grassroots networks, and community-based initiatives have sprouted across the country (Sandvig 2003; Bar and Galperin 2004; FCC 2005a). However, no national statistical data is available and it is difficult to separate the hype from the actual state of deployment. According to estimates, by 2009, about 180,000 commercial hotspots will be deployed in the United States (ON World Inc. 2004). This figure does not include the rapidly proliferating home, office, campus, and free community networks. About one-third of the commercial hotspots are expected to be in cafes, another third in hotels, and the remainder in trains, retail, and other establishments. This market is currently populated by several service providers and, using conventional measures of market concentration, such as the Herfindahl-Hirschman Index (HHI),15 in each location is less concentrated than the fixed segment of broadband access.16 Wireless Internet Services Providers (WISPs) may use WiFi technology, such as Tropos Networks’ equipment, to build mesh networks. Several firms, including Boingo, offer turnkey WiFi solutions. Moreover, wireless deployments are often based on proprietary platforms such as Motorola’s “Canopy.” This system is available for

145

johannes m. bauer

146

licensed 900 MHz as well as license-exempt 2.4 GHz and 5 GHz bands. Canopy delivers 6 mbps aggregate data rates in point-to-multipoint node with a reach of up to 15 miles. In point-to-point mode the system has a delivery range of up to 80 miles with an aggregate data rate of 33 mbps.17 In the near future, it is anticipated that WiMax (802.16) and Mobile-Fi (802.20) will be used to provide broadband Internet access. In April 2005 Intel announced the availability of the first WiMax chipset. Mass market production and diffusion will likely not start until sometime during 2006, limiting the present role of the platform.18 performance Several government agencies and private sector research and consulting firms collect data according to different metrics. The heterogeneity in the available information renders it difficult to generate consistent time series of data, although the overall trends and the magnitude of numbers seem to be relatively robust. After a somewhat sluggish start, since 2000 the number of broadband access lines has been growing rather rapidly at a compound annual rate of 65.1%. According to Census data, in August 2000, 51.6% of U.S. households declared having a computer at home; 41.5% reported an Internet connection; and 10.6% of these households (or 4.4% of all households) relied on broadband access. By October 2003, the year with the latest available Census data, the share of computer owners had increased to 61.8%; the share of Internet households had expanded to 54.6%; and 36.4% of these households (or 19.9% of all households) were served by broadband. More recent data from other sources indicate continued growth. In-Stat / MDR, at the end of 2004, estimated the share of online households to nearly 70% and the share of households with broadband as 39% of these Internet households (or 27.2% of all households) (USTA 2004, p. 29).19 The number of Americans online via broadband expanded from 1.6% of the population in 2000 to 11.4% in 2004. During the past few years, growth in the online population has been slower, leading some to believe that the saturation level for Internet use will be lower than that for telephone service. From the time it had reached a 5% threshold, the diffusion of broadband is comparable to that of color television and exceeds the initial diffusion of dial-up Internet access, VCRs, and PCs (OECD 2004b). Only limited data is available documenting the average attained speed of broadband connections. In a survey during the first quarter of 2004, consultancy comScore Networks found that the average download speed for DSL was 861 kbps and that for cable modem service 2,178 kbps.20 Performance data collected by the CNET Bandwidth Meter—based on a smaller sample and likely heavier weighted toward home users—reports somewhat slower average

Broadband in the United States

speeds. During December 2004, the average cable modem speed among the 18 service providers whose subscribers had submitted data was 1,317 kbps.21 Companies using fiber networks, such as Insight Broadband, were able to attain higher speeds. During the same time period, the average DSL connection allowed data rates of 775 kbps.22 In both surveys, the difference between DSL and cable download speeds is in the same order of magnitude, with cable approximately twice as fast as DSL. Table 5.3 indicates the broad range of prices offered, but no systematic data as to the price level of services is available. Surveys by Pew indicate that the average monthly price for broadband has fallen moderately from $46 in January 2002 to $39 in February 2004 (Horrigan 2004). With $38 per month, DSL was on average slightly cheaper than cable, whose average price was $41. By comparison, the average price for a dial-up connection in February 2004 was $23 (Horrigan 2004). So far, price competition between service providers has been relatively limited although there are signs that it might intensify. Service providers differentiated their services mainly through improvements in the download speed and the bundling of Internet access with other services. This is reflected in the fact that between 2003 and 2005 none of the larger incumbent service providers included in Table 5.3 reduced prices for residential users. However, several of the new entrants reduced their prices. When telephone companies introduced DSL at a monthly price of $39.95, cable systems initially reacted by offering higher speeds at unchanged prices, often $49.95. Many service providers offer introductory discounts (after which prices go up to the regular level) or lower prices if customers sign one-year service contracts. Moreover, in certain areas wireless Internet service providers are entering with low-priced basic offers. During the past year cable companies and telephone companies have introduced bundled packages. For example, SBC offers DSL Internet access in bundles including local and long distance voice service and/or mobile voice service. Customers realize savings between $4 and $11 per month compared to the unbundled rates. SBC also offers entertainment services in partnership with DBS service provider Dish Network. Many cable companies have adopted “triple play” bundling models, offering cable television, Internet access, and cable telephony, some at relatively aggressive discounted prices.23 Per unit of bandwidth, in 2004, representative prices for leading DSL and cable modem service were nearly identical at $28.6 per mbps per month.24 Table 5.3 further indicates that a broad range of prices per unit of bandwidth exists, indicating a market environment with strong differentiation. However, U.S. prices per unit of bandwidth are still much higher than prices in nations with a higher penetration.

147

Type

ADSL ADSL ADSL

ADSL ADSL

Cable Cable

Wireless Wireless Wireless Wireless Wireless Satellite

Clear valuef Premiumg Basic Preferred Premium EarthLink Satellite

Road Runner High Speed Internet

TeleSoho (Business) DSL Internet

Residential Business DSL SBC Yahoo

Plan

512 1500 512 768 2000 500

1500 1500

1500 500 –1000

1500 or 768a 1500 or 768a 384–1500

49.95 N/A 37.00 57.00 N/A 69.95

44.95 42.95

69.95 49.95

29.95 or 34.95b 59.95c 26.95 or 49.95d

2003

2005

27.99 37.99 37.00 49.00 67.00 69.95

44.95 42.95

69.95 39.95

29.95 or 34.95 39.95c 26.95 or 49.95e

Monthly charge (USD)

97.56 N/A 72.27 74.22 N/A 139.90

29.97 28.63

46.63 49.95–99.90

19.97– 45.51 39.97–78.06 17.97–130.08

2003h

h

h

2005h

54.68 25.33 72.27 63.80 33.50 139.90

29.97 28.63

46.63 39.95–79.90

19.97– 45.51 26.63–52.02 17.97–130.08

Per Mbps, month (USD)

sources: Data for 2003 based on OECD 2004a, p. 60; data for 2005 own research. aDownload speed dependent on distance from central office: under 2.2 miles loop length will receive faster service; above 2.2 miles slower service. b$29.95 with one-year service commitment; $34.95 month-to-month; online orders receive one month free. cOne-year agreement required. dStandard package $26.95 with one-year commitment; $49.95 month-to-month. eExpress package. fPrices for St. Cloud, MN; first three months for $17.99. gPrice for St. Cloud, MN; first three months for $27.99. hRanges give the prices per Mbps and month for the fastest and cheapest and the slowest and most expensive option, which define the range of prices.

Clearwire Clearwire Wheatland Wheatland Wheatland Earthlink

Fixed wireless service providers

Time Warner Comcast

Cable service providers

Covad EarthLink

ISP using incumbent local loops

Verizon Verizon SBC

Incumbent local exchange carriers

Company

Download speed (kbps)

Selected Broadband Prices, October 2003 and February 2005

table 5.3

Broadband in the United States

broadband policy American broadband policy evolves in a piecemeal fashion, driven by political agendas, corporate strategy, and legal and regulatory battles. Policymakers at three levels are involved and proceed in diverse, sometimes contradictory, ways. Congress set the general direction in Section 706 of the Telecommunications Act of 1996, which directed the FCC to ascertain that the Internet would evolve “unfettered by state and federal regulation.” Subsequent administrations have emphasized their commitment to support advanced information infrastructure. However, no overarching national policy blueprint was developed. In June 2001, Democrats developed a “Congressional Democratic E-Strategy for Economic Growth.” During the 2004 presidential election both President Bush and challenger John Kerry came forward with broadband policy proposals, although they had starkly different visions, with Bush emphasizing market forces and Kerry universal service obligations. In April 2004 President Bush announced a deregulatory agenda with the goal of bringing broadband to every American by 2007, utilizing measures such as tax relief, additional spectrum allocations, and easier access to federal rights of way to stimulate private investment.25 At the federal level, the FCC is in charge of developing rules governing broadband markets. Over the past eight years an extended legal struggle unfolded over the specific framework that would guide advanced networks. For more than 100 years the U.S. telephone network had been operated as a common carrier network, requiring telephone companies to make their services available at non-discriminatory conditions. The deregulation of customer premises equipment (CPE) and value-added services since the late 1950s had increased the openness of the network. Thus, when the dial-up Internet emerged service providers could build their business models on a transparent platform, which is thought to be one of the main reasons for the rapid growth of the Internet in the U.S. In contrast, cable television systems had historically been operated as private contract carriers, relatively closed networks fully controlled by their owners.26 Wireless communications, while formally considered a common carrier, was relieved by FCC forbearance from many obligations, giving network owners strong control over their networks as well. Thus, when broadband emerged the question arose whether cable modem service should also be treated as a common carriage activity, whether DSL should be freed from common carriage obligations or whether these legal models should continue to co-exist. Although several cable franchise authorities at the city level, such as Portland, opted for the first approach, it was subsequently overturned by the courts. In March 2002, the FCC declared broadband access

149

johannes m. bauer

150

provided via cable modem as an interstate information service under FCC (and not local or state) jurisdiction and as such not subject to regulation. In contrast, incumbent local exchange carriers (ILECs) were mandated— under certain conditions—by sections 251 and 252 of the amended Communications Act to unbundle network elements.27 In determining which elements should be unbundled, the FCC was instructed to evaluate whether a competitor was “impaired” without access to it. In its Local Competition Order in August 1996, the FCC declared that seven network elements, including, inter alia, local loops, switching, and interoffice transportation, had to be unbundled. It further established that these network elements had to be priced at forward-looking incremental cost prices, for which the agency developed the Total Element Long-Run Incremental Cost (TELRIC) method.28 In a separate Order in 1999, the FCC expanded the unbundling provisions to include line sharing, requiring ILECs to also sell the high-frequency portion of the local loop at non-discriminatory rates. As a result of these rules, after 1999 competitive local exchange carriers (CLECs) could buy access to the highfrequency portion of the local loop at very low or even zero cost as regulatory agencies found that the cost of the loop was already recovered through retail rates (or the loop wholesale rates in cases where the loop had been rented by another CLEC to provide voice service) (Nuechterlein and Weiser 2005, p. 182). ILECs also had to make their retail DSL services available to their competitors at non-discriminatory wholesale prices. These arrangements lowered barriers to entry into the broadband access market for CLECs. However, several experts and the ILECs argued that the line sharing provisions distorted their investment incentives (e.g., Pindyck 2004). It was argued that whereas unbundling did not affect sunk investment (i.e., already existing copper loops), it did matter differently for new investment including DSL upgrades and fiber roll out, which would be retarded. The ILECs used this line of argument in addition to the claim that the FCC’s “impairment” standard was overly broad to challenge the line sharing and other unbundling provisions before regulatory agencies and in the courts. Over a period of several years, they gradually were able to change them (see FCC 2005b for a detailed account of the legal history). Of particular importance for the present rules were the FCC’s Triennial Review Order, issued in February 2003 (FCC 2003b), and the Triennial Remand Order, issued in February 2005 (FCC 2005b) in response to the decision by the D.C. Court of Appeals of March 2004 that had overturned several provisions of the Triennial Review Order.29 In the former, the Commissioners, forced to seek a compromise to reach the required majority of three votes, agreed to retain the narrowband (voice) unbundling provisions but phase-out line sharing over a period of three years. During that transition period, ending

Broadband in the United States

in 2006, prices to be paid by CLECs for existing line sharing agreements would increase in successive years to 25%, 50%, and 75%, respectively, of the full loop cost. Ironically, four of the five Commissioners (all but Commissioner Martin) expressed disagreement with the elimination of line sharing (although for different reasons) but attempts to re-introduce them were in vain. Central narrowband unbundling rules, most importantly the mandate to make switching available on an unbundled basis, were subsequently eliminated by the D.C. Court of Appeals. Other rules were made contingent on more stringent competitive conditions. After the transition period competitive service providers seeking to provide DSL will have to: (1) lease full loops, which continue to be subject to unbundling at TELRIC prices; (2) split a loop with another competitive voice service provider who might be willing to take the voice portion of the loop; or (3) build their own facilities. Most likely, any of these scenarios will increase the costs of new entrants compared to the status quo ante of being able to lease high-frequency loops at low prices. The Triennial Review Order also had eliminated unbundling requirements for fiber deployment to the premises (FTTP) in new developments (“greenfield” projects). In October 2004, in response to a request by Bell South and other ILECs, the Commission clarified that this exemption would also apply to fiber-to-the-curb (FTTC) projects in which fiber extends to within 500 feet of all the customers served by that loop (FCC 2005b). If an ILEC overbuilds copper loops, it will either have to keep the copper loop in service or make a narrowband channel available on an unbundled basis if the copper loop is retired. More specifically, ILECs must provide access to a voice grade channel via time division multiplexing (TDM) technology or, if no TDM is available, make a 64 kbps channel available. In the Triennial Order, the FCC had also eliminated the broadband sharing requirement for hybrid loops. In hybrid networks fiber is deployed to points that do not qualify as FTTP or FTTC. In such cases, CLECs may deploy their own networks to the fiber termination point of the ILEC (“remote terminal”) and then lease the remaining copper loop (called “subloop”). Overall, interpreting the instruction in the Telecommunications Act to facilitate the deployment of advanced communications infrastructure and services, unbundling obligations in the broadband markets have been essentially eliminated. With these Orders the legal frameworks for cable modems and DSL were made more compatible, although many experts believe that the FCC should have moved the other way, subjecting cable to open access provisions. Taken together, the short-term effect of the new rules is likely an increase in the cost of inputs for broadband ISPs and competitive DSL providers. Overall, the new framework strengthens the investment incentives of the ILECs but it may reduce the investment incentives of the CLECs. However, it will provide weaker

151

johannes m. bauer

152

incentives for fiber deployment than a framework that would have maintained line sharing obligations for cooper loops and thus created an additional incentive to migrate away from DSL (as FCC Chairman Powell had envisioned). It is difficult to predict whether the claimed positive effects of these rules on ILEC advanced infrastructure investment will dominate or whether the negative effects via reduced competition and the mix of DSL and fiber will be stronger. As was discussed above, ILECs have already backed away from some of their earlier promises. Three of the five FCC Commissioners seem to believe that the positive effects will prevail and outweigh higher wholesale and input prices for competitors.30 Whereas the empirical record indicates strong DSL and fiber growth during the past two years (which would support dominance of the positive effects) it is too early for a conclusive assessment. In addition to its unbundling policies the FCC has adopted a number of other measures that affect broadband markets. In 2002 the FCC formed a Spectrum Policy Task Force to review its approach to spectrum management. The Task Force recommended the expansion of license-exempt bands and the creation of bands allowing ownership of spectrum (FCC 2002). The FCC has since allocated additional spectrum to license-exempt bands—now totaling more than 680 MHz—that can be used to roll out wireless broadband. Most of these frequencies are shared with other applications and essentially create underlay rights for advanced wireless. In October 2004, the FCC released an Order adopting rules for broadband over power line (BPL) intended to increase competition and promote the availability of broadband services.31 In February 2005 the report by the Wireless Broadband Task Force, which had been initiated in 2004, was released (FCC 2005a), recommending, inter alia, improving access to spectrum, more flexible licensing, and more market incentives in spectrum management. The federal government also funds programs to support advanced Internet access for schools and libraries (“E-rate” program) and for rural health care providers. However, general broadband access is not part of universal service funding mechanisms. Between 1998 and 2003 the Universal Service Administrative Company, a private non-profit organization that administers the E-rate initiative, spent nearly $13 billion subsidizing advanced network access for schools. Although there is evidence of the success of these programs (Hudson 2004), they are also criticized for their inefficiency and questionable impact and were temporarily suspended to allow promulgation of new accounting safeguards.32 Given the high expectations associated with broadband, several states and many communities are concerned that the investment strategies of the leading suppliers, many of which target low cost–high profit urban areas first and only gradually deploy services to smaller communities and rural areas, might not establish connectivity fast enough. Consequently, they have adopted measures

Broadband in the United States

to accelerate the diffusion of advanced information technology ranging from legal reform to low-cost financing and direct investment in community networks. In January 2005 Tropos Networks, the largest municipal WiFi solutions provider, announced that it had acquired contracts with several large and small cities as customers, including Philadelphia, Oklahoma City, New Orleans, and Washington, DC to establish hotspot zones.33 Somewhat counter the dominant trend toward deregulation and expansion of market forces, cooperatives and the public sector are taking on a renewed role as infrastructure providers often carried out by municipally owned electric or water utilities. According to the American Public Power Association (APPA), at the end of 2004, 621 public power systems34 offered some kind of broadband service in their communities (APPA 2004; Gillett, Lehr et al. 2004). However, the dominant private service providers have mounted legal and court challenges to prevent or at least delay projects, often with the argument that public networks distort competition.35 Fourteen states, including Nevada, Nebraska, and Pennsylvania prohibit or put limits on municipalities’ ability to invest in broadband networks. Despite this animosity between parts of the private and public sectors there are many partnerships between government, non-profit organizations, and industry. For example, Abilene, a 10-Gbps network within the Internet 2 project that aims at 100 mbps connectivity between every Abilene connected desktop, is a collaboration including private vendors Qwest, Nortel, and Juniper Networks along with public and non-profit partners. the u.s. broadband system of innovation Before possible answers to the puzzle of a comparatively low broadband penetration in the U.S. shall be developed, it may be helpful to recall that the U.S. lagged behind other nations before. A recent example is digital mobile communications. During the 1980s, in analog mobile service, the U.S. had opted for a single national standard and soon could boast the fastest diffusion of service. However, it had taken quite some time to agree on a standard. Thus, when the next generation of digital mobile technology emerged, a decision was made to allow standards competition. This resulted in four industry groups deploying initially incompatible network platforms, a longer rollout process, and presumably a more fragmented and less differentiated market in terminal equipment. Consequently, from the early 1990s, the diffusion rate of mobile service in the U.S. lagged behind the EU, with the gap narrowing only recently. Nonetheless, by 2004, despite an overall lower diffusion rate, Americans were using their cell phones more than customers in most European nations. Americans were generally paying lower real prices, not least because of the more intense level of competition, which had contributed, among other

153

johannes m. bauer

154

things, to the introduction of free evening and weekend calling by U.S. service providers. This example raises the question as to whether in multimedia industries and broadband the U.S. might, once again, be on a slower “glide path” toward eventual leadership. After all, the nation can boast global leaders in hardware (e.g., Cisco, Motorola, Qualcomm, and Intel), a vibrant content industry, application service providers, network service providers with strong management skills, and a culture of entrepreneurship. This seems a possible scenario although, for reasons that will be discussed in more detail, closing the gap may be more difficult than in the mobile industry. Several comparative statistical studies suggest that broadband diffusion does not follow one pattern across nations but that it is highly dependent on national circumstances (see Bauer, Kim and Wildman 2005, which also provides a review of the literature). As has been discussed widely for other industries and economic development in general, markets, policy, and technology co-evolve in a dynamic interaction embedded in the respective national institutional environment (e.g., North 1990, 2005; Levy and Spiller 1996). Relevant aspects of the U.S. broadband system of innovation are the industry-government nexus, the status quo ante of Internet access platforms, the structure and evolution of competition in important broadband market segments, the choice to allow relatively closed network platforms with their possible negative repercussions for content provision, and supplemental policies affecting intellectual property rights, privacy, and security. The American communications policy system is populated by many players at the federal, state, and municipal level, related to each other through multiple channels of influence as well as many checks and balances. In addition to 50 state legislatures and U.S. Congress, 50 state public utility commissions and the FCC, it includes courts with considerable powers to review regulatory decisions on procedural and substantive grounds, antitrust authorities, and a broad range of business associations and advocacy groups. The system is complicated but flexible and typically produces incremental rather than sweeping policy change. Its complexity is often seen as a necessary safeguard against the dominance of any single interest group; however, it also offers many options for strategic gaming by stakeholders. All these factors have played out in the broadband arena. Cable companies were successful in defending their status as private contract carriers not subject to unbundling rules like the ILECs. As mentioned, ILECs may have used broadband as a lever to get relief from unbundling rules, an approach that finally succeeded in 2003. The extended legal struggle and the associated regulatory uncertainty may have contributed to the initially slow ILEC investment in DSL.

Broadband in the United States

A further factor is the availability to most Americans of a reliable and cheap dial-up platform. Unlike consumers in other nations, Americans benefited from the flat pricing structure of local telephone service. In that environment, while the access platform may be slow the incremental cost of Internet access is zero. Flat pricing and the presence of switching costs from the installed base to a new generation of access technology may thus have slowed the transition to broadband. Moreover, the specific unbundling framework put in place in the implementation of the Act may have (inadvertently) reduced the incentives for facilities-based competition. Not only did the rules create incentives to use unbundled network elements leased from the incumbent service providers rather than invest in new network facilities,36 they also channeled significant investment funds in the roll out of middle-mile fiber networks (which were initially rewarded with high stock market gains). The line sharing obligation of ILECs was not effective until 1999 and thus may have come too late to allow configuring the complementary access facilities required for the middle-mile investments to succeed. Eventually, the stock market bubble deflated in early 2000 devaluing CLECs drastically. All these factors contributed to a unique pattern of competition in the U.S. market. Whereas outside of the U.S. competition unfolds mostly between incumbents and new entrants in the DSL market (and thus line sharing and bitstream access are important policy instruments), the new U.S. unbundling framework dampens this form of rivalry. As a consequence, the dominant form of competition is inter-modal rivalry between cable and DSL and perhaps in the future between fixed and wireless broadband access. Due to the spatial structure of the broadband market competition unfolded unevenly. Moreover, perhaps as consequence of inter-model rivalry, price competition was less important than service competition, leading to a lower intensity of competition in the broadband access markets than in other countries. It seems that price competition is slowly intensifying, perhaps also accelerating the diffusion of broadband. The recent trend toward bundled services, such as “triple” and “quadruple play” offers by cable service providers and telephone companies, might boost competition, although this could be only a temporary phenomenon. The realignment of the policy framework has strengthened the position of the incumbent services providers. The vertical integration between local and long distance service providers as well as the integration between fixed and wireless service providers may further reduce the intensity of competition. Incumbents have also been successful in a growing number of states to get laws enacted restricting public sector participation in advanced infrastructure. All these factors may result in continued weak competition.

155

johannes m. bauer

156

Although this is more contested, a further contributing factor might be the closed architecture of cable systems—until recently the main access platform—and since the latest reforms of DSL platforms. This is in sharp contrast to the open and transparent telephone network platform upon which dial-up Internet access and a multitude of services could be configured. The debate on these issues is split between pundits arguing that the owners of broadband access networks would voluntarily enter into agreements with broadband ISPs and content providers as this would increase their profitability (e.g., Farrell and Weiser 2003) and those claiming that end-to-end open access is an important precondition for complementary innovation in content and applications (e.g., Lemley and Lessig 2001). Indeed, if content and knowledge production are cumulative and characterized by positive feedback effects, open platforms might stimulate content and application markets. The gap between broadband supply and demand seems to suggest that consumers other than heavy Internet users do not yet see a clear value proposition in broadband services. It is therefore possible that the effects of limited competition on the network infrastructure level are compounded by the limited openness and its potentially negative effect on content and applications. Lastly, a lack or poor design of supplementary policies and measures also seems to contribute to the present state of broadband in the U.S. Empirical studies reveal the existence of unequal access that resemble the patterns known from Internet access (e.g., Grubesic 2003). However, policies directed to supporting access may fall short of addressing digital illiteracy, which is likely more deeply rooted in shortcomings of the educational system. Likewise, the U.S. decision to implement an opt-out approach to privacy has contributed to the accumulation of vast amounts of personal, commercially traded data and increased risks that the information is abused for illegal purposes. With identity theft on the rise many users may remain skeptical as to the net benefits of broadband. Finally, with the passage of the Digital Millennium Copyright Act of 1998 (DMCA) and subsequent measures, the intellectual property rights framework has been tightened considerably, often criminalizing acts carried out in a digital environment that used to be legal in an analog world (Lessig 2004). Whereas some experts argue that even more stringent intellectual property protection is needed to accelerate the creation of broadband content many opposing views exist. In particular, it is argued that in information industries with sequential innovation processes, strong intellectual property rights might create a “thicket” of rights and very high transaction costs, possibly slowing the process of innovation. The institutional diversity created by initiatives to diversify intellectual property rights, such as the Creative Commons Project, should create a learning environment to better understand these relations.

Broadband in the United States

re´sume´ and outlook Although the American broadband market is expanding continuously, it lags behind other nations such as the global leaders South Korea, Japan, and Canada. The gap is more modest with regard to the Nordic and Benelux countries, which also are ahead of the U.S. The disparity can be explained by the coevolution of economic, policy, and demographic factors. No single driver seems to propel demand; broadband users typically spend more time online and do more of all online activities. It could be that the ubiquitous availability of dial-up Internet access, combined with relatively low flat pricing for local telephone service, have contributed to a slower adoption of broadband access. Ceteris paribus, in the short term, the recent elimination of mandatory line sharing will increase wholesale prices for access to high-speed loops and reduce competition from new entrants. On the other hand, it may accelerate investment in more advanced services by the ILECs and other players. Moreover, the recent homogenization of the policy framework may be at odds with the economic structure of the industry, which would require more spatial differentiation. The overall effect of these contradictory developments is difficult to predict but the most likely scenario is one in which the U.S. will gradually overcome the gap to its European peer nations. At this point it is questionable whether it will be able to fully eliminate the gap to its South East Asian peers anytime soon. Furthermore, although this is commonly assumed, it is not selfevident that ubiquitous broadband access is a necessary let alone sufficient condition to future economic growth and prosperity. At least temporarily, the American entrepreneurial culture and willingness to take risk might be able to compensate the lower availability of advanced networks and achieve dynamic growth regardless. acknowledgments The author wishes to thank Junghyun Kim for diligent research assistance in collecting some of the raw data for this chapter and Martin Fransman for helpful comments on earlier versions of the manuscript.

notes 1. For example Federal Communications Commissioner Michael J. Copps, “Disruptive Technology . . . Disruptive Regulation,” remarks at the Fifth Quello Communications Law and Policy Symposium, Washington, D.C., February 25, 2004, available at http://hraunfoss.fcc.gov/edocs_public/attachmatch/DOC-244356A1.pdf (last visited

157

johannes m. bauer

158

February 7, 2005). For a recent, very critical analysis see Bleha (2005). In a comparison that includes non-OECD countries, the International Telecommunication Union (ITU) ranks the U.S. as number 16 (see http://www.itu.int). 2. The Telecommunications Act of 1996 amended the Communications Act of 1934, to which section numbers refer unless otherwise indicated. 3. See National Cable and Telecommunications Association (NCTA) Statistics and Resources, available at http://www.ncta.com/Docs/PageContent.cfm?pageID⫽86 (last visited April 17, 2005). 4. In FCC terminology “high-speed” lines provide 200 kbps at least in one direction whereas “advanced” lines permit speeds exceeding 200 kbps in each direction. 5. For March 2004, comScore Media Metrix reported more than 11 million unique visitors to such services. See “The State of Music Downloading and File-Sharing Online,” Pew Internet Project and comScore Media Metrix Data Memo, Washington, DC, April 2004, p. 4. Available at http://www.pewinternet.org/pdfs/PIP_Filesharing _April_04.pdf (last visited January 28, 2005). 6. See “27% of Online Americans have Heard of VoIP Telephone Service, 4 Million are Considering Getting it at Home,” Data Memo—Pew Internet Project and New Millennium Council, Washington, DC, June 2004. Available at http://www .pewinternet.org/pdfs/PIP_VOIP_DataMemo.pdf (last visited January 28, 2005). 7. Similarly, Ting and Wildman (2002) illustrate the cost disadvantages of Internet radio. 8. See “Percent of American Adults Online 1995–2004,” Pew Internet & American Life Project, available at http://www.pewinternet.org/trends/InternetAdoption.jpg and “Demographics of Internet Users,” Pew Internet & American Life Project, available at http://www.pewinternet.org/trends/DemographicsofInternetUsers_12.20.04.htm (both visited January 28, 2005). 9. See “Demographics of Internet Users,” Pew Internet & American Life Project, available at http://www.pewinternet.org/trends/DemographicsofInternetUsers_12.20 .04.htm (last visited January 28, 2005). 10. See “BellSouth, SBC Communications, and Verizon adopt common technical requirements for fiber to the premises, will seek equipment proposals for potential network deployment,” available at http://www.findarticles.com/p/articles/mi_m0NVN/ is_22_23/ai_103240395 (last visited February 4, 2005). 11. See “SBC to invest $4 billion in fiber upgrade,” CNET news.com, November 11, 2004, available at http://news.com.com/SBC+to+invest+4+billion+in+fiber+upgrade/ 2100-1034_3-5449219.html (last visited February 5, 2005). 12. See In the Matter of Amendment of Part 2 of the Commission’s Rules to Allocate Spectrum Below 3 GHz for Mobile and Fixed Services to Support the Introduction of New Advanced Wireless Services, Including Third Generation Wireless Systems; Amendments to Parts 1, 2, 27 and 90 of the Commission’s Rules to License Services in the 216 –220 MHz, 1390 –1395 MHz, 1427–1429 MHz, 1429 –1432 MHz, 1432–1435 MHz, 1670 –1675 MHz, and 2385–2390 MHz Government Transfer Bands, Seventh Report and Order, WT Docket No. 02-8 and ET Docket No. 00-258, October 14, 2004, Washington, D.C., Federal Communications Commission. 13. See “FCC to commence spectrum auction that will provide American consumers new wireless broadband services,” Press Release, December 29, 2004, Federal

Broadband in the United States Communications Commission, available at http://hraunfoss.fcc.gov/edocs_public/ attachmatch/DOC-255802A1.pdf (last visited February 5, 2005). 14. See “Sprint Nextel to Pursue on CDMA-EVDO Vision, Spinoff Wireline Divisions,” Converge Network Digest, December 15, 2004, available at http://www .convergedigest.com/ Mergers/financialarticle.asp?ID⫽13261 (last visited February 5, 2005). 15. The Herfindahl-Hirschman Index is calculated as the sum of the squares of the market shares of all market participants. It ranges from 0 for a perfectly competitive market to 1 for a monopoly. U.S. antitrust authorities consider markets with an HHI below 0.1 as “unconcentrated”; markets falling in the range between 0.1 and 0.18 as “moderately concentrated”; and markets with an HHI above 0.18 as “concentrated.” 16. For example, in 2004, according to market share data by ON World (2004), in San Francisco the HHI was 1,716, which, in U.S. antitrust terms, is considered a “moderately concentrated” market. 17. See Motorola, “Solutions for WISPs,” available at http://motorola.canopywireless .com/prod_wisp/index.php (last visited February 5, 2005). 18. See “World domination postponed,” The Economist (U.S. edition), January 29, 2005, pp. 62– 63. 19. Most of the estimates converge to the cited figures. Nielsen NetRatings consistently reports higher numbers. For May 2004 it reported that 49% of all home Internet users were connecting via broadband. See “Broadband Internet access reaches 49% of all homes, NetRatings reports,” http://www.internetretailer.com/dailyNews.asp?id ⫽12383 (last visited January 22, 2005). 20. See “Cable Connection Speed More than Twice that of DSL,” media advisory, http://www.comscore.com/press/release.asp?id⫽451 (last visited January 26, 2005). 21. See CNET, “Top Cable Providers,” http://reviews.cnet.com/7020-9032_7-0 .html?tag⫽bbw (last visited January 26, 2005). 22. See CNET, “Top DSL Providers,” http://reviews.cnet.com/7020-9031_7-0 .html?tag⫽bbw (last visited January 26, 2005). 23. Cablevision is offering a triple play bundle consisting of cable service, Internet access, and VoIP for just below $90 per month before taxes, a significant saving compared to the individual prices. 24. Cable modem service provided by Comcast at $42.95 per month (1.5 mbps); DSL service provided by SBC Communications at $26.95 per month (942 kbps). 25. See “Bush: Broadband for the people by 2007,” CNET news.com, April 26, 2004, available at http://news.com.com/Bush+Broadband+for+the+people+by+2007/ 2100-1028_3-5200196.html (last visited February 5, 2005). 26. Congress, the FCC and franchise authorities have imposed some quasicommon carriage obligations in the form of limited must-carry, leased access, and public/educational/government access channel requirements. 27. For a more detailed discussion of the evolution of U.S. unbundling policy see Bauer (2005). 28. TELRIC is a “greenfield” approach in that it sets prices based on the assumption that the most efficient and modern available technology (and not the embedded plant used by the ILECs) is implemented. The ILECs claimed that TELRIC forced them to sell unbundled network elements at a 50 – 60% discount.

159

johannes m. bauer

160

29. In the Triennial Review Order, the FCC had delegated the task of assessing impairment of competitors on a more granular basis to the state regulatory agencies and had found that the mass market for switching was impaired on a national level. The D.C. Court of Appeals overturned both provisions and also found the impairment standard used by the FCC left wanting (Bauer 2005). 30. Commissioners Michael J. Copps and Jonathan Adelstein, the two Democrats on the FCC, strongly dissented from these respective parts of the Triennial Review Order and the subsequent Triennial Review Remand Order. 31. See In the Matter of Amendment of Part 15 regarding new requirements and measurement guidelines for Access Broadband over Power Line Systems; Carrier Current Systems, including Broadband over Power Line Systems, ET Dockets No. 04-37 and 03-104, Washington, D.C., Federal Communications Commission. However, the prospects of BPL are at best uncertain (see Tongia 2004). 32. See “Fraud threatens Internet program for U.S. schools,” CNET news.com, June 17, 2004, available at http://news.com.com/ Eroding+E-rate/2009-1028_35236723 .html (last visited February 8, 2005). 33. See MuniWireless.com, a portal providing information on municipal wireless and broadband projects, available at http://www.muniwireless.com/reports/index.html (last visited February 5, 2005). 34. See “Powering the 21st Century Through Community Broadband Services,” American Public Power Association, December 2004, available at http://www.appanet .org/files/PDFs/ Telecom-Flyer.pdf?sn.ItemNumber⫽2042&tn.ItemNumber⫽10002 (last visited February 5, 2005). 35. In some cases, municipally owned utilities may in fact enjoy advantages, such as tax exemption, access to low-cost financing, that are not available to private competitors. However, contrary to the claims of private industry, not all municipal enterprises do enjoy such advantages. 36. Of the 32 million competitive access lines in place in June 2004, only 23% were installed by the new entrants; 67% are based on network elements leased from incumbents; and 10% on resale.

references American Public Power Association (APPA). 2004. Community broadband: Separating fact from fiction. Washington, D.C.: APPA. Bar, F., and H. Galperin. 2004. Building the wireless Internet infrastructure: From cordless Ethernet archipelagos to wireless grids. Communications & Strategies 54: 54 – 68. Bauer, J. M. 2005. Unbundling policy in the United States: Players, outcomes and effects. Communications & Strategies 57: 59-82. Bauer, J. M., J. Kim, and S. S. Wildman. 2005. Towards an integrative framework for assessing broadband policy options. Michigan State University Law Review Spring (1): 21–50. Bleha, T. 2005. Down to the wire. Foreign Affairs, May/June: 111–24.

Broadband in the United States Burnstein, D. E., and D. J. Aron. 2003. Broadband adoption in the United States: An empirical analysis. In Down to the wire: Studies in the diffusion and regulation of telecommunications technologies. Edited by A. L. Shampine. Hauppauge, NY: Nova Science Publishers, 119–38. Crandall, R. W., and J. H. Alleman, eds. 2002. Broadband: Should we regulate highspeed Internet access? Washington, D.C.: AEI-Brookings Joint Center for Regulatory Studies. Computer Science and Telecommunications Board (CSTB). 2002. Bringing home the bits. Washington, D.C.: National Research Council, CSTB. Farrell, J., and P. J. Weiser. 2003. Modularity, vertical integration, and open access policies: Towards a convergence of antitrust and regulation in the Internet age. Harvard Journal of Law and Technology 17, 1: 85–134. Faulhaber, G., and C. Hogendorn. 2000. The market structure of broadband telecommunications. Journal of Industrial Economics 48, 3: 305–29. FCC. 2002. Spectrum policy task force report, ET Docket No. 02-135, Federal Communications Commission. ———. 2003a. In the Matter of Implementation of Section 6002(b) of the Omnibus Budget Reconciliation Act of 1993 Annual Report and Analysis of Competitive Market Conditions With Respect to Commercial Mobile Services, WT Docket No. 02-379, Eight Report, Federal Communications Commission, June 26, 2003 (Eighth CMRS Report). ———. 2003b. In the Matter of Review of the Section 251 Unbundling Obligations of Incumbent Local Exchange Carriers; Implementation of the Local Competition Provisions of the Telecommunications Act of 1996; Deployment of Wireline Services Offering Advanced Telecommunications Capability, CC Dockets Nos. 01-338, 96-98, 98-14, Report and Order and Order on Remand and Further Notice of Proposed Rulemaking, Federal Communications Commission, February 20, 2003 (Triennial Review Order). ———. 2004a. Availability of advanced telecommunications capability in the United States. Washington, D.C.: U.S. Federal Communications Commission. ———. 2004b. High-speed services for Internet access: Status as of June 30, 2004. Washington, D.C.: Industry Analysis and Technology Division, Wireline Competition Bureau, Federal Communications Commission. ———. 2005a. Connected on the go: Broadband goes wireless. Report by the Wireless Broadband Access Task Force. Washington, D.C.: Federal Communications Commission. ———. 2005b. In the Matter of Unbundled Access to Network Elements; Review of the Section 251Unbundling Obligations of Incumbent Local Exchange Carriers, WC Docket No. 04-313 and CC Docket No. 01-338 (Triennial Remand Order). Ferguson, C. H. 2004. The broadband problem: Anatomy of a market failure and a policy dilemma. Washington, D.C.: Brookings Institution Press. Forman, C., A. Goldfarb, and S. Greenstein. 2003. The geographic dispersion of commercial Internet use. In Rethinking rights and regulations: Institutional responses to new communication technologies. Edited by L. Cranor, and S. S. Wildman. Cambridge, MA: MIT Press, 113– 45.

161

johannes m. bauer

162

Gillett, S. E., W. H. Lehr, and C. Osorio. 2004. Local government broadband initiatives. Telecommunications Policy 28: 537–58. Greenstein, S. 2004. Broadband and Internet adoption by enterprises. Presentation at the Federal Communications Commission workshop, Formulating a Research Agenda for Communications Policy, in Washington, D.C. Grubesic, T. H. 2003. Inequities in the broadband revolution. The Annals of Regional Science 37: 263– 89. Grubesic, T. H., and A. T. Murray. 2004. Waiting for broadband: Local competition and the spatial distribution of advanced telecommunication services in the United States. Growth and Change 35, 2: 139– 65. Horrigan, J. 2004. Broadband adoption at home: Trends and prospects. Presentation at the Federal Communications Commission workshop, Formulating a Research Agenda for Communications Policy, in Washington, D.C. ———. 2005. Broadband adoption at home: Trends and prospects. Presentation at the PURC Annual Conference, Emerging Technologies and Trends, in Gainesville, FL. Hudson, H. 2004. Universal access: What have we learned from the e-rate? Telecommunications Policy 28, 3– 4: 309–21. Lemley, M. A., and L. Lessig. 2001. The end of end-to-end: Preserving the architecture of the Internet in the broadband era. UCLA Law Review 48, 4: 925–72. Lessig, L. 2004. Free culture: How big media uses technology and the law to lock down culture and control creativity. New York: Penguin. Levy, B., and P. T. Spiller, eds. 1996. Regulations, institutions, and commitments: Comparative studies of telecommunications. Cambridge: Cambridge University Press. National Cable and Telecommunications Association (NCTA). 2005. 2004 year-end industry overview. Washington, D.C.: NCTA. North, D. C. 1990. Institutions, institutional change and economic performance. Cambridge: Cambridge University Press. ———. 2005. Understanding the process of economic change. Princeton, NJ: Princeton University Press. Nuechterlein, J. E., and P. J. Weiser. 2005. Digital crossroads: American telecommunications policy in the Internet age. Cambridge, MA: MIT Press. NTIA. 2004. A nation online: Entering the broadband age. Washington, D.C.: U.S. Department of Commerce, Economics and Statistics Administration, National Telecommunications and Information Administration. OECD. 2003. Communications outlook. Paris: Organisation for Economic Cooperation and Development. ———. 2004a. Benchmarking broadband prices in the OECD. Paris: Organisation for Economic Co-operation and Development. ———. 2004b. Information technology outlook. Paris: Organisation for Economic Cooperation and Development. ———. 2004c. The development of broadband access in rural and remote areas. Paris: Organisation for Economic Co-operation and Development. ———. 2005. Communications outlook. Paris: Organisation for Economic Co-operation and Development.

Broadband in the United States ON World Inc. 2004. North American hotspots: A rapidly growing ecosystem. San Diego, CA: ON World. Owen, B. M. 1999. The Internet challenge to television. Cambridge, MA: Harvard University Press. Pelkovits, M. D., and V. G. Cerf. 2003. Economics of the Internet. In Emerging telecommunications networks. The international handbook of telecommunications economics, vol II. Edited by G. Madden. Cheltenham: Edward Elgar, 27–54. Pindyck, R. S. 2004. Mandatory unbundling and irreversible investment in telecom networks. NBER Working Paper 10287, Cambridge, MA: National Bureau of Economic Research, Inc. Rappoport, P. N., D. J. Kridel, L. D. Taylor, J. H. Alleman, and K. T. Duffy-Deno. 2003. Residential demand for access to the Internet. In Emerging telecommunications networks. The international handbook of telecommunications economics, vol II. Edited by G. Madden. Cheltenham: Edward Elgar, 55–72. Sandvig, C. 2003. Assessing cooperative action in 802.11 networks. Paper presented at the 31st Annual Conference on Communications, Information and Internet Policy, September, in Arlington, VA. Ting, C., and S. S. Wildman. 2002. The economics of Internet radio. Unpublished manuscript, East Lansing, MI: Michigan State University. Tongia, R. 2004. Can broadband over powerline carrier (PLC) compete? A technoeconomic analysis. Telecommunications Policy 28, 7– 8: 559–78. United States Telecom Association (USTA). 2004. Broadband facts. Washington, D.C.: USTA. Woroch, G. A. 2002. Open access rules and the broadband race. The Law Review of Michigan State University Detroit College of Law Fall, 3: 719– 42.

163

part three. europe: france, germany, italy, and sweden

part iii contains chapters on broadband in four European Union (EU) countries: France, Germany, Italy, and Sweden. Comparative information is also provided, both here and in Chapter 1, on the United Kingdom. The uninformed outsider might suppose that there are significant commonalities between these five countries, particularly in view of the existence of the European Union’s New Regulatory Framework, which applies to broadband communications as well as other areas of electronic communications. Such a supposition, however, would be wrong as will shortly be shown in this introduction to Part III. But first, by way of background, a brief account is provided of the evolution of the New Regulatory Framework. the european union’s new regulatory framework In 1990, with the adoption of rules on open access (contained in the ONP [open network provision] Framework Directive), a comprehensive regulatory framework was introduced for the liberalisation of the telecoms market and the achievement of a single market for telecoms services and equipment within the European Union. Liberalisation was achieved in three steps. Step 1 involved the partial liberalisation of telecoms services through the harmonisation of network interfaces and rules on open access. In Step 2 partial liberalisation of infrastructures was implemented, involving mobile communications networks and cable television networks. Full liberalisation came with Step 3 introduced from 1 January 1998. This included the liberalisation of all services, including voice telephony, and all infrastructure provision. With the full liberalisation of services and infrastructure provision achieved formally from January 1998, discussion began on a New Regulatory

martin fransman

166

Framework that would establish an appropriate context for the development of telecoms in the EU in the era of competition and for achieving EU objectives in this field. On 10 November 1999 the European Commission presented the 1999 Communications Review to the European Parliament. On 26 April 2000 the Commission presented a communication to the European Parliament on the results of the public consultation on the 1999 Communications Review. On 7 March 2002 the five Directives were issued comprising the main part of the New Regulatory Framework. Member states were required to comply with these measures from 25 July 2003. In the 1999 Communications Review it was already concluded that the establishment of an EU regulatory authority would not “add sufficient value to justify the likely costs” (EC 1999:14). Accordingly, it proposed that “primary responsibility for achieving objectives set out in sector-specific Community legislation should rest with the independent national regulators (NRAs) who are best placed to take account of the different levels of competition and market development in Member States” (EC 1999:14). But if the NRAs were to be given “primary responsibility,” how would harmonisation of regulation across member states be achieved? (Without harmonisation, the objectives of a single market would be frustrated.) In the 1999 Communications Review three mechanisms were proposed to achieve harmonisation. First, the Commission, in its directives underpinning the New Regulatory Framework, would provide a “framework of principles” for NRAs to follow (EC 1999:ix). It was hoped that this framework would produce a significant degree of consistency among member states. Second, the Commission proposed the creation of what was called a High Level Communications Group consisting of NRAs and the Commission that would discuss consistency problems and propose solutions on the basis of rules agreed at Community level. Third, the Commission proposed that it itself would have to undertake “monitoring and quality assessment activities to ensure consistent and effective implementation of regulation at national level” (EC 1999:14). On 7 March 2002 five Directives were issued comprising the main part of the New Regulatory Framework with which member states were required to comply from 25 July 2003: the Framework Directive, the Authorisation Directive, the Access and Interconnection Directive, the Users’ Rights Directive, and the Data Protection Directive. broadband in the five european countries Although supposedly regulated by a common European Union regulatory framework, there are significant differences between the five European

Europe: France, Germany, Italy, and Sweden

countries under review here. The reasons for these differences are straightforward: industrial structure, institutional specificities (including regulatory institutions), and historical path-dependencies have all been shaped primarily at national level. Although it is conceivable (but by no means inevitable) that in the future European institutions such as the New Regulatory Framework may influence these determinants of broadband communications, for the foreseeable future national differences together with the delegation of powers of implementation to national regulatory authorities (under the concept of subsidiarity) will ensure that such differences continue. The differences are most apparent if we focus on the main incumbent provider of broadband connectivity and services in each of the five countries and examine some of the most important challenges that this incumbent faces. In so doing it is worth keeping in mind the corresponding challenges facing the major incumbents in Japan, Korea, and the United States examined in Chapter 1 and in the introductions to Parts I and II. At one extreme lies Germany, the largest European country in terms of both economy and population. The outstanding broadband fact about Germany’s incumbent, Deutsche Telekom, as emerges from the picture that Franz Büllingen paints in his chapter on Germany, is the overwhelming dominance of this company. As Büllingen points out, in 2002 Deutsche Telekom controlled 94% of the broadband market. By 2004 this figure had decreased to 88%. How has Deutsche Telekom been able to control such a large share of the market? Two reasons emerge from Büllingen’s account. The first is the relative absence of inter-modal competition coming from cable companies (the complex causes of which are examined in Chapter 7). As OECD data shows (see Figure 1.1 in Chapter 1), Germany and Italy stand out amongst the highincome countries in terms of the insignificance of cable as a source of broadband (although Iceland and Ireland also have low cable visibility). Furthermore, competition coming from other technologies/networks than cable is also virtually non-existent (as is true for most of the countries in Figure 1.1). Second, Büllingen’s story also shows that intra-modal competitors, also using DSL like Deutsche Telekom, are relatively weak (as is confirmed by their small collective market share). In part their weakness is to be explained by the relative insignificance of DSL competitors providing broadband connections over lines unbundled by Deutsche Telekom. The relative absence of competition in the German broadband market has become an important political issue as Büllingen makes clear: “Lack of competition in the DSL market is one of the most crucial points in the current regulation debates. . . . Competitors have complained to the regulator that most of the regulatory measures proposed by the European Commission have been set out far too late and are too weak.”

167

martin fransman

168

But this second reason moves the question one step back: Why has Deutsche Telekom been able to get away with so little entry from determined competitors using unbundled access to its network? In reply to this question the accusing finger points inevitably to the German regulator, RegTP. The case of Japan, examined in detail in Chapter 2, shows just how much competition a regulator can facilitate if it is determined to create low entry barriers for competitors willing and able to use unbundled access. The clear implication of the logic of the present discussion is that RegTP has been either unwilling or unable to impose competition on Deutsche Telekom. But again this moves the question one step further back: If this is indeed the case, how is it to be explained? This raises a host of further questions. For example, does the fault lie with the regulator or with the German government, which may have overridden or constrained the regulator in Deutsche Telekom’s favour? And whoever is to blame, what are the precise political and ideological circumstances that have produced this outcome? As in the case of some of the other countries examined in this book, our inability to address important questions such as these is a reflection of the inadequate attention paid by analysts and social scientists (particularly economists) to what regulators actually do and why in favour of recommendations about what they should be doing based on notions of optimal regulatory outcomes.1 The result, unfortunately, is that institutions such as these are not taken seriously. The broadband situation in the UK makes an illuminating contrast to that in Germany. As noted in Chapter 1, much has been made in the UK of the strong competition that exists in the broadband market and the incumbent BT’s relatively low market share. Recent fall in broadband access prices have been used to indicate the extent of satisfactory competition. However, such assertions fail to acknowledge other important features of the British broadband story. Most importantly, these assessments fail to emphasise that this competition exists in the retail market. The reality is that BT has dominated the wholesale DSL market with the company’s retail competitors buying and reselling BT’s wholesale products in one form or another. Furthermore, inter-modal competition from cable companies, although high relative to the other large European countries (see Figure 1.1 in Chapter 1), has been muted as a result of the financial difficulties faced by the two mail cable competitors, ntl and Telewest. A further important feature of the British broadband story has been the virtual absence of competition to BT based on unbundled lines (as is graphically shown in Tables 1.6 and 1.7 in Chapter 1). This has until now prevented the kinds of new entrants that as we saw earlier played such an important

Europe: France, Germany, Italy, and Sweden

role in Japan and Korea and more recently have begun to play in France (see below). While competition in the retail broadband market may have served to reduce retail prices, prices have been constrained by BT’s (regulated) wholesale price. Furthermore, the absence of unbundling has meant that new entrants have been unable to compete through innovation in the portion of the network to customers that they control, a key benefit of unbundled access in countries like Japan and more recently Korea. Asking the same question of the UK as was asked of Germany, namely why the regulator has let the incumbent get away with limited effective unbundling, yields a little more light than in Germany. As was noted in Chapter 1 (in the section on the Japan/Korea puzzle) the Director General of the then-regulator Oftel, in an unusually frank interview, said that if he had “realised earlier that BT was playing a long game” he would have “handled local loop unbundling differently.” More specifically, he would have been “more directive.” Oftel’s successor from 2004, OFCOM, has apparently taken these issues on board as is evident in its strategic telecoms review and the steps it has taken to increase unbundling substantially. The situation in France, as Jackie Krafft shows in Chapter 6, provides a further contrast both to Germany and to the UK. Krafft describes how the incumbent, France Telecom, until 2002 enjoyed a fairly easy ride from its regulator, ART. Furthermore, Krafft highlights the remarkable fact that France Telecom owns 40 percent of the capital of the cable companies that supposedly provide a degree of inter-modal competition in the French broadband market. This, she suggests, has resulted in a “distortion of competition” that has “induced very low incentives to develop [cable networks capable of providing broadband competition].” But Krafft also shows that the situation changed significantly in 2002 when ART became “increasingly inspired by the broadband success observed in countries such as Japan and Korea.” As a result, “the regulator introduced aggressive policies [regarding] unbundling and insisted that the incumbent had to charge very low prices to competitors wanting to use its local access network.” The result, through the entry of aggressive new entrants such as Free and Neuf (discussed in Chapter 1), has been a significant improvement in France’s broadband performance, making it one of the bestperforming in Europe. This outcome has explicitly influenced the thinking of OFCOM in the UK and challenged conventional wisdoms in Germany. Italy provides a further contrast. As Cristiano Antonelli and Pier Paolo Patrucco point out, the Italian incumbent, Telecom Italia, faces no intermodal competition from cable as a result of a “complete lack of co-axial cable” in Italy. In turn, although the authors do not go into this in detail, the absence of cable competition is in part the result of the institutional conse-

169

martin fransman

170

quences of Italy’s broadcasting system, a system influenced by the fact that Berlusconi, the Italian Prime Minister, is also the country’s billionaire media mogul. Instead, competition comes from new entrants, the most notable of which is a firm called Fastweb. Fastweb was established in 1999 in Milan, a joint venture between the city’s main power company, AEM, and a firm called e.Biscom established by an entrepreneur by the name of Silvio Scaglia. A notable characteristic of Fastweb is its extensive use of FTTH (although it also makes use of DSL). This largely accounts for Italy’s relatively large ratio of other technologies compared to DSL plus cable (as shown in Figure 1.1 in Chapter 1). The final European country, Sweden, is one of the best-performing in broadband as Figure 1.1 shows, coming only slightly behind Japan in terms of penetration and well-ahead of the other European countries included in this book. Furthermore, Sweden has one of the highest ratios of “other” broadband technologies compared to DSL plus cable, largely a reflection of this country’s relatively large diffusion of fibre to the home. Why has Sweden performed relatively well? In Chapter 9, Sven Lindmark and Per Björstedt suggest several reasons. Perhaps the most prominent— and the most unusual—is the significant role that has been played by the Swedish government. Indeed, of the countries examined here it is only Korea (and, of course, China) where government has played such a prominent role in the broadband area. One of the main consequences of substantial government involvement has been that the incumbent, TeliaSonera (the merged incumbents of Sweden and Finland), has had its wings clipped by the presence of significant government-owned and controlled broadband infrastructure. As the authors point out, “the Swedish government has played a large role in the Swedish Internet and broadband development, both directly and indirectly.” They note that through “the ownership of infrastructure companies such as Telia[Sonera], Vattenfall (power generation), Svenska Kraftnät (power grid), Banverket (railroad), and Teracom (broadcasting), the state still controls a very large share of the total communication backbone systems. In 2003, the state controlled 78% of the high-speed network infrastructure, and municipalities and municipality-owned corporations controlled another 11%.” Through these companies—which include electricity and railway companies—the Swedish government has attempted to achieve its goal of an efficient broadband infrastructure for all. But it is not only government that constrains the broadband activities of TeliaSonera. From 2003 strong inter-modal competition has come from the cable company, Com hem. Significantly, until this time Com hem was the cable subsidiary of Telia but a regulatory condition of the merger between

Europe: France, Germany, Italy, and Sweden

Telia and Sonera of Finland was that the cable company be spun off. It has since become a significant broadband competitor. However, TeliaSonera also faces strong competition from aggressive new entrants, the largest of which were Bredbandsbolaget and Bostream. In 2003, as the authors show, these companies had 14% and 9% of the Swedish broadband market, respectively. In 2004 the former acquired the latter. Importantly, Bredbandsbolaget is also a significant deployer of fibre to the home. In 2003 the authors estimate that this company had a 65% share of the FTTH market. Its activities, together with those of various levels of the Swedish government, explain Sweden’s relatively high ratio of “other technologies” to DSL plus cable shown in Figure 1.1 in Chapter 1. In short, in Sweden a relatively high intensity of competition has resulted from the deployment of government-owned infrastructure, the activities of the cable company Com hem, and the rivalry of aggressive competitors such as Bredbandsbolaget. In turn, this has produced Sweden’s relatively good broadband performance. Although good comparative international data on FTTH do not exist, it is likely that Japan and Sweden are the world’s leaders in FTTH. The reasons for this leadership are examined in this book, in Chapter 1, the introduction to Part I, and in the Japan and Sweden chapters. However, it is clear that Europe as a whole, including the European countries examined in this book (Sweden and to some degree Italy excepted) is lagging significantly behind Japan and also, though to a lesser extent, the U.S. in the deployment of FTTH. This is a further indicator of the European lag in broadband. Martin Fransman

note 1. In my own formal and informal interviews with numbers of Europeans well-informed and well-placed in key companies and public sector organisations, my questions about the reasons for these outcomes in Germany failed to produce satisfactory answers. While there is a clear perception in European circles that regulation of the incumbent in Germany is relatively “weak,” there is no coherent or consistent view regarding why this is the case.

reference European Commission (EC). 1999. Towards a new framework for electronic communications infrastructure and associated services: The 1999 Communications Review. (COM [1999] 539).

171

6 Emergence and Growth of Broadband in the French Infocommunications System of Innovation Jackie Krafft

in this chapter, we focus on the emergence and growth of broadband in France. The issue deals with important questions in the domain of industrial dynamics. Broadband is, of course, related to the mature telecommunications industry, since one of the major competitors in this activity is the incumbent telecoms operator, France Telecom (FT). In this perspective, the emergence and growth of broadband is deemed to be highly determined by the role that France Telecom has played and still plays within the French Infocommunications System of Innovation (FISI). But broadband is also a radically new activity, supported by a new technological trajectory that significantly changes the FISI, and thus requires a deeper investigation. Emergence and growth of broadband is also connected to important challenges in the domain of international competitiveness. The one who takes the lead in this global competition will play a decisive role for the future of the information society. If we consider various indicators given in the introduction of the book1 on the development of broadband, on the one hand, France has a medium ranking for penetration and price but, on the other hand, French operators (incumbents or entrants) are absent in the ranking of the top ten providing the highest capacity/speed. This unequal development of broadband is a general characteristic of European countries and helps to explain why Europe lags behind Asia and North America. Essentially, the fact that major European companies (except for Swedish and Belgian companies) are not at the core of the global competition appears as a major weakness in the position of Europe. In that respect, France is an interesting case; though within the medium ranking of Europe, France performs nevertheless better than Germany, the UK, or Italy, and still progresses up in the ranking.

Broadband in the French Infocommunications System of Innovation

In this chapter, we will essentially dissect the industrial dynamics of the French broadband industry. We argue, however, that innovation and the emergence of a new industrial dynamics do not come only from companies, but are also highly influenced by the infocommunications system of innovation that prevails at the national level. In the line of a series of recent contributions, we think that innovation and industrial dynamics are linked with the development of a combined process of change involving supporting institutions such as regulation, standardisation, and competition authorities, as well as government agencies and policies (Pavitt, 2001; Nelson, 2004; Metcalfe, 1995; Fransman, 1994, 2002, 2004; Antonelli, 2001, 2003; Saviotti, 1996, 2001; Witt, 2003; Krafft, 2003, 2004). The purpose of this chapter is thus to analyse how industrial dynamics in the domain of broadband operates within the FISI. We first define broadband, in order to specify the boundaries of the industry (in reference to a good— broadband access—and its associated services), as well as the relative strengths and weaknesses of broadband in France over time, compared to other European countries (Section 1). We then concentrate on one of the key features of industrial dynamics, namely the competition that lies between technologies and between firms (Section 2). We further relate this industrial dynamics to one key dimension of the FISI, the regulatory framework (Section 3). We consider the fact that this regulatory framework is itself highly dependent on the legal framework and judicial procedure that makes all the FISI gradually develop and change (Section 4). In all these sections, we will also note, however, that there are still elements in the FISI that presumably cannot change in the near future, since they are characterized by important path-dependencies, historically or geographically related, that strongly shape how this innovation is developed and used. The last section will conclude our discussion. defining broadband Broadband today still has a very large definition. By broadband, French regulation and competition authorities generally mean Internet connections that are higher than 128 kbps. Of course, the choice of the threshold between broadband and narrowband Internet can greatly affect the boundaries of the industry as well as results in comparative performance, especially if other countries fix the threshold at a higher level.2 Finally, this definition is greatly evolving over time. Since technology is changing so rapidly, what was considered as broadband yesterday can turn out to be narrowband today. Difficulties in defining broadband involves that reference is now more often made to the spectrum of related services and Internet-user applications.

173

jackie krafft

174

Only a few Mbps are necessary for real-time video, some tens of kbps are sufficient to listen to music, especially radios over the Internet. Navigation on the web and e-mail require also just a few Mbps, but broadband greatly enhances downloading ease. In practice, however, traditional criteria still help to characterise broadband, such as availability, penetration, capacity/speed, tariffs, and so on. In this section, these different criteria are presented and informed on the basis of more recent data collected from annual reports of the French regulator ARCEP (Autorité de Régulation des Communications Electroniques et des Postes; formerly ART, Autorité de Régulation des Télécommunications3) and other complementary information sources (Observatory of Internet Uses, Médiamétrie, etc.). All the data tend to show that though France is not one of the top leaders, it gradually progresses to the position of a European leader. Availability Availability can characterise essentially two different things. Availability can express the fact that anyone willing to connect to the Internet can do it whatever the place of connection. Availability in this sense is progressing very quickly, since 45% of the French population can be considered as Internet users. In 2004, Observatory of Internet Uses, Observatory of Multimedia equipment, and Médiamétrie converge to show that 23.1 million (20.2 million in 2003) French citizens aged 11 and over claimed to have connected to the Internet in the last month. But availability also means that Internet users have a personal connection at home (29.2% of the French population compared to 27.7% in 2003). This, of course, requires them to own a computer (29% in 2004 compared to 27.7% in 2003). Thus, almost anyone having a computer also has Internet access. The type of Internet connection also has changed very quickly. Narrowband Internet users first reached a peak at the end of 2002 and in the beginning of 2003, but are gradually decreasing thereafter. ARCEP/ART, in its annual report in 2004, shows that narrowband subscriptions declined by 15% compared to 2003 data. On the other hand, the number of broadband users has risen at spectacular growth rates until the very last months, involving a sixfold increase in broadband subscriptions in the period 2002–2004. Penetration We now consider more precisely broadband penetration, namely the percentage of households having a broadband connection at home. Though France used to be below the European average in terms of broadband penetration

Broadband in the French Infocommunications System of Innovation

175 7% 6% 5% 4% 3% 2% 1% 0% 2002 Q3

2002 Q4

2003 Q1

2003 Q2

2003 Q3

2003 Q4

2004 Q1

Broadband penetration rate in Europe Broadband penetration rate in France

figure 6.1 Broadband penetration rate in France and Europe, 2002–2004 Source: ARCEP/ART.

rate, demand for broadband services started to rise very sharply in 2004 (see Figure 6.1). The number of broadband Internet subscribers actually grew by more than 120% to over 3.5 million Internet users versus 1.6 million earlier. ADSL (Asymmetric Digital Subscriber Line) has the lion’s share of this market with over 3 million customers at the end of 2003. The numbers of cable subscribers rose by only one-third as much, by around 40%. This sudden public enthusiasm is partly due to a very appreciable fall in access price (a fall of over 30% for a 512 kbps connection, for instance), following a price war initiated by the different Internet service providers in the broadband market and the diversification of offers (higher bit rates for an equivalent price). The emergence of “triple play” offers associating Internet access, TV over ADSL, and voice over ADSL, can also be mentioned as an important element that favoured broadband subscriptions. Capacity/Speed Capacity and speed depend on the Internet applications one can access (e-mail, real-time video, music), but also on the Internet users themselves

jackie krafft

176

(business or residential). On the one hand, the business market is highly segmented. In 2004, it generated revenues estimated at $0.62 billion (PPP $0.897), excluding very high bit rates provided on optical fibre. On the other hand, the residential broadband market generates revenues of around $1.86 billion (PPP $0.897), of which 85% comes from ADSL technology and 10% from cable. In any case, demand for bandwidth has literally exploded since 2003, jumping from 30% to 90% of demands in 2004 (cf. ARCEP/ART annual report 2004), and confirming thus an increasing use of broadband access. Tariffs Broadband tariffs are often unrelated to the temporal length of connection. This is one of the basic differences with narrowband that was based on the classic commuted telephone network, and was thus dependent on conventional modes of pricing. Nevertheless, these fixed-price contracts for unlimited connection also co-exist with other offers that are limited in time (for instance 20 hours connection) and or in volume of exchanged data (5 Gb of downloaded data). These offers are dedicated to bring the former customer base of narrowband to broadband, and to manage the technological shift smoothly. The price for narrowband has thus declined from $50 in 1999 (PPP $0.928) to $18 in 2004 (PPP $0.897). In the meantime, 1024 kbps broadband access is offered at $40 (PPP $0.897) compared to $68 (PPP $0.910) in 2003. Variety of Suppliers: Shared Access or Unbundling? The rapid development of broadband Internet in 2003 is also due to the rise of shared access in France, which is also present in the Netherlands and the Scandinavian countries—some of the most advanced European countries in broadband. Alternatively, full unbundling is higher in Germany, Italy, Denmark, and Finland than in France (see Figure 6.2). competition We know that competition, and especially competition between technologies, is a key element shaping industrial dynamics. Important contributions in the domain have extensively referred to technological innovation (product and process innovation) as a key driver in the evolution of industries (Klepper, 1997). But, in a larger context, we also have to acknowledge that competition

Shared access Austria Belgium Denmark Finland France Germany Greece Ireland Nothern Ireland Luxembourg Netherlands Portugal England Spain United Kingdom 0

50

100

150

200

250

300

Full unbundling Austria Belgium Denmark Finland France Germany Greece Ireland Nothern Ireland Luxembourg Netherlands Portugal England Spain United Kingdom 0

figure 6.2

200

400

600

800

1,000

1,200

1,400

1,600

Shared access and full unbundling in Europe, 2004, in millions of lines

Source: ARCEP/ART.

jackie krafft

178

primarily takes place among heterogeneous firms using a large range of actions aimed at ensuring the realisation of the choices of a given firm, while restraining at the same time the sphere of actions of its rivals. These actions can be related to technology, but can also concern productive resources (human, physical, financial), as well as the implementation of new forms of organisation in which customers, suppliers, partners, and even competitors may be involved (Krafft, 2000; Fransman, 2001). In the present analysis of the emergence and growth of broadband in France, it will be argued that the disruptive competition that took place between technologies, and more importantly between companies supporting these technologies, generated increasing financial difficulties and turbulences, as well as the decline of the initial developers of broadband, the first (cable) movers. This situation also drove the unequal development of broadband, both geographically and technologically, that stimulated the emergence of a new actor in the process of competition, namely the intervention of local public authorities. Competition among Technologies Broadband Internet can be provided via different networks, and one of the motivations to be expected from the emergence of this innovation was to increase competition between technologies. More precisely, what was intended was a decisive contest of the incumbent’s technology (DSL networks) with other, alternative technologies supported by new entrants (such as cable, one of the initial technologies providing broadband; or more advanced technologies such as satellite, optical fibres, WiFi , WiMax, Radio local loop, 3G). Fierce competition thus occurred among these technologies. The outcome of this fierce competition is that in France, as in half of the European countries, technologies other than DSL still account for a very low share of the market. ARCEP/ART’s annual report shows that in 2004 only five countries (among which are the UK, the Netherlands, and Austria) out of fifteen rely on alternative technologies (other than DSL) as a dominant source for broadband access. This fact can be interpreted as the definitive domination of DSL, and that at the end of the day competition provided by other technologies is not so great in France (which uses DSL at 85%, and alternative technologies at 15%), as well as in most of the European countries. This can also suggest that the process of disruptive competition undertaken so far urgently has to be inflected toward more coordination to sustain the development of alternative, emergent technologies. The introduction of a fierce

Broadband in the French Infocommunications System of Innovation

competition between technologies led to the following contradiction. The incumbent’s technology was intended to be contested, but this technology finally dominates all the other competing technologies. This de facto dominant technology is neither the optimal one (such as the large spectrum of advanced technologies) nor the initial one (such as cable). Today, the outcome of this competition in technologies can thus be summarised by the following features. ADSL Develops Faster

DSL technologies, with 3,043,800 broadband subscribers, have developed faster, presumably because (1) DSL was the main technology of the incumbent, the company that has developed the larger broadband network; (2) major competitors essentially tried to copy the incumbent’s strategy in the early phases of the introduction of competition on broadband in the phase of disruptive competition; (3) DSL was also the main technology of the French equipment supplier, Alcatel, and more generally of all the traditional equipment suppliers. For these reasons, DSL technologies finally won the competition in technologies. Thus, while the initial motivation of regulation was to increase the variety in technologies, there has been quite rapidly a convergence toward the domination of this technology. Cable Started Quickly but Stagnates Today

Cable, with 393,800 broadband subscribers, is the second means of large-scale broadband access. In France, cable networks have always been considered as a major alternative infrastructure for the supply of telecoms services, and this is why, since 1998, they were largely developed all over the country. Cable technology has three major handicaps that have slowed down its gradual progression: (1) cable operators are still carrying unpaid debt from the cable plan initiated in 1982 and need to invest heavily to develop new services and upgrade old networks; (2) allocation of network operation licences led to a fragmentation of operating areas throughout the country, preventing economies of scale; (3) regulatory constraints, such as a maximum coverage fixed at 8 million customers have further exacerbated the difficulties of the sector. Today, cable clearly stagnates in France compared to other countries. The central position of FT in cable activities, combined with the fact that it faces disruptive competition on ADSL, contributes to explain why cable did not develop to a sufficient level to compete with other infrastructures owned by the historical operator.

179

jackie krafft

180

UMTS Was Significantly Delayed

Back in January 1998, France was insisting on the vital importance of UMTS (Universal Mobile Telecommunications System) for promoting the information society and the telecoms industry. By pressing for candidate selection using the “beauty contest” method with a detailed technological and service agenda instead of the “auction” system that was implemented in other countries, France encouraged the emergence of different operators to apply for a licence. Hence, at the end of 2002, three operators competed in 3G: France Telecom-Orange, Cegetel-SFR, and Bouygues Telecom. This four-year process was marked with profound hesitations, and uncertainties, that still delay the development of the technology. Very early on, in fact in 2001, France officially declared that the timetables for launch of UMTS services were overoptimistic. GPRS (General Packet Radio Service) was seen as the essential link in the transition between 2G and 3G, since it allowed users, business or private, to be acquainted with new services combining mobility and data transmission. But, in the meantime, GPRS offered low speeds and relatively poor services, and the relatively slow takeoff that was finally observed for this technology did not favour the subsequent success of 3G. WiFi and WWL Are Lagging Behind

Other technologies, such as WiFi, WWL, satellite, and FTTH are lagging behind. First, the decision to liberalise the rolling-out of some of these technologies was rather late. It was only in 2002 that France took steps in the development of wireless local area (WLAN), commonly called WiFi. France only confirmed in 2003 the objective to complete gradually the national coverage, so as to allow wider use in both urban and rural environments. Prior to that decision, 2.4 GHz band were exclusively used for military/defence purposes. Second, difficult economic conditions in the telecoms sector in 2002, the lack of maturity in the 3.5 GHz band, and the cost of equipment in the 26 GHz band have all impeded WLL development. WLL, however, is complementary to WiFi, notably for connection of access point, since it can be used to transport high speed data and as a means of Internet access for small and medium-sized companies. Third, the economic model that will give private users access to a broadband offering under satisfactory tariff conditions on the basis of satellite is still to be developed. Finally, other technologies, such as FTTH (Fibre-to-the-Home) and powerline carrier systems, play only a marginal role in terms of broadband subscriptions.

Broadband in the French Infocommunications System of Innovation

Competition between Companies In the distribution between the incumbent and the new entrants’ market shares, France occupies a medium range in the classification of European countries. More precisely, France Telecom is dominant (with 57%), but new entrants can be considered as active players (with 43%). Compared to the UK, Sweden, or Belgium, in which the incumbent’s market share ranges from 30% to 50%, penetration of new entrants in France is thus much lower. In the meantime, France performs better than Germany or Italy where the dominance of the incumbent is in general superior to 60% in market shares. However, neither the French incumbent nor the entrants can be considered as key players in the global competition of the broadband Internet. How can we explain this paradox: liberalisation and competition is progressing within the national frontiers but does not foster the position of the French companies in international competition? Various reasons can be exposed, some related to the disruptive competition that prevailed amongst players in ADSL, others related to a certain kind of distortion of competition in cable due to the participation of the incumbent into the capital of its competitors. Disruptive Competition Involved Financial Difficulties

Today, the wholesale market is composed of five essential operators: France Telecom (57%), Iliad-Free (17%), Telecom Italia-Alice (8%), LDCom (5%), and Cegetel (5%). These five essential operators are thus facilities-based and provide broadband access. At the retail level, we have eight essential companies: France Telecom-Wanadoo (52%), Free (17%), AOL France (8%), Tiscali (5%), Noos (5%), T-online/Club-Internet (4%), Neuf Telecom (2%), and Tele 2 (2%). Some of these companies are facilities-based (the ones that also operate at the wholesale level), but their essential activity is to resell broadband access and services. The essential strategy of these companies was driven by the introduction of a disruptive competition. All the companies (the first mover was Iliad-Free) initially followed a pricing strategy where price does not cover short-run marginal costs. More recently, some of the companies (Free, again) proposed an advanced offer for broadband access and services (called the Free Box) combining broadband Internet access, Voice on IP, and access to information content (television and video on demand). At the moment, the result of this disruptive competition, whatever the form that it has taken, is that, except Free, most of the companies are running into financial difficulties that highly question their long-term viability.

181

jackie krafft

182

140 120 100 80 60 40 20 0 ⫺20 1997

1998

1999 Number of firms

2000

2001

2002

Net entry

figure 6.3 Number of firms and net entry in the French telecommunications industry, 1997–2002

Disruptive Competition Generated Increasing Turbulence

During the first two years of liberalisation, entry was an easy, common phenomenon in the French telecommunications industry. However, the number of firms started to stagnate in 2000, and even declined in 2001 (see Figure 6.3). In 2002, exits were largely superior to entries, leading to a negative net entry. This period 2000 –2002 corresponds to the development of broadband activities in France, and especially to the introduction of competition in the local loop access for voice, data, and the Internet. The procedure of allocation of licences for radio local loop operators that occurred in 2000 clearly attracted a lot of new companies (many of them entered in 1999). Most of these companies (more than 60%) exited only a few months after entry, either because they did not obtain the licence or because they went into important financial troubles with the financial crash.

Broadband in the French Infocommunications System of Innovation

The process of entry that occurred on the local loop access was intended to develop an alternative technology (namely, radio local loop) that could be considered at that time as one of the more promising alternative technologies, and also as one possibility to limit the monopoly exerted by the incumbent on this activity. However, the largely uncoordinated process of entry that actually took place—combined with the fact that this process of entry coincided with the financial crash—clearly resulted in an increasing turbulence at the firm and industry level. Distortion of Competition and the Decline of the First (Cable) Movers

At the end of the 1990s, cable was considered in most advanced countries as one of the main architectures that could compete on the local loop with the historical operator. In this perspective, large mergers and acquisitions (M&As) occurred in most countries. For instance, in Europe, France Telecom invested in the UK cable company ntl and in Casema in the Netherlands. In the U.S., AT&T acquired a couple of cable operators to restore its own local loop, which was sold in 1984. These costly acquisitions occurred during the boom period of the telecoms industry and, when the bubble deflated in 2000, were clearly perceived as inefficient strategies. AT&T invested more than $100 billion in these M&As and, with the weight of debt, soon had to sell the cable assets to its competitor Comcast. France Telecom also had to get out of Casema’s capital for the same reasons. Often, countries in which Internet penetration was the largest are also countries where competition between ADSL and cable was stimulated. For instance, in the U.S., the UK, Belgium, and most of the north European countries (Sweden, Denmark, the Netherlands), cable operators have kept large market shares, at least comparable, and to some extent superior to ADSL companies. France appears as a paradoxical case. Though cable was one of the first technologies supporting broadband Internet access, the use of this technology is still very limited today. In 1998, 13,464 Internet subscribers connected to their Internet through cable operators. Since then, the number of subscribers has grown regularly but at a much lower rate compared to ADSL (+40% for cable in 2003, against +100% for ADSL). Alternatively, in the Netherlands, penetration rate of cable represents 22% of Internet connections, and ADSL 20%. In Sweden, 10% of the connections occur via cable, and 33% via ADSL. Global coverage in these countries ranges from 70% to 95%, while it is only 25% in France. One argument explaining why cable is lagging behind ADSL technologies is certainly the lack of technical maturity. Many experts and actors of the

183

jackie krafft

184

industry say that this could change in the near future with the development of new offers. End users often kept in mind a separation with, on the one hand, telecommunications services (telephone and Internet) offered by telecoms operators and, on the other hand, value added television services offered by cable operators. New offers, such as “triple play,” could favour the convergence between telecommunications and television. Another argument is that the French incumbent operator, France Telecom, has a large share in the capital of competitors Noos, NC Numéricable, and UPC France. France Telecom has a dominant position in cable in France; it owns 100% of France Telecom Cable, 70% of the essential facilities used by NC Numéricable, and 27% of Noos. This means that FT owns either directly or indirectly 40% of the capital of cable operators. This reveals that this certain type of distortion of competition induced very low incentives to develop the cable technology. Innovation, Disruptive Competition, and Distorted Competition

Competition among players in the broadband industry either seems to be overamplified (in the case ADSL) or underdeveloped (in the case of cable). In any case, this involves inadequate incentives for the development of innovation in the industry. First, disruptive competition tends to expose companies to an excessive financial risk, especially in the period of the deflated bubble. Second, distorted competition exists in the domain of alternative technologies, leading to a very low probability to stimulate the development of these technologies. The assessment of the current situation in the broadband industry shows that “more competition” in terms of an increasing number of companies and technologies, often prone in conventional economic models, is not necessarily beneficial to the society in the long run, and especially to the emergence of innovation. Moreover, it seems that leaving competition in the industry to work out for itself is not a sensible strategy. Rather, the type of interaction that exists among companies has to be checked and eventually corrected by regulation or competition authorities. Moreover, the effective development of a technological diversity or variety has to be guaranteed. Emergence of a Potential New Actor: Local Public Authorities Activities related to infocommunications have exploded in the last few years in France and have generated a multiplicity of new companies. The geographic location of these companies, however, remains highly unequally dis-

Broadband in the French Infocommunications System of Innovation

185

table 6.1 Concentration of Infocommunications Firms and Employment in France, 2002 Regions/ Areas/ Clusters

France Ile de France PACA Rhône Alpes Aquitaine Bretagne

Infocommunications Firms

Total Firms

45,694 8,870 4,500 4,000 2,018 719

2,870,888 662,674 283,106 292,885 148,959 119,497

Firm InfocommunSpecialiications sation Employment

1.5 1.3 1.5 1.3 1.3 0.6

876,325 255,502 32,000 60,000 21,000 22,483

Total Employment

Employment Specialisation

15,443,000 5,042,724 1,576,085 2,263,018 1,105,435 1,112,583

5.6 5.7 2.0 2.6 1.9 2.0

tributed and reveals the irregular development of broadband. There is large dominance by Ile de France (Paris) where most of the broadband activities (as well as other related activities such as hardware and software, electronics, test and control, etc.) are concentrated. Then there is a group of medium-sized areas, such as Provence Alpes Côte d’Azur (PACA), and Rhône Alpes; and a group of smaller areas, such as Aquitaine and Bretagne (see Table 6.1). This disparity thus calls for a strong political intervention, such as the recent measures on the “digital divide” promoted by the government and the regulator. The challenge is to create conditions encouraging geographical expansion of broadband access offerings based on various types of technologies, and at affordable prices to allow all users, even those living in isolated rural regions, to benefit from the competitive dynamism observed in the more densely populated urban areas (Krafft, 2004). Preliminary reflections to consider local public authorities as potential new actors in the broadband telecommunications industry emerged in 2002, at a time when market players were having increasing difficulty with financial constraints. Telecommunications services are crucial for enhancing the attractiveness of cities and regions. Local authorities are now fully aware of the importance of providing digital technologies to citizens, particularly as a way of fostering economic growth and development. Given this structural trend but also the specific economic difficulties facing broadband companies today that forced them to select more cautiously their investments, many local authorities have been motivated to test their capacity to establish and operate telecommunications networks. Some already have taken initiatives, notably in the field of broadband access using alternative technologies such as cable, WiFi, WiMax, optical fibre, or FTTH. In a sense, to avoid the unequal development of digital technology and services throughout the country, public intervention in both offers and services

jackie krafft

186

is becoming a necessity in the ongoing—sometimes chaotic—process of competition. With the government programme against the “digital divide,”4 the legal framework allows, since 2003, that local authorities can assist private businesses in developing their activities by taking financial responsibility for most onerous network construction costs. This new legislation states that local authorities can create, but also operate, telecommunications networks. Further developments would allow local authorities to provide telecommunications services where private initiatives are inadequate to satisfy the requirements of business or private users. Of course, such a new player could not come out of the sole market forces that are traditionally intended to shape industrial dynamics. More adequate, important decisions at the regulation and institutional level were necessary, because industrial dynamics of the French broadband Internet now has to be intrinsically related to the FISI. regulation Competition and regulation are distinct, yet connected, fields of economic policy. Competition generally applies in industries the functioning of which tends to be compatible with “normal” (i.e., fair and equal) rivalry. Regulation operates in sectors where normal rivalry cannot be guaranteed, for instance, in sectors were high fixed/sunk costs dominate, or in sectors that were recently liberalised. Competition and regulation also differ in terms of procedures. While competition authorities are limited to the compatibility of firms’ activities in respect with the law, regulation authorities presumably have a greater scope of intervention. For instance, they can control and impose certain prices, investment strategies, and productive choices to firms. Moreover, most of the time they act ex ante, and not ex post as in competition policy; and their action is not occasional but rather long term. Finally, while competition is often based on oligopoly theory, regulation is mostly inspired by agency models where the principal is the regulation authority and the agent the incumbent company. In this perspective, one key difficulty is to reduce the gap between the interests of the regulator (which may be resolute or weak), and the interests of the incumbent (which may be pro-active or inert, see Motta, 2004; Hausman, 1997). In various countries, the assessment today is that regulation was often based on wrong assumptions about the future development of the telecoms and broadband activities. Regulation thus either imposed unrealistic purposes to

Broadband in the French Infocommunications System of Innovation

the incumbent leading to increasing difficulties (uncoordinated strategies, high level of debt, massive process of mergers and acquisition, etc.), or left the incumbent in a comfortable situation with very few requirements (Fransman, 2002). In France, the regulator was increasingly inspired by the broadband success observed in countries such as Japan and Korea. In that perspective, the French regulator (ARCEP/ART) took a first step toward full facilities-based competition (2000 –2002), encouraging the development of alternative, competing networks. In a second step (2002–present), the regulator introduced aggressive policies on unbundling and insisted that the incumbent had to charge very low prices to competitors wanting to use its local access network. Full Facilities-based Competition The first step in the regulator’s plan was to favour a full facilities-based competition, by the entry of a large number of new firms (around 50) that were supposed to develop their own network. For most of these companies, the usual strategy was to obtain a licence to operate at the local or regional level, for instance in large urban centres, that could be extended later to a national licence. Many new U.S. or European operators were among these new entering companies, such as 21st Century, 360 Networks, Winstar, Titan, Iaxis, and Dynegy, and thus applied for a local or experimental licence. This wave of entry was also composed of a number of cable operators that could also provide alternative networks and were thus considered as credible competitors. From the regulator’s side, the large entry process was considered as a positive thing, since entry is often used as a vehicle for introducing new innovations, while the absence of entry is generally associated with technological stagnation. Also, entry is considered as a strong mechanism for getting prices right in markets, as well as for getting product and process specifications right. From the companies’ side, entry in the French broadband industry was a logical step in their expansion in Europe, especially because at the time of entry (2001–2002), broadband was still underdeveloped in France and could thus be characterised by high potential profit opportunities in the near future. This first step in the regulator’s strategy did not really provide the expected results, however. First, the demand for broadband did not really take off at this time. Until the end of 2002, demand for broadband still remained stagnant and operators that entered two or three years earlier could not develop adequately their business strategy. Second, the financial crash that occurred in

187

jackie krafft

188

2000 clearly questioned the viability of these companies that relied on stock markets or bank loans to finance their infrastructure project. Aggressive Policy on Unbundling The turning point for broadband access came thus in the second half of 2002, with acceleration in growth of the number of residential subscribers. In one year, the number of subscribers multiplied by 2.5, making the French market the fastest growing in Europe, with 1.7 million broadband Internet users (ADSL + cable). At the end of 2002, broadband access represented 18% of all Internet access subscriptions. In 2004 (see ARCEP/ART, 2004), the total number of broadband subscribers was 3,346,800: 393,800 for cable and 3,043,800 for ADSL. Three different options (termed “option 1,” “option 3,” and “option 5”5) were defined by ARCEP/ART concerning ADSL unbundling, leaving different roles in the access and collection of broadband traffic to France Telecom and alternative operators. These options reflect the increasing introduction of competition in the broadband market, and correspond to the aggressive policy on unbundling developed by ARCEP/ART. Option 1 corresponds to unbundled local loop, where all the access and collection of DSL traffic is operated by competitors of FT. This option concerns, in 2004, 275,600 subscriptions. Option 3 corresponds to “shared access,” and characterises the situation where France Telecom is the essential provider of the local loop, but the access and collection of DSL traffic at the national level is left to alternative operators. Here 100,000 subscribers were concerned in 2004. Option 5 corresponds to “wholesale,” which is a situation of monopoly by France Telecom on the local loop and also at the national level of the access and collection of DSL traffic. This represents 1,738,000 subscribers for France Telecom-Wanadoo and 928,600 for other ISPs. Decisions by ARCEP/ART in April and July 2002 concerning copper pair unbundling (option 1), collection and transport (option 3), and France Telecom resale offers (option 5) had a helpful impact. These decisions helped to create favourable conditions for the emergence of a more diversified offering and a decline in prices, leading a larger number of customers to use the Internet. ARCEP/ART focused particularly on ADSL, in both retail and wholesale markets, for two reasons. First, it was the dominant broadband access technology, and second, strong market growth on a competitive basis implied that alternative operators could use the access network (copper pairs) owned by the incumbent operator, France Telecom, through the local loop unbundling. The main objective was to improve operating conditions for unbundling

Broadband in the French Infocommunications System of Innovation

and service quality in densely populated areas and encourage access migration from option 5 to option 1, and from option 3 to option 1. Another objective was to expand the geographical coverage of ADSL to serve the greatest number of users in large towns and in less densely populated areas. The last objective was to give new impetus to resale offers by improving economic terms of option 3, particularly paying attention to the risk of a price squeeze with option 5. ARCEP/ART’s actions in this field were reinforced by the European Commission’s New Regulatory Framework created by transposition of European directives. Wholesale supply of unbundled access (including shared access) to loops and sub-loops on copper pairs is in fact one of the relevant markets predefined by the European Commission in which ARCEP/ART can impose interconnection and access obligations on operators with significant market power. legal framework and judicial procedures The new legal framework created by the European directives defines the general objectives and regulatory principles that should guide the action of the member states and the national regulatory authorities (NRAs). Each NISI (National Infocommunication System of Innovation) in Europe had thus to adapt to this common legal framework, and to find its own way to respect European requirements. For the regulator, this essentially led to important changes in the sphere of action, as well as major adjustments both internally and externally. Europe’s New Regulatory Framework: From Regulation to Competition The goal of the new legal framework created by the European directives is to foster and consolidate competition in the electronic communications sector markets. To perform this task, regulation must adapt to changes in each market, increasingly using the tools and concepts of competition law. Hence, while maintaining sector-specific regulation, this new framework lays the groundwork for the transition toward gradual application of competition law as the sole reference governing the sector. Once a market has become competitive, a priori sector regulation has to be replaced by a posteriori competition law. This means, first, that regulation has to apply in “relevant markets.”6 Regulators have then to designate the operator(s) exercising significant influence in a market when it enjoys a position equivalent to dominance within the meaning of competition law. If the answer is yes, then the operator(s) is considered as having a significant market power. In this case the regulator has to decide whether to maintain, eliminate, or modify the obligations of these operators so

189

jackie krafft

190

as to remedy failures of competition in the market. Regulation is a priori confined to sectors where competition is not yet sufficiently well established and concerns primarily the wholesale markets. Retail markets are only subject to regulation at a later stage if regulation of the wholesale markets upstream is not sufficient to ensure competition in the downstream retail markets. Second, the objective of the “telecom package” is that regulation has to take technological convergence into consideration and examines under unique notification procedure the establishment of public network and supply of a public electronic communications service regardless of the technical platform used (cable, microwave, wireline). The other aspect regulation has to adapt is the increasing concentration in the sector and the emergence of players with a European or international dimension. This naturally leads the regulator to re-orientate its scope of activities. The attendant risk of this movement is the creation of virtual oligopolies that the regulator has now to tackle seriously. Likewise, technological changes, and notably the convergence between fixed, mobile telephony, the Internet, and the audiovisual sector are opening up new market segments and bringing about a new distribution of network and service functions. The internationalisation of service providers, and the fact that many of them are outside national boundaries, calls for a European or even international approach. The regulator cannot confine its analysis to the domestic market without running the risk of leaving out a whole section of the market. Nevertheless, national regulatory authorities must remain attentive to the specific features of their own domestic markets and seek to uphold competition while continuing to encourage investments in new networks. This apparent contradiction actually illustrates the importance of the harmonisation/subsidiarity tandem, which is the base on which regulation will have to be set up for the next few years. Internal and External Adjustments for the NRA Recent changes at ARCEP/ART essentially concerned the internal organisation, which is now supported by an increased budget and a new team. Especially, the new direction implemented a specific vision of regulation, increasingly influenced by recent competition and economic developments. These internal changes were also developed in a context of increasing embeddedness within international NRA’s networks. A New Organisation

ARCEP/ART’s annual budget is currently set by the Minister in charge of the Economy, Finance, and Industry. In 2004, the budget increased by 10%

Broadband in the French Infocommunications System of Innovation

($22 million, PPP $0.897 compared to $20.77 million, PPP $0.910 in 2003). The breakdown of the budget is 56% for fixed expenses and 44% for variable expenses. The staff increased in the period 2003–2004 from 139 to 146 and kept a relatively low average age (41). In September 2003, ARCEP/ART considered possible changes to its organisational structure with a view to improving efficiency and adapting the organisation to the new regulatory framework resulting from the transposition of the European directives into national law. The objective was to strengthen the economic skills and ability to adapt to the changing situation on the one hand, and to structure its activity around the key process of market analysis on the other. The new organisation was put in place in February 2004. Together with these internal changes, professional training programmes and conference participation increased by 35% in 2003 for ARCEP/ART staff. Finally, by a decree of the President of the Republic, Paul Champsaur has recently been appointed chairman of the Regulation authority. He replaced Jean-Michel Hubert, whose six-year mandate ended on 3 January 2003. As a new director, Paul Champsaur inspired a new policy at ARCEP/ART, largely influenced by recent economic developments in the domain of regulation and competition economics. A New Vision of Regulation

New developments in competition (Laffont and Tirole, 2000) tend to sustain the belief that competition in the telecommunications market has to favour the development of long-distance backbone frameworks to the customer. The simple explanation for this is the existence of a high degree of service sharing and data flows in the network backbone and high tariffs for long-distance calls. Hence, investment costs in network backbone can be amortised over many services and in high margin segments. Two major consequences on the role of the NRA were involved from this proposition, and implemented at ARCEP/ ART. First, ARCEP/ART has to guarantee that the new entrant operator must be able to buy the services they cannot provide themselves from the incumbent operator. Second, ARCEP/ART has to check that the incumbent operator’s retail tariffs must be gradually adjusted so as to create adequate economic space for the development of alternative operators. In practice (see ARCEP/ ART 2004 annual report), this means that the control of wholesale tariffs is based on the principle of alignment of tariffs with costs (cost-oriented tariffs) and is carried out by annual approval of the interconnection catalogue. This control rests on transparency of the price charged by the incumbent, as well as multi-annual agreements between the incumbent and the regulator. Control in prices also operates at the retail level and is based on prevention of anti-competitive practises, particularly those of dominant position. ARCEP/

191

jackie krafft

192

ART thus systematically prevents predatory pricing, price discrimination, or pre-emption of a new market. The other domain that influenced a new vision of regulation at ARCEP/ART concerns the development of differentiation in services, or the development of proprietary access networks that can be interpreted as purely competitive behaviours, but in the meantime can also turn into segmentation, exclusivity, and foreclosure. ARCEP/ART thus developed new procedures, in conjunction with competition authorities, to systematically detect the emergence of these anti-competitive behaviours. Increasing Embeddedness Within International NRAs’ Networks

External exchanges are also fostered by an increasing embeddedness within networks composed of national regulatory authorities. Within these networks, NRAs can confront their past experience of liberalisation, and also enlarge their vision of regulation by a better knowledge of companies, technologies, and associated services. ARCEP/ART is involved in two major networks: COCOM (Electronic Communications Committee), and IRG/ ERG (Independent Regulators Group/ European Regulators Group). These networks regroup a large number of NRAs and are involved in three essential tasks: standardisation, in conjunction with ETSI (European Telecommunications Standardization Institute); access to electronic communications for disabled people; and definition of relevant markets and competitive behaviours on these markets. conclusion In this chapter, we have described various factors that, to us, contribute to explain the current performance of France in broadband. These factors explicitly include industrial dynamics into the FISI. Alternatively, one could logically consider that the absence of these factors should also explain the lower/higher performance of other countries. In what follows, we advance some (non-exhaustive) arguments on this issue. Among the factors that have tended to stimulate performances in the development and growth of broadband in France is certainly the early introduction of competition, both at the level of technologies and companies. In the meantime, however, the fierce and uncoordinated process of competition that was introduced generated the dominance of the incumbent technology, and the underdevelopment of alternative technologies. Thus, in the French case, competition favoured innovation, but not necessarily varied innovation, and this may limit opportunities for future growth.

Broadband in the French Infocommunications System of Innovation

In this perspective, the general tendency to develop a new vision of regulation, highly connected to new recommendations in the domain of competition policy, may also lead to the assimilation of the broadband industry, which is still in its development phase, to a mature industry. Over the recent years, this assimilation has not been yet completed in France, since ARCEP/ART was largely inspired by the broadband success observed in countries such as Japan and Korea. The development of alternative, competing networks and technologies was encouraged, though sometimes with unsuccessful results, together with the implementation of drastic obligations from the incumbent concerning the access to local network. However, the compatibility and sustainability of this regulation policy in France is still largely dependent on the European new regulatory framework, and on how the different NRAs representing different countries with different stages of development in broadband may themselves integrate the legal framework and judicial procedures. In view of recent changes implemented in France, the recurrent dilemma of promoting competition or protecting the (national or European) competitors still apparently applies. This also shows that in each country industrial dynamics in the domain of broadband shapes and is shaped by the NISI, yet the European infocommunications system of innovation (EISI), which is still under construction, will definitely transform the growth of broadband industry of member states. The way in which this transformation will operate is, however, complex and still difficult to anticipate.

notes 1. See Chapter 1. 2. More technically, narrowband Internet regroups 0.5 Mbps for a traffic volume of 1 million minutes, while broadband Internet requires 23 Mbps to meet the requirements of 1,000 subscribers. 3. In May 2005, the initial domain of activity of ART (the regulation of the telecommunications sector) was enlarged to include media and postal mail sectors. The name of the regulator was changed to ARCEP at this occasion. In this chapter, we will refer to the double acronym “ARCEP/ART” to characterise the French regulator. 4. In French, the “fracture numérique.” 5. Note that there are no options 2 and 4. 6. This includes demand side and supply side substitutability for the definition of market in terms of product and services, as well as for the geographical definition of the market.

193

jackie krafft

194

references Antonelli, C. 2001. The microdynamics of technological systems. Oxford: Oxford University Press. ———. 2003. The economics of innovation, new technologies and structural change. London: Routledge. ART. Annual Reports 2000 –2004. Fransman, M. 1994. Information, knowledge, vision and theories of the firm. Industrial and Corporate Change 3, 1: 1– 45. ———. 2001. Analysing the evolution of industry: The relevance of the telecommunications industry. Economics of Innovation and New Technology 10, 2–3: 109– 40. ———. 2002. Telecoms in the Internet age: From boom to bust to . . . ? Oxford: Oxford University Press. ———. 2004. The telecoms boom and bust 1996 –2003 and the role of financial markets. Journal of Evolutionary Economics 14, 4: 369– 406. Hausman, J. 1997. Valuing the effect of regulation on new services in telecommunications. Brooking Papers on Economic Activity, Special Issue on Microeconomics, 1–54. Klepper, S. 1997. Industry life cycles. Industrial and Corporate Change 6, 1: 379– 460. Krafft, J., ed. 2000. The process of competition. Cheltenham: Edward Elgar. Krafft, J. 2003. Vertical structure of the industry and competition: An analysis of the evolution of the infocommunications industry. Telecommunications Policy 27: 625– 49. ———. 2004. Entry, exit and knowledge: Evidence from a cluster in the infocommunications industry. Research Policy 33, 10: 1687–706. Laffont, J. J., and J. Tirole. 2000. Competition in telecommunications. Cambridge: Cambridge University Press. Médiamétrie. Reports 2000 –2004. Metcalfe, S. 1995. Technology systems and technology policy in an evolutionary framework. Cambridge Journal of Economics 19, 1: 25– 46. Motta, M. 2004. Competition policy: Theory and practice. Cambridge, Mass.: Cambridge University Press. Nelson, R. 2004. The co-evolution of technology, industrial structure, and supporting institutions. Industrial and Corporate Change 3, 1: 47– 63. Observatory of Internet Uses. Reports 2000 –2004. Observatory of Multimedia Equipment. Reports 2000 –2004. Pavitt, K. 2001. Public policies to support basic research: What can the rest of the world learn from U.S. theory and practice? (and what they should not learn). Industrial and Corporate Change 10, 3: 761–79. Saviotti, P. 1996. Technological evolution, variety and the economy. Cheltenham: Edward Elgar. ———. 2001. Variety, growth and demand. Journal of Evolutionary Economics 11, 1: 119– 42. Witt, U. 2003. The evolving economy: Essays on the evolutionary approach to economics. Cheltenham: Edward Elgar.

7 Development of the Broadband Market in Germany Franz Büllingen

although the broadband market is still in its early stages, in all industrialised countries fast Internet access is increasingly becoming an emerging reality for enterprises, public institutions, and private households. Looking at the supply side, nearly 100% of business customers and 80% to 90% of private customers potentially can achieve broadband Internet access in Europe. While in Germany nearly 80% of enterprises are already connected to broadband Internet, the penetration rate for private households was much lower at only 16% at the end of 2004. Because prices for broadband Internet access are quite low in Germany compared to many other industrialised countries, there must be other reasons for this slow uptake. It is an astonishing fact that, though cable networks have a high penetration rate, nearly 95% of all broadband subscribers are connected to DSL. In addition, although the DSL market is the main driving force for growth within the fixed telecommunications network, Germany fell back in 2004 in comparison to other European countries. Because a high impact of broadband Internet penetration is assumed with regard to general economic growth, the increase in productivity of certain branches, the general educational level, and the creation of new services, broad political discussion has focused on how to boost the overall penetration rate of broadband Internet. While in other countries like South Korea or Japan concrete measures in regard to industrial policies have been taken to promote broadband Internet, in Germany the political strategy mainly comprises information and discussion initiatives. In 2002 a network of experts from industry, science, and the German government called the “D21 Initiative” was established. Therefore, it is meaningful to analyse the state of both (modal)

franz büllingen

196

competition within the DSL market and inter-modal competition between the different distribution platforms such as DSL, cable, PWLAN, FTTH, WiMax, and Powerline, and how they deal with the regulatory and legal framework. structures of the broadband market The Development of the DSL Market Infrastructure and Availability

In Germany, as in most European countries, the development of broadband Internet access is based on DSL technology. The different types of DSL technologies are summarised by the term xDSL and comprise, for example, ADSL, SDSL, VDSL, and HDSL. Levels of access are differentiated by the bandwidth provided to customers and by their division into down- and upstream. The spectrum of services offered is based on the quality of the existing telecommunications network, which is regarded to be of high quality in Germany due to large investments in the 1990s and the short range of technical transmission on average. This means that there are no real technical or infrastructural obstacles in most of the local area networks if suppliers want to migrate their customers to higher bandwidth. Though there are no technical problems in providing broadband Internet access to all households, it can be observed that network suppliers concentrate their investment activities on areas with highly concentrated population due to economies of density. According to the German incumbent, Deutsche Telekom AG (DTAG), nearly 95% of the 2.9 million enterprises and about 90% of the 38.7 million households in Germany can generally be provided with DSL access at economically reasonable costs. In the mid-term (until 2008) the incumbent is aiming to connect 95% of all households by DSL. While in urban centres DSL coverage has reached almost 100% and coverage in suburban areas is nearly 90%, rural areas lag far behind these figures. Experts assume that the availability of DSL, which currently reaches about 50%, will not exceed 60% in the long run. Enterprises and private households in these areas that want to be connected to fast Internet only have the option to subscribe to the generally achievable, but much more expensive, alternative “Internet via satellite.” Structure of Suppliers

The DSL market, which has emerged since 2000, has been and is still strongly dominated by DTAG (see Figure 7.1). In 2001 and 2002, its market share was 97% and 94%, respectively. Since then this share has declined due to the in-

Development of the Broadband Market in Germany

197

7 6

Millions

5 4 3 2 1 0

2000

2001 Deutsche Telekom AG

2002

2003

T-DSL-Resale

2004 Competitors

figure 7.1 Development of DSL market shares, 2000 –2004 Source: RegTP, WIK.

tensified marketing activities of both competitive city carriers like Net Cologne, HanseNet, and BerliKomm, and specialised suppliers like QSC AG and ISIS (Arcor). As a result, DTAG’s market share fell to 91% in 2003 and decreased again in 2004 to approximately 88%. These enterprises operate mainly in local markets, often using their own local area access networks and offer, besides DSL access, telephony services to their customers. Some of these city carriers, like for example HanseNet, are quite successful and have achieved a 40% share of the DSL market in the last three years. Despite this minor success, DTAG’s market share is still very high compared to other European markets. The ISP of the incumbent, T-Online, which is the leading European ISP, has an extraordinarily strong market position and has so far successfully defended its first mover advantage. At the end of 2003, only about 50 competitors offered broadband Internet access services in Germany. Facing the market power of the incumbent, some of the successful competitors mainly offer their services to business customers and enterprises. QSC AG, for example, is focusing its business model on broadband access services for enterprises. It offers not only fast Internet access like SDSL but also com-

franz büllingen

198

bines its products with other services like VPN, hosting, and VoIP. According to QSC, it has a 2% to 3% market share of small and medium enterprises (SMEs) and about 1% of the large enterprises market segment. Moreover, about 800 ISPs offer services in the German telecommunications market. Because until recently there was no opportunity for resale, these ISPs were only able to offer fast Internet access services in combination with the T-DSL service of DTAG. Resale means that competitors buy the access from the incumbent and sell it to customers. Their additional value creation is the distribution of this service. At the end of 2003, about 700,000 DSL subscribers bought their service from one of DTAG’s competitors. In comparison to other DSL markets, the supply of DSL and Internet service provision is not single-handed. This means that in many cases a subscriber has to sign a contract with the network provider (usually the incumbent) as well as with the ISP. To avoid this multi-stop buying most users decided to subscribe to DSL from the incumbent. Competitors, therefore, are hopeful that the introduction of bitstream access to the market will give them the opportunity to create new product bundles in order to strengthen their market position. Bitstream products will enable competitors to offer broadband access services defining new products with different bandwidth offers according to the needs of customers, giving them access to the co-location infrastructure of the incumbent and using their own infrastructure as far as available. Products and Prices

Though the supply of a broad variety of products with different bandwidths can be observed within the DSL market, ADSL technology still holds a market share of more than 90%. This means that Germany holds a medium position with regard to generally subscribed bandwidth in comparison to other European countries. Until the end of 2003, the 768 KB downstream service dominated the product spectrum of all service providers. Since April 2004, with the change to the incumbent’s product portfolio, the ADSL offer has been upgraded to 1 Mbps. Furthermore, DTAG offers broadband Internet services of 2 and 3 Mbps for the mass market. This means that three years after the introduction of DSL to the market a broad migration to higher bandwidths can now be observed. In line with the general market trend, competitors have started to offer DSL with higher bandwidths. QSC AG, for example, doubled the upload bandwidth for private customers to 512 KB from June 2004. Toplink-plannet has created a new product with four bundled DSL channels of 3 Mbps in order to

Development of the Broadband Market in Germany

reach transmission rates of 12 Mbps for a monthly fee of €29, which is marketed as “SuperDSL” and is aimed mainly at business customers. The overall migration to higher bandwidths in the DSL market also puts pressure on the cable network providers, which have become much more involved in the market for broadband Internet access services. Kabel Deutschland (KDG), for instance, has offered a new service with a doubled transmission rate of 4 Mbps since June 2004. In the last two years, a great variety of tariff models has emerged in the DSL market, where bandwidth and related prices is the differentiating factor. It is significant that only service providers with their own infrastructure have created their own products and pricing models. The other providers resell the TDSL product of DTAG, only differentiating with regard to flat rates, volume and time-based rates, and DSL-by-call. Services with the highest market penetration are those with flat rates and volume-based Internet access. Generally, it can be observed, that price structures have become very confusing for customers because of the bewildering variety in the particulars of different service offers. The Internet portal onlinekosten.de, for example, provides an overview of more than 262 different tariffs for volume-based services and 234 tariffs for flat rate services. Looking at the price level for broadband Internet access, Germany’s position is in the lower middle field. With regard to the average income, German households spend 1.8% of their monthly income on fast Internet, which is quite low compared to other industrialised countries according to an ITU benchmark (2003). Since DSL services were introduced to the market, prices have remained almost the same, decreasing by only 1.3% between 2002 and 2003. The greatest change in prices was in Internet-by-call tariffs for private customers, which reduced by 79% between February 1999 and December 2003. At the same time prices for business customers decreased by 18% due to more intensive competition in this market segment. Regulation

The lack of competition in the DSL market is one of the most crucial points in the current regulation debate. With the introduction of unbundled local loop (ULL) in 1998 and line sharing the German Regulatory Authority (RegTP) hoped to encourage sustainable competition in the broadband market. At the end of 2003, 1.3 million connections based on ULL had been registered, a large percentage of which consists of DSL services offered by competitors. So ULL has proved to be fairly significant with regard to competition because most DSL connections provided by competitors are based on ULL.

199

franz büllingen

200

Line sharing, however, has played almost no role in competing DSL service offers because its high prices did not allow competitors to create and offer attractive services. In April 2004, following the regulator’s decision, the prices for line sharing had to be reduced significantly by the incumbent. The second focus of the current debate on intensifying competition is that customers who want to subscribe to DSL have to subscribe to the telephony service as well. Discussions are under way with the regulator to try to change this situation by unbundling telephone and DSL access. The third topic under discussion deals with the introduction of bitstream access in order to allow competitors to create and offer broadband Internet (and other) services while using their own telecommunications network elements as much as possible. It seems likely that this matter will be resolved sometime in 2005. Demand Aspects

The penetration dynamics of DSL connection services are not as intensive as in many other European countries. Although nearly 61% of all private households already have Internet access, the demand for higher speed is obviously growing quite slowly. According to several private customer surveys, most households consider prices to be still very high. As highly attractive services like video-on-demand, audio-on-demand, and online gaming services are not available for mass market demand, the willingness to pay for fast Internet is not very pronounced. The Development of the Broadband Cable Market Both the telecommunications and cable network infrastructure were owned in the past by the incumbent DTAG. It was a political decision of the German government in the 1970s to roll out cable networks offering access to TV programmes to the majority of German households. As it became clear two decades later, because the incumbent had to give up its cable network involvement due the liberalisation politics of the European Commission, with the exception of Berlin, any investment to modernise cable networks stopped and a long period of stagnation began. Infrastructure and Availability

In terms of the number of cable households worldwide, Germany has the second largest cable market behind the U.S. With a turnover of over €3.3 billion, Germany’s cable industry generates more revenues than any other cable market in Europe. In 2003, more than 21 million, or approximately 57%, of all

Development of the Broadband Market in Germany

Level 3 Operators

Kabel Deutschland; ish; Kabel BW; iesy 17.5 Mil. HH

EndCustomer

201

Free Housing operators (ANGA u.a.) association 5.5 Mil. HH 6 Mil. HH 6 Mil. HH Level 4 operators

15%

33%

57%

figure 7.2

Free cable network operators (ANGA u.a.) with own level 3

Satellite

Terrestrial antenna

13.5 Mil. HH

2 Mil. HH

37%

5%

3.5 Mil. HH

9%

TV-cable households

Operators and their market share of the broadband cable market

Source: ANGA 2004.

German households with TV (36.5 million) were connected to broadband cable networks. Of these 21 million, 17.5 million were connected directly (5.5 million) or indirectly (12 million) to the dominant regional operators KDG, ish, iesy, and Kabel Baden-Württemberg, which run DTAG’s former cable network (see Figure 7.2). The other 3.5 million households were supplied by competing network operators with independent TV-signal feeds using their own fibre networks or by satellite platforms. This means that 85% of all homes passed are connected to cable television. This high penetration rate reflects not only the great importance of cable networks in Germany but also promises opportunities for operators to offer a competitive full service broadband network. However, it becomes clear that these cable networks, which were built more than 20 years ago, are narrowband, one-way 450 MHz coaxial networks that are still mainly in use for analogue broadcast, a few digital channels, and the distribution of analogue radio signals. Experts estimate that it will take at least €1 billion in order to digitise and modernise these networks to an 862 MHz full service bi-directional network. This 862 MHz network upgrade investment is much more time-consuming than the modular digitisation strategy due to unknown demand and willingness to pay for more programmes and interactive services like broadband Internet access.

franz büllingen

202

Furthermore, cable faces very stiff competition, particularly from satellite reception, leaving little scope for price increases for television. Nearly 13.5 million households receive TV programmes by satellite and more than 2 million already use T-DAB receivers. So households with a twin satellite dish can simultaneously receive 30 analogue and almost 600 digital programmes from the competing Eutelsat system platform without paying any charges. As the transmission of TV programmes is the main revenue source for cable owners, it is clear that making large investments in the existing cable network infrastructure is a risk. Structure of Market Operators

Since the roll-out of cable networks during the 1970s the structure of the German cable market has shown some peculiarities that are still of fundamental importance for its regulatory framework and its development perspectives nowadays. When the cable networks were built it was one of the main tasks of DTAG to manage the so-called network level 3 in order to transmit broadcast signals from their play-out centres to network level 4 operators over public ground. Within private grounds and buildings it was the task of the network level 4 players—mainly SMEs—to distribute these signals to their customers. This dual build-up and maintenance strategy was a result of federal economic policies that aimed to provide an additional economic basis for many small electro-technical firms. For this reason, there still exist today between 4,000 and 5,000 enterprises, which in some cases provide broadcast signals to some hundreds of households only. Because they still have this relationship with more than one-third of all cable network end-customers, a structural conflict has arisen between level 3 and level 4 cable network operators. As it became clear in the mid-1990s that DTAG had to give up their cable network activities, most further investment ceased and a period of uncertainty began. Even now, this heterogeneous market structure between level 3 and level 4 operators is the main reason why it is so difficult to build a nationwide integrated cable network that allows market players to realise economies of scale as well as economies of scope. When the European Commission ordered DTAG to sell their networks in 1996 it took two to three years before negotiations began between DTAG and potential investors to sell regional cable networks. Eventually, Callahan Associates bought the cable networks in North Rhine/ Westphalia and BadenWürttemberg, marketed under the new brand “ish,” and the investment

Development of the Broadband Market in Germany

company Klesh & Co Ltd became the new owner of the cable networks in Hesse (“iesy”). All the other regional networks were put together in a new enterprise named Kabel Deutschland Gesellschaft (KDG). In 2001 fundamental changes seemed to be ahead when the U.S. investor Liberty Media made an offer to buy KDG. With this merger, it had been expected that the long phase of uncertainty and delayed investment would come to an end, as Liberty was intending to invest several billion euros on network upgrades and the creation of new services. But at the end of February 2002 the German Cartel Office vetoed the merger claiming that this would strengthen Liberty’s dominant position in the supply market for broadcasting signals to end customers, the market for feeding signals into broadband cable networks, and the market for the supply of network level 3 signals to network level 4 operators. In 2003 KDG finally was sold to the financial investors Goldman Sachs, Apax Partners, and Providence Equity. In 2004 KDG started to merge with ish and iesy but this was again vetoed by the German Cartel Office due to their narrow market definition. Prices and Products

The narrow regulatory framework and the resulting uncertainty for investors is one of the main reasons for the standstill in and non-modernisation of German cable networks. Investment was cut back due to financial difficulties and the network upgrades proved to be more expensive than anticipated. As a result, investments were made by the new owners only in a step-by-step fashion and had to be financed by their own cash flow. This means that, at present, cable networks are mainly used for the transmission of broadcast signals and most of the programmes transmitted are still analogue. It is astonishing that even in areas with modernised cable network infrastructure the marketing activities of cable network operators are quite low-key. Only in some highly populated areas such as, for example, Berlin, Cologne, Leipzig, and Stuttgart have cable networks been slowly upgraded and provided with a back channel. Private households in these areas can subscribe to broadband Internet access as well as to IP-based telephony services. The foremost operators are Kabel Baden-Württemberg, ish, and EWT in Berlin. However, there are indications that some small level 4 operators have started to invest and upgrade their networks in order to offer triple-play services. At the end of 2004, about 145,000 (compared to 70,000 in 2003) households have subscribed to broadband Internet services.

203

franz büllingen

204

Generally, these Internet by cable network offers are orientated by DSL price structures. Most of the cable network operators like PrimaCom, Kabel BW, or KDG offer flat rates for their large spectrum of services. Iesy, for instance, offers 2 Mbps for €19.98 and ish customers can subscribe to 5 Mbps for €74.90. Regulation

The very complex system of cable market regulation is strongly influenced by Germany’s federal political system, and it is assumed to be one of the major barriers to the modernisation of the cable network infrastructure. While the central government is responsible for the regulation of networks and the allocation of frequencies (which is executed by RegTP), the regulation of content and the related use of frequencies is left to the 16 federal member states. While in the past cable network operators were only paid for their transport and distribution services related to programme signals, they now want to change this business model by playing a more active role in packaging and marketing not only TV but also telecommunications services like broadband Internet access and telephony (triple play services). Furthermore, they are keen to maintain an exclusive billing relationship with private households. In pursuing this business model as successfully as possible, the cable network operators face three major obstacles. First, they have to fulfil “must carry” obligations that determine the transmission of all programmes of publicly financed broadcast stations as well as regional programmes. Moreover, access to one-third of the digital transmission capacity is regulated on the basis of criteria such as plurality of content based on federal law due to Germany’s history.1 This regulatory framework has a strong impact on programme packaging and the overall allocation of transmission capacity by the network operators. The second problem relates to other competitive broadcasting platforms, which will continue to offer a wide selection of free TV programmes, financed either by public television fees, paid by all private households owning radios and TVs, or through advertising. Experts assume that the availability of high quality free TV is one of the major obstacles for private TV investors because while households are able to receive basic TV packages at little or no cost via satellite or DVB-T, their willingness to pay for extended cable programming is low. The demand for extended cable programming can be assumed to be very cost-dependent as even small price increases would be likely to provoke high rates of change to satellite or DVB-T. Added to this, the mandatory television fee reduces the demand for extra cable pay-TV packages.

Development of the Broadband Market in Germany

Third, content and service providers want to keep their close relationships with their customers and to hold on to the old “transport model” using cable network services only as a distribution platform. While cable networks are upgraded step-by-step by investors, content and service providers like AOL are putting growing pressure on the regulator to gain open access to cable networks by demanding an open network provision regime (ONP). ONP, however, would constitute a major investment disincentive for cable network operators to modernise cable infrastructure. Because it would be economically the most favourable way for both groups of players, cable network operators and content and service providers, to find the middle ground on their respective interests, it can be assumed that the most likely future business model will be a hybrid. Demand for Cable Network Services

Twenty-two million or approximately 60% of all households were connected by cable at the end of 2003. As it stands, the low degree of digitisation is hindering the operators’ ability to offer and to create new services. So the demand for broadband Internet services, for telephone services over cable, and for value-added services like video-on-demand, is just beginning. The main payTV provider, Premiere, reached a figure of 2.9 million subscribers in 2004 and made a profit for the first time since its market entrance. The Development of the FTTX Market Fibre-to-the-home, fibre-to-the-curb, or fibre-to-x is regarded as the most efficient means of providing business and private customers with overflowing bandwidth. Therefore, in many countries, such as, for example, Japan, South Korea, France, the U.S., and Canada, there has been intensive activity in applying FTTX not only in trans-national and national backbone networks but also in local area networks. While in countries like Japan and the U.S. commercial offers have already been introduced to the broadband access mass market, no concrete plans or pilot projects are presently underway in Germany. When the telecommunications networks were modernised in East Germany at the beginning of the 1990s in the reunification process, some 1.2 million households were connected to a so-called OPAL-fibre network, but after being implemented these networks were not used to offer broadband Internet access services. Because these networks are not compatible with DSL technology, they cannot be used for broadband Internet access over DSL. Excluding some idiosyncratic offers for broadband connection for enterprises, there

205

franz büllingen

206

is currently no activity in Germany to start or intensify investment to connect private or business customers by FTTX networks. The reasons for this lack of activity are not clear and are not a major political or regulatory issue. Therefore, it can only be suggested that there is a distinct lack of interest on the part of the incumbent to invest in its LAN infrastructure, which is an area with a high ROI and is not threatened by competition or regulatory intervention. A second aspect might be that the relevant political institutions and, in particular, the regulator will not take action in order to boost investment in FTTX for the mass market while competition in the sector is regarded as unsustainable. Because any kind of fibre networks have enormous economies of scale, broad activities to implement FTTX would be regarded as a strategy of re-monopolisation. Therefore, some voices have been heard in the last year in favour of a socalled provider model. Such a model is based on the establishment of an independent infrastructure society that builds FTTX networks and rents passive transmission capacity to any telecommunications service providers. The capacity of FTTX networks offers plenty of opportunities for potential service providers to use this infrastructure simultaneously. However, the chance of realising such a provider model and its economic feasibility is not yet clear. The Market for PWLAN Services Infrastructure and Availability

The market for public wireless LAN (PWLAN) services is an emerging market that started about three years ago. Since then varied service providers, such as start-up enterprises, mobile and fixed network operators, ISPs, and owners of regularly frequented public locations like airports, hotels, train stations, and cafes, have offered PWLAN services in order to provide fast Internet access to their customers via wireless networks. Because the international standardisation organisation, IEEE, quickly released the protocol family 802.11 and the hardware components have become inexpensive, market penetration has forged ahead. At the end of 2003, there were almost 2,300 hotspots in Germany. By the end of 2004, there were more than 6,000. Some forecast studies assume that PWLAN market penetration in Germany has not yet reached its peak and that in a further one to two years more than 12,000 hotspots will offer fast Internet access. The above figures pinpoint that nearly 80% of all PWLAN hotspots are located at hotels and gastronomic enterprises. At present especially petrol stations and train stations are equipped with PWLAN infrastructure.

Development of the Broadband Market in Germany Structure of Suppliers

Looking more closely at the structure of suppliers, four different kinds of player groups can be distinguished by their business models (see Figure 7.3). The first group comprises mobile and fixed network operators as well as ISPs, and they handle all segments of the value chain, have a relationship to a critical mass of customers, and are familiar with the organisation of roaming and billing processes. The incumbent’s subsidiary, T-Mobile, is the biggest supplier of PWLAN hotspot services, while the other mobile network providers like Vodafone, E-Plus GmbH, and O2 are a long way behind. This group has the largest proportion of all PWLAN hotspots. In the long run the mobile network providers will merge their PWLAN services with their other mobile communication networks, GSM (Global System for Mobile Communication) and UMTS (Universal Mobile Telecommunications System), into a “mobile seamless web.” This means that customers using PWLAN and other mobile communication services will then not be able to identify by which technical platform they communicate their data. The migration of SIM (Subscriber Identity Module) card technology to PWLAN services will contribute to solving the still existing authentication problems as well as problems of authorisation, accounting, and international roaming. Therefore, it has to be concluded that this group of PWLAN players not only has the most self-sustaining business model but also will dominate this service in the future. The second group of PWLAN service providers consists of enablers and resellers that seek to be economically successful with complementary business models at those hotspots that presumably will be profitable due to intensive customer demand and high data traffic. This group is quite heterogeneous and comprises start-up enterprises as well as technology and service suppliers like Monzoon, Netchekin, Cisco, Greenspot, and iPass. This group of players has its business focus on the rollout of PWLAN networks, the technical management of hotspots, and the organisation of roaming and billing platforms. The only chance the vast majority of these enterprises have to survive in the long run lies in co-operation with roaming and billing platforms due to their lack of critical mass. So far, none of these platforms has been successful in the PWLAN market. Those initiatives that were, for example, promoted under the label “Greenspot” by the ECO-Verband, one of Germany’s important telecommunications service provider associations, first have to reach a critical mass of associated members of hotspot service providers before they become attractive enough

207

franz büllingen

208 Mobile network operators

City carrier

T-Com BerlinNet Isis O2 T-Mobile Swisscom Eurospot Pu rch ase Talkline of E-Plus C ho Ganag oope Megabeam WLAN Group ration tspot s Monzoon Start-up enterprises Netcheckin Sofa Networks Personal WLAN Others STSN M3 Connect iPass Greenspot Acoreus Garderos ISPs Cisco Intel AOL T-Online Marconi Vodafone

Debitel

figure 7.3

Groups of suppliers and consolidation processes in the PWLAN market

Source: WIK.

for others to join. Because most of these hotspots generate a low turnover, their billing processes and the maintenance of their infrastructure have to be strongly standardised and rationalised in order to keep costs low. In summary, most experts are convinced that broad consolidation processes will be seen in this second group of PWLAN service providers within the next two to three years. The third group of PWLAN service providers comprises resellers like Debitel, Talkline, and O2, which do not own their network infrastructure but offer fast Internet access to their customers under their own brand. These players will only offer their services at hotspots that are profitable. If it is not, most experts assume that they will close it down. The last group of market players consists of location owners. In the past, it was supposed that most of these players would step out of the market after a short time because it would be difficult for them to run their hotspots profitably. However, it is now apparent that more and more members of this group offer free PWLAN services in order to enhance their key activities with an attractive costless add-on service. The idea is that people will stay longer in convenient places with PWLAN and therefore will consume more. This group includes, for example, international hotel chains, petrol stations, and cafes and restaurants.

Development of the Broadband Market in Germany Products and Prices

Due to the complex dynamics and great variety in the PWLAN market, it will not be particularly helpful to give an overview of the available products and services. As already mentioned, many hotspot owners have switched their business model and offer free Internet access services, because the task of applying conditional access and billing users is very costly. On the other hand, most of the other players pursue a clear turnover strategy. For instance, T-Mobile at present let their customers choose between three different time-related tariffs. These prices, introduced in March 2004, were lower than those published in January that year in an attempt to stimulate increased demand. Subscribers to T-Mobile services were also offered the opportunity to pay for PWLAN services with their monthly bill, while users of PWLAN services who did not subscribe to T-Mobile’s or T-Coms Services could pay directly by credit card. Regulation

PWLAN systems, which apply for 2.4 GHz frequencies, transmit their signals within the ISM-spectrum. The ISM-spectrum is allocated on the basis of an internationally agreed regulatory regime and can be used for Industrial, Scientific, and Medical purposes. The advantage of the ISM-spectrum lies in its unrestricted use by anyone without being obliged to pay licence fees. The bureaucratic processes that are necessary for the allocation of other frequency spectrum would have led to an unacceptable increase in the costs of WLAN technologies. Because it could be foreseen that the demand for PWLAN services would quickly exceed the given ISM frequency spectrum, the German regulator decided at a very early stage of market development (in November 2002) to provide more bandwidth at the 5 GHz spectrum by general assignment. Since that decision, the spectrum between 5150 MHz to 5350 MHz and 5470 MHz to 5725 MHz can be freely used for PWLAN services without individual permission from the regulator. The only problem with the given PWLAN spectrum is that it is also used for many other applications and services like satellite transmission, position finding (GPS, global positioning system), and amateur radio operating. This synchronic use can cause radio interference. Providers of PWLAN services are subject to the same regulatory rules as all other telecommunications service providers. Previously, PWLAN service providers had to be licensed under §6 of the German telecommunications law (TKG). Since licensing is no longer necessary in line with the EU Regulatory Guidelines, this provision has not been applicable since July 2003, although providers do have to be registered by the regulator.

209

franz büllingen

210

Furthermore, PWLAN service providers are obliged to take adequate measures to ensure the integrity and confidentiality of data communication. Under the TKG (§107) all service providers are obliged to take technical and organisational measures to protect personal data and to secure their telecommunications systems and data storage against intrusion and interception. Demand for PWLAN Services

Most of the recently published market studies indicate that the general public are becoming much more aware of PWLAN services and that the share of laptops and notebooks technically capable of interchange with wireless networks is growing very fast.2 However, the use of PWLAN services is quite restricted and limited mainly to business customers. A market study by Jupiter Research found that while 70% of all Internet users have heard about PWLAN services, only 6% have actually experienced this kind of wireless Internet access. Furthermore, the willingness to pay for fast Internet access seems, in general, to be very low. Detecon & Diebold Consultants published another market study in November 2003 on the opinions of an expert panel about future PWLAN market trends. The main consensus was that laptops and notebooks will be the dominant access devices over the next five years. At present, more than 70% of notebooks in Germany are used by business customers. This implies that it will take some time for the penetration rate of notebooks in the hands of private customers to grow significantly. Looking at applications, the use of e-mail services is still the killer application for business users, while private users prefer information research and entertainment services. At the end of 2003 T-Mobile published the results of a market survey on the preferences of customers in order to find out their selection criteria in distinguishing the services of different providers. From these results, it is clear that price sensitivity is the crucial factor in selecting a service provider followed by the quality of service and the transparency of price structures. As the number of hotspots continues to grow rapidly, customers are demanding much more information not only about the locations but also on the quality of services. Therefore, a self-regulatory certification initiative was established two years ago by PWLAN suppliers in order to set up a clear classifications regime related, for example, to speed, billing, roaming, IT security, service, and support. Customers who want to use PWLAN services are guided by various signs (from 1 W ⫽ simple service for private customers up to 5 W ⫽ first class service for business users). In addition, the logo of the Greenspot initiative and the label of the WiFi-Alliance are also a guide for customers.

Development of the Broadband Market in Germany

The Market for Power Line–Based Telecommunications Services Infrastructure and Availability

No network is as ubiquitous as the power supply network. Since the early 1930s, it has been used for the transportation of signals for different management purposes of power consumers—the synchronisation of watches and so on. Consequently, it was a fascinating idea to use the existing electricity power lines for bi-directional transmission of (broadband) telecommunication signals. The costs of implementing new telecommunications networks are the major hurdle for intensified competition on the last mile level (e.g. local access network), but with such a technology it seemed that it would be possible to introduce an alternative LAN infrastructure in a short space of time and at comparatively low cost. So in the 1990s, a great deal of research was carried out and pilot projects were started to realise this idea in Germany. The main players at that time were hardware suppliers like Siemens AG and Alcatel. Power Line Communication (PLC) was regarded as having great potential to enforce competition on the local loop level. Since then increased efforts have been made—mainly in Germany and Switzerland—to solve technical and regulatory problems. When the first suppliers started to offer their still very limited broadband services at the end of the 1990s transmission rates of 2 to 3 Mbps could be realised on the low voltage level. At that time, it was expected that usable bandwidth would increase from 5 up to 10 Mbps within the next five years. The theoretical capacity was assumed to exceed 30 MHz, so PLC technology seemed to be a very promising way of addressing mass market needs with new services. The power supply industry had the biggest incentives to invest in PLC systems. For major players like RWE or E.ON, the success of this technology would have meant a tremendous upward revaluation of their low voltage networks. PLC would have offered a unique platform not only for telecommunications services but also for innovative energy-related services and applications. The power supply industry would have been able to create additional operating areas that were close to their core businesses. However, while research and the pilot projects were still under way, it became clear that the major obstacle facing the realisation of this new business model was the electromagnetic interference radiated and received by power lines, which can lead to mutual interference between PLC and other radio services. Therefore, the German regulator took a leading role in Europe to find a

211

franz büllingen

212

solution to this problem. In 2002 very low limits had been fixed by CENELEC that were used as guidelines for the German regulator’s action. Since then it has become very difficult for technology suppliers to develop efficient PLC systems that keep to cost benchmarks and ERC’s (European Radiocommunications Committee) limits while simultaneously performing at high bit rates. To a certain degree, power lines technically can be made suitable for the transmission of telecommunication data by installing additional technological equipment to prevent these lines from radiating interference. But because the power line networks originated at different times and are of different qualities, it is very expensive to upgrade them for fast Internet access services in most areas. Structure of Suppliers

Due to the aforementioned technological and regulatory problems PLC currently is a quite limited telecommunications service only offered by some regional energy suppliers. The energy supplier EnBW, for example, has offered PLC since April 2002 in the small town of Ellwangen, which is situated in Baden-Württemberg. MVV Energie AG has been a PLC service provider in the city of Mannheim since 2001. There are also some other PLC service providers in small towns in the northern part of Germany and in Dresden, without really addressing the mass market. In summary, therefore, PLC is not a competitive access technology for fast Internet and the number of users did not exceed 20,000 in 2004. Facing the diffusion of new and much cheaper wireless technologies like PWLAN or WiMax it would appear that PLC has missed its “window of opportunity.” Prices and Tariff Structures

As PLC is a shared medium all connected users have to share provided bandwidth. Bandwidths offered by providers are always related to the bruttotransmission rate, which benchmarks the theoretical maximum. Tariff structures for PLC services are orientated to DSL prices. Service providers like Vype or DREWAG offer a 2 Mbps symmetric up- and download rate for private and business subscribers at almost the same rates as DSL providers. Regulation

Embedded in international and European regulatory frameworks, the Ministry of Economy and Labour (BMWA) developed a national plan for the use of frequencies in electrical lines in the late 1990s. This national plan for the use

Development of the Broadband Market in Germany

of frequencies, the so-called Frequenzbereichszuweisungsplan 30 (NB 30), comprises concrete technical parameters, defines the rules for using frequencies, and fixes the limits of interference with other radio services. This regulatory framework was enacted in July 2001. When the limits of interference were published, it became clear that they posed a significant technological and financial challenge to potential PLC suppliers, and it is regarded as one of the main reasons why many players stopped their involvement in PLC. Demand

PLC services for fast Internet access so far is a niche market. PLC service bundles consisting of telephone, Internet, and energy-related services have been mainly marketed to private households with higher bandwidth needs such as teleworkers and those self-employed in creative businesses, for example, as well as small and medium-sized enterprises. However, as access alternatives like DSL penetrate the market, it is unlikely that PLC has the potential to become a mass-market service, though it is an interesting feature for in-house use. With PLC technology, electric devices like PCs, printers, cameras, and other electronic equipment can be connected easily by power line networks. So PLC might positively contribute to the penetration of fast Internet in households that have to apply to a new cable infrastructure in order to connect their PC to the PSTN. The Market for WiMax-based Services Infrastructure and Availability

Ever since the new standard for fast Internet access—WiMax (World Interoperability for Microwave Access, IEEE-802.16)—was introduced in 2002 by the international standardisation organisation IEEE, it has been regarded as a highly promising wireless infrastructure for players to offer broadband access services to the mass market. According to the suppliers, WiMax should be able to provide bandwidths from 1 to 100 Mbps using the frequency band from 2 to 6 GHz in an area of 10km diameter without being connected by sight. It is believed that WiMax could contribute enormously in upgrading and completing line-based telecommunications infrastructure, especially in those countries that suffer from inadequate telecommunications networks. Experts believe that even in Germany there will be a broad field of applications for WiMax as, in their opinion, it will intensify competition with DSL, WLAN, and cable networks. The first WiMax services are expected to be launched in the German market in 2005.

213

franz büllingen

214

Several international forecast studies offer a very optimistic view on the future business potential of WiMax, and producers of electronic equipment like Intel, Motorola, Siemens, and Alcatel have strong interests in WiMax. Especially in countries with wide rural distances, like Eastern Germany, this new transmission platform is expected to be quickly rolled out in areas with less dense population. However, previous studies with similar optimistic expectations did not always accurately forecast market realities, so assumptions of prosperous growth should be treated with caution. Even so, potential suppliers of WiMax services have already started to analyse the business potential of this technology in order to design different scenarios of competition with other transmission platforms. Regulation

The availability of frequencies needed to realise WiMax services is regarded as one of the main bottlenecks in Germany. The regulatory authority has taken no decision so far as to whether the frequency spectrum of wireless local loop (WLL) services could be used for WiMax. Demand Side

Taking a closer look at the demand side, it is not easy to make coherent assumptions about consumers’ behaviour because new investments in technical equipment have to be made. Furthermore, it is still not clear if WiMax radio signals are only available for outdoor use or if they are also available for inhouse connectivity. If they are not, additional installations will be required deploying outside antennas. Last but not least, WiMax is a shared medium. This means that users will have to share available bandwidth, which might be very low in the event that WiMax becomes a mass-market service. In order to avoid such a bottleneck a dense network of wireless cells has to be implemented, which leads experts to believe that rolling out wireless networks based on WiMax technology will be as expensive as UMTS networks. As a result, potential investors will face big financial risks for their business models. conclusion: the future development of germany’s broadband market The strong demand for DSL services will continue in subsequent years. According to our forecast on future broadband market dynamics in Germany, we expect that at the end of 2005 more than 8 million households will be con-

Development of the Broadband Market in Germany

215

Broadband penetration in 2010: about 56% of all households Broadband penetration in 2005: about 20% of all households

30,000 Internet access in thousands

Broadband penetration in 2015: about 72% of all households

25,000 20,000 15,000 10,000

CAGR: about 20%

5,000

DSL

figure 7.4 2015

Cable

Satellite

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

0

Others

Forecast of the broadband Internet penetration rate in Germany until

Source: WIK.

nected to fast Internet. We assume that market penetration will reach 56%, or nearly 20 million households, in 2010. Within this forecast, the development of fast Internet access over cable networks is the greatest area of uncertainty. In the long run it seems to be realistic to assume that market growth will have come to an end by 2015 (see Figure 7.4). We believe that the overall penetration rate of broadband Internet access will not exceed 80% of all households due to the reluctance phenomenon of many households. This development will be encouraged by the growing use of Internet services generally, which reached 61% in Germany in 2004. The adoption rate of narrowband Internet is important because it will be substituted step-by-step by fast Internet access. The additional value of higher bandwidth at lower cost which, in turn, leads to more convenient use are the main driving factors stimulating demand for fast Internet. This is especially true regarding the introduction of VoIP to the mass market within the last two years. VoIP has become an additional driving force to boost the DSL and cable modem penetration rates. Another strong incentive for the market growth of DSL and the substitution of narrowband Internet access will be the structure of tariffs. Volume tariffs,

franz büllingen

216

especially, will be extremely popular because they can be easily controlled by subscribers and offer always-on connectivity at comparatively low rates. On the supply side, it is presumed that the market share of competitive players will grow slightly. Intensified marketing activities based on partially “aggressive” bundled service packets of alternative carriers will contribute to their better market position as well as the resale offer of the incumbent DTAG that is now available. Though only a small part of value creation will remain with competitive carriers, the service input of the incumbent’s DSL resale will be in strong demand. These alternative carriers will use DSL resale as a complementary product in order to complete their product portfolio. This will make their services more attractive because subscribers will have to sign contracts with only one service provider. In fact, this has already been proved by the strongly growing demand for DSL services based on the incumbent’s DSL resale offer, which had already reached 3% of the entire DSL market within six month of being introduced. Additionally, it can be assumed that due to declining prices for IP transport, DSL volume tariffs will decrease slightly in 2005 while the transmission capacity will increase steadily. To summarise the observed market trends, it can be concluded that competition between DSL service providers will intensify within the next few years. City and regional carriers, in particular, anticipate gaining larger market shares with new bundled service packages. But not only modal competition will reach a new level; inter-modal competition between the different transmission platforms will also increase. Most of the cable network operators not only have intensified their investments in upgrading their infrastructure step-by-step but have also strengthened their marketing activities in order to create higher ROI. The biggest cable network operator, KDG, recently announced a new marketing initiative for 2005 to offer fast Internet access services in many of Germany’s major cities such as, for example, Dresden, Hamburg, and Saarbrücken. In mid-2006 experts expect fast Internet access services to be offered to the mass market over the WiMax platform. Though the regulatory framework has still not been settled with regard to the use of frequencies, it is assumed that this market will be developed in a short space of time. As far as development of the regulatory framework is concerned, many experts believe that in most market segments competition is still not yet stable and self-sustaining, which means that competitors will not be able to run their businesses independent from continuous regulatory interventions. While many other countries in Europe have DSL resale and unbundled bitstream access, Germany is far behind in that respect. Competitors have complained

Development of the Broadband Market in Germany

to the regulator that most of the regulatory measures proposed by the European Commission have been set out far too late and are too weak. There has been much speculation about the reasons for the incumbent’s strong position in the broadband Internet access market. One argument often to be heard is that because the German government still holds a 34% share of DTAG it tries to keep the price as high as possible in case the political administration wants to sell it. The other, maybe more reasonable argument, is that Deutsche Telekom is an important component of German industrial policy for the following reasons. First, a highly rated and powerful incumbent is protected against hostile mergers and has itself, with its large turnover, the potential to intervene and grow in foreign telecommunications markets like the U.S. or Eastern Europe by taking over minor players. Second, the German government has a significant influence on DTAG’s labour policy, which means that the incumbent is not going to set labour power free in line with their own strategic policies. Besides industrial and regulatory policies, some political activities can be observed promoting fast Internet in the mass market linked with the “eEurope 2005” framework. In 1999 the so-called D21 Initiative association was set up in Germany as a public-private partnership in order to stimulate discussion between governmental, industrial, and societal institutions. This “Network for the Future” has since been joined by more than 400 firms, political institutions, and associations. It aims at promoting a broad educational and cultural background to strengthen general qualifications and skills for general Internet use. So far, the D21 Initiative has mainly issued strategic studies and position papers as well as organising conferences and workshops to distribute and provide information to the general public. Concrete government action similar to that taken in many other countries like Japan or South Korea (for example, the deployment of industrial policy instruments, consumer grants and subsidies, financial support for training and awareness, or use of regulatory instruments like special allowable rates of return) is not part of this initiative and it is not expected that any of these measures will be introduced in Germany within the next few years.

notes 1. Since mass media have been misused by the National Socialists for propaganda purposes, federal German law imposes obligations on broadcasters that they have to provide a broad variety of opinions to the public. 2. Cf. Jupiter Research: Jupiter Research reports smart public hotspots business models look beyond consumers for positive ROI, Press Release, 2 December 2003.

217

franz büllingen

218

references Büllingen, F. 2004. Die Entwicklung der breitbandigen Internetnutzung privater Haushalte in Deutschland bis zum Jahr 2015. In Breitband-Perspektiven. Edited by J. Eberspächer, and H.-P. Quadt. Berlin: Schneller Zugang zu innovativen Anwendungen, 64 –74. Büllingen, F., Chr. Gries, and P. Stamm. 2004. Der Markt für Public Wireless LAN in Deutschland. WIK Diskussionsbeitrag Nr. 252. Bad Honnef: WIK. Gries, Chr. 2003. Die Entwicklung der Nachfrage nach breitbandigem InternetZugang. WIK-Diskussionsbeitrag Nr. 242. Bad Honnef: WIK. ———. 2004. Entwicklung der DSL-Märkte im internationalen Vergleich. WIKDiskussionsbeitrag Nr. 257. Bad Honnef: WIK. ITU. 2003. Birth of Broadband. ITU Internet Reports. Geneva: ITU. Stamm, P. 2000. Entwicklungsstand und Perspektiven von Powerline Communication. WIK-Diskussionsbeitrag Nr. 201. Bad Honnef: WIK. ———. 2002. The Singularities of Germany’s Cable Industry. In Förderung der Marktperspektiven und der Wettbewerbsentwicklung der Breitbandkommunikationsnetze in Deutschland. Edited by Franz Büllingen et al. Bad Honnef: WIK.

8 Broadband in Italy Timing in Inter-Modal Rivalry Cristiano Antonelli and Pier Paolo Patrucco

the characteristics and evolution of broadband in Italy are the results of the systemic interactions between overlapping needs and niches of consumers, competing and yet complementary technologies, as well as between learning and heterogeneous actors involved in such technologies. More specifically, the structural and dynamic characteristics of broadband in Italy seem affected by three classes of factors: (1) the overlapping of the characteristics of demand; (2) the competition and complementarity between the incumbent’s strategy (i.e., Telecom Italia), aiming at rejuvenating the old copper access infrastructure with ADSL combined with optical fibre in transmission (Antonelli, 2001 and 2003); and (3) the new entrants’ strategies aimed at (partially) bypassing1 Telecom Italia infrastructure by using their own proprietary infrastructure based on optical fibre to the curb (FTTC) and to the home (FTTH) in access and the Internet Protocol (IP). In such a context, the diffusion of broadband in Italy can be viewed as the outcome of an interesting form of inter-modal competition that can be analyzed by means of an extension of monopolistic competition between the incumbent Telecom Italia characterised by financial weakness and a quasiirreversible huge stock of copper cables infrastructure, and a small number of new entrants, very often operating in local markets that represent geographical and product niches. Among these, the major new entrant is Fastweb, which introduced radical technological changes in broadband in Italy, investing exclusively in proprietary optical fibre network combined with IP, and providing advanced and integrated services for voice, video, and data. Fastweb is eventually emerging as the major competitor to Telecom Italia at the national level.

cristiano antonelli and pier paolo patrucco

220

The outcome of such interactions in terms of the diffusion of broadband in Italy characterises the Italian society in an extremely uneven way. The diffusion of broadband is characterised by a strong heterogeneity, from both the technological and the geographical viewpoint with important consequences on the characteristics of the demand for broadband services. In this regard, the so-called digital divide is much strong in Italy and most evident first of all from the geographical viewpoint, not only reinforcing the social and economic distance between north and south, but also between metropolitan and rural areas. Moreover, it is most evident also from the demand viewpoint, with a minor portion of innovative firms and consumers that adopt complex broadband services and support the diffusion of broadband at large, and a larger portion of industrial and private users that are following or lagging behind. Finally, it should be clear that within a systemic perspective where broadband is the result of the interplay between complementary and competing actors and technologies, broadband diffusion cannot be regarded only as a matter of physical infrastructure and connectivity. The provision of services, content, and applications matter and, in turn, the characteristics of demand do also interact with the supply side factors in order to explain the evolution of broadband in Italy. The appreciation of the preferences and behaviours of users of broadband infrastructures and content should be regarded as indispensable for the understanding of the rate and direction of broadband diffusion. This chapter is structured as follows. Section 2 characterises the overlapping characteristics of the demand for high-speed communication services (Madden, 2003). Section 3 combines the analyses of the characteristics of demand and supply to analysing the diffusion of broadband services. Section 4 describes the structural elements of the broadband system in Italy with a special emphasis to the specific market characteristics and the technological infrastructures in use, their diffusion and their implications in terms of barriers to entry deterring competition to the incumbent from new potential entrants (Geroski, 1995). It also shows that the interactions between different technologies and actors and the positive effect on the diffusion of advanced services and broadband at large are limited to specific geographical areas and technologies. This has important consequences in terms of the understanding of the digital divide in broadband in Italy. Section 5 qualitatively shows the way in which a great part of the radical changes in such an infrastructure and more general in the way of providing broadband services is due to the innovations introduced by Fastweb. Section 6, through the cases of COLT and TOP-IX, provides evidence for the way in which localized players can segment the market both geographically and technologically in order to identify specific niches where they

Broadband in Italy

can be competitive providing customized and ad hoc solutions for their clients. Conclusions summarize the main results. the demand for high speed communication services The demand for high speed communication services is highly heterogeneous both with respect to the needs and characteristics of consumers and their geographical location. From the viewpoint of geographical distribution, the demand for high speed communication services is highly uneven across the country. Demand is quite strong in a few metropolitan areas and in the northern part of the country, while the south and the countryside are lagging behind. The historical features of the nation in terms of low levels of metropolitan concentration play an important role. Italy has been for centuries and still is the hundred-towns country; population density is spread through the country. Rome, the capital city, accounts for less than 3% of total population. On the other hand, Lombardy, Emilia, and Veneto exhibit high levels of homogeneous distribution of affluent population and advanced economic activities, and one-third of the national population is concentrated in three regions. These characteristics have played a major role in shaping the strategies of actors involved. From the viewpoint of the characteristics of consumers’ needs it can be defined as an instable bundle of interdependent and overlapping demand schedules for an array of specific services. Such a broad bundle includes: (1) the demand for high speed data communication such as access to Internet services in a fixed location; (2) the demand for high speed data communication services such as Internet services by mobile users; (3) the demand for high quality telephony or VOIP; (4) the demand for television services; and (5) the demand for entertainment services. These five specific demand schedules can merge into a broader aggregate demand provided a number of qualified conditions, both on the supply and demand side, in terms of quality and prices apply. As a matter of fact, however, each of these specific demand schedules exists and exhibits quite specific characteristics. Consumers may be induced to use the services provided by the same basic technological platform or may find the supply of each service provided by quite well distinct technologies and firms operating in markets with their own idiosyncratic features. The merging of such demand schedules, in other words, is the result of a process where both consumers and suppliers play an active role. Access to Internet services may be operated by means of Bluetooth technologies that work via cellular phones either in a fixed location or in mobile

221

cristiano antonelli and pier paolo patrucco

222

conditions. ADSL technology, however, makes it possible to supply mediumspeed data communication via copper cables. Optical fibre technology indeed provides faster data communication services, but only in a fixed location. Access to television and entertainment can take place via analogue broadcasting, digital broadcasting, narrowcasting on cellular phones, or delivered by means of coaxial copper cables or optical fibre. High quality interactive entertainment instead can only take place effectively by means of mobile narrowcasting or optical fibre, while the provision of IPTV over ADSL is still characterised by lower quality that can harm adoption and diffusion. Low quality telephony, like Internet access, can be provided via ADSL although the negative gradient in quality is larger. The same applies to television and entertainment narrowcasting. Product enrichment based upon the implementation of the broadband platform generates important effects. Product enrichment makes it possible to climb the quality ladder and increase the spectrum of needs and uses that can be delivered by means of a universal platform. This can be the case of the broadband infrastructure, but also of an aggressive wireless platform able to deliver the full range of services. Progressively the demand for the services delivered by the universal platform moves upward and the room for alternative services provided by rival platforms can be squeezed. In the long term more than one rival universal platform can survive, especially if switching costs for consumers are relevant. In the Italian case, as for instance in areas of other countries such as North Sweden, it is clear that the combination of geographical dispersion and product heterogeneity has exerted a strong impact on the strategies of the firms involved in the process. The creation of a national broadband infrastructure stretched to the curb in fact has been substantially delayed by the long-term prospects necessary for an appropriate payback to take place. In turn, geographical dispersion and heterogeneity has induced increasing product heterogeneity and enhanced the levels of inter-modal competition. the analytics of inter-modal competition in high speed data communication Traditional models of monopolistic competition can be stretched in order to understand the logic of the strategic interaction upon which the diffusion of broadband in Italy is based. The diffusion of broadband in Italy, in fact, depends upon the outcome of inter-modal rivalry, where on the demand side different niches—geographically clustered—of customers with different needs exist and yet overlap

Broadband in Italy

systematically and exhibit high levels of cross-price and revenue elasticity, and on the supply side different categories of firms, each characterised by different strategies and different resources, compete. The differences with respect to standard monopolistic competition must be stressed. A group of markets, as opposed to a single market, needs to be identified. Each market is characterised by porous borders with high levels of mobility both on the demand and on the supply side. A large number of firms can be active in each market on the supply side. Each of them, however, must face a strong competition that takes place both within each market and among markets, as it is both intra-product and inter-product. Finally, on the supply side, firms exhibit high levels of product diversification; each firm elaborates a strategy to try and operate across markets in order to take advantage of economies of scope stemming from joint production in production and fidelisation of customers. Many classes of technological and historical factors shape such a competition process, namely: (1) the total lack of coaxial cables in distribution due to the institutional imitations to TV broadcasting that impeded back in the 1970s the use of cables as an alternative method of broadcasting; (2) the systematic use of optical fibre in core network transmission, that is communication between switches (Telecom Italia, actually SIP before privatization, had implemented a general strategy of modernization of all transmission and distribution networks); (3) the strong support to ADSL in distribution in order to rejuvenate to copper last mile by the incumbent; (4) the diffusion of optical fibre to the curb and to the home combined with IP and the emergence of VOIP (Voice Over the Internet Protocol) as a new disruptive paradigm for voice communication by the new entrants; and (5) the transition from satellite pay-TV to terrestrial digital TV based upon digital transmission capabilities. In this context, broadband provides the opportunity to supply an array of high speed communication services based upon the same technological platform. The bundling of such heterogeneous services on the same platform is extremely convenient both on the demand and the supply side. Specifically, on the supply side a universal broadband platform can be very effective because of the strong effects of economies of density. Optical fibre displays almost unlimited capacity, no variable costs and almost no wear and tear, at least related to usage intensity. Last mile broadband, however, requires strong upfront costs for instalment. In such conditions it is clear that the larger the number of users and their usage intensity, the lower the unit costs. In this case unit costs consist of long-term average fixed costs. If each niche of customers remains separated and the bundling does not take place, the advantages of a broadband supply stretched to the curb may

223

cristiano antonelli and pier paolo patrucco

224

become less obvious. The advantages of broadband in fact become evident only if different niches of heterogeneous customers are brought together and induced to purchase the services based upon the same technological platform. This is a long-term process where historic time is relevant both for the delays in the diffusion of communication services stemming from the resilience of users at large and for the effects of inter-modal rivalry. In order for broadband to be successful, the bundling must take place since the beginning of the process. Timing is relevant in assessing the outcome of the inter-modal competition. A major window of opportunity emerged in the early 1990s for the incumbent who could take advantage of the combination of the new technology with the lack of all coaxial cables for TV in the country due to prior legislation in favour of other broadcasting. The project SOCRATES elaborated by SIP-STET in the years preceding privatisation consisted exactly in an aggressive strategy of mass supply of optical fibre to the curb across the country. The aim was clear; the supply of fibreto-the-curb would have prevented the entry of all kinds of competitors in voice telephony as well as in all adjacent markets. At the same time the strong financial solidity of the company would have easily supported the investment, although increasing debt and reducing profitability in the medium term (Fransman, 2002). The Project SOCRATES would have, however, created long-term barriers to entry and to mobility and hence secured long-term profits. At the same time, however, SOCRATES would have provided the country with the unique opportunity to take advantage of an advanced digital infrastructure although partly unjustified from a short-term payback viewpoint. Privatisation (Fransman, 2003) eventually took place in 1999 and a sort of management buy-out brought the company under the control of Pirelli in 2001. The take-over, conducted firstly by Olivetti and eventually by Pirelli, was almost completely funded with the financial resources of the telecommunications company, in the while renamed Telecom Italia. The level of debt increased dramatically reducing the room for long-term investment. Geographical dispersion of the country and the large number of small metropolitan areas scattered throughout the peninsula played a major role in imposing a dramatic stop to the broadband diffusion at least to reach consumers. The optical fibre network was completed just for the backbone network since the mid-1990s. This paved the way to a variety of alternative strategies implemented by newcomers and the impoverished incumbent. Each newcomer has tried to take advantage of the financial weakness of the incumbent and to take the control of a well-identified niche of customers in the attempt to reduce the

Broadband in Italy

mobility of customers across products and hence secure a final demand. In this context the widespread diffusion of mobile communication by Italian consumers can become a factor of irreversibility in favour of the implementation of a mobile platform alternative to broadband. The active role of consumers and their taste for ad hoc consumption technologies closer to the habits of advanced users may play a role (Lancaster, 1971; Bianchi, 1998). Telecom Italia, which also controls TIM (Telecom Italia Mobile, the largest wireless company) has chosen to implement ADSL in order to rejuvenate the capillary copper network and at the same time to try and push the wireless scenario. Consistently the incumbent has systematically delayed the diffusion of the new low-quality telephony (VOIP over ADSL). The quality of the oral communication services provided by ADSL over copper networks remains indeed much lower than dedicated telecommunications services. The duopolistic structure of the television industry (i.e., MEDIASET and RAI) has gradually agreed to try and defend broadcasting, taking advantage of the new digital transmission technology that has indeed increased widely spectrum capacity. The chances to introduce interactive services, however, have suffered this outcome and there is potential for high-quality entertainment based upon fibre-to-the-curb. Finally, mobile companies have systematically tried to push the so-called convergence scenario by means of which mobile wireless access to Internet services based upon Bluetooth technologies implemented by means of the systematic use of WiFi and recently WiMax technology becomes customary. In such a scenario the advent of UMTS technologies, the third generation of mobile telecommunications systems, has been regarded by a few players as the opportunity to implement a multi-service mobile platform able to deliver voice, data, and images including direct access to broadcast TV at the same time. The strong segmentation of the broad bundle of advanced communication services into well-defined niches, each defended by a group of firms with conflicting strategies, has emerged as a result of the financial weakness and the irreversible endowment of Telecom Italia in the years between the end of the 20th century and the first years of the 21st century. Let us analyse such a process in more detail. the market and technologies for broadband in italy The diffusion and penetration of broadband access in Italy is to a great extent due to the integration between the traditional copper infrastructure and new compression technologies.

225

cristiano antonelli and pier paolo patrucco

226

table 8.1 Diffusion of Broadband Access in Italy, First Halves of 2003 and 2004

xDSL Optical fibre Total

2003 Number of Users

2004 Number of Users

2003 (%)

2004 (%)

1,472,000 136,000 1,563,000

3,099,000 201,000 3,300,000

91.3 8.7 100.0

93.9 6.1 100.0

source: Assinform 2004 (www.assinform.it).

table 8.2 Penetration of Broadband in Italy, Percentage of Population

xDSL Optical fibre Total

2003 (%)

2004 (%)

2.48 0.24 2.72

5.38 0.35 5.73

sources: Data from ISTAT (www.istat.it) and Assinform (www.assinform.it).

Tables 8.1 and 8.2 show that in 2003 only 8.7% of the diffusion of broadband access was due to the implementation and use of optical fibre, consisting in 0.24% of the total population. On the contrary, 91.3% of the total accesses in broadband were due to access via xDSL technologies. This represents 2.48% of the total population. In 2004, although we see an increase in the penetration of broadband, which has reached 5.73% of the population, the gap between the use of xDSL technologies and optical fibre is even greater. Accesses via xDSL technologies are, in fact, more than doubled reaching around 3 million of accesses, representing 93.9% of the total accesses in broadband, 5.38% of the total population, and accounting for most of the growth in broadband access. The relative contribution of optical fibre is instead decreased from 8.7% to 6.1% of the total diffusion of broadband, reaching 0.35% of the total population. Such a situation, in which the diffusion and penetration of broadband is mainly due to the use of ADSL, able to rejuvenate the traditional copper network, is affected and characterised by market, technological, and institutional factors, which are important to understand the structure and evolution of broadband in Italy. Let us analyse them in turn. From the market viewpoint, Table 8.3 shows the distribution of market shares in the Italian broadband market between the incumbent Telecom Italia,

Broadband in Italy

227

table 8.3 Market Shares in the Italian Broadband Market

Revenues (million $) Revenues (PPP) Access Lines (thousands) % Revenues % Access Lines Revenues Growth 2004/2003 Access Lines Growth 2004/2003

31.12.03 31.12.04 31.12.03 31.12.04 31.12.03 31.12.04 31.12.03 31.12.04 31.12.03 31.12.04

Telecom Italia Wireline

Fastweb

OLOs (estimated)

Total

18,050.98 23,787.62 14,511.84 14,747.94 2,040 4,010 90.67 87.73 79.08 80.90 31.78

554.66 974.74 445.91 604.32 331 496 2.79 3.59 12.83 10.01 75.74

1,302.39 2,352.42 1,047.04 1,458.46 209 451 6.54 8.68 8.09 9.09 80.62

19,908.03 27,114.78 16,004.79 16,810.72 2,580 4,957 100.00 100.00 100.00 100.00 36.20

96.57

49.85

115.96

92.14

source: Telecom Italia and Fastweb 2004 Financial Year Reports; U.S.$/ Euro exchange rate: Federal Reserve; PPP: OECD. U.S. dollars; PPP (purchasing power parity); 2003, 2004; values, %, 2004/2003 growth rates. Telecom Italia Wireline is Telecom Italia’s business unit for the broadband market.

the main alternative carrier Fastweb, and the other local operators (OLOs). The Italian market can be clearly characterised as a quasi-monopolistic context where Telecom Italia with the business unit Telecom Italia Wireline is by far the major operator in terms of both revenues and access lines (in 2004, 87.73% and 80.90% of the total market shares in terms of revenues and access lines, respectively). Fastweb and the other local operators share only a minor part of the whole Italian broadband market. In fact, in 2004, Fastweb’s share of the Italian broadband market is 3.59% in terms of revenues and 10% in terms of access lines, while in 2004 OLOs’ shares of the Italian broadband market amount to 8.68% and 9.09% in terms of revenues and access lines, respectively. Telecom Italia is clearly benefiting from its previous monopolistic position in traditional telecommunications and the diffused copper infrastructure implemented and updated through the country, by providing broadband services mainly over ADSL. On the contrary, it should be clear that alternative carriers can compete on the Italian broadband market only by adopting segmentation strategies that allow them to satisfy specific users’ needs, focusing on added value services mainly for business users and advanced consumers, and ultimately implementing niche strategies.2

cristiano antonelli and pier paolo patrucco

228

The telecommunication infrastructure already implemented and in use in the country plays also a major role in shaping the technological characteristics of broadband in Italy. The wide diffusion of the traditional copper network and the huge investments in traditional telecommunication access infrastructures based on copper network stimulated investments in new compression technologies especially by the incumbent Telecom Italia. Telecom Italia owns basically the total copper infrastructure in access (consisting in 104 million km of copper cable) and this represents huge sunk costs. New ADSL also stimulated new investments by the new entrants, in order to use the traditional copper infrastructure for digital transmission at high speed without investing heavily in new infrastructures such as proprietary networks and optical fibre through local loop unbundling. In particular, xDSL (ADSL, HDSL, VDSL, etc.) compression technologies have been developed in order to rejuvenate copper networks and the copper last mile as an access system to broadband. In Italy, strictly considering the technological viewpoint, compression technologies find in the traditional copper infrastructure a particularly favourable environment to be applied for broadband access. In fact, in Italy—compared to Germany, the U.S., and the UK—a greater proportion of users are located closer to the local switch, which increases the efficiency of DSL. Moreover, it can guarantee a widespread potential diffusion through users given the diffusion of the copper network. However, in Italy most important technological and institutional barriers that can harm market entries in general and xDSL technologies in particular are in place, limiting market and technological opportunities for the new entrants and eventually the development of broadband. These can be specified as follows. First, limitations to the access to the incumbent’s infrastructure may harm market entry and the diffusion of broadband (Geroski, 2001). Second, new entrants must have access to the Telecom Italia proprietary copper network. Moreover, the use of ADSL implies the availability of broadband metropolitan infrastructure in distribution, in order to connect the new entrant’s network to Telecom Italia sites. New entrants cannot implement their proprietary metropolitan network easily and without very important fixed and sunk costs. Without such a proprietary infrastructure, new entrants need to licence connections from the incumbent, with a substantial increase in the operational costs. Finally, it is clear that the institutional development of appropriate conditions of unbundling local loop (ULL) is most important to support market entry together with the diffusion of ADSL. Monopolistic rights delivered to the incumbent reinforced natural barriers to entry due to the technological

Broadband in Italy

characteristics of the integration between compression technologies and the existing copper network. In sum, the diffusion of broadband in Italy is strongly biased by the presence of the incumbent Telecom Italia, characterising the Italian broadband market as a quasi-monopolistic context. The integration between ADSL and the traditional copper network accounts for the greatest part of the diffusion of broadband in access. Although broadband access based on xDSL technologies is easier and less costly than broadband based on optical fibre because it does not require new infrastructural investments, innovation in techniques and services is limited and constrained by the market, technological and institutional characteristics of the “old” system. Limitations in market entry and competition in turn are likely to harm the development and diffusion of broadband, both qualitatively and quantitatively. It is clear in fact that limitation in the technical variety due to limitation in competition and market entry affects not only the diffusion of single specific techniques (i.e., optical fibre, wireless, satellite, etc.) and related services in broadband but also and consequently the quantitative overall diffusion of broadband. The following section, in fact, will show that, although different types of technologies, services, operators, and consumers coexist at the national level, advanced broadband services and consumers are concentrated in very few areas. The diffusion of broadband is uneven and the digital divide strong at different levels. the digital divide in italy The analysis of the diffusion of broadband shows a picture that is characterised by strong heterogeneity from both the geographical, technological, and market viewpoint (see Table 8.4). The first area is represented mainly by the north-western regions (Lombardy, Liguria, and Emilia-Romagna, and to a lesser extent Piedmont) and the Lazio and is characterised by the fact that broadband is here provided through a variety of technologies where the conditions for the unbundling local loop are most important to enable competition between the national incumbent and a variety of especially local players, active in specific market niches. Optical fibre is present and complementary to xDSL, but only to a minor extent and in practise limited to metropolitan areas. From both the technological and market viewpoint this area is the one in which broadband services are provided by a variety of competitors and through a variety of technologies. However, this is the narrow area when considering either the number of municipalities covered (205), the percentage of population and firms reached

229

cristiano antonelli and pier paolo patrucco

230

table 8.4 The Digital Divide in Broadband in Italy, 2003 Number of % % MuniciPopu% GeoOptical palities lation Firms graphy WLL Fibre ULL xDSL ADSL Sat

1st area

205

25

29

3

2nd area

1,869

50

50

36

3rd area

6,027

25

21

62

X

X

X

X

X

X

X

X

X

Competition

Technological Variety Local Competitors Technological Dualism Wholesale Bitstream Single Technology Very Limited Competition

source: Formez and Osservatorio sulla Banda Larga 2003.

(25% and 29%), and the extent of the territory covered (3%). Here, users and consumers are innovative and proactive and adopt complex broadband technologies and services. The second area is represented by the north-eastern regions or the so-called third Italy (with the exception of the Emilia-Romagna) (Veneto, Friuli Venezia Giulia, Toscana, and Marche) and in part by the notable exception of the Campania and Calabria in the south of Italy. Here broadband is mainly provided through ADSL by the incumbent and by those local players operating in wholesale bitstream, which are reaching 50% of both the population and the firms, but covering only 23% of the Italian municipalities and 36% of the territory. Finally, the largest area in terms of the municipalities (74%) and the territory (62%) covered, but relatively minor in terms of firms (21%) and population (25%) reached, is one in which competition is most limited and broadband services can be accessed only through satellite technologies. Here, it is the case of the regions in southern Italy, but also of those peripheral and mountain areas such as the Valle d’Aosta and the Trentino. The second and third areas represent the larger part of backward industrial and private users, generally adopting less advanced technologies and services, lagging behind the innovative users. Such an uneven diffusion of broadband can be explained as the result of the intertwining effect of different factors. On the one hand, new investments in optical fibre require important expenditures characterised by long-term

Broadband in Italy

paybacks. On the opposite, such investments are now taking place in an increasingly competitive context, especially at the local level with a limited time horizon. Actors are often seeking a short run payback for their investments and thus mainly investing in traditional technologies. The picture is therefore characterised by strong inertia due to the characteristics of the competitors and the stock of investments already in place. In a way, technological innovation and the introduction of competition seem to have reinforced the divide between regions. Investments in advanced infrastructures and technologies for broadband focus on those richer and more innovative regions and areas (such as the northern regions and metropolitan areas) that can guarantee a shorter payback time for those investments, thus attracting new entrants. On the opposite, investments in those poorer and peripheral areas (such as rural or mountain areas and southern regions) are slower and focused on incremental adjustments in technologies and services, in turn deterring market entry. It is clear that in a context where resources are limited and infrastructures cannot be duplicated, two factors are most important. First, access to the incumbent’s infrastructures and the development of the conditions of ULL is crucial to favour market entry and the opportunity of introducing innovations. Second, the development of appropriate services should parallel the evolution in technologies and infrastructure in order to provide a variety of content able to stimulate the demand of broadband. The next section will show in which way technological complementarities, different from those between copper networks and ADSL, can generate innovation in both technologies and services, introducing important changes in the characteristics and evolution of broadband. In such a process where new technologies are used and new services are introduced, Fastweb is emerging as the key player and the main alternative carrier. radical innovations in broadband: the role of fastweb Fastweb has been originally set up in Milan in 1999 as a joint venture between AEM, Milan’s main power utility, and e.Biscom, the Italian leader in broadband telecommunications and media-related services established by Silvio Scaglia. In December 2004, e.Biscom and Fastweb merged following the boards’ approval in April. The merger represented the natural outcome of a streamlining process started in 2002 when the group began to focus on its core business of fixed broadband telecommunications in Italy. Following the merger, the new company is named Fastweb, whose shareholding structure is as follows: Silvio Scaglia (Chairman), 30.1%; Carlo Micheli (Vice Chairman), 10%; Fidelity Investments, 5.3%; and Free float, 54.6%.3

231

cristiano antonelli and pier paolo patrucco

232

Fastweb is now the main alternative carrier and competitor to Telecom Italia. It can be characterised as mainly a metropolitan carrier with, at the end of 2004, about 500,000 telephone lines activated distributed in the six main Italian big metropolitan areas (Milan, Rome, Turin, Genoa, Naples, and Bologna), representing 10% of the Italian broadband market. Moreover, revenues at the end of 2004 amount to $974.74 million. Revenues and access lines growth rate between 2003 and 2004 are 76% and 50%, respectively (data showed in Table 8.3). However, even though from a market and quantitative viewpoint Fastweb can be described as a niche operator, from the technological viewpoint the strategy behind such market segmentation and specialization led Fastweb to be the major innovator in broadband in Italy. In fact, Fastweb’s strategy is based on investment in radically new technologies in broadband, which make use of either proprietary fibre connections or unbundled copper lines from Telecom Italia and which, in turn, allow the provision of new and high value services to both residential and business users. In particular, Fastweb completely invested in an infrastructure alternative to and thus unbounded from Telecom Italia’s one. This infrastructure is based on the integration between the Internet Protocol (IP) and the implementation of a proprietary fibre network able to reach apartments (FTTH), in turn allowing the so-called triple play services, that is, broadband Internet access, high quality voice over IP (VOIP), and broadband video services (both pay-TV and pure video-on-demand). While it is clear that for Telecom Italia new investment in fibre connections would have been highly expensive (because of the important irreversibility due to the previous endowment in copper lines and because of the financial weakness following the privatisation), different factors made the integration between IP and FTTH affordable and profitable for Fastweb. First, the complete lack of coaxial cables and thus the total lack of cable TV in Italy stimulated the diffusion of advanced broadcasting over alternative channels, such as satellite and the Internet. Second, in Italy, as well as in most of industrialized countries, residential users in cities together with business users such as SOHO and business service firms are mostly concentrated in condominiums. Third, in Milan in particular the local utility, AEM, had already implemented a widespread ducts infrastructure for optical fibre as part of its infrastructure of underground ducts for gas and electricity. The same ducts were used to lay optical fibre cables, allowing Fastweb to build a metropolitan area network more easily and without starting investment from scratch. More-

Broadband in Italy

over, similar infrastructures had been partially developed also in the main Italian metropolitan areas in the 1990s by Telecom Italia as part of the project SOCRATES, aimed at developing optical fibre but which was never completed. In turn, demand concentration and the availability of network infrastructures allowed Fastweb to reach scale economies high enough to implement fibre-to-the-home (FTTH) and fibre-to-the-office (FTTO) connections that allowed a bidirectional date rate of 10 Mbps. Moreover, in those rural regions and small cities and towns where optical fibre is not yet available, Fastweb is providing broadband via ADSL integrating unbundled last miles from Telecom Italia in access and Fastweb’s technology in distribution, allowing a speed of 2 Mbps. In this regard, Fastweb is the only carrier that uses optical fibre in access to the home and the office, and this may be regarded as the first radical technological change introduced by Fastweb. However, Fastweb has been able to introduce innovations also in services, such as the triple play services, only through the integration between optical fibre in access and the use of the Internet Protocol in distribution. Here, the use of IP as a distribution technology can be regarded as the second and crucial radical innovation introduced by Fastweb. The different actors providing broadband services based on the integration between ADSL and the traditional copper network, both the incumbent and the new local entrants, had to face the problem of changing from the traditional Public Switched Telephone Network (PSTN) to the packet switching (mainly ATM). In fact, PSTN is today almost entirely digital in technology except for the final link from the central (local) telephone office to the user. On the contrary, the development of a proprietary and independent fibre network represented twice a technological opportunity for Fastweb. On the one hand, it allowed Fastweb to bypass Telecom Italia infrastructure and to overcome the related barriers to entry. On the other hand, from the technological viewpoint Fastweb has not been bounded to the use of the traditional network for the development of broadband services, in turn being able to base broadband services on a single technology, namely the Internet Protocol, and to bypass also the shift from PSTN to the digital system. The integration between IP technologies in distribution and optical fibre to the home and the office in access is now the key source of advantage for Fastweb (Guerci and Marcolongo, 2003). It makes possible the provision of new services, alternative to both narrowband telecommunications and ADSL, with much shorter time-to-market when compared to those of the incumbent and minor new entrants. The latter are constrained by the characteristics and

233

cristiano antonelli and pier paolo patrucco

234

limitations of the traditional network and longer time for testing, adjusting, and finalizing the provision of new services. Fastweb can instead capitalise on the investments in new technology because of the important economies of scope between voice, data, and video; in fact the same infrastructure is shared to provide different services at high quality, that is, data, voice, and video over the Internet. The integration between IP and optical fibre can emerge as a standard in the provision of broadband services. Triple play services may be regarded as the third kind of radical innovation introduced by Fastweb. Let us analyse more specifically the two most important services that may be regarded as key services for both the development of broadband and the understanding of competition between Fastweb and Telecom Italia; namely, VOIP and digital TV. First, in Italy VOIP can be a killer application for operators in the traditional copper infrastructure, including the incumbent Telecom Italia, to different extents. On one hand, Italy still has a very low rate of PC penetration and Internet usage for data transmission. On the other hand, voice is still the main source of revenues for the great part of operators in telecommunications. Broadband services based on IP and data transmission alone would represent only a very small market. In this regard, Fastweb’s choice is to use the IP network to carry also primary voice and to provide the basic Internet service integrated with data, voice, as well as additional services such as video-ondemand and TV-based conferencing. VOIP reinforces the independence of new entrants, and Fastweb in particular, from Telecom Italia and from its traditional infrastructure. On the contrary, traditional operators such as Telecom Italia are bounded to change toward VOIP because of the coupled effect of the high costs of previous investment in the traditional infrastructure and the new investment necessary to change infrastructure. In this regard, Telecom Italia is not only facing the risk of increasing competition and decreases in the market share, especially in those innovative segments characterised by higher added value. Telecom Italia is also facing the risks for the high levels of indebtedness due to the investments in both the traditional and eventually the new infrastructure. Such risks are currently limiting the incumbent with regard to the introduction of VOIP, which is in fact provided by Telecom Italia through ADSL with significant lower quality in the service. Second, digital TV is also most important to understand the direction toward which innovative service in broadband can evolve and the factors that affect such an evolution. The convergence between the traditional telecommunication copper infrastructure and television is in Italy especially at the base of strong institutional support to the transition from the traditional broadcasting system to the digital TV and pay-TV. Strong support to the transition to

Broadband in Italy

digital TV has been in fact given both through fiscal incentives for consumers buying decoders and through a law fixing a deadline for the complete transition to digital TV and the switch-off of the analogue transmission of 31 December 2006. Competition between different technologies through which digital TV can be provided is most important to understand the way in which digital TV could evolve. On the one hand, digital terrestrial TV can take advantage from lower costs due to the use of technologies and infrastructures already in place and in use; that is, the copper telecommunication infrastructure and the traditional TV screens in Italy are in practise covering the total population. However, terrestrial digital TV using the copper cable for the uploading is providing lower quality interactive services. On the contrary, Internet digital TV based on broadband and optical fibre can be more costly but can also guarantee more reliable interactive services and triple play services in particular. In this regard Fastweb is again an important innovator. Relying upon the proprietary fibre network and taking advantage of the important indivisibilities in capital in place, Fastweb is providing triple play services with digital TV on demand over the Internet based on a 5,000 titles movie catalogue and around 300,000 regular subscribers to TV on demand. Moreover, Fastweb is also adding new services such as answering machine and fax services as well as distant video recording. Here it is clear that the provision of appropriate contents will be the key to support the diffusion of digital TV. Digital TV can provide a better quality transmission and the multiplication of TV channels over the current frequencies because of the information compression through digital technologies. However, the risk is to have very scarce content, imposing to the consumers the costs of a fast transition to the new system without providing them the appropriate content. market segmentation and local niches in broadband: the cases of COLT and TOP-IX The cases of COLT Telecom Plc and the consortium TOP-IX represent two different and yet complementary examples of market segmentation based on the identification of localized niches of broadband users and the provision of high-value and advanced broadband services. Market strategies are here localized both geographically and technologically. First, from the geographical viewpoint, the strategies of COLT and TOPIX find in metropolitan areas appropriate local markets because of the concentration of large firms and research centres, which are most important potential

235

cristiano antonelli and pier paolo patrucco

236

users of the broadband services provided by COLT and TOP-IX. Second, from the technological viewpoint the strategies of COLT and TOP-IX are localized because they are customized according to the users’ needs. Large firms and research centres can be in fact the typical users of the advanced broadband services provided and able to pay for ad hoc and value-added services and technologies based on broadband. Let us analyse the characteristics of COLT and TOP-IX broadband strategies more specifically. COLT Telecom Group plc is a European provider of broadband telecommunication business services with its Italian headquarters in Milan. Although COLT is an international player, its presence in Italy, and especially northern Italy, well represents the kind of strategies that players alternative to the incumbent can put in place in the Italian broadband market. Given the quasimonopolistic power retained by the incumbent, Telecom Italia, it is in fact clear the alternative players are induced to adopt segmentation strategies aimed at identifying rich niches not covered by the incumbent services. The location in Milan is in this perspective strategic for the development of COLT’s Italian market and the selling of its services to business users, in turn configuring COLT mainly as a metropolitan operator.4 COLT in fact found in Milan the appropriate environment in terms of both partners and preferential users, such as knowledge-intensive-business-service firms and high-tech firms. Consultants and business services are much more concentrated in this area than in the rest of Italy. In terms of finance, management, and support to R&D, the province of Milan business services amount to 18.7% of Italian business services. In terms of telecommunications and new technology business services, almost 30% of Italian business services are concentrated in this area. Finally, from the qualitative viewpoint, in the Milan metropolitan area large firms in the high-tech sectors such as IBM, Alcatel, ST Microelectronic, and Cisco can represent excellent partners and users for the services provided by COLT. In this context, the broadband infrastructures and services provided by COLT can be the means for important organizational opportunities for its clients, especially when considering connections to other European metropolitan areas. In fact, COLT is providing the first pan-European proprietary network on optical fibre and based on the Internet Protocol (COLT EuroLAN) connecting 32 metropolitan areas. Firms located in the Milan metropolitan area through this network can be connected Europe-wide, conveying voice, data, video, and Internet over a single network. E-business services are also provided through the COLT network and the hosting service provided by the Internet Solution Centre (ISC) in Milan, connected to 10 other ISCs in Europe.

Broadband in Italy

A different rationale is behind the experience of TOP-IX—Torino Piemonte Internet Exchange. TOP-IX is a consortium promoted by the CSI Piemonte (public centre for research, development, and experimentation in advanced Information and Communication Technologies and gathering the support of some of the main economic and research institutions at the regional and national level, such as Atlanet (Fiat Group), Sanpaolo IMI, Fastweb, Telecom Italia, Colt Telecom, and Noicom. TOP-IX is born with a strong orientation toward future technologies, in particular multimedia applications in multicast and IPv6. The consortium implemented the only one NAP (Neutral Access Point—first level Internet access point) in Italy, in order to favour the local concentration of Internet traffic. The NAP provides virtually unlimited band, very high-speed capacity, distributed access, and multimedia applications. TOP-IX adopted the WDM technology (Wave Division Multiplexing) as the suitable networking layer for the interconnection of the different nodes that represent the distribution backbone model. TOP-IX is constituted by different nodes distributed on the territory and linked by means of high speed backbones (>2.5 Gbps) that, exploiting the WDM technology, allow to share over the same fibre different technological platforms. Each one of these nodes is equipped with 10/100/1000 Mbps switches (Ethernet). The nodes are connected through a high-speed connection constituted by Gigabit Ethernet links assembled in a single logical channel (GigaEtherChannel) so as to fully exploit the available band. This approach allows coexistence on a common physical infrastructure of the three TOP-IX platforms for unicast Internet traffic, multicast Internet traffic (for the provision and transmission of multimedia contents5) and, on a totally experimental level, IPv6-based traffic, respectively. In this context, public administrations, research institutions, and large firms can be identified as the main potential users of the services provided by TOP-IX. In particular, two main applications of the services provided by the consortium can be more relevant. First, applications in the so-called eGovernment, such as the implementation of a unified desk for different administrative activities and online tax register and online tax payment services, reduce the complexity of administrative processes and increase the capabilities to share information between public and private users. Second, applications devoted to create a common layer enforce the co-operation between the Polytechnic School of Engineering, the University of Torino, and private firms. In this regard, the start-up of the experimentation of the IPv6 platform benefited from the agreement between two main local research institutions such as the CSP and TILAB (Telecom Italia Lab). Particularly the agreement between

237

cristiano antonelli and pier paolo patrucco

238

Topix and TILAB will provide firms and research institutions the access to the 6BONE at a first stage and subsequently to the EUROSIX (of which TILAB is a promoter company) and the U.S.A. IPv6 backbones (vBNS+, Abilene, 6TAP). conclusions The financial weakness of Telecom Italia and the geographic dispersion of the country have caused a major delay in supplying the country with a broadband infrastructure through the last years of the 20th century. This, in turn, has paved the way to a strong inter-modal competition where different groups of players have made systematic efforts to implement their own alternative technological platforms with conflicting innovation and economic strategies. Digital television seems more and more entrenched with little opportunity for broadband to become the basic distribution infrastructure. On the other hand, the slow diffusion of UMTS has reduced the prospect for a wireless scenario. The progress in VOIP technology and the fast penetration of Internet services in households and business have increased the levels of the demand for broadband services stretched to the curb. The increasing financial solidity of Telecom Italia provides the final element. After a decade of delays, the prospect for a rapid and extensive coverage of the country of broadband is now gaining momentum. The implications are important both in terms of a reduction in the digital divide within the country and also in terms of the prospect for a shakeout and new consolidation of the communication industry into a general multi-service technological platform. The role of antitrust may become decisive in the immediate future in order to preserve competition.

notes 1. Some new entrants do not bypass (e.g., they buy wholesale products from the incumbent or LLU). 2. See Section 5 for Fastweb’s strategy and Section 6 for two examples of OLOs’ strategies. 3. Shares at 25 January 2005. 4. The Italian headquarters of COLT is located in Milan, which represents a COLT Metropolitan Area Network (MAN) and also hosts an Internet Solution Centre. In Italy, COLT is also present in Turin, with the Turin MAN and Internet Solution Centre, while Rome is covered as a MAN. It is clear from this solution that COLT’s strategy in Italy is particularly aimed at covering main metropolitan areas. This is not the case, for instance, in Germany or France, where, as nodes of wider networks, relatively

Broadband in Italy smaller cities such as Essen, Leipzig, Nuremberg, Mannheim, Karlsruhe, Toulouse, Bordeaux, and Nantes are also part of the network. 5. The embedded multicast technology provides optimal delivery of streaming events, video on demand, and of multimedia applications in general, allowing the management and prevention of network overload.

references Antonelli, Cristiano. 2001. The microeconomics of technological systems. Oxford: Oxford University Press. ———. 2003. The economics of innovation, new technologies and structural change. London: Routledge. Bianchi, Marina, ed. 1998. The active consumer. London: Routledge. Formez and Osservatorio Banda Larga. 2003. La Banda Larga: Politiche, Utilizzi e Modelli di Sviluppo. Rome: Formez and Osservatorio Banda Larga. Fransman, Martin. 2002. Telecoms in the Internet age: From to boom to bust to . . . ? Oxford: Oxford University Press. ———. 2003. Telecom Italia: Italy’s high tech old-style telecoms incumbent. ICER Working Paper Series, no. 6/2003. Geroski, Paul. 1995. What do we know about entry? International Journal of Industrial Organization 13: 421-440. ———. 2001. Models of technology diffusion. Research Policy 29: 603– 625. Guerci, Carlo Maria, and Vito Marcolongo. 2003. Modelli di Business a Confronto nelle Telecomunicazioni: Il Caso eBiscom. Milan: Evidenze. Lancaster, Kelvin. 1971. Consumer demand: A new approach. New York: Columbia University Press. Madden, Gary, ed. 2003. Handbook on telecommunications. Cheltenham: Edward Elgar.

239

9 The Swedish Broadband Market Sven Lindmark and Per Björstedt

setting the swedish broadband stage This chapter describes the evolution and state of the Swedish broadband market. It is structured as follows. The first section introduces some aspects of Swedish demography, which has influenced government policy and in turn the development of the broadband market. Then an account of the evolution of Internet and emergence of broadband services is given, followed by relevant Swedish regulatory and policy initiatives. Section 2 introduces the broadband operators and shows how competition has developed, as well as the accompanying performance indicators (prices, data rates, and diffusion). Section 3 investigates how broadband services are used by Swedish households and businesses, including a discussion of digital divide. Section 4 looks at the supply of broadband, in terms of different network categories and technologies. Finally, Section 5 summarizes and discusses the main observations of the chapter. A first introductory note concerns terminology. There is no strict, generally accepted definition of broadband, either in Sweden or elsewhere. However, in practice, and for the purposes of this chapter, broadband has currently come to denote a fixed (always-on) Internet access with data rates higher than those offered through dial-up modem or ISDN access, that is, higher than (56 and) 128 kbps, in reality more than 500 kbps.1 Furthermore, dial-up modems and ISDN access typically require the user to call the number of an Internet services provider. When using PSTN (Public Switched Telephone Network), the telephone usually cannot be used for voice calls and Internet access simultaneously. The most commonly used broadband access techniques in Sweden are xDSL,2 CATV, and fiber LANs.

The Swedish Broadband Market

A Note on Swedish Demographics Sweden is a small nation population-wise, but a rather large country when it comes to surface area. Naturally, this has important implications for the structure of the broadband market in Sweden. It is therefore worthwhile to begin a country-specific analysis by providing some general background on the country’s geography as well as its demographic situation. To put things into perspective, consider the following comparison between Sweden and the state of California in the U.S. Sweden’s area is slightly larger than California’s. At the same time, the population of Sweden is around 9 million, whereas roughly 35 million live in California. Consequently, the Swedish population density is only 20 persons per km2, less than half of the world’s average and far below Japan and major European countries (which are above 100 persons per km2). It is also below the relatively sparsely populated United States. Moreover, major parts of northern Sweden, especially in the inlands and mountains (i.e., roughly half the total area), are very sparsely populated, and presumably unattractive from most network investors’ point of view. Obviously, constructing a nationwide broadband infrastructure in a sparsely populated region such as Sweden poses problems when it comes to investment costs and profitability. As we will see later on in this chapter, the Swedish government has taken measures to mitigate these problems by providing incentives for broadband construction and tax rebates for access, so as to reduce the risk of widening digital gaps in the country. In order to get an idea of the possible size and makeup of the market for broadband connections, one can note that the number of households in Sweden is around 4 million. Roughly half of these (54%) are located in apartment buildings and the rest are in detached houses and terrace houses (SCB 2003). The relatively high proportion of households in apartment buildings does to a certain extent counteract Sweden’s low population density in setting the conditions for the diffusion of broadband technology. On the same note, a majority of the population is found in the southern parts of the country, especially in and around the three major cities of Stockholm, Göteborg, and Malmö. As we will see, the diffusion of Internet access was initially higher in these areas than in the rest of the country. In recent years, however, this gap has diminished and, at the time of writing, the proportion of people with some sort of access to the Internet is basically equal across the nation.

241

sven lindmark and per björstedt

242

Evolution of Commercial Internet Services Among the aspects that have influenced the adoption of broadband technology in Sweden is the fact that the country had a particularly high penetration of Internet access even before broadband became available. This situation has come about in spite of the fact that commercial Internet service providers emerged later in Sweden than for example in the U.S. By the end of the 1980s, there were no commercial actors selling Internet access to companies and households. A user society called SNUS (Swedish Network User Society) developed plans for an IP network connecting universities and companies. The network was dubbed Basnät 90 (Mattson and Carrwik 1998) and the stateowned telecom operator Televerket was contacted to build it. Televerket offered to do so, but insisted on coordinating it with their communication service Datapak. SNUS, wanting to build a pure IP network, then turned to the small satellite communications company Comvik Skyport (later renamed Tele2) with their specifications. Comvik agreed to build the network if SNUS ran it for three years (with an option for Comvik to take over the network operations after three years). The network, then renamed Swipnet (Swedish IP Network), connected its first customers in March 1991. In total, around 30 companies purchased fixed Internet connections from Swipnet in the first year (Mattson and Carrwik 1998). During the latter part of 1991, Televerket (the Swedish PTT) introduced an IP network called TipNet. Later, a third commercial Internet operator, Transpac, owned by France Telecom, entered the Swedish arena (Sturmark et al. 1997; Statskontoret 1994:11). In 1994 Algonet, then a part of the Semic publishing company, started providing Internet access through modems for consumers/private users. Until then, all Internet providers had catered exclusively to academic and corporate users. In early 1994, Tele2/Swipnet charged SEK 1,500 ($190, PPP $162) per month for a dial-up connection, a price far too high for most consumers. Algonet used a fixed connection from Tele2, who welcomed the consumer initiative since they could profit from Algonet buying capacity while catering to customers not included in the Tele2 strategy at the time (Mattson and Carrwik 1998). Six months after the launch Algonet had 1,500 users, and it was soon followed by a number of other ISP start-up companies. When Tele2 noted the success of Algonet, it started offering Internet access to consumers as well, and soon Telia (Televerket was renamed Telia in 1993), followed suit. By then, entrepreneurial activities flourished in the emerging ISP industry. In 1996 there were 75–100 Internet service providers, with a combined market share of around 1% (Stelacon 1997). During 1995 the number of dial-up Internet connections reached 54,000 in Sweden, with a strong increase

The Swedish Broadband Market

in 1996. Since then the number of Internet connections has increased rapidly, until 2003 when it levelled off at about 3 million, corresponding to more than half the Swedish households. In a few years, Sweden became the leading country when it came to Internet penetration. The relatively rapid diffusion was aided by the service providers’ heavy subsidies on modems in return for one- or two-year contracts with customers. In addition to this, the rapid diffusion can to quite some extent be explained by a rapid diffusion of PCs, in turn aided by government initiatives. Parliament introduced tax reductions on computers bought by companies for their employees’ private usage (Steen 2002). This program quickly led to an increase in both absolute computer numbers and usage in Sweden, and Swedish PC penetration rates outpaced most other countries’ in the following years. A recent study by ITPS (2003), the Swedish Institute for Growth Policy Studies, indicates that the employee computer program has played an important role in accelerating diffusion rates of PCs in Sweden. A comparison with initiatives and diffusion rates in other Nordic countries suggests 10 percentage units lower penetration today without the program. In addition, the program has created a more even distribution of computers among income classes. The ITPS report also emphasizes the decreasing effect on diffusion rates in recent times and concurs with the recent government proposal of a reformed subsidy program. In line with the Internet development, more and more of the computers came equipped with modems and networking capabilities. Thus, the diffusion of computers, in turn propelled by government tax subsidies, paved the way for rapid diffusion of Internet usage. An international comparison indicates that Internet adoption levels were lower in Sweden than in the most advanced Internet nations (Finland and the U.S.A.) in 1995. In 1997, however, Swedish Internet diffusion had outpaced even the U.S.A. and Finland, and it has since stayed in a top position concerning Internet penetration levels in the world. In the early years of the 21st century, Sweden was still among the top three countries. In the late 1990s, intense competition led to price reductions in Internet access. Between 1998 and 2000, access fees were reduced by two-thirds, and “free” Internet subscriptions (no fixed, only variable charge) appeared in great numbers, offered by most ISPs. Although the number of ISPs has not increased dramatically since the mid-1990s,3 the combined market shares of the three largest actors, Telia, Tele2, and Telenordia/Algonet /Glocalnet,4 have gradually decreased in line with stronger competition from broadband Internet service providers, for example, Utfors, Bostream, Tiscali, and UPC. The market share of the three largest Internet service providers has decreased from 99% in 1996

243

sven lindmark and per björstedt

244

table 9.1 Internet Services Market Shares ISP

TeliaSonera Tele2 Algonet / TeleNordia/ Glocalnet Spray Bredbandsbolaget Tiscali Com hem Direct2Internet Bostream UPC Others

1996 (%)

1997 (%)

1998 (%)

1999 (%)

2000 (%)

2001 (%)

2002 (%)

2003 (%)

42 43 14

36 40 15

33 33 18

31 27 14

34 26 11

35 27 11

38 26 10

38 N/A 8

8

9

9 2 2

9 3 4

0 2 2 10

1 2 2 5

9 4 4 3 3 3 2 26

1

9

16

20

1 19

sources: PTS 2003a, 2004a:79. Tele2, having a market share of 26% in 2002 (PTS, 2003a), is not included in recent statistics.

to 56% in 2003. By mid-2004 the number of ISPs was estimated to be 150 (PTS 2004b). About 90 of these companies offer Internet access to private households (SIKA 2004). This implies a more intense competitive situation in the marketplace. Apart from the incumbent TeliaSonera (having a 30 – 40% market share over the last seven years) the market is fragmented (see Table 9.1). The Emergence of the Swedish Broadband Market The first commercial broadband services were launched in Sweden in the late 1990s. Already in 1995 Telia held its first broadband trials via ADSL. Later trials were conducted by cable TV operators and new entrants (PTS et al. 2002:13). However, 1999 was in many respects the year that marked the inception of the Swedish broadband market.5 The term “broadband to households” was introduced that year when, among others, Jonas Birgersson, part-owner and founder of the company Framfab, launched the idea of building new access networks to households in order to be able to offer high transfer rates and—at a later stage—broadband services. The same year, the Swedish government expressed ambitious goals for broadband development. It was established in the government’s IT Bill that the goal for Swedish IT policy would be that Sweden, as the first country, should become an information society for all.6 In the first instance this was to take place under the auspices of the market, but the state had an overall responsibility to ensure through various measures that an IT infrastructure

The Swedish Broadband Market

with a high data rate would become available throughout Sweden (see more below).7 The exact phrasing of the bill was somewhat vague, and consequently the bill was later perceived by the public as a promise of “broadband for all” (ITPS 2003). As a result, many were disappointed since they have still not received a broadband connection, and may not be able to do so for quite some time. While this is unfortunate, one should bear in mind that the goals in the bill were formulated at a time when everything that was even remotely related to information technology seemed possible. Since then, however, the market conditions have changed and unfettered optimism has been replaced by harsh reality as investment in communications technology has diminished together with stock market value. The development of broadband also enjoyed great attention from the media during 2000, and during the spring of that year hopes were expressed that at the end of 2000 there would be up to 1 million households connected. It turned out to be approximately 165,000 which, although lower than desired, was still a substantial increase from the year before.8 According to PTS (2001a:32) the growth could have been even stronger had the operators not established a de facto ceiling for how high the monthly charges should be, namely around SEK 200 ($24, PPP $22). Framfab founder Jonas Birgersson played an important role as a carrier and speaker of this vision and the SEK 200, “the price of a monthly bus ticket,” became something of an axiom (Engren and Kronberg 2003). As it turned out, however, this price was too low, and the operators felt compelled to compensate with long-term exclusive contracts with the customers, creating lock-in effects for the consumers.9 By the end of 2000, there were more than 15 operators offering broadband services. Two main groups of undertakings that offer broadband connections could be distinguished during that year: (1) cable TV operators (e.g., UPC, Com hem, Kabelvision); and (2) those operators who offer optical cable via newly laid property networks (e.g., Bredbandsbolaget). In addition, Telia started offering ADSL services by autumn 2000.10 Still, the market take-up was quite limited. By 2003, the number of broadband subscriptions had increased to almost 1 million, and the number of ISPs offering high-speed Internet access now exceeds those offering dial-up connections (PTS 2004a:46). This most recent development will be more closely examined in the following sections. Regulatory Developments During the 1980s and 1990s, the Swedish telecommunications market experienced several reforms and went through a number of changes. One of these

245

sven lindmark and per björstedt

246

was the commencing right to connect your own equipment to the network; and while this may not seem earth-shattering in itself, it paved the way for various new services that would rapidly gain importance. Among these were voice mail, fax, data traffic, and eventually the Internet. Furthermore, it meant that a line in the sand had been crossed since new actors were now allowed access and could start positioning themselves for future developments.11 A major milestone in Swedish telecom regulations was ultimately the telecom law (Telelagen) of 1993. Until then, the Swedish telecom market had essentially been an unregulated monopoly. In an international perspective, Sweden was an early starter in liberalizing its telecom market, and hence there was little prior experience to draw from when formulating the new legislation. As a result, the legislation was developed on a best-effort basis and it was then modified and improved as new needs arose. One year after the law was passed, the incumbent operator Televerket was turned from a public enterprise into Telia and in 2000 the company was listed on the Stockholm Stock Exchange when the state let go of 30% of its shares. From 1999 onwards, the government activity in the field of regulation has increased and Telelagen was replaced in 2003 by a law on electronic communication (Lagen om elektronisk kommunikation, EkomL). The new law takes account of telecom networks as well as cable TV and other broadband networks, and it contains relatively few general obligations for the actors; instead it enables the regulator, National Post and Telecom Agency (Post och TeleStyrelsen, PTS), to use certain tools if competition does not work properly or if consumers’ interests are neglected by the market. Regarding the access network, an EU directive of January 2001 states that actors with significant market power (SMP) are obliged to offer other operators access under non-discriminating conditions and cost-based prices. Access for other operators to the traditional access network can take place either via regulated access in the form of LLUB (full or shared access) or via bitstream (PTS 2004e, 2004f ). Recently, we have seen some examples of PTS taking action, primarily against TeliaSonera for not complying with the obligations on an SMP. During the second half of the year 2003, Bredbandsbolaget, among others, started to offer households fixed connection using LLUB. The operators that competed with TeliaSonera in the market for ADSL and other DSL technologies for households had to date primarily demanded wholesale-based, complete ADSL products, rather than access to the physical lines themselves (PTS 2004a:47). In November 2004, the markets for LLUB and bitstream access were considered by PTS to be relevant markets subject to regulation according to

The Swedish Broadband Market

EkomL, and PTS made a regulatory change for these markets. The major change was that TeliaSonera was required to provide access via bitstream (requiring smaller investments and thus expected to lower the entry barriers, particularly in sparsely populated areas). Also prices will be set by PTS. TeliaSonera immediately announced that they would oppose this decision in court.12 In conclusion, the Swedish regulator has had a history of being both lenient and tough with the incumbent, roughly corresponding to two time periods. In the first, the incumbent Telia was able to control the market, largely due to weak capacities of the regulator to enforce interconnection rates. In the second, the new legislation (since 2003) has given the regulator new powers over interconnection. Various filings and complaints against the incumbent have increased as well. The tendency to stiffen the grip on the incumbent started earlier, however, as symbolized by the fact that Telia was excluded from the 3G licenses in the beauty contest (in 2001). Government Initiatives and Broadband Policy Overview of Policy Initiatives

The Swedish government has played a large role in the Swedish Internet and broadband development, both directly and indirectly. Through the ownership of infrastructure companies such as Telia, Vattenfall (power generation), Svenska Kraftnät (power grid), Banverket (railroad), and Teracom (broadcasting), the state still controls a very large share of the total communication backbone systems. In 2003, the state controlled 78% of the high-speed network infrastructure, and municipalities and municipality-owned corporations controlled another 11% (PTS 2003b). IT has received an important place on the political agenda in Sweden in recent years. As mentioned earlier, the overall goal for the Swedish IT policy is to be the first country to become an information society for all (Ministry of Industry, Employment and Communication 2002). More specifically, three areas are prioritized in Swedish IT policy: IT trust; competence to use IT; and accessibility to IT services. In recent years, stimulating activities have been conducted in all three areas, summarized in Table 9.2. In the following section, we will concentrate on the ones most relevant for broadband development. Broadband Policy

The present broadband policy has its roots in the IT infrastructure study (Government Official Report 1999b:85) and the IT Commission’s vision report

247

sven lindmark and per björstedt

248

table 9.2 Selection of IT Initiatives by Public Authorities IT Trust

Competency to Use IT

Accessibility of IT Services

– Jointly developed standards for electronic signatures by public administration authorities

– Tax reductions on company computers bought for employees’ private usage

– Tax reductions for broadband access installations

– ITiS project (SEK 1.7 billion during 3 years) to raise IT literacy among schoolteachers

– Proposal of opening up channel space in stateowned infrastructure for telecom operators

– New electronic communications laws, placing Internet communications alongside television and radio

– State-owned companies obliged to build high-speed backbone infrastructure Implementation of the “24/7 authorities”

source: Compiled from various government reports and proposals.

(Government Official Report 1999a:134). The assignment for the IT infrastructure study included providing proposals for the role that the central government should adopt, partly in order to promote industry policy goals and partly to promote the regional and social goals in respect of access for everyone (ITPS 2004:67). The approach proposed by the IT infrastructure study was that the government should assume a special responsibility for access to broadband in the sparsely populated parts of the country. The proposals had the effect, among other things, that a national infrastructure program was drawn up and that the government would enter into discussions with the Swedish Association of Local Authorities and the Swedish Federation of County Councils on voluntary measures for the supply of broadband within their borders (ITPS 2004:67). In order to achieve the ambitions declared in the IT bill, a national infrastructure program was prepared that, among other things, contains proposals for a number of state-supported measures aimed at facilitating investment in, above all else, sparsely populated parts of Sweden (PTS 2001a:32). The proposed measures that were associated with financial undertakings included (1) support for regional and local telecommunications networks in areas that were not supplied with ICT infrastructure through the market, (2) tax allowances for natural and legal persons for broadband connections, and (3) a basic network with high levels of access for all municipalities in the country (ITPS 2004:67). After the Parliament had made its decision, a national IT infrastructure

The Swedish Broadband Market

program was drawn up and the first ordinance entered into force on 1 July 2001. This and subsequent ordinances laid down, among other things, the conditions (co-financing and an approved IT infrastructure program) for receiving support for networks that link different places together and for area networks. Initially, support was intended to continue until 2004, but this has been extended to 2005 (ITPS 2004:67– 68). In August 2000, the government commissioned the Swedish National Grid (Svenska Kraftnät, the public utility that operates Sweden’s electricity grid) to construct a high-capacity national backbone network. It was to link all the municipalities in the country (290 in 2001) and be open to all operators. The utility was instructed to install fiber optic cable on its own national electricity grid. Where necessary, the work was to be undertaken in co-operation with other regional network owners and be completed by 31 December 2002. Network capacity in the form of “dark fibre” (unlit optic cable for leasing) was to be made available at market prices. By March 2003, a total of 215 municipalities had hooked up to the national network. The network was to supplement existing backbone networks owned by both private and public stakeholders. The expansion work is market-based and does not draw on public funds. Since February 2003, those municipalities not connected by the National Grid via market channels have been able to apply for government funding for a backbone network connection (ITPS 2004). By October 2003, only 23% of the total SEK 3.1 billion ($400 million, PPP $333 million) subsidy funds had been distributed to municipalities (ITPS 2003). Another form of support is tax reductions of up to SEK 5,000 ($650, PPP $531) per house or apartment for certain costs associated with communication access installations (PTS 2003b). Future policy will be laid down in a new government bill planned for 2005. Currently (January 2005) a lot of preparatory work is being conducted, not least by the IT Policy Strategy Group, appointed in 2003. It is too early to speculate on what this bill will include. To the best judgment of the authors of this chapter, the magnitude of government engagement is not likely to change. When it comes to direction, indications are that the development of IT-based services will become a new focus area.13 the operator/isp market Overview Currently, there are almost 100 broadband service providers in Sweden. A majority of these are small and local, and also differentiated in terms of market

249

sven lindmark and per björstedt

250

table 9.3 Broadband Market Shares ISP

TeliaSonera Bredbandsbolaget Com hem Bostream UPC Glocalnet Spray Others

2000 (%)

2001 (%)

2002 (%)

2003 (%)

41

54 15

54 12

20

9 11

39

11

9 8 0.1 1 16

42 14 11 9 7 3 2 12

source: PTS 2004a:79. Broadband includes xDSL, PLC, cable TV, radio, satellite, and other fixed access forms. During 2004 Bostream and Bredbandsbolaget merged. Tele2 is also excluded from these statistics but had only a 3% market share in 2002.

segments served and access technology used (PTS 2004a:46). Still, as can be seen by comparing Table 9.1 and Table 9.3, the market for broadband access is less fragmented than the total ISP market. The three largest broadband suppliers, that is, TeliaSonera, Bredbandsbolaget (BB), and Com hem, together hold more than 75% of the market. This market share has been fairly stable over the last three years. The major broadband players are briefly summarized below. TeliaSonera, the state-controlled14 but partly privatized former monopoly, dominates the broadband market with a share of more than 40%, down from more than 50% in 2002, largely due to the fact that Telia had to divest Com hem when merging with Sonera.15 Bredbandsbolaget was started in 1998 as a development project in Framfab (in turn founded by the Swedish Internet entrepreneur and broadband evangelist Jonas Birgersson). For a number of years it has been the primary competitor to Telia. In 2004, it acquired Bostream and strengthened its position as the number two broadband provider with a 23% market share. Bredbandsbolaget has a strong position particularly in the larger cities (Stockholm, Göteborg, and Malmö), where according to some sources it has a higher market share than TeliaSonera. Currently the main owners of Bredbandsbolaget are one private equity firm and two investment companies, among them Investor controlled by the Wallenberg family. Bredbandsbolaget started showing operating profits in 2004. The CATV operator Com hem, previously a part of Telia, started providing broadband in 1999 and was divested from Telia as a result of the merger with

The Swedish Broadband Market

Sonera. It is now owned by the venture capital firm EQT. Today Com hem offers CATV, digital TV, broadband, and telephony. UPC Sverige AB (UPC Sweden) provides CATV services and, in addition, broadband communications services under the “Chello” brand. It is a subsidiary of UGC Europe, Inc., and part of its UPC Broadband division. UGC Europe in turn is a wholly owned subsidiary of United GlobalCom, Inc., a broadband communications and entertainment company. TDC Song Networks started out as a virtual operator named Tele1 Europe. It began building out its own network in 2000. When Sonera and Telia merged, Telia’s activities in Finland and Sonera’s in Sweden were acquired by Tele1 Europe, which thus became a major Nordic player, renamed Song. It runs an integrated network (backbone network and local access networks) in the Nordic countries. Song has focused on the business and capacity resale markets. Recently the company was acquired by TDC, after a bidding process where also Tele2 placed a bid on Song. Among the minor competitors Tele2 is worth mentioning. The company was established in 1993 and is the major incumbent attacker and by far the second largest operator in Sweden. However, to date it has had a small impact on the broadband market. The reluctance as regards ADSL may have been a result of unjustified excess pricing of the ADSL services charged by Skanova (Engren and Kronberg 2003:49). Yet another minor but aggressive player is Glocalnet. The company was formed in 1998 and has taken over TeleNordia’s private customers. Its main owner is Norwegian Telenor, which has taken over Utfors. Glocalnet is targeting the broadband market, with an outspoken aim to become the third largest player. Rivalry The Swedish broadband market is polarized—with one and recently two dominant ISPs (TeliaSonera and Bredbandsbolaget), a few other large players (see above), and a multitude of small ones (ADL 2003). According to PTS, the dial-up Internet access market has been rather competitive; but with the gradual transfer to broadband, the control over the access network has increased in importance and, as a consequence, broadband competition is more limited. This applies especially to xDSL services, as these make use of TeliaSonera’s copper access network (PTS 2004a:47). Competitors claim that Telia’s pricing strategy effectively shuts them out from competing profitably in the ADSL broadband market—that Telia set resale prices too high compared with their prices to end-customers, and thus effectively squeezed competing operators’ margins (PTS et al. 2002:34 –35). In addition to pricing, Telia has been ac-

251

sven lindmark and per björstedt

252

cused of other anti-competitive behavior, such as deliberately delaying other operators’ installation of broadband. Other factors restraining competition have been house owners’ limited choice regarding network providers and customers’ switching costs for ADSL (technical difficulties, long subscription contracts, high equipment investment costs) (PTS et al. 2002:36 –39). Not all operators are dependent on accessing TeliaSonera’s network. The cable television operators currently have the opportunity of exclusively offering Internet access in their networks. Similar preconditions apply to those operators who offer Internet access via LAN networks, that is, Internet access with high transmission capacity via newly installed property networks in multiple-occupancy buildings (PTS 2004a:47). Rivalry in the capacity resale market has so far been limited but is gradually increasing. In Sweden, almost 80% of radio and fiber networks are owned by state-owned companies (including TeliaSonera, which accounts for more than 60%). Of the remainder, municipal companies and municipalities own just over 10%. A large number of the municipalities that own networks are found in sparsely populated areas (e.g., in the north of Sweden) (SIKA 2004). TeliaSonera and Song Network are nationwide backbone network operators. Other operators that own networks throughout a large part of Sweden include Banverket (whose principal activity is operating railroad infrastructure), Teracom (broadcasting infrastructure), Vattenfall (electricity), Svenska Kraftnät (electricity), and Utfors (Internet, owned by Telenor). In addition, there are a number of smaller operators (SIKA 2004). During 2003, Song started offering access to its network for operators active in the broadband consumer market and signed agreements with, among others, Vattenfall and Universal telecom. Thereby it broke TeliaSonera’s (through Skanova) long-term monopoly on this market.16 Diffusion, Tariffs, and Data Rates Sweden was an early leader in the take-up of broadband networks, having the third highest penetration in the OECD (Organization for Economic Cooperation and Development) by 2001. Two years later, by late 2003, Sweden had fallen to eighth place. The early rapid take-up of broadband coincided with Sweden having among the lowest broadband access prices in the OECD in 2001. Prices subsequently rose at a time when they were generally falling in other countries (OECD 2004b:53–54). As indicated in Table 9.4, diffusion rates slowed down slightly in 2002 and the first half of 2003, to gain pace again later in 2003. In 1998, when broadband access was only delivered on a trial basis, uncertainties were high concerning customers’ willingness to pay, as well as concern-

The Swedish Broadband Market

253

table 9.4 Internet Subscribers Per Access Technology (Thousands) 6/30/00

12/31/00 6/30/01 12/31/01 6/30/02 12/31/02

6/30/03 12/31/03 6/30/04

PSTN 1,903 1,987 2,093 2,233 2,264 2,235 2,187 2,149 1,925 ISDN 107 126 135 133 123 117 102 90 76 xDSL 4.7 26 123 242 337 421 484 581 661 – ADSL 417 478 566 641 – VDSL 7.8 12.3 PLC (power 0.0 0.3 0.4 0.4 0.3 0.1 line) Cable TV 34 56 92 112 128 157 179 212 229 Radio 0.2 0.6 1.4 2.5 3.0 4.3 6.2 6.3 Satellite 0.5 0.7 0.8 Other 10.7 46 64 106 136.0 151 177 204 217 – of which 141 167 194 208 LAN Total Broadband 4 49 128 280 461 603 732 845 1,004 Total 2,059 2,242 2,508 2,827 2,990 3,084 3,134 3,243 3,115 Internet source: PTS 2004b.

ing pricing models and to what degree operators would be able to distribute and charge for content, such as video-on-demand. It could be mentioned that Telia contemplated tariffs of SEK 600 per month ($75, PPP $63) and, in addition, SEK 0.5 (6.35 cents, PPP 0.052 cents) per Mbyte in both directions or minute charges. During the early competitive situation in 1999, in particular the rivalry between Telia and Bredbandsbolaget, operators concentrated on closing broadband deals with house owners. This often meant that they helped to finance build-out of networks, in return for an exclusive deal to deliver services over these networks. This rivalry became largely based on price. In July 1999, Bredbandsbolaget closed one of the first broadband-to-household agreements, a deal with HSB, one of the largest real estate owners in Sweden. The terms for consumers were SEK 2,000 ($235, PPP $214) for installation and a flat rate of SEK 200 ($23.5, PPP $21.4) per month. This deal more or less set the standard for pricing in Sweden. A few months later, Telia closed a deal in Stockholm for 35,000 tenants with SEK 995 ($117) for installation and a monthly fee of SEK 200 ($23.5, PPP $21.4) (PTS et al. 2002:19,36 –37). One main driver in Sweden’s early success with broadband diffusion was the new entrant (Bredbandsbolaget) that began offering high-speed access at relatively low prices, which in turn made the incumbent Telia’s own offers far more competitive. Bredbandsbolaget began by offering broadband access (10 Mbps)

sven lindmark and per björstedt

254

to apartment buildings, using Ethernet LANs, at what by then was the lowest price in the OECD. In response, Telia also offered relatively inexpensive broadband access compared to operators in other countries. As the market developed, Telia began to differentiate its prices for apartment buildings and individual dwellings that were not served by Bredbandsbolaget (OECD 2004b:53–54). As mentioned before, prices were initially based on operators’ expectations of revenue streams coming from content and of more rapid adoptions (PTS et al. 2002). These prices were too low, and the operators felt compelled to compensate with long-term exclusive contracts with the customers, creating lock-in effects for consumers.17 In the spring of 2001 several of the network operators increased their monthly charges.18 Telia has increased its charge for ADSL from SEK 250 per month to SEK 325 in September 2001 (corresponding to $24 and $31, respectively, at that time, PPP $27.8 and $34.8), and then to SEK 375 ($36, PPP $40.1) in March 2002. Telia claimed that price rises were necessary to cover the cost of expanding DSL access to areas not covered by their initial deployments (OECD 2004b:53–54). UPC increased its monthly charge from SEK 250 ($24, PPP $26.7) to SEK 299 ($29, PPP $31.9) and Bredbandsbolaget from SEK 200 ($19, PPP $21.4) to SEK 320 ($31, PPP $34.1) in March 2002 (OECD 2004b:53–54). The problem with many of these changes was that they affected not only new subscriptions but existing ones. Consumers questioned whether the service providers had the legal right to increase prices in this way (PTS et al. 2002:22). In spite of these tariff rises, Swedish broadband tariffs remained rather low in an international perspective. During the second half of 2003, the Swedish market showed signs of becoming increasingly competitive again. One influencing factor was probably that Telia had to divest its CATV company Com hem following its merger with Sonera. Com hem then quickly started to offer broadband services, significantly undercutting the prices of TeliaSonera (OECD 2004b:53–54). There has also been a trend toward offerings with higher data rates. For example, Bredbandsbolaget is primarily offering services with high data rates. Bostream offers Scream, a VDSL service with up to 26 Mbps, although available only if located within 300 metres of an exchange. For long TeliaSonera did not seek to offer exceptional broadband access speeds. In late 2003 they launched a new range of higher-speed broadband access offers, available in major cities in Sweden, and where municipal authorities are investing in ADSL technology as part of the Swedish government’s broadband initiative or (for the VDSL service) in Stockholm only (OECD 2004b:53–54). Here it should be mentioned that the capacities of the broadband connections have been claimed to be exaggerated. For the purpose of counteracting faulty claims and raising customer access to correct performance measures, the

The Swedish Broadband Market

Swedish IT Commission developed software that allows users to measure the performance of their own connections. The software is called TPTEST and was made available on a number of authorities’ websites (PTS et al. 2002 : 25). Before turning to the most recent events, we take a brief pause to compare the broadband penetration in Sweden with other countries at this point in time. Although Sweden’s position had declined somewhat since 2000 –2001, Sweden was still ranked number eight among the OECD countries in late 2003. It is not clear how these figures translate into household penetration for the other countries, but for Sweden the household penetration was estimated to be 20% by then (PTS 2004a:48). Sweden also has an extremely high LAN access penetration while cable access is relatively low.19 During the autumn of 2004 competition seemed to intensify further. Several operators either lowered their installation charges (e.g., Telia and Bredbandsbolaget), reduced them to zero (Glocalnet), offered free modems (e.g., Spray), raised offered services with higher data rates (Bostream), or competed with other seemingly attractive offers (one month for free, etc.). During this period, broadband services were extensively marketed in the media. These are all indications that competition intensified, although the effects on broadband diffusion remain to be seen since no data for the second half of 2004 were available at the time of writing.20 How competitive are these offerings in an international perspective? It has been notoriously difficult to obtain any comparative tariff data. However, according to a study made by IDC for the Danish IT and Telecom Agency, in September 2004, Sweden ranked among the cheaper European countries, when it comes to broadband tariffs. The study compiled average tariffs for three broadband categories (512 kbps, 2 Mbps, and 4 Mbps). Swedish tariffs were 273 Dkr ($44.8, PPP $32.2), 338 Dkr ($55.4, PPP $39.9), and 359 Dkr ($58.9, PPP $42.4) per month, respectively (including taxes). This ranked Sweden as the third to fifth cheapest broadband country in Europe, with only France and Belgium being less expensive in all three classes.21 Thus it seems fair to claim that Sweden’s prices are comparatively low. Why this is the case, and how rivalry may evolve in the future, will be further discussed in the last section of this chapter. the demand side Households’ Broadband Usage In 2001, Swedes used broadband for essentially the same kind of services as dial-up access users, only more intensively. Between 80% and 90% of clients used broadband access for e-mail and general information search, more than 50% for gaming, over 40% for music listening, and so on. For some services,

255

sven lindmark and per björstedt

256

differences were larger. Differences in use between broadband and dial-up access were highest for web-TV/video downloading (roughly five times more), web radio/music downloading (almost three times more), and instant messaging services (more than two times more).22 In a more recent survey for 2003, similar patterns were revealed. Internet in general was primarily used for (1) email, (2) hobby/personal interest-related search for information, (3) banking, (4) general surfing, (5) a group of activities related to news and entertainment, and (6) a group of activities related to authorities. Also in 2003, Swedish broadband users were more active in all service categories than were dial-up users, the discrepancy being largest for music, dating, radio, TV, e-commerce, and entertainment. The difference was smallest when it came to more “useful” services such as e-mail, access to government authorities, medicine, stock trading, and working at home (WII 2004). According to a survey carried out by the National Post and Telecom Agency in collaboration with TEMO,23 two out of ten dial-up users claimed they might sign up for broadband access within the forthcoming three months. The respondents were also asked to state their reasons for switching from dialup to broadband access. The most common argument was the desire to be able to make and receive phone calls while connected to the Internet. Next on the list was the wish for higher speed for Internet surfing. Moreover, one-third of the respondents believed they used Internet to a sufficiently high degree to be able to reduce costs by switching to fixed-rate broadband access. Among those who expressed no plans for ordering broadband, the two most common reasons were that they do not use the Internet so much and that broadband access is too expensive (PTS and TEMO 2004). As far as drivers for broadband access are concerned, one application currently in focus among broadband suppliers is IP telephony. During late 2002 some broadband providers, among them Bredbandsbolaget and Digisip, launched IP telephony subscriptions. Thereby, customers could discontinue their analogue Telia subscriptions (PTS 2004b). According to PTS there were around 20,000 IP telephony subscriptions in Sweden in mid-2003, up from around zero six months earlier (PTS 2003a). One year later the number of subscribers had increased to 56,000 (PTS 2004b). The vast majority of those belonged to Bredbandsbolaget, and 49,000 were LAN subscribers (PTS 2004b). Telia only launched its services in February 2004 (no subscriber figures available) and cable TV provider Com hem entered the market for IP telephony during the latter part of that same year. Also worth mentioning are some smaller companies, including DigiSip and IP-only. However, considering the price of DigiSip when it was acquired by Labs2 in May 2003 (SEK 2 million,

The Swedish Broadband Market

$250,000, PPP $212,000), it can be assumed that the company has very few subscribers. Com hem, on the other hand, has the potential to become a significant player in this field due to its relatively large existing customer base of 1.4 million Swedish households. The company will start by offering IP telephony to subscribers in Linköping and Lund, and move on to Stockholm in 2005. In addition to the ISPs offering IP telephony, rapidly growing Skype and other VoIP networks further drove demand for broadband access (Svenska Dagbladet 2004). When it comes to future services, it should be noted that service providers have long been exaggerating what new kinds of bandwidth-consuming content services will be driving the demand for broadband in the future. A more nuanced picture has lately emerged, where broadband is used first and foremost for traditional Internet services. Still the perception among several broadband ISPs seems to be that television (part of the recent “triple play” concept) may become an important application of broadband in the future. Broadband Business Usage As is the case for households, Internet penetration among businesses is elevated in Sweden. In 2001, the latest year for which we could find comparable data, 90% of Swedish companies had Internet access, placing Sweden fourth among OECD countries, trailing only Denmark, Japan, and Finland. In 2003, 95% of all companies with 10 or more employees used the Internet. Differences between various sectors were not very large, with the transport sector having the lowest percentage (86%) and the energy and financial sectors the highest (99%). Company size seemed to play an important role in determining whether or not companies use the Internet. The smaller a company is, the lower the percentage that uses the Internet. It is used by 92% of companies with 10 –19 employees, 95% of companies with 20 – 49 employees, and 100% of large companies. About half (52%) of those employed work with computers that are connected to the Internet.24 The majority of companies connect to the Internet via broadband of various speeds: 18% with a speed less than 2Mbps; 23% more than 2 Mbps; and 32% with ADSL. Both modems and ISDN are used by more than one-third of companies. There is also a diminishing share of companies (1%) that connect via satellite. Compared to the previous year, fixed broadband connections have increased at the expense of modems and ISDN.25 When it comes to broadband connections, the numbers vary significantly between different sectors, with the energy and financial sectors having the largest broadband penetration (90% and 88% in January 2003). The difference

257

sven lindmark and per björstedt

258

is also noticeable between large and small firms, where companies with more than 500 employees had 97% broadband penetration as opposed to the average of 62%.26 As far as Internet usage is concerned, 94% of Swedish companies reported that they used the Internet as a means of searching for general information. Carrying out financial transactions was second on the list (80%), while obtaining service and support came in third with 75%.27 To sum up, household users tend to emphasize the desire to be connected to the Internet while still having access to the phone line as a reason for switching to broadband. Neither households nor business clients seem to use particularly bandwidth-consuming services to the extent that this drives their demand for broadband. On the contrary, Swedes use broadband for basically the same services as they do dial-up access, but they use these services more intensively if they have broadband connections. Regarding future developments and potential killer applications driving the demand for broadband in Sweden, the situation is difficult to assess. On the one hand, suppliers of broadband connections stress the diffusion of IP telephony as a possible future money-maker. On the other hand, consumer demand for broadband seems to rely more on convenience issues such as surfspeed and cost. Digital Divides In the debate on digital divides in Sweden, one of the more recent contributions comes from the workgroup for IT and democracy at the Ministry of Justice. In a report published in February 2004, the workgroup provides an assessment of the present situation when it comes to digital gaps and suggests a strategy to decrease them. One of the more interesting single ingredients of this strategy is the proposal to include broadband access to the Internet and computer ownership in the social security norm (Justitiedepartementet 2004). The present situation has also been analyzed by ITPS in an attempt to pinpoint the factors that create digital divides and to evaluate their size and development over time (ITPS 2004). The factors that have been investigated are gender, income, education, place of origin, and age. Among those who had access to the Internet in 2000, there were just as many men as women. On the other hand, age seemed to play a major role as a creator of digital divides. However, the greatest impact on Internet access and use seemed to come from the level of education. Three years later, in 2003, the differences are considerable and education still causes the largest gap in Internet access.28 Moreover, age differences play a significant role, while the importance of income is less

The Swedish Broadband Market

and of the same size as in 2000. As a matter of fact, according to an OECD outlook examining ICT diffusion in general, Sweden is somewhat of an exception on this point. While the gap related to income has widened in six countries analyzed by the OECD, the trend has been reversed in Sweden where it has declined by comparison with the late 1990s (OECD 2004a). Gender differences with regard to access to computers and the Internet are generally low in Sweden; but when it comes to place of origin, native Swedes seem to have a certain head-start as compared to immigrants (SIKA 2004:135). Indications are that at least the “gender divide” is smaller in Sweden than in many other countries. Proceeding to Internet use, there are indications that the largest divide refers to age and that the importance of this factor has diminished only marginally over the period in question. The education divide is the second largest in both years and shows only a minute decrease. The divide between the sexes has been halved, with the effect that in 2003 it is the smallest divide. The income divide is relatively small but is almost the same in both years. The divide explained by ethnic origin has increased considerably; from being the smallest in 2000, it now ranks among the factors in the middle of the list (ITPS 2004:31). Thus far in our analysis of the digital divide in Sweden, we have mainly addressed gaps in Internet access and Internet use in general. In fact, for the purposes of this book, the problem with the ITPS analysis is that it does not explicitly treat broadband gaps. Furthermore, and equally relevant, the study does not treat gaps between different parts of the country. In order to attend to this, we turn to a study carried out by Swedish Statistics (SCB 2004a). According to this study, the digital divide is actually much higher when it comes to broadband. For example, while there is no gender divide for Internet access in general, it is significant for broadband. Moreover, the age divide is considerably larger for broadband. As far as geography is concerned, Nordicom provides data indicating that Internet access penetration was higher in Stockholm and other major cities than in the rest of Sweden in the mid-1990s. In recent years, however, the gap has decreased and in 2003 the numbers were basically the same for the different regions (Nordicom 2004:11). Finally, on the subject of digital divide in Sweden it is worth mentioning that by 2003, more than half of all the built-up areas in Sweden had an area network in some part of the community, a substantial increase from the year before. Of the built-up areas that lack area networks, 85% had fewer than 1,000 inhabitants (SIKA 2004).

259

sven lindmark and per björstedt

260

supplying broadband: network build-out Networks based on fiber optic cable and radio links now total 175,000 network kilometers. This is an increase of 10% compared with the previous year. The increase has been in fiber. In addition to fiber optic cable and radio links, there are other techniques, such as coaxial cable and copper wire (where they have been upgraded), that can also be used for high-speed networks (PTS 2003b). In the access network, initially broadband connections were fairly evenly distributed between ADSL, cable TV, and LAN. Several operators, Bredbandsbolaget in particular, aimed for the LAN, due to the uncertainties involved in ADSL access to Telia’s networks and the prices Telia would charge through its network capacity wholesaler company Skanova. Broadband access through cable TV networks was reserved to the cable TV companies. As a result, presumptive broadband operators were rushing in to close deals with landlords and real estate companies. Recently xDSL has been taking market shares from the other technologies. In the second quarter of 2004, for example, some 75% of new subscriptions were xDSL (PTS 2004c:10), of which a large majority was ADSL. Since mid2003, xDSL via LLUB has grown rapidly, and by now it constitutes the most rapidly growing form of broadband access, still at low level (PTS 2004c:10). The FTTH market has been notoriously difficult to map. Table 9.4 indicates that there were less than 217,000 subscribers in June 2004. Telia claims to have 80% market share in fiber access connections (physical cables). They do not know how the remaining shares are distributed.29 Alternative network technologies (besides XDSL, LAN, and cable) have from time to time been given much attention but have been adopted to a very low degree. For instance, during 1998, much attention was paid to the use of the power lines as an access network for telecommunications services (PTS 2000a:63). However, the technology has taken longer than anticipated to develop, and is currently (late 2004) offered on a limited basis only. Satellite is also used only on a limited scale, to customers with no other option for broadband access. FWA (fixed wireless access) allows large broadband coverage at a relatively limited infrastructural cost. PTS (2004c) states that, from the consumers’ viewpoint, the investment cost is rather large, and FWA is therefore likely to address the corporate segment initially. The technology that in recent years has gained most attention is WLAN. WLAN access is offered in several ways and with different business models. The most common ones seem to be (1) public access through an operator; (2) broadband access through an ISP; and (3) private access. In 2003, there were six public WLAN operators: Telia (more than 700 surfzones all over Sweden);

The Swedish Broadband Market

Powernet (around 50 in South Sweden); Firstnet (circa 10 in three towns); Default (7 in Stockholm); Amazing Ports (Stockholm); and Wirelessbolaget (1 in Stockholm). In addition, there are a number of enthusiast establishments (e.g., Stockholm Open, Elektrosmog, and Nora Wireless) offering WLAN access for free. Telia’s Homerun service, launched already in 1999, is by far the most ambitious one in Sweden, including also roaming agreements with nine countries in Europe. Fixed WLAN-based broadband access is offered to households by a handful of ISPs, including some of the abovementioned.30 Even more recently the WiMax technology has been much hyped. In the autumn of 2004, it was offered for the first time on a trial basis to households in villages outside Skellefteå in northern Sweden. The trial project is conducted by Mobile City, Skellefteå municipality, Skellefteå Kraft, Avarion, Teracom, and Intel.31 conclusions Most international rankings place Sweden among the top nations when it comes to ICT penetration, usage, readiness, skills, and so on. Broadband has been no exception, although the position has deteriorated somewhat since 2002. In sum, the Swedish broadband market displays: (1) an early start with initially very low and still relatively low prices; (2) extensive backbone networks owned by public entities; (3) a strong political will and measures taken to implement broadband; (4) a relatively high penetration rate; (5) ADSL access dominance, with relatively high proportion of LAN access and relatively low proportion of cable access; (6) limited adoption of alternative technologies so far; (7) a polarized market with a strong incumbent, a few major national players and numerous smaller, often local, players; (8) recent regulatory change lowering entry barriers; (9) a high Internet and PC penetration when broadband emerged; and (10) an ongoing substitution of dial-up modem for broadband access. Clearly, historical path-dependencies have influenced the outcome so far. One could say that Sweden had a high broadband readiness, not least because of the outstanding PC penetration, which first facilitated Internet and later broadband adoption. The incumbent’s extensive build-out of a modern backbone and access network also contributed to this readiness. The presence of other actors’ backbone networks also mattered. The practice of using public infrastructure for private services is not new in Sweden. Tele2 closed a deal with the Swedish Railway Administration (Banverket) already in 1989 to install and operate a national fiber optic backbone network.

261

sven lindmark and per björstedt

262

The private sector, here represented by Tele2, thus used public infrastructure to challenge the telecom monopolist (Televerket / Telia). Government also played a key role as a proactive player, not least when it came to network build-out. The established policy to make Sweden the first “information society for all” manifested itself most clearly in the aim to ensure through various measures that an IT infrastructure with a high data rate would become available throughout Sweden—in particular through the commission of the Swedish National Grid to construct a high-capacity national backbone network to link the municipalities, as well as through the concomitant financial support and tax reductions provided. These actions mitigated, and were also triggered by, the profitability problems in providing universal broadband services in sparsely populated parts of the country. The presence of agile, possibly disruptive, new entrants clearly contributed to the dynamism. In particular Bredbandsbolaget, which set the initial very low price standard, probably contributed to the ambitious policy and media hype as well. Although the low tariff structure could not be sustained, it kickstarted the Swedish broadband market. The incumbent’s (Telia, later TeliaSonera) position has been very strong in Sweden, partly as a result of the virtual monopoly of the local loop. In some sense, this strong presence may have counteracted the development of broadband, where competition has been less intense than in the modem Internet market. Through its network company Skanova, Telia has been claimed to have squeezed the margins of competitors. On the other hand, Telia has a long tradition of being an early adopter of new technologies. It also responded swiftly to threats of new entrants, by implementing and offering ADSL at low prices. The fact that Telia was not hit very hard by the telecom bust also aided a more aggressive ADSL strategy. After two years (2002–2003) that could be characterized by less intense competition, it seems to have increased again. Telia’s divestment of Com hem, as well as recent regulatory decisions, has apparently contributed to this. The general opinion seems to be that it is vital to establish a strong market position soon, or the competitive race may be lost. We can therefore expect that companies like Tele2 will try to strengthen their positions. Demand for broadband appears to have been less driven by unique killer applications, although broadband users are more active than modem users in some bandwidth and usage-intensive applications like music downloading and dating. The main services are still traditional ones (e-mail, etc.) and the main reasons for adopting broadband instead of modem access are freedom to make simultaneous phone calls, general need for higher bandwidth, and in some cases cost reductions. A recent driver for broadband adoption is IP telephony, or at least this seems to be the suppliers’ view.

The Swedish Broadband Market

Finally, speculating somewhat into the future, we believe that the Swedish broadband industry will become increasingly dynamic, as new technologies are introduced and old ones are improved. The Swedish market may be especially receptive to new wireless broadband technologies and services (higherspeed 3G, WiMax, etc.). The competitive landscape is likely to change, with Telia’s strong position in ADSL being challenged. Mergers and acquisitions are also likely to take place. A mixture of need for convenience, increased competition in voice services, triple play, as well as new bandwidth-hungry services will drive the transition from narrow to broadband Internet access and usage. It is not a daring prophecy to say that the future of communications in Sweden will be broadband (and mobile). acknowledgments We are grateful to Erik Andersson and Erik Bohlin for providing useful input and comments.

notes 1. Other, stricter definitions have been used such as the one proposed by the Swedish Broadband Committee (in Swedish, IT-infrastrukturutredningen), which defined broadband as a tele- and data-communication transfer capacity of at least 2 Mbps in both directions (to and from the user) (see, e.g., Government Official Report [1999a]:164 –165). However, this stricter definition seems to have been more or less abandoned by the time this chapter was being written. 2. xDSL is used to denote various DSL techniques including the most frequent ADSL and VDSL. 3. In 2004 there were around 150 companies providing Internet access to companies and consumers (PTS 2004b). 4. Telenordia acquired Algonet in 1996. 5. At the end of 1999, there were in Sweden approximately 3,800 connections via leased lines to the Internet, so-called fixed access. All of these are commercial subscriptions. The three largest parties, Telia, Tele2, and Telenordia, were even more dominant in the market for fixed access than for dial-up access, together having some 80% of the market. The other companies included Sonera, WorldCom, Global One, and also IT companies such as WM-data, Enator, and others normally not acting as ISPs (PTS 2000:59– 69). 6. Regeringens proposition (Government bill) 1999/2000:86, Ett informationssamhälle för alla (An information society for all). 7. This paragraph is based on PTS (2001:32–33). 8. This paragraph is based on PTS (2001:32–33).

263

sven lindmark and per björstedt

264

9. This paragraph is based on PTS (2001:32–33). 10. This paragraph is based on PTS (2001:33). 11. This paragraph is based on Ewertsson and Hultkrantz (2004). 12. See www.pts.se and www.svd.se, accessed 28 November 2004. 13. See http://www.regeringen.se/sb/d/2156/a/22411. 14. The Swedish government holds 46% and the Finnish government 19.6% of TeliaSonera. 15. This and the following paragraphs are based on company homepages. 16. See www.tdc.com/tdcsong/ and Song annual reports. 17. This paragraph is based on PTS (2001:32–33). 18. This paragraph is based on PTS (2001:32–33). 19. This discussion is based on statistics as provided in OECD (2004a:147). 20. The current service offerings of the major ISPs can be found at http://www .internetworld.idg.se/tjanster/ bredband/. 21. See http://www.itst.dk /static/ Markedsundersøgelser/ EuroprisBenchmark.pdf (accessed 28 December 2004). 22. PTS et al (2002:18) referring to Stelacon’s Hushållbussen 2002 survey of 2002. 23. The survey was directed toward 3,000 Swedish citizens aged 16 –75 years. 24. SIKA (2004:161) based on SCB statistics. 25. SCB as presented in SIKA (2004:161) and SCB (2004b). 26. SIKA (2004:159–163) based on SCB statistics. 27. SCB as presented in SIKA (2004:66). 28. Note: Access for 2003 refers to access in the home. 29. Interview with a senior representative at Telia Research, 2 April 2004. 30. Based on PTS (2004d) and http://www.internetworld.idg.se/. 31. See http://www.internetworld.idg.se/.

references ADL. 2003. The broadband industry: Industry and technology analysis. Presentation by consulting firm Arthur D. Little. Engren, J., and T. Kronberg. 2003. The market for broadband communications in Sweden: Its features and development. Master’s thesis, Royal Institute of Technology, Stockholm. Ewertsson, L., and L. Hultkrantz, L. 2004. Informationssamhällets institutioner (in Swedish). Stockholm: SNS Förlag. Government Official Report. 1999a. Secure IT infrastructure for Sweden in the future (in Swedish), 134. ———. 1999b. Broadband for growth throughout the country (in Swedish), 85. ITPS. 2003. En lärande IT-politik för tillväxt och välfärd: ITPS slutrapportering av uppdraget att utvärdera den svenska IT-politiken (in Swedish), A2003:015. ———. 2004. A learning ICT policy for growth and welfare (in Swedish). Final report on ITPS’s assignment of evaluating the Swedish ICT policy, A2003:015. Available online at www.itps.se.

The Swedish Broadband Market Justitiedepartementet (Swedish Ministry of Justice). 2004. Förslag till strategi för att minska de digitala klyftorna, Ju 2002:E (in Swedish). Available online at www .regeringen.se. Mattson, Nicklas, and Christian Carrwik. 1998. Internetrevolutionen: 1000 dagar som förändrade Sverige (in Swedish). Falun, Sweden: Bonnier Icon Publishing. Ministry of Industry, Employment and Communication. 2002. Follow-up of Swedish Government IT policy. Article no. N2002.002. Nordicom. 2004. Sveriges Internet barometer 2003 (in Swedish). Available online at http://www.nordicom.gu.se/mt /filer/inetbar_2003.pdf. OECD. 2004a. OECD Information Technology Outlook. Available online at www .oecd.org. ———. 2004b. Benchmarking broadband prices in the OECD. Available online at www.oecd.org. PTS. 2000. The Swedish telecommunications market 1999. 16 June. Available online at www.pts.se. ———. 2001. The Swedish telecommunications market 2000. Available online at www.pts.se. ———. 2003a. The Swedish telecommunications market 2002. 5 June, PTS-ER2003:21. Available online at www.pts.se. ———. 2003b. Bredband i Sverige 2003: Tillgänglighet till IT infrastruktur med hög överföringskapacitet (in Swedish). 15 August, PTS-ER-2003-27. Available online at www.pts.se. ———. 2004a. The Swedish telecommunications market 2003. PTS-ER-2004:24. Available online at www.pts.se. ———. 2004b. Svensk telemarknad första halvåret 2004 (in Swedish). 22 December, PTS-ER-2004:43. Available online at www.pts.se. ———. 2004c. Konkurrensen i accessnätet (in Swedish). 29 December, PTS-ER2004:42. Available online at www.pts.se. ———. 2004d. Trådlösa LAN, En teknisk marknadsbeskrivning (in Swedish). PTS-ER2004:12. Available online at www.pts.se. ———. 2004e. Beslut. 24 November, 04-6949/23, a. Available online at www.pts.se. ———. 2004f. Beslut. 24 November, 04-6948/23, b. Available online at www.pts.se. PTS and TEMO. 2004. Så efterfrågar vi elektronisk kommunikation: en individundersökning (in Swedish). 10 December, PTS-ER-2004:40. PTS, KKV and KV. 2002. Alltid på! Bredbandsmarknaden ur ett konsumentperspektiv (in Swedish). Report from the Post and Telecom Agency, the Swedish Competition Authority and Konsumentverket. Available online at the home pages of the respective authorities. SCB. 2003. Befolkningens boende 2003, Sammanställning över nationell och internationell boendestatistik (in Swedish). Available online at www.scb.se. ———. 2004a. SCB Privatpersoners användning av datorer och internet 2004 (in Swedish). (Statistics Sweden, Use of computers and the Internet by private persons in 2004). Available online at www.scb.se.

265

sven lindmark and per björstedt

266

———. 2004b. SCB Företagens användning av datorer och internet 2004 (in Swedish). (Statistics Sweden, Use of computers and the Internet by companies in 2004). Available online at www.scb.se. SIKA. 2004. Facts about information and communications technology in Sweden 2004. Available online at www.sika.se. Statskontoret. 1994. Internet, [email protected]. Solna, Sweden: Alliance AB. Steen, Roland. 2002. Personaldatorer: en utvärdering av arbetsmarknadseffekter (in Swedish). IT-kommissionen. Stelacon. 1997. Marknaden för telekommunikation i Sverige 1996 (in Swedish). Report for PTS. Sturmark, C., T. Borg, J. Jonasson, M. Sandström, and J. Stjärne. 1997. Internet Bonnier DataMedia/DataMedia IT AB. Svenska Dagbladet. 2004. Com hem har siktet på Telia (Com hem aims for Telia) (in Swedish). 6 October. WII. 2004. Svenskarna och Internet 2003. Available online at www.worldinternet institute.se.

Conclusion: The Importance of Institutions Martin Fransman

The advent of broadband communications has not only resulted in the outbreak of various kinds of battles around the world, it has also presented us with the opportunity to conduct an experiment. In this experiment alternative new technologies (i.e., broadband technologies) become available to several countries (i.e., the nine countries examined in this book). These new technologies are key technologies because they facilitate the next stage of development of the Internet, a crucial information and communication infrastructure for all countries. The experiment involves observing how the countries will use the technologies and with what effect. The ultimate value of the experiment lies in the explanation that it facilitates regarding why the countries have used the technologies they have, in the way they have, to produce the effects that they have. The resulting effects were analysed in Chapter 1 in terms of a number of performance indicators. In studying these performance indicators three puzzles became apparent: the Japan/Korea puzzle, the U.S. puzzle, and the European puzzle, all of which were examined in Chapter 1 and in the introductions to Parts I, II, and III. In explaining these puzzles (and the inter-country differences in performance that they contain), two related causes were highlighted. The first was the intensity of competition, while the second was the strength of regulation. Although other causes were also identified (such as the role of government and geography and demography), these were shown to provide insufficient explanations on their own. However, the intensity of competition and the strength of regulation are only intermediate causes. The reason is that they themselves need to be

martin fransman

268

explained. That is, we need to ask why the intensity of competition and the strength of regulation are what they are in the countries concerned. The studies in this book show that the intensity of competition and the strength of regulation are themselves outcomes of complex processes. For example, intensity of competition depends on competition between networks, technologies, and firms. It also depends on entry, and potential entry, of new entrants willing and able to challenge the dominance of the incumbents. The intensity of competition also depends on firm-level determinants, such as the willingness of firms to become what is called in this book disruptive competitors. Particularly importantly, in the case of broadband as shown here, the intensity of competition also depends on the regulatory regime that has evolved in the country concerned. It is this regulatory regime that has determined the strength of regulation in the country. In taking stock of the country studies and attempting to “stand back” and explain country performance, most difficult has been coming to grips with a country’s regulatory regime. (Intensity of competition was more easily analysed by taking account of factors such as the extent of inter- and intra-modal competition, the disruptiveness of competitors to the incumbent as measured by their profits or losses, and the threat that they posed to the incumbent.) The problem with “regulatory regime” is that it is only the tip of the iceberg. While it is a relatively visible tip (although a degree of contestable interpretation is needed in making a judgment regarding overall how “strong” or “weak” regulation has been), less visible is the foundation on which the regulatory regime tip sits. It is this foundation that may be thought of as the institutional basis for regulation. What should we understand by “institutional basis” in this context? Institutions have been defined as “the rules of the game,” the factors that constrain and define the choices that are available and therefore shape the decisions that are made. In the present context these institutions are not only things such as the regulations that are passed by the regulatory authorities that obviously constitute some of the rules that affect the game; they are also the processes that have made the regulations what they are and not other than what they are. These processes include political and ideological processes and the dominant interests that they express. But, as we have seen in this book, such processes are usually very difficult to analyse, justifying the metaphor of the submerged part of the iceberg. To make this abstract discussion more concrete, what are the institutional processes that explain why in Japan the regulator has insisted that the incumbent telecoms company, NTT, make its fibre-to-the-home network accessible at low cost to competitors while in the U.S. the FCC has decided to forbear,

Conclusion

that is, to refrain from regulating FTTH? It will hopefully be apparent to the reader from the material presented in this book that questions such as these pose significant analytical problems. Unfortunately, the difficulties have not been helped by the dominant approach in economics, which pays more attention to what regulators should do (based on concepts of optimal regulatory outcomes) than to what they actually do and why. This approach, it should be clear, does not take institutions seriously. It may be concluded, therefore, that the main analytical implication of this book highlights the importance of institutions as determinants of intercountry differences, including differences in performance. The short answer to the question posed in the subtitle of this book—why the U.S. and Europe lag behind while Asia leads—is that the countries concerned have different institutions, a fact which has resulted in different strength of regulation and intensities of competition and, therefore, different performances in broadband. However, as this book suggests, we still have some way to go to develop more rigorous institution-based explanations of economic phenomena.

269

Contributors

Martin Fransman is Professor of Economics and Founder-Director of the Institute for Japanese-European Technology Studies at the University of Edinburgh in Scotland. His books include Telecoms in the Internet Age: From Boom to Bust to . . . ? (2002), winner of the Wadsworth Prize for the best business history book published in the UK in 2002, and The Market and Beyond: Information Technology in Japan (1990), winner of the Masayoshi Ohira Memorial Prize. Takanori Ida is an Associate Professor of Economics at Kyoto University in Japan. His main academic interests are in micro-economics and industrial economics, particularly in the network industries such as telecommunications, electricity, gas, and railways. His publications include “The Broadband Market in Japan” in Japanese Telecommunications Market and Policy in Transition (2005), “Analysis of Internet Topology with a Three-Components Model” in Managerial and Decision Economics 26.7 (2005), “The Regional Cost Gap of the Japanese Local Telecommunications Services,” Papers in Regional Science 84.2 (2005) with S. Asai, and “Discrete Choice Analysis of Demand for Broadband in Japan,” Journal of Regulatory Economics 29.1 (2005) with T. Kuroda. Inho Chung was born in Seoul, Korea, in 1961. He graduated from Sogang University, Korea, with a bachelor’s degree in art and literature in 1985. He received his master’s degree in economics from Oklahoma State University in 1987 and his PhD in economics from the University of California, Davis, in 2001. He has worked for KT since 1991 and is currently a director of the Management Research Lab in KT. Manqiang Liu is a Professor and Doctor of Economics and Deputy Director, Center for Informatization Study, Chinese Academy of Social Sciences

Contributors

272

(CASS). His research focuses on the theory of technical change, science and technology policies, industrial policies, informatization in China, telecommunications development and reform in China, enterprise informatization, and electronic commerce. Johannes M. Bauer is a Professor in the Department of Telecommunication, Information Studies, and Media at Michigan State University and the Co-Director of the Quello Center for Telecommunication Management and Law. He joined Michigan State University in 1990 after receiving his doctorate in economics from the Vienna University of Economics and Business Administration, Vienna, Austria. He has written and edited five books and more than fifty articles on issues of regulatory reform in telecommunications and energy. Jackie Krafft has a PhD in economics and has been a researcher at the National Centre for Scientific Research in France (CNRS) since 1995. Her main research interests are the dynamics of innovative industries; processes of integration/specialisation; and inter-firm co-operation, competition, and regulation policies, especially in the domain of the infocommunications industry. She has published a number of recent articles on these topics in academic journals such as Research Policy, Economics of Innovation and New Technology, and Telecommunications Policy. She has also edited a book on the process of competition and has contributed to collective volumes. Franz Büllingen is a graduate of the University of Bielefeld, where he also obtained a PhD in Social Sciences. He is head of the department of Communication and Innovation at WIK in Bad Honnef. Franz Büllingen has 20 years’ working experience in advising political and regulatory institutions as well as enterprises. He has carried out many research projects for government institutions as well as private entities in the field of diffusion and innovation research, analysis, and prognosis of market structure, regulatory matters in connection with the problem of convergence, the Internet economy, the production and demand of telecommunication services, as well as IT security. Major publications are VoIP—Marktentwicklungen und regulatorische Herausforderungen (et al. 2005); Regelungen zum Telekommunikations-Kundenschutz im internationalen Vergleich (et al. 2005); Stand und Perspektiven der Vorratsdatenspeicherung im internationalen Vergleich (et al. 2005); “Die Entwicklung breitbandiger Internetnutzung privater Haushalte in Deutschland bis 2015,” in: Eberspächer, Quadt (Hg.), Breitbandperspektiven. Schneller Zugang zu innovativen Anwendungen (2004); Mobile Multimedia-Dienste—Deutschlands Chance im globalen Wettbewerb (et al. 2004); “Development perspectives, firm strategies and applications in mobile commerce,” in: Journal of Business Research (et al. 2004).

Contributors

Cristiano Antonelli is Professor of Economics at the University of Turin, Italy. He has been Rockefeller Fellow at the MIT and Economist at the OECD. He founded the Doctoral Program in Economics of Complexity and Creativity of the University of Turin, and he is Director of the Laboratory for the Economics of Innovation “Franco Momigliano.” He is the managing editor of Economics of Innovation and New Technology and he has widely published in the field of the economics of innovation. His most recent books are The Microeconomics of Technological Systems (2001) and The Economics of Innovation, New Technologies and Structural Change (2003). Pier Paolo Patrucco is Assistant Professor of Economics at the University of Turin, Italy. He obtained a PhD in economics at the University of Nice– Sophia Antipolis (France) and he has been Marie Curie Fellow at the Victoria University of Manchester (UK), Full Visiting Fellow at SPRU–The University of Brighton (UK), and Jean Monnet Fellow at the Robert Schuman Centre for Advanced Studies of the European University Institute in Fiesole (Italy). His main research interests are in the economics of innovation and the economics of knowledge. He recently published papers in Information Economics and Policy (2002), Regional Studies (2003), and the Cambridge Journal of Economics (2005). Dr Sven Lindmark is currently head of the Division of Innovation Engineering and Management, Department of Technology Management and Economics at Chalmers University of Technology. His research interests include a broad range of innovation-related issues, including industrial and technological evolution, strategic management of innovation, standardisation, innovation systems, diffusion of innovations, and forecasting. He has worked extensively in the field of mobile communications, including studies on the history of mobile communications, mobile Internet services in Japan and Europe, 3G and 4G mobile systems, barriers to diffusion of mobile data services, and the evolution and state of the Swedish telecommunications innovation system. Per Björstedt holds a master of science degree from Chalmers University of Technology and is currently working for the Swedish Trade Council in Milan, Italy. Until recently, he held a position as research assistant at the Department of Technology Management and Economics at Chalmers University of Technology and engaged in several projects related to the telecommunications industry.

273

Index

Boldface page numbers indicate material in tables or figures. Page numbers followed by “n” indicate notes. ACCA Networks, 76 ADSL (asynchronous digital subscriber line): described, 15–16, 54nn14 –15; in France, 175, 179; in Germany, 196 –200; in Italy, 222, 223, 225, 226 –229; in Japan, 65– 66, 70, 75, 76 –78, 77; in Korea, 88; in Sweden, 240, 245, 246; in U.S., 39, 134, 135, 137, 140 –143 Algonet, 242 American Public Powers Association (APPA), 153 ARCEP/ART (France), 46 – 48, 169, 174, 190 – 191 Asia, 2. See also China; Japan; Korea Asia Pacific Information Infrastructure (APII), 98, 107n3 asynchronous digital subscriber line. See ADSL AT&T, 39 B2 (Bredbandsbolaget), 47, 250 BcN (Broadband convergent Network Plan), 105 Beijing Communication, 116 Beijing Gehua Cable, 118 Beijing IPTV, 119 Beijing Olympics, 124 –125 Belgium, 10, 11 Bellcore Laboratories, 15 Bell South, 144 Birgersson, Jonas, 244, 245 Boingo, 145

Bostream, 48 Bredbandsbolaget (B2), 47, 250 Bright House Networks, 42, 134 broadband: access spread, 5– 6, 7, 54n8; ADSL described, 15–16, 54nn14 –15; cable described, 16 –17; characteristics influencing adoption, 25; in China. See China; commercialization of the Internet and, 3, 4; competition between companies, 19–24, 55nn21–26; described, 2; economic impact, 51–53; European puzzle. See European Union; geography and demography factors, 27–28; governments’ influence on diffusion, 27; institutional basis for regulation, 268 –269; intensity of competition and, 267–268; Japan/Korea puzzle. See Japan; Korea; laser access, 19; mobile communications growth, 3, 5, 53n6; national performance measures, 7– 8, 54n8; new entrants’ impact, 11–13, 12; optical fibre, 17, 54n17; penetration per country, 8, 9; powerline, 19; price rankings, 10, 13–15, 14; regulatory framework and, 19–20, 27, 55n21, 267–268; satellite, 19; services provision, 6 –7, 25; speed offerings by incumbents, 10, 11; “triple play” for consumers, 26 –27; U.S. puzzle. See United States; VoIP, 25–26, 26; wireless, 18, 54nn18 –20 Broadband convergent Network Plan (BcN), 105 BT (British Telecom), 45, 168

Index

276

cable networks: in China, 113–114; in the European Union, 44 – 45; in France, 179, 183– 184; in Germany, 200 –205, 217n1; in Korea, 62; reasons for limited reliance on, 16 –17; in U.S., 39, 134 –137 Canada: new entrants’ impact, 12; penetration of broadband, 8, 9; price rankings, 13–15 Canopy, 145–146 Cegetel, 181 Change Bell Services, 107n2 China: applications of broadband, 122; Beijing Olympics and, 124 –125; broadband access development phases, 114 –118, 117; competition between companies, 113–114; content development, 118 –120; cooperation between companies, 63– 64; domestic networks provisions, 112; economic status of country, 109–110, 111, 125nn2–3; evolution of telecommunications market, 111–112; FTTH progress, 124; government’s role in broadband, 50, 63; growth of broadband users, 120 –122, 121, 126n14; market trends, 122–123; penetration and market position, 49–50, 50, 57nn57–58, 121; pricing factors in access, 115, 116, 118; SARS impact on broadband use, 116; wireless broadband access, 124 China International Trust and Investment Corporation (CITIC), 64 China Mobile, 111 China Network Communications (CNC), 50, 111, 113, 118, 123 China Satcom, 111 China Telecom, 50, 111, 113, 117, 118, 123 China Tietong, 111, 117, 118, 123 China Unicom, 111 Chinavnet, 119 Chung, Inho, 31 CITIC (China International Trust and Investment Corporation), 64 CLECs (competitive local exchange carriers), 150 CNC (China Network Communications), 50, 111, 113, 118, 123 COCOM (Electronic Communications Committee), 192 COLT broadband, 236, 238n4 Com hem, 250 –251 Communications Review (1999, EU), 166 competitive local exchange carriers (CLECs), 150 Covad, 42, 134 Cyworld, 94 Daum, 94 Denmark: penetration of broadband, 8, 9; price rankings, 13–15; speed offerings, 10, 11

Deutsche Telekom (DTAG), 167, 196, 202 digital divide: in France, 185–186; in Italy, 229–231; in Japan, 81– 82, 83– 85, 86nn9–10; in Sweden, 258 –259 Digital Millennium Copyright Act (1998, U.S.)(DMCA), 156 digital TV in Italy, 234 –235 DSL (digital subscriber line), 10, 15–16, 44 – 45, 54nn14 –15. See also ADSL DTAG (Deutsche Telekom), 167, 196, 202 eAccess in Japan, 76 economies of scale, 19 Electronic Communications Committee (COCOM), 192 European Telecommunications Standardization Institute (ETSI), 192 European Union: broadband puzzle, 2; compared to Japan/Korea, 48; disruptive competitors, 45– 48, 56nn49–51; DSL versus cable modem, 44 – 45; France overview, 169; Germany overview, 167–168, 171n1; Italy overview, 169–170; regulation of broadband, 49, 56n54, 57n56; regulatory framework, 165–166, 189–190, 193n6; Sweden overview, 170 –171; UK overview, 168 – 169. See also France; Germany; Italy; Sweden; United Kingdom Fastweb, 227, 231–234 Federal Communications Commission (FCC), 40, 41, 133, 149–152, 160nn29–30 fibre to the home. See FTTH FibrLINK, 123 films, Internet, 95 FISI (French Infocommunications System of Innovation), 172 Framfab, 244, 245 France: broadband status, 173–176, 193n2; competition between technologies, 178 – 180; disparities in availability and service, 185, 185–186; disruptive competitors, 181– 183, 182; distortion of competition, 183–184; innovation and new entrants, 184; intensity of competition and, 45– 47; overview of broadband industry, 169; penetration of broadband, 8, 9, 174, 175; pricing factors in access, 175; regulatory framework, 186 –192, 193n6; speed offerings, 10, 11; unbundling, 22, 23, 176, 177. See also European Union France Telecom (FT), 172, 181, 184 French Infocommunications System of Innovation (FISI), 172 FTTH (fibre to the home): China, 124; defined, 17; France, 180; Germany, 205–206; Italy, 232–233; Japan, 10, 12, 59– 60, 78 – 81,

Index 79, 80, 82– 85; Korea, 63, 104 –106; Sweden, 260; U.S., 129–130, 143 FTTX, 17 games, online: in China, 119–120; in Korea, 95–96 German regulatory authority (RegTP), 168, 199 Germany: broadband cable market development, 200 –205, 201, 217n1; DSL market development, 196 –200, 197; FTTX market development, 205–206; future outlook, 214 –217, 215; market trends, 215–216; overview of broadband industry, 167–168, 171n1; penetration of broadband, 8, 9; powerlinebased services market, 211–213; PWLAN services market development, 206 –211; regulatory framework, 216 –217; speed offerings, 10, 11; unbundling, 22, 23; WiMax-based services, 213–214. See also European Union Global System for Mobile Communication (GSM), 207 Glocalnet, 251 government regulation of broadband: China, 50, 63; deregulation in Japan, 33–36, 55n34, 129, 132n2; Europe, 49, 56n54, 57n56; European Union, 165–166, 189–190, 193n6; France, 186 –191; Germany, 199–200, 204 – 205, 209–210, 212–213, 217n1; influence on diffusion, 27; Japan, 68 – 69; Korea, 36 –37, 62– 63, 90, 92, 97, 97–102, 102; overview, 19–20, 27, 55n21, 267– 8; Sweden, 245–247, 248, 262; telecoms in Asia, 33–36, 55n34 government regulation of broadband in U.S.: complexity of, 154; diffusion outside urban areas, 152–153; FCC and, 149–152, 160nn29–30; FTTH development and, 143; overview, 127; policymakers’ inputs, 148 – 149; spectrum management, 152; strength of, 40 – 42, 56n44; unbundling policies, 149–152 Great Wall Broadband (GWB), 50, 64, 114, 118, 123 GSM (Global System for Mobile Communication), 207 Hanaro Telecom, 31–33, 88, 89, 100 Herfindahl-Hirschman Index (HHI), 145, 159n15 Home Networking Plan, 106 home networks: broadband demand in U.S., 138 –141; broadband use in Sweden, 255– 257 IEEE (Institute of Electrical and Electronic Engineering), 18

Iliad/ Free, 46 – 48, 181 i-mode service, 5 incumbent local exchange carriers (ILECs), 128 –129, 149–150 Independent Regulators Group/ European Regulators Group (IRG/ ERG), 192 instant messaging in China, 120 Institute of Electrical and Electronic Engineering (IEEE), 18 Internet, 3, 4. See also broadband IP television, 119 ISM-spectrum, 209 Italy: competition between companies, 223, 224 –225; current state of broadband, 219– 220; demand for services, 221–222; diffusion and penetration of broadband, 225–227, 226, 227; digital TV, 234 –235; disparities in availability and service, 229–231, 230; infrastructure and availability, 228 –229; innovations from Fastweb, 227, 231–234; intermodal competition between technologies, 223–224; new entrants’ impact, 224 –225; overview of broadband industry, 169–170; penetration of broadband, 8, 9; speed offerings, 10, 11; strategy from COLT broadband, 236, 238n4; strategy from TOP-IX broadband, 237–238; VoIP, 232, 234. See also European Union Japan: ADSL and, 65– 66, 70, 75, 76 –78, 77; benefits of policies, 67– 69, 68, 85n2; broadband puzzle, 2; competition between companies, 61, 66; conflict between NTT and ministry, 60 – 61; degree of openness of networks, 69; economic state before 2000, 65; FTTH and the digital divide, 81– 85, 86nn9–10; FTTH current status, 78 – 81, 79, 80; FTTH diffusion, 59– 60; geography and demography factors, 37; government intervention, 36 –37; government’s role in broadband, 33–36, 55n34, 68 – 69, 129, 132n2; incumbents strong position, 28 –29; intensity of competition, 29–31, 55nn28 –32; market analysis, 74 –75; market demand, 37; new entrants’ impact, 12, 70 –71, 85– 86nn4 –5; NTT’s regrouping by management, 71–73, 73; penetration of broadband, 8, 9; pricing factors, 13–15, 75; speed offerings, 10, 11; trends in users, 73–74, 74; unbundling, 22, 23; VoIP lead, 26, 37 Kabel Deutschland (KDG), 199 KCC (Korean Communications Commission), 100 KDDI, 80 K-Opticom, 70, 71, 78, 80, 85nn4 –5

277

Index

278

Korea: broadband puzzle, 2; economic effects of broadband, 93; emerging technologies and services, 63, 103–106; geography and demography factors, 37, 62; government’s role in broadband, 36 –37, 62– 63, 90, 92, 97, 97–102, 102; incumbents’ strong position, 28 –29; intensity of competition, 31– 33; lead in broadband, 6; market demand, 37; market structure, 88 – 89, 89; market trends, 92–93; new entrants’ impact, 12; penetration of broadband, 8, 9, 87, 91–92; price competition, 89–91, 90, 107n1; price rankings, 13–15; pricing system changes, 102–103; regulation of telecoms, 33–36, 55n34; service types, 94 –96, 107n2; speed offerings, 10, 11; unbundling, 22, 23 Korean Advanced Research Network (KOREN), 97, 98 Korean Communications Commission (KCC), 100 Korean Information Infrastructure-Government (KII-G), 97–98 Korean Information Infrastructure-Private (KII-P), 97, 98 Korea Telecom (KT), 28, 31, 88 LANs (local area networks), 18, 103–104. See also PWLAN (public wireless local access network), Germany laser access, 19 LDCom, 181 Lechleider, Joseph W., 15 local loop unbundling (LLU): competition between companies and, 20 –22, 23, 55n25; countries overview, 22–23; in France, 188 – 189; in Korea, 62– 63, 100 –101; in U.S., 128 –129 Marconi, Guglielmo, 18 MCI / WorldCom, 39 Mediacom, 42, 135 Ministry of Communication (Japan), 30 Ministry of Communication (Korea), 31 Ministry of Internal Affairs and Communications (Japan), 67, 85n3 Motion Picture Association of America (MPAA), 138 music, Internet, 94 –95, 107n2 Naver, 94 Nespot, 103–104 network externalities, 19 New Bridge Capital, 107n1 newspapers, Internet, 96 New Zealand, 10, 11

ntl (UK), 45, 183 NTT: ADSL market in Japan and, 76 –78, 77; battle with government, 33–34; competition with SoftBank, 66; FTTH market in Japan and, 78 – 81, 79, 80; impact in Japan, 28; proposed breakup, 67– 68, 85n2; regrouping by management, 71–73, 73 NTT DoCoMo, 5, 145 NTT East, 10 OECD (Organisation of Economic Co-operation and Development), 8 OFCOM, 20, 33–36, 36 –37 Oftel, 35 Olivetti, 224 online education, 96 online games, 95–96 optical fibre, 17, 54n17. See also FTTH (fibre to the home) Organisation of Economic Co-operation and Development (OECD), 8 peer-to-peer (P2P) file sharing, 138 penetration of broadband: in China, 49–50, 50, 57nn57–58, 121; in France, 174, 175; in Germany, 201; in Korea, 8, 9, 87, 91–92; per country, 7, 8, 9; in Sweden, 243, 252–255, 253. See also specific countries Pew Internet and American Life Project, 138 Pirelli, 224 Post och TeleStyrelsen (PTS), 246 Powell, Michael, 40, 41 powerline-based services, Germany: demand, 213; infrastructure and availability, 211–212; prices and tariff structures, 212; regulation, 212–213; structure of suppliers, 212 powerline-based services market, 19 PTS (Post och TeleStyrelsen), 246 PWLAN (public wireless local access network), Germany: demand for services, 210 – 211; infrastructure and availability, 206; products and prices, 209; regulation, 209– 210; structure of suppliers, 207–208, 208 QSC AG, 197 quality of service, 8 Recording Industry Association of America (RIAA), 138 Regional Bell Operating Companies (RBOCs), 39, 143 RegTP (German regulatory authority), 168, 199 regulation of broadband. See government regulation of broadband

Index SARFT (State Administration of Radio, Film, and Television), 119, 123 SARS in China, 116 satellite, 19 SBC, 39, 144 Shanghai Cable Network, 114 Shanghai Oriental IPTV, 119 SIM (Subscriber Identity Module), 207 SIP-STET, 224 SK Telecom, 94 SNU.S. (Swedish Network User Society), 242 SOCRATES, 224 SoftBank, 29–30, 65– 66, 70, 76 –78, 77 Son, Masayoshi, 29–30 South Korea. See Korea Sprint, 39 State Administration of Radio, Film, and Television (SARFT), 119, 123 Subscriber Identity Module (SIM), 207 Sweden: broadband market characteristics, 261; broadband market emergence, 244 – 245, 263n5; business broadband use, 257– 258; competition between companies, 262; current state of broadband, 240, 263n1; diffusion and penetration of broadband, 243; digital divide, 258 –259; evolution of commercial services, 241–243, 244; geography and demography factors, 241; government broadband policy, 247–249; government policy initiatives, 247; government’s role in broadband, 262; household broadband use, 255–257; intensity of competition and, 47– 48; network build-out, 260 –261; new entrants’ impact, 12, 262; operator/ISP market, 249–255, 250; overview of broadband industry, 170 –171; penetration of broadband, 8, 9, 252–255, 253; price rankings, 13–15; pricing factors, 243, 263nn3– 4; regulatory framework, 245–247, 248; speed offerings, 10, 11; unbundling, 22, 23. See also European Union Swedish National Grid, 249 Swedish Network User Society (SNU.S.), 242 Swipnet, 242 TDC Song Networks, 251 Tele2, 242 Telecom Italia, 169–170, 181, 225, 227 Telecommunications Act (1996, U.S.), 134, 148 –149 Televerket, 242 Telewest, 45 Telia, 246 TeliaSonera, 170 –171, 246, 250

TELRIC (Total Element Long-Run Incremental Cost), 150, 151, 159n28 Thrunet, 31, 88, 89 Tiantian Zaixian, 119 TipNet, 242 T-Mobile, 207 T-Online, 197 TOP-IX broadband, 237–238 Total Element Long-Run Incremental Cost (TELRIC), 150, 151, 159n28 Transpac, 242 Triennial Review by FCC, 40, 150 –151, 160nn29–30 UMTS (Universal Mobile Telecommunications System), 180, 207 United Kingdom: intensity of competition and, 45– 48, 56nn49–51; overview of broadband industry, 168 –169; penetration of broadband, 8, 9; regulation of broadband, 20; speed offerings, 10, 11; unbundling, 22, 23. See also European Union United States: broadband puzzle, 2; business broadband demand, 141–142; common carrier network issues, 128; companies’ investments in broadband, 143–144, 144; competition between companies, 127, 136; current state of broadband, 133; digital divide, 156; FTTH and the FCC, 129–130; ILEC network unbundling issues, 128 –129; intensity of competition, 38 – 40, 130 –131, 132n2, 136 – 137, 155–156; lag behind other countries, 157; market description, 134 –137, 135; market share data, 39– 40; mobile communications success, 153–154; new entrants, 42– 43, 56n48; new entrants’ impact, 12; penetration of broadband, 8, 9; price and performance factors, 146 –147, 148, 155; price rankings, 13–15; regulatory framework. See government regulation of broadband in U.S.; residential broadband demand, 138 – 141, 139; service availability, 142–143, 144; speed offerings, 10, 11; system innovation factors, 154 –156; unbundling, 23; unlicensed services, 145–146; wireless broadband access, 144, 145 Universal Mobile Telecommunications System (UMTS), 180, 207 UPC Sverige AB, 251 VDSL and LAN in Korea, 103 Verizon, 39, 143–144 video-on-demand (VoD), 16 VoIP (voice over Internet protocol): competition between technologies, 19; in Italy, 232,

279

Index

280

234; in Japan, 25–26, 37; in Korea, 106; in U.S., 138 –139 web portals, 94 WiFi (wireless fidelity), 18, 180 WiMax-based services, 18, 213–214 wireless broadband: in China, 124; described, 18, 54nn18 –20; in Korea, 104; United States, 144, 145 WISPs (wireless Internet service providers), 145

WLAN (wireless local area network): in France, 180; in Korea, 103– 4; in Sweden, 260 –261 WorldCom, 5 Xingkong Alliance, 119 Yahoo!, 94 Yahoo! BB, 29–31, 55nn28 –32, 70