*Pages 1--24 from Microsoft Word - 32004* A Staff Report of the Office of Strategic Planning and Policy Analysis and International Bureau * October 2003 “Broadband Internet Access in OECD Countries: A Comparative Analysis” Sherille Ismail ** Senior Counsel Office of Strategic Planning and Policy Analysis Irene Wu** Assistant Chief Regional and Industry Analysis Branch Strategic Analysis and Negotiations Division International Bureau * This report represents the individual views of the authors and does not necessarily reflect the views of the FCC, any FCC commissioner, or other staff. ** The authors gratefully acknowledge the help of the following FCC colleagues: Anita Dey, Regional Specialist for Asia, International Bureau (IB); Barbara Esbin, Associate Chief, Media Bureau; Scott Marcus, Senior Advisor on Internet Issues, Office of Strategic Planning and Policy Analysis (OSP); Carol Mattey, Deputy Chief, Wireline Competition Bureau; Robert Pepper, Chief, Policy Development, FCC; Donald Stockdale, OSP; Sean Wang, intern, OSP; Douglas Webbink, Chief Economist, IB; Simon Wilkie, Chief Economist, FCC. We also thank Dr. Sam Paltridge, Directorate of Science, Technology and Industry, OECD, the author of several OECD reports on broadband issues. 1 - 1 - I. Introduction This paper reviews the broadband policy experiences of selected OECD countries. 1 These countries have adopted a variety of strategies to promote broadband growth. 2 Because many different strategies have been tried, it is possible to examine specific policy proposals (e. g., open access requirements, local loop unbundling, line sharing, duopoly competition) and evaluate which have been effective. Beyond the scope of this paper are other factors which might be influential such as population density, tax incentives, and subsidy programs. In section II, we provide an overview of broadband internet access in selected OECD countries: South Korea, Canada, Belgium, Denmark, Sweden, the United States, Switzerland, Japan, Germany, and the United Kingdom. Each of the countries we discuss has taken a somewhat different approach to encouraging the development of broadband access. In the interest of brevity, we do not discuss all the approaches taken internationally. Other studies (including those listed in the “references’ section at the end of this paper) comprehensively review the broadband experiences of all OECD members. We also do not discuss advanced broadband countries that do not belong to the OECD. In section III, we offer a synopsis of the development of broadband in each of the selected OECD countries (excluding the U. S.). In each country, we identify key developments, including regulatory policy, and assess the state of competition. Finally, in section IV, we offer our conclusions. We caution that any conclusions must be tentative because broadband growth is still in its very early stages, even in the most advanced markets. Moreover, market share among different broadband platforms, and within a platform, among incumbents and non- incumbents, is quite fluid. 1 OECD refers to the Organization for Economic Cooperation and Development, which consists of 30 of the world’s industrialized countries, including the United States. For more information, please see www. oecd. org/ about. 2 The OECD definition of broadband is at least 256 kbps downstream and at least 64 kbps upstream, which differs from the FCC definition of high- speed lines as faster than 200 kbps in at least one direction. OECD data on the U. S. conforms to the OECD definition, which, except as noted, is used throughout this paper. 2 - 6 - C. Broadband Pricing There is no simple way to compare the variety of broadband service packages available in different countries. We have chosen to calculate the monthly cost per 1 mpbs. While this method produces numbers that may be useful for the purpose of making broad comparisons, one limitation is that it requires an assumption that each additional mbps is equally valuable to the consumer. Economists may question this assumption, noting that experience in the U. S. has shown that there is no linear relationship between broadband capacity and price. Figure 3 12 Selected International Broadband Prices (PPP adjusted) March - May 2003 Source: ITU Promoting Broadband: Background Paper $1.57 $3.88 $9. 50 $23. 71 $25. 24 $28.85 $29.43 $71.95 $77.44 $198. 83 0.02 0.08 0.06 0.07 0.05 0.06 0.07 0.08 0.04 0.04 $0. 00 $50.00 $100.00 $150.00 $200.00 $250.00 Japan - YahooBB S. Korea - Hanaro + KT Belg ium - Belg a c om Germany - Deutsche Teleko m Canada - Bel l Sympatico Switzerland - Bluewin US - Comcast U K - Pipex ADSL Sweden - Te le2 Denmark - Tele2 Monthly US$/ 1mbps downstream 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 Broadband price as % monthly household income US$/ 1 mbps downstream % monthly household income 12 Prices for broadband packages are from March 2003, as cited in the ITU report, Promoting Broadband: Background Paper, except for Germany, which was drawn from Deutsche Telekom’s website in May 2003. Purchasing power parity conversion factors (2001) are from the World Bank, World Development Indicators, an on- line resource, accessed in May 2003. GDP and household data (2002 or most recently available) are from ITU Telecommunications Indicators, also an on- line resource, accessed May 2003. 7 - 10 - less incentive to switch to broadband. Yet, these countries include the highest and lowest broadband subscriber levels in our sample. Thus, we note that while price is always a factor in demand for a service, this evidence suggests that other factors may also have a strong influence. For example, in some markets there may be technical hurdles faced by consumers wishing to subscribe that serve as a barrier. In other markets, broadband service speeds may be so much higher and, therefore, offer so many more applications than dial- up Internet service, that the two services may no longer be comparable. Under those circumstances, the relative prices of broadband and dial- up may not be the most important factor in a consumer’s decision to subscribe to broadband. Further research is needed to clarify which factors have the most impact on consumer decisions to switch to broadband service. Second, we also address whether broadband subscriber levels are influenced by differences in the rate structure for local telecommunications and dial- up Internet service. For instance, if a country has metered pricing for dial- up internet access and flat- rated pricing for broadband access, this suggests that dial- up users will face higher marginal costs of usage than broadband users. In such a case, heavy users of internet access may be more likely to choose broadband access over dial- up access. In contrast, if both forms of internet access are charged on a flat- rated basis, economic theory suggests that consumers would have less incentive to switch to broadband. What does the data show? The chart below shows the underlying rate structure for local telephony, local telecom access for dial- up Internet service, and DSL. Several countries, such as Belgium, Denmark, Sweden, Switzerland and Germany, have metered dial up access and unmetered broadband access. These countries have among the highest and lowest subcribership rates in the countries in our sample. 11 - 11 - Korea and Canada both have unmetered telecom access for dial- up Internet service and broadband access, although Korea has metered local telephony rates. These two countries have the highest level of broadband subscribership in our sample. Of the two countries with the lowest level of broadband subscribership, Germany has metered and the UK has unmetered telecom access for dial- up Internet service. Rate Structure for Dial- up Internet and DSL Service Source: OECD, 2003 Broadband subscribers per 100 Local telephony rate structure Unmetered telecom access for dial- up Internet? Internet access pricing structure DSL pricing structure Korea 21.4 Metered Yes Metered Flat Canada 11.7 Unmetered Yes Flat Flat Belgium 8.5 Metered Metered Flat Denmark 8.3 Metered Metered Flat Sweden 8.1 Metered Metered Flat U. S. 6.9 Metered/ Flat/ Unmetered Yes Metered/ Flat* Flat Switzerland 6.3 Metered Metered Flat Japan 6.1 Metered Yes Metered Flat Germany 4.0 Metered Metered Flat U. K. 2.3 Metered Yes Metered/ Flat Flat * Information provided by FCC. Further research is needed to clarify whether rate structure has an impact on consumer decisions to switch to broadband service. In particular, research should focus on what happens in a particular market over time. Cross- market analyses may be less relevant on this issue. 12 - 12 - III. A synopsis of broadband competition in selected countries The following is a brief discussion of the broadband growth and competition in the nine OECD countries in our sample. A. South Korea: Competitors’ access to cable networks spurs early broadband deployment. Incumbent telco aggressively gains market share. When cable service began in 1995, the Korean government required structural separation of conduit and content. The two state- owned cable infrastructure owners – Powercomm and Korea Telecom were not permitted to offer services, but instead leased capacity to programmers. Therefore, new entrants to the broadband market, such as Thrunet in 1998 and Hanaro in 1999, initially leased cable capacity to reach their earliest customers. Subsequently, the structural separation rules in cable were relaxed. Incumbent telco Korea Telecom sold its cable infrastructure to cable service providers in 2000. Hanaro, in addition to leasing cable capacity, also provides DSL services over its own facilities- based network. In 2002, 48% of Hanaro’s revenues were from cable modems (1.5 million subscribers) and 44% from DSL (1.3 million subscribers). Thrunet, sold parts of the company in recent years and filed for bankruptcy in 2003. Korea Telecom, the telecom incumbent, entered the broadband services market in 2000, in response to the challenge presented by companies like Hanaro, which had by then signed up more than a million customers. KT rapidly increased subscribership, reaching 2 million in 2001 and growing to more than 4 million customers in 2002. In 2002, local loop unbundling rules went into effect. In 2002, there were 6.3 million DSL subscribers (63%) and 3. 7 million cable modem subscribers (37%). 13 - 13 - B. Canada: Cable operators led the way on broadband access. Cable “open access” required on the books but not yet implemented due to court challenges. DSL gains market share after slow start. As of 2002, broadband by DSL or cable is available in communities which account for 85% of the population. However, reflecting the large number of widely dispersed small communities in Canada, broadband is available in only 24% of Canada’s 1,281 communities. Cable companies became global pioneers, offering cable modem services as early as 1996. Canadian cable companies pass 93% of all homes and have the potential to provide broadband access to 6 million homes. In 2002, 26% of Shaw cable subscribers and 21% of Rogers cable subscribers signed up for broadband access. Incumbent telecommunications carriers, like Bell Canada, Telus, and Sasktel, also offered DSL services ahead of other countries. Sasktel, for instance, became the first carrier in the OECD to do so by offering DSL in November 1996. Although the regulator has required local loop unbundling since 1997, alternative providers have used local loop unbundling to provide service primarily to business (not residential) customers. In 2003, Canadian regulators clarified rules requiring incumbents to offer DSL unbundling for carriers serving residential customers. Similarly, although cable open access has been required since 1999, legal challenges have delayed implementation of this rule. Some cable operators have voluntary agreements to provide ISPs access to their networks. In 2002, there were 2 million cable modem subscribers (55%) and 1.6 million DSL subscribers (45%). 14 - 14 - C. Belgium: Cable spurs incumbent telco to provide broadband. Offers 3- 4 mbps, the best baseline speeds in Europe. In 1997, Telenet became one of the first cable companies in Europe to begin offering high speed Internet access. Other smaller cable companies followed suit in 1998 and 1999. In 1999, Belgacom, the incumbent telecommunications carrier responded by offering DSL services. Competition has been driven by the fact that almost 100% of Belgian household have access to cable services. Today, 98% of households also have access to DSL services, prices are low, and consumers have among the highest speeds of bandwidth availability in Europe. Cable companies are offering 4 mbps downstream and DSL companies are offering 3 mbps downstream. Local loop unbundling has been required since 2000, but competitive providers have been unable to gain any share of the broadband access market. Belgacom has 85% of all DSL subscribers, with the remaining 15% using an unaffiliated ISP who obtains broadband access from Belgacom at wholesale rates. In 2002, there were 517 thousand DSL subscribers (59%), 326 thousand cable modem subscribers (38%), and 26 thousand other subscribers (3%). D. Denmark: Unbundling local loop helps new entrants offer DSL, but incumbent regains lost market share. Since 1998, Denmark has required the incumbent telecommunications carrier to unbundle the local loop. Line sharing has been required since 2001. As a result, new entrants gained a market share of 44% of DSL lines in October 2001. By December, 2002, however, this percentage had declined to 21% after the incumbent captured market share by lowering prices. Even with the decline, Denmark has one of the highest percentages of DSL lines sold by new entrants. 17 17 Competitive carriers market share of DSL lines in other European Union countries range from 0% to 4%. 15 - 15 - Sixty percent of households are passed by cable and half of these households are passed by upgraded cable television networks capable of providing broadband service. There are two large cable providers: TDC, which is owned by the incumbent telecommunications carrier, and TeliaSonera. Only 3% of TDC’s potential customers have signed up for broadband access, compared to 13% of Telia Sonera’s customers. Some have drawn comparisons between Denmark and Belgium, noting that the markets often have comparable subscriber levels for telecommunications and Internet. In the broadband market, while about 70% of the Danish population has access to DSL that provides at least 2 mbps downstream service, in Belgium higher speeds such as 4 mbps service are more widely available. One possible explanation is that, unlike in Denmark, the Belgian incumbent telecommunications carrier does not own any cable networks, and is better able to compete, e. g., by offering more capacity. In 2002, there were 307 thousand DSL subscribers (69%), 133 thousand cable modem subscribers (30%), and 6 thousand other subscribers (1%). E. Sweden: Ethernet LANS start broadband competition. Competition from cable weaker because owned by telco. The leading technology for broadband access in 2000 was neither DSL nor cable modem access. Instead, Ethernet LANS connected more broadband customers than any other technology. 18 The leading provider, Bredbandsbolaget (B2) offers 10 mbps broadband access utilizing its own fiber optic network and switched Ethernet networks within large apartment buildings. By the end of 2001, however, DSL subscribers outnumbered subscribers from all other technologies. TeliaSonera, the incumbent telecommunications carrier, provides an integrated DSL service to 75% of all DSL subscribers and has a substantial wholesale DSL business accounting for 24% of all DSL subscribers. TeliaSonera raised prices of 18 See, supra. n. 11, for definition of “Ethernet LANS.” 16 - 16 - broadband services to individual homes (which are not served by B2) by 30% in 2001. Local loop unbundling has been required since 2000 and line sharing since 2001. At the end of 2002, there were only 2282 unbundled loops, out of a total of seven million local loops. Competitors complain that TeliaSonera’s local loop unbundling prices is higher than its DSL retail price. About 65% of Swedish homes are passed by cable. TeliaSonera also, until recently, owned ComHem, the largest cable operator with over 60% of total cable subscribers. In 2002, only 2.7% of homes passed by ComHem were cable modem subscribers. By contrast, 15% of homes passed by UPC, the other large cable operator, were broadband access subscribers. In April 2003, the European Commission directed TeliaSonera to divest its cable networks. In 2002, there were 424 thousand DSL subscribers (59%), 153 thousand cable modem subscribers (21%), and 142 thousand other subscribers (20%). F. Switzerland: Neither unbundling nor open access required, nevertheless competition between cable modem and DSL has been strong. In 1995, Swisscom, the incumbent telecommunications carrier, acquired 32% of Cablecom, the largest cable operator with over half of all subscribers. In 1998, despite the opposition of Swisscom, Cablecom began to build its own broadband network. Swisscom subsequently sold its stake in Cablecom in 1999 and began to offer its own DSL services in 2000. There is now strong competition between the two providers. In the fourth quarter of 2002, both Swisscom and Cablecom added 60,000 new subscribers. This is one of the highest per capita rates of growth in OECD countries. Local loop unbundling was introduced in April 2003. Swisscom offers DSL at wholesale rates to other ISPs and an integrated DSL package directly to consumers 17 - 17 - through its own ISP. Approximately 40% of DSL subscribers buy the service from ISPs unaffiliated with Swisscom. In 2002, there were 260 thousand cable modem subscribers (57%) and 195 thousand DSL subscribers (43%). G. Japan: Unbundling local loop the primary method of competitor entry into broadband. Fiber networks are extensive. Until December 2000, NTT, the incumbent telecommunications carrier, was still marketing ISDN lines instead of developing a DSL offering. At that time, there were fewer than 70,000 DSL subscribers in Japan, two- thirds with new entrants. NTT was required to offer local loop unbundling and line- sharing (but not bit stream access). In December 2000, NTT was also required to unbundle fiber- to- the- home lines. Softbank Group, with its subsidiaries Yahoo- Japan and BB Technologies, took advantage of the unbundling rules to launch a new DSL offering in September 2001 that was hugely successful, and NTT also began to compete by cutting prices for its DSL services. By the end of 2001, the number of DSL subscribers had increased to 2.3 million. This number grew to 5.6 million at the end of 2002, and over 7 million by March 2003. NTT, the incumbent, has a market share of less than 40% of DSL lines. Non- incumbents, of which the largest is Yahoo BB (with 1.5 million customers), have a market share of around 60%. Cable operators played a significant role in spurring broadband growth in Japan in the earlier years. In 2000, for instance, Jupiter had 141,000 cable modem subscribers, well ahead of the DSL numbers. Since the, however, growth in cable modem access has been far slower. From 2001 to 2002, for instance, cable modem subscribers grew from 1.3 million to 1.95 million. One factor may be that cable networks pass only one third of Japanese households, compared to the four- fifths coverage of DSL. Japanese cable companies, such as J- Com, however, drove competition by offering higher bandwidths of 8 mbps downstream and 2 mbps upstream— speeds which are now matched or exceeded by the leading DSL providers. 18 - 18 - Fiber optic cable services, which offer 100 mbps access, are available to 43% of Japan and had more than 200,000 subscribers in 2002. A leading company is USEN. In 2002, there were 5.6 million DSL subscribers (72%), 1.9 million cable modem subscribers (( 25%), and 200 thousand other subscribers (3%). H. Germany: Incumbent telco owned cable operator until 2003. Broadband slow to develop. Local loop unbundling has been required since 1996, and line sharing since 2001. New entrants, however, have leased only about 2% of local loops and are not significant providers of broadband access in competition with Deutsche Telecom. DT signed up 1.6 million customers in 2001, but then raised its prices. New subscribers in 2002 were less than a million. A small number of subscribers [less than 6%] buy broadband access from DT and choose an unaffiliated ISP. DT does not offer DSL at wholesale rates to unaffiliated ISPs. Although 86% of German households are passed by cable networks, cable modem service is available only to around 260,000 households, and less than 60,000 are subscribers. More than 98% of broadband access subscribers use DSL services, and less than 2% use cable modem services. These low numbers can be explained in part by the fact that DT, until recently, owned the cable backbone networks and had little incentive to develop cable modem services. In March 2003, DT had sold its majority ownership stakes in all the cable networks. HanseNet Telekommunikation offers broadband access through its fiber optic network in Hamburg to about 60,000 customers. Although more than 860 fixed wireless licenses were awarded to 12 different operators, few subscribers have signed up. 19 - 19 - In 2002, there were 3.1 million DSL subscribers (96%), 56 thousand cable modem subscribers (2%), and 70 thousand other subscribers (2%). I. United Kingdom: Cable operators late to offering broadband service, telco also slow to act. Broadband access has developed slowly in the U. K., although cable operators and telecommunications carriers have had the longest experience with infrastructure competition of any OECD country. Cable operators focused more on offering telephony and digital television services than cable modem services. Cable operators, primarily NTL and Telewest, began offering cable modem services in 1999 and 2000. Cable passes about 50% of UK homes as of March 2003. BT, the incumbent telecommunications carrier did not offer DSL services until May 2001, making the UK among the last major developed countries to have DSL. Although local loop unbundling is required, fewer than 2000 lines were unbundled in 2002. BT has a significant wholesale DSL business, amounting to 49% of all subscribers in March 2003. Sub- broadband services, i. e., cable modem access at 128 kbps, are popular to a greater degree than in other OECD countries. As of March 2003 about 57% of the population has access to broadband either through DSL or cable modem service. About 25% can choose between either DSL or cable modem service. In 2002, there were 779 thousand cable modem subscribers (57%) and 590 thousand DSL subscribers (43%). 20 - 22 - Figure 6 19 Broadband Access Platforms 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1999 2000 2001 2002 1999 2000 2001 2002 1999 2000 2001 2002 1999 2000 2001 2002 1999 2000 2001 2002 Percentage Korea Canada Belgium Denmark Sweden 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1999 2000 2001 2002 1999 2000 2001 2002 1999 2000 2001 2002 1999 2000 2001 2002 1999 2000 2001 2002 Percentage DSL Subscribers Cable Modem Subscribers Other Subscribers U. S. A Switzerland Japan Germany U. K 19 Broadband Over Cable Television at table 4. 23 - 23 - References Aizu, Izumi, “A Comparative Study of Broadband in Asia: Deployment and Policy” (Asia Network Research, September 29, 2002). European Cable Television Association, ECTA DSL Scorecard, at www. ectaportal. com/ ectauploads/ dsl_ apr03. xls. FCC, High- Speed Services for Internet Access: Status as of December 31, 2002 (Industry Analysis and Technology Division, Wireline Competition Bureau, June 2003). FCC, High- Speed Services for Internet Access: Status of June 30, 2002 (Industry Analysis and Technology Division, Wireline Competition Bureau, December 2002). ITU, Promoting Broadband: Background Paper (April 2003) ITU, ITU Telecommunications Indicators 2003. OECD, Broadband Access in OECD countries (2001). OECD, Broadband Access for Business (2002). OECD, Communications Outlook 2003. (2003) OECD, Developments in Local Loop Unbundling (2003). OECD, Broadband and Telephony Services Over Cable Television Networks (2003) (unpublished draft, May 9, 2003). World Bank. World Development Indicators Online. 2003. Wu, Irene, Canada, South Korea, Netherlands and Sweden: Leading Markets in the Convergence of Telecommunications, Broadcasting and Internet Services (2003) (forthcoming in Telecommunications Policy). www. point- topic. com [website on DSL services]. 24