1328

MOBILE VIRTUAL NETWORK OPERATORS (MVNOS) IN ISRAEL
ECONOMIC ASSESSMENT AND POLICY RECOMMENDATION
Prepared for the State of Israel, Ministry of Communications and Ministry of Finance
Final Report: 1 August 2007
Note: "[]" indicates confidential content that has been removed
Project Team Nigel Attenborough, Director, NERA London Christian Dippon, Vice President, NERA San Francisco Soren Sorensen, Consultant, NERA London
NERA Economic Consulting 15 Stratford Place London W1C 1BE United Kingdom Tel: +44 20 7659 8500 Fax: +44 20 7659 8501 www.nera.com
NERA Economic Consulting One Front Street, Suite 2600 San Francisco, CA 94111 USA Tel: +1.415.291.1000 Fax: +1.415.291.1020 www.nera.com
Contents
Page 1. Executive Summary ................................................................................................................. 1 2. Introduction .............................................................................................................................. 5 2.1. MVNO Definitions ........................................................................................................... 6 2.2. Historical Overview.......................................................................................................... 7 2.3. Technical Considerations of MVNOs .............................................................................. 7 2.4. MVNO Integration ........................................................................................................... 8 2.5. MVNO Operational Components..................................................................................... 8 2.6. MVNO Business Models................................................................................................ 10 2.6.1. Classification of MVNO business models ............................................................. 11 2.6.2. Challenges facing MVNOs .................................................................................... 12 3. Review of the Activity and Impact of MVNO Players Worldwide ....................................... 14 3.1. United States................................................................................................................... 14 3.1.1. Available technological infrastructure ................................................................... 14 3.1.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 15 3.1.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale, and impact)............................................................................................. 18 3.1.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 21 3.1.5. MVNO business models (positive and negative aspects and implications)........... 24 3.1.6. Pricing Trends before and after MVNO entry ....................................................... 25 3.2. United Kingdom ............................................................................................................. 28 3.2.1. Available technological infrastructure ................................................................... 28 3.2.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 29 3.2.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact).............................................................................................. 34 3.2.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 34 3.2.5. MVNO business model (positive and negative aspects and implications) ............ 35 3.2.6. Pricing Trends before and after MVNO entry ....................................................... 35 3.3. Ireland............................................................................................................................. 38 3.3.1. Available technological infrastructure ................................................................... 38 3.3.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 38 3.3.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale, and impact)............................................................................................. 40 3.3.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 45 3.3.5. MVNO business model (positive and negative aspects and implications) ............ 45 3.3.6. Pricing Trends before and after MVNO entry ....................................................... 45 3.4. France ............................................................................................................................. 46 3.4.1. Available technological infrastructure ................................................................... 46 3.4.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 46 3.4.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact).............................................................................................. 48 3.4.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 48
NERA Economic Consulting
3.4.5. MVNO business model (positive and negative aspects and implications) ............ 49 3.4.6. Pricing Trends before and after MVNO entry ....................................................... 49 3.5. Denmark ......................................................................................................................... 50 3.5.1. Available technological infrastructure ................................................................... 50 3.5.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 50 3.5.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact).............................................................................................. 53 3.5.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 53 3.5.5. MVNO business model (positive and negative aspects and implications) ............ 54 3.5.6. Pricing Trends before and after MVNO entry ....................................................... 54 3.6. Spain ............................................................................................................................... 55 3.6.1. Available technological infrastructure ................................................................... 55 3.6.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 55 3.6.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact).............................................................................................. 59 3.6.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 60 3.6.5. MVNO business model (positive and negative aspects and implications) ............ 61 3.6.6. Pricing Trends before and after MVNO entry ....................................................... 61 3.7. Hong Kong ..................................................................................................................... 61 3.7.1. Available technological infrastructure ................................................................... 62 3.7.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry........................................................................... 62 3.7.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact).............................................................................................. 65 3.7.4. MVNO entry (current and forecasted market shares of MVNOs)......................... 68 3.7.5. MVNO business model (positive and negative aspects and implications) ............ 69 3.7.6. Pricing Trends before and after MVNO entry ....................................................... 69 3.8. Other Countries .............................................................................................................. 69 3.8.1. Japan....................................................................................................................... 69 3.8.2. South Korea............................................................................................................ 70 3.8.3. Jordan ..................................................................................................................... 70 3.8.4. Norway................................................................................................................... 71 3.8.5. Finland ................................................................................................................... 72 3.8.6. Canada.................................................................................................................... 73 3.9. Lessons Learned from the Benchmark Countries .......................................................... 73 4. Assessing the Need for MVNO-Related Regulation in Israel ............................................... 75 4.1. Competitive Conditions in the Israeli Mobile Market.................................................... 75 4.1.1. Market definition.................................................................................................... 75 4.1.2. Analysis of the level of competitiveness of the Israel retail market with respect to MVNOs ................................................................................................. 76 4.1.3. Market shares and market-share trends.................................................................. 77 4.1.4. Pricing and Profitability ......................................................................................... 84 4.1.5. Barriers to entry and expansion ............................................................................. 99 4.1.6. Historical behavior............................................................................................... 110 4.1.7. Likelihood of collusion ........................................................................................ 110 4.1.8. Summary of competitive conditions in the Israeli wireless market ..................... 117 4.2. Business Models for Israeli MVNOs ........................................................................... 119
ii
4.2.1. Access to Infrastructure ....................................................................................... 119 4.2.2. Value Proposition................................................................................................. 127 4.2.3. Customer relations ............................................................................................... 127 4.2.4. Finances ............................................................................................................... 127 4.3. Understanding the Absence of MVNOs in Israel......................................................... 127 4.4. Effects of MOC-Mandated MVNO Players in Israel ................................................... 131 4.5. Regulatory Principles of Competition Policy............................................................... 132 4.6. Is Regulatory Intervention Necessary for MVNO Entry into the Mobile Market in Israel? .......................................................................................................................... 134 4.7. Are There Policies or Regulatory Measures for MVNO Activity to Have a Positive Effect or in Furthering the MOC Mandate? ...................................................... 134 4.8. MVNO Policy Recommendation for Israel .................................................................. 134 4.9. Recommended Business Models for MVNO Success and Furtherance of the MOC Mandate................................................................................................................. 136 Appendix A Outgoing RPM analysis .................................................................................... 137 Appendix B - Profitability of Israeli Mobile Operators............................................................ 140 Appendix C ROCE Rates in Excess of WACC Do Not Necessarily Indicate Excessive Returns ................................................................................................................ 142 Appendix D Sector-Specific ROCE Methodology ................................................................ 145 Appendix E Normalized Israeli Churn Rates......................................................................... 146 Appendix F Discounted Cash Flow Analysis ........................................................................ 147
iii
Figures
Page Figure 1 MVNO Operational Components ............................................................................ 10 Figure 2 US Wireless Subscribers 19942006 ...................................................................... 15 Figure 3 US 2G Wireless Market Shares (by subscribers) .................................................... 16 Figure 4 US Wireless Subscriber Growth.............................................................................. 17 Figure 5 Projected MVNO Revenues in US$ 20022010 ..................................................... 24 Figure 6 US Average Revenue Per User 19932005............................................................. 26 Figure 7 US Average Revenue Per Minute 19932005......................................................... 27 Figure 8 Monthly Consumer Price Index Wireless Telephony 19982005........................... 28 Figure 9 UK Wireless Subscribers 20002006...................................................................... 29 Figure 10 UK Total Wireless Market Shares (by subscribers) .............................................. 30 Figure 11 UK Average Revenue Per User 20012005 .......................................................... 36 Figure 12 UK Average Revenue Per Minute 20012005 ...................................................... 37 Figure 13 Wireless Subscribers Growth in Ireland 20002006 ............................................. 39 Figure 14 Wireless Network Operator Market Shares Ireland............................................... 40 Figure 15 France MVNO Subscribers June 2005December 2006....................................... 49 Figure 16 Denmark Wireless Subscribers 19962006........................................................... 51 Figure 17 Denmark MNO Market Shares (by subscribers) ................................................... 52 Figure 18 Spain Wireless Subscribers 20002006................................................................. 56 Figure 19 Spain Total Wireless Shares (by subscribers) ....................................................... 57 Figure 20 Wireless Subscribers Growth in Hong Kong 19902006...................................... 63 Figure 21 Subscriber Market Shares 2001 vs. 2006 .............................................................. 64 Figure 22 Israel Total Wireless Market Shares 2006............................................................. 78 Figure 23 Global Comparison of HHIs 2006......................................................................... 79 Figure 24 Israel Total Wireless Market Shares 19992006................................................... 80 Figure 25 Israel Wireless Market Shares Private Sector 20022006..................................... 82 Figure 26 Israel Wireless Market Shares Business Sector 20022006.................................. 83 Figure 27 Israel Wireless Market Shares Prepaid Sector 20022006.................................... 84
iv
Figure 28 PPI Adjusted RPM of Outgoing and Incoming Voice Traffic 2006 ..................... 85 Figure 29 Change in Revenue Per Minute from 1Q06 to 4Q06 ............................................ 86 Figure 30 Outgoing RPM (US$) Trends Postpaid Private Sector Israel 2004-2006 ............. 88 Figure 31 Outgoing RPM (US$) Trends Business Sector Israel 2004-2006 ......................... 88 Figure 32 Outgoing RPM (US$) Trends Prepaid Sector Israel 2004-2006 ........................... 88 Figure 33 EBITDA as a Percentage of Revenue 4Q06.......................................................... 95 Figure 34 EBITDA Margin vs. HHI 2006 ............................................................................. 96 Figure 35 Profit per Minute vs. GDP 4Q06 ........................................................................... 97 Figure 36 EBITDA Margins by Region 20022006.............................................................. 98 Figure 37 EBITDA per User vs. GDP 4Q06.......................................................................... 99 Figure 38 Average Monthly Churn Rates by Region 20022006 ....................................... 108 Figure 39 Average Churn Rates vs. Share of Prepaid Subscribers ...................................... 109 Figure 40 MVNO Policy Recommendation for Israel ......................................................... 135
Tables
Page Table 1 US MVNO Subscribers as a Percent of Wireless Subscribers 19992010............... 22 Table 2 Technological Infrastructure Ireland......................................................................... 38 Table 3 Denmark Total Wireless Market Shares (by subscribers) ........................................ 54 Table 4 Technological Infrastructure Hong Kong ................................................................. 62 Table 5 Israel MNO Subscriber Growth 20032006 ............................................................. 81 Table 6 Cumulative Change in Revenue Per Minute 20022006.......................................... 87 Table 7 Outgoing RPMs (US$) by Sector and Company 2006 ............................................. 87 Table 8 ROCE of Israeli MNOs 20032006.......................................................................... 91 Table 9 Profitability and CC Findings ................................................................................... 92 Table 10 ROCE by Sector Largest Israeli MNOs 2003-2006 ............................................ 93 Table 11 ROCE by Sector Partner Communications 2003-2006 ....................................... 93 Table 12 ROCE by Sector Cellcom 2003-2006.................................................................. 94 Table 13 ROCE by Sector Pelephone 2003-2006............................................................... 94 Table 14 UK SIM Locking Policies November 2002.......................................................... 105 Table 15 Considerations of Collusive Behavior in Israel Summary.................................... 116 Table 16 Generic DCF Model Assumptions Israel .............................................................. 124
vi
1.
EXECUTIVE SUMMARY
Mobile virtual network operators (MVNOs) are non-facility-based wireless carriers that lease network capacity from facility-based carriers (MNOs) and then resell wireless subscriptions using their own branding and value-added services. MVNOs have entered the market in the US, the UK, and many other countries with voluntary, commercially negotiated wholesale contracts with a host MNO. The questions we seek to answer in this report are why MVNOs have not entered the Israeli wireless market, whether such entry would increase competition and consumer welfare, and what the Ministry of Communications should do, if anything, to encourage MVNO entry. To answer these complex questions, we conducted three general types of analyses. First, to gain a thorough understanding of how MVNO entry occurred in different countries and to determine its impact on competition and consumer welfare, we studied MVNO entry in 13 different countries. Second, we examined the competitive conditions in Israel with respect to MVNO entry to determine whether the current lack of MVNOs was the result of market failure or simply the consequences of market forces. Third, in order to assess whether MVNO entry is economically profitable, we studied possible business models for Israeli MVNOs and detailed short-run and long-run survival challenges. The benchmarking study revealed a number of interesting findingsthe first was that all of the countries used as benchmarks had at least one MVNO. Second, MVNOs tend not to compete directly with their host MNO or with MNOs in general. Rather, they aim to widen and deepen the market through brand appeal, the targeting of niche markets, and the presence of alternative distribution channels. Third, the consumer-welfare impact of MVNOs is in their extended and innovative service offerings as opposed to lower prices. Fourth, while some countries have many MVNOs, typically only three in each country have the majority of MVNO customers. Moreover, the combined MVNO market share per country is less than 10 percent. Fifth, MVNOs generally use older second generation technologies and have not upgraded to third generation technologies as have most MNOs. Finally, the regulatory agencies in the benchmark countries did not intervene in MNO-MVNO relationships with the exception of Spain and Hong Kong. However, in all cases, the regulatory agencies based their decision to intervene or not on a competitive review of the relevant market. If market forces work properly (i.e., there is no market failure) regulators usually refrain from intervention, although they do continue to carefully monitor the MNO-MVNO relationship. Based on these findings, we analyzed the competitive conditions in the Israeli retail wireless market with respect to MVNO entry and found: Overall market shares: The largest operators have relatively low market shares. This implies that no particular MNO has a dominant position. Specifically, the two largest firms, Cellcom and Partner, have a smaller combined market share than is the case in most other countries. Similarly, the HHI index in Israel is lower than in most other countries. Sector-specific market shares: The limited role of MIRS in the private sector results in a higher concentration in this sector relative to the overall market. However, even at this elevated level, the concentration is still low by international standards. Thus, the near absence of MIRS in the private sector is not necessarily of competitive concern.
Overall market share trends: While market shares have changed less in recent years, this might be a sign of a maturing market. Similar patterns can be observed in the US and the UK. More importantly though, the MNOs overall growth patterns from 20032006 differ, which is consistent with competition rather than market failure. Sector-specific market share trends: Israeli market shares by customer segment do differ and have fluctuated over the last five years, which is not consistent with the existence of market power or collusion. However, given the decreasing difference in market shares among the three largest MNOs, we do recommend that the Ministry of Communications continue to monitor the developments in sector-specific market shares as in a competitive market we would expect market shares to continue to change over the next few years rather than converge to the same level. Overall Prices: Israeli prices, measured by total revenue per minute, appear low by international standards. While we have a number of concerns about any conclusions drawn from such a comparison, it appears that the available evidence on pricing does not of itself indicate market failure. Pricing trends: Recently Israeli prices overall have fallen less than in other nations. However, prices for postpaid and business services have decreased while prices for prepaid services have increased. The large variance in the MNOs price changes is an indicator that they have not acted in a concerted manner. Hence, neither pricing trends nor current price levels by themselves suggest market failure. Service-specific prices: Examining the prices charged by the Israeli MNOs for prepaid, private postpaid and business services reveals different prices and different pricing strategies. They also indicate different responses to the MOCs mandated decrease in interconnection revenue, confirming that prices by themselves are not indicative of market failure. Overall profitability: Profitability levels of Israeli MNOs when measured by ROCE and EBITDA margins: (1) have fluctuated over time, (2) are asymmetric, (3) are below what UK regulators have historically considered as excessive, and (4) are lower than in most countries. Furthermore, profits per minute are low relative to countries with similar GDP per capita. Profits per user are comparable or on the low side relative to countries with similar GDP per capita. Hence, the profitability of Israeli MNOs by itself provides no evidence of market failure. Sector-specific profitability: A sector-specific ROCE approximation for Cellcom, Partner, and Pelephone indicates higher profitability in the private sector than in the business sector. This, in turn, could indicate less competition in the private sector, relative to the business sector. However, given the margin of error typically associated with this type of analysis and the fact that the three MNOs private ROCEs do not appear excessive by international standards, a firm conclusion cannot be drawn. Barrier to entry and expansion: Spectrum in Israel appears not to be a significant barrier as spare spectrum is available for one additional MNO (to provide both 2G and 3G services). Regulatory approval for both MNOs and MVNOs might amount to a barrier to entry. Access to capital markets should not be an entry barrier, at least for larger companies already in the communications industry. However, municipal restrictions on 2
new cell sites, as well as the lengthy approval process associated with them, can represent a barrier to entry. Similarly, economies of scale represent a barrier to entry, although they could be the result of first-mover advantage and market saturation. Switching barriers do exist in the Israeli market. While in many cases they may result from competitive behavior rather than attempts to restrict competition, we are concerned about the predominance and uniform offering of 36 months contracts. Typically, term contracts in other countries range from 12 to 24 months in length. We conclude that some of the existing barriers to entry and expansion are not signs of market failure. However, the current predominance of 36-month contracts along with some of the terms and conditions and potentially the promotion of seemingly inferior contracts appear not to be consistent with competitive markets. Churn rates: Customer churn in Israel is low by international standards, although we do question the correlation of customer churn and competition. Given the impending introduction of mobile number portability, we do not find current churn rates to be evidence of market failure. Historic behavior: In addition to excessively long term contracts, various parties, including the Ministry of Communications, have alleged historic behavior by the Israeli MNOs, which is suggestive of the existence of market power. It is not within our remit to review the evidence of such alleged behavior as they do not directly related to MVNO entry. If substantiated, it would not necessarily prevent market forces from working (i.e., it is not necessarily a sign of market failure). It would, however, suggest that the level of competition in Israel could still be improved. Collusive behavior: While we find that collusion would be difficult in the Israeli wireless market, we cannot entirely rule out the possibility of it occurring particularly with respect to MVNOs.
Our review of the competitive conditions in the Israeli wireless retail market with respect to MVNO entry does not suggest the presence of endemic market failure that would be sufficient to preclude the entry of MVNOs. Thus, the absence of MVNOs in Israel seems to be primarily the result of market forces and/or the absence of an established licensing process for MVNOs. Notwithstanding, there are aspects of the market that are seemingly inconsistent with competitive markets. Therefore, while we find no presence of endemic market failure, we cannot entirely rule out the possibility of such, at least with regard to wireless wholesale access. With respect to possible business models for Israeli MVNOs, we find that the successful MVNO will be partially facilities based, have its own distribution channels, and will not compete directly with the MNOs. Given the high saturation in the Israeli mobile market, the potential lack of demand for prepaid services, relatively low existing RPMs and ROCEs, and the existence of customer-switching barriers, the business case for MVNOs in Israeli is weak. However, for carriers with existing retail distribution channels and experience in the communications industry, a possible business case does exist. In light of these findings, we recommend against the adoption of an MVNO policy at this point in time. Instead, we recommend allowing market forces to work by licensing MVNOs and encouraging them to negotiate with the MNOs. Should these negotiations be successful,
then we recommend that the Ministry of Communications continue to refrain from intervening in the wholesale market. Should no wholesale agreements be reached within a given period of time (i.e., 6 months after the first licensed MVNO has contacted an MNO in writing expressing its interest in negotiating a wholesale access contract or 9 months after the MOC gives public notification that it intends to issue MVNO licenses, whichever is longer), then we recommend that the Ministry of Communications carefully review the reasons behind each of these failed negotiations. If this analysis reveals that the failed negotiations are due to anticompetitive tactics by the MNOs, we recommend that the Ministry of Communications consider implementing wholesale access regulations. If the analysis determines that the failed negotiations are not a result of anticompetitive behavior, we recommend that the Ministry of Communications continue refraining from regulation. In the meantime, to further improve competitive conditions in Israel, we recommend that the Ministry of Communications encourage mobile number portability and the introduction of shorter term contracts. We also recommend that the Ministry of Communications ensure that early termination fees are commensurate with the economic damages caused by early termination and that pricing plans are transparent and consumers understand the different plans and options available to them. In addition, we encourage the Ministry of Communications to open up new markets, such as WiFi and WiMAX to increase competition and to investigate its own list of issues to ensure that the Israeli market is consistent with competitive markets and thus open to potential MVNO entry.
2.
INTRODUCTION
This study has three main objectives. The first is to analyze the likely effects of MVNO entry into the mobile market in Israel. Second, it assesses the degree to which Israeli regulators should mandate or regulate MVNO entry. Third, this study can be used as a reference document for policy design and implementation. We have divided the study into the following analytical tasks: A comprehensive overview of the policy issues germane to the MVNO question, including economic, technological, regulatory, and consumer issues A comprehensive review of the introduction and activity of MVNOs in a range of benchmark countries An assessment of the need for MVNO-related regulation in Israel, including a review of the competitive conditions in the Israeli mobile with regards to MVNO entry. An analysis of the possible competitive effects of regulatory-mandated MVNOs on the Israeli mobile telephony market An evaluation of whether the introduction of regulatory-mandated MVNOs in Israel will contribute to the fulfillment of the MOC mandate
In the remainder of this section, we review some generally accepted definitions of MVNOs and merge them into a working definition that we use for this study. In this section, we also provide an overview of the technical aspects of MVNOs, their integration with the underlying mobile network operators (MNOs), and a discussion of the various components of MVNO operations. The section ends with descriptions of some of the most commonly encountered business models for MVNOs. Section 3 applies the theoretical concepts from the previous section and provides a comprehensive review of the introduction and activity of MVNO players in a number of benchmark countries. In particular, our review includes MVNO activity in the United States, the United Kingdom, Hong Kong, Ireland, France, Denmark, and Spain. We have also added anecdotal evidence from other countries where appropriate. In particular, we address the issues that were most controversial when deciding on regulatory intervention or its absence. To the extent possible, we also provide insights into how the regulatory policies affected competition, retail prices, and overall consumer welfare. A summary of the lessons learned from the benchmark countries concludes this section. In Section 4 we assess the need for MVNO-related regulation in Israel. Specifically, we examine the Israeli mobile market for signs of market failure, study possible business models for Israeli MVNOs, and attempt to understand the absence of such operators in Israel. Furthermore, this section reviews from an economic perspective whether the introduction of regulatory-mandated MVNOs in Israel will contribute to the fulfillment of the MOC mandate. We address the costs and benefits of such regulation, mainly from the consumer perspective, to determine the degree to which the MOC should mandate or regulate MVNO entry. We also address many of the comments we received from different parties during our
visit to Israel in March 2007. We conclude this section with a policy recommendation to the MOC regarding MVNOs in Israel.
2.1. MVNO Definitions

A standard definition of an MVNO does not exist. The US Federal Communications Commission (FCC) defines MVNOs as resellers that purchase airtime from facilities-based providers and resell service to the public for profit.1 Under this definition, simple resellers that sell mobile services under an MNOs brand name would be classified as MVNOs. Industry analysts, on the other hand, define MVNOs as an arrangement where a network operator acts as a wholesaler of airtime to another firm, which then markets itself to users just like an independent operator with its own network infrastructure.2 This definition distinguishes MVNOs from resellers because they add value such as brand appeal, distribution channels, and other items, such as different pricing options, to the resale of mobile services. The UK regulator, Ofcom, offers probably the most general definition of an MVNO as an organisation which provides mobile telephony services to its customers, but does not have allocation of spectrum.3 Like the FCC, Ofcom uses the terms resellers and MVNOs as synonyms, but Ofcom does not require that airtime be purchased from facilities-based providers, and it does not require that such service be sold at a profit. Finally, the European Commission (EC) defines an MVNO as a mobile operator, which does not have a license to use radio spectrum, but has access to the radio infrastructure of one or more mobile operators and is able to offer services to customers using that infrastructure and its own network.4 The ECs definition expands on the FCCs and Ofcoms definitions by adding that resellers must be partially facilities-based in order to be considered MVNOs.5 Unlike simple resellers of telecommunications services, such as long distance, local exchange, and mobile network services, MVNOs typically add value to the resale of mobile services. Most MVNOs do not own facilities, although some leading providers have their own mobile switching centers (MSCs) and service control points (SCPs).6 Consequently, we
2 3
Before the Federal Communications Commission, 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, Eleventh Report, rel. September 29, 2006, 27 (FCC Eleventh CMRS Report). Tom Standage, The Internet, Undeterred, The Economist, October 11, 2001, p. 16. Office of Communications, Annex F Glossary, Statement on Wholesale Mobile Voice Call Termination consultation, http://www.ofcom.org.uk/consult/condocs/mobile_call_termination/ wmvct/annexf/ (accessed May 1, 2007). European Commission, Towards a New Framework for Electronic Communications infrastructure and associated services, The 1999 Communications Review, 1999, http://ec.europa.eu/comm/ information_society/policy/telecom/review99/pdf/review_en.pdf (accessed May 1, 2007). Interestingly though, in reporting the number of MVNOs in Europe, the EC seems to include pure resellers. See European Commission, 12th Report on the Implementation of the Telecommunications Regulatory Package 2006, Final Report 2006, March 29, 2007, p. 7. An MSC is a telephone switch, similar to a central office switch, which bridges a mobile telephone network with another telephone networksuch as the public switched telephone network. An SCP is a database residing in a Signaling System 7 network that is queried to determine how a call should be handled (e.g., an
base the definition of MVNOs for the purpose of this report on actual market observations and define MVNOs as: Companies that buy network capacity from at least one mobile network operator in order to offer their own branded mobile subscriptions and value-added services.
2.2. Historical Overview

Historically, the MVNO business model originated around the same time in both the US and Japan. In May 1996, TracFone Wireless, a subsidiary of Amrica Mvil, launched its pay-asyou-go wireless service in the US, while at the same time Japan Communications Inc. (JCI) introduced the data MVNO business model to Japan.7 The first MVNO in Europe was Virgin Mobile UK, which launched its mobile services in November 1999. Currently, over 50 MVNOs are in operation in the European Union (EU), serving a diverse set of customers and offering a range of voice and data services. Similarly, the mobile market in the US offers customers a choice of over 40 MVNOs.8 According to the FCC, approximately 6 percent of all US mobile subscribers were served by resellers, including MVNOs, and analysts find that the 13.4 million wireless subscribers (end of June 2005) served by resellers has nearly tripled since year-end 2003.9 The number of US MVNOs is likely to grow even more as retailers, such as Wal-Mart, contemplate launching MVNO offerings that tie their established brands to mobile services.10 The initial hype about MVNOs, however, seems to have played out. With the failure of ESPN Mobile, the bankruptcy of Ampd Mobile, and the limited success of other players (such as Disney Mobile), potential MVNOs are exercising extra caution and entering the market with more realistic expectations.11 The success stories of some MVNOs in the US and the EU, however, have caused similar business models to emerge in other countries where MVNOs seek to offer new and innovative product and service offerings to customers.
2.3. Technical Considerations of MVNOs

Fundamentally, there are no technical considerations of MVNOs as they are feasible on any network or network technology. Nevertheless, there are some technical issues that might impact the MVNOs service management, the availability of handsets and other devices and its service offerings.
9 10 11
SCP is consulted to provide the translation of an 800 number to an actual phone number and to bill the owner of the 800 number for the call). Simple wireless resale has been around for much longer than this and, using a broader definition of MVNOs, one could argue that the simple reseller of the early 1980s brought about the MVNO trends we observe today. Due to the constant entry and exit of MVNOs and the increased blurring between what constitutes an MVNO, a mobile virtual network enabler (MVNE), and a simple reseller, these numbers are approximations only. FCC Eleventh CMRS Report, 27. For a comprehensive list of US MVNOs, see http://www.takashimobile.com/mvno.html. For instance, while there were a lot of rumors and hype about Apple Computer, Inc. potentially becoming an MVNO, the company announced earlier this year that rather than entering as an MVNO, it will license its upcoming iPhone to AT&T Mobility instead.
MVNO service management comprises customer enabling, customer service variation, customer assistance, and billing. In practice, an MVNO can do as much or as little of these as it wishes, subject to the willingness of the underlying operator to provide control access to these functions. In selecting a host MNO, an MVNO will consider both the accessibility of the host MNOs service management system, as well as the MNOs reputation in this area. For instance, an MVNO commercial success will be negatively impacted if it selects an MNO with an inferior market reputation in terms of service management. A significant selling point of mobile services is the range and type of handsets offered. Thus, an MVNO might want to consider selecting a host network technology that is generally supported in the country it wishes to operate. For instance, GSM dominates much of the European markets and thus might be the prudent choice for a European MVNO. In the US or South Korea, on the other hand, CDMA handsets are widely available could offer an additional technical alternative to an MVNO in these countries. Finally in considering the various technical aspects, an MVNO must carefully select the desired network technology. Depending on the type of service offered, the level of competition in the market place, an MVNO might want to select a GSM, CDMA, or iDen network. For instance, if push-to-talk is of importance to the MVNOs strategy, iDen might be the preferred network. Alternatively, if one technology is least represented in a market, an MVNO might be able to negotiate more favorable wholesale access deals.
2.4. MVNO Integration

At a minimum, the launch and operation of an MVNO requires (1) a wholesale agreement with an MNO, (2) start-up capital, and (3) capital to cover operational expenditures and consumer acquisition costs. It is estimated that it takes, on average, US$25US$50 million and two years to launch an MVNO.12 Critical to the launch of an MVNO is the negotiation and maintenance of a wholesale agreement. Furthermore, if the MVNO decides to manage all aspects of its own business, it typically has to purchase a customer-relationship software application, a data platform, and billing software. Alternatively, an MVNO can outsource the operational aspects of its business to third parties, so called mobile virtual network enablers (MVNEs).
2.5. MVNO Operational Components

The main operational components of the MVNO model consist of (1) the access network, (2) the core network, (3) the service platform, (4) pricing and billing, (5) customer care, and (6) marketing and sales. The access network consists of the base stations and transceiver equipment that provides access to the spectrum. Because MVNOs do not have spectrum licenses, they are required to use the access network of at least one MNO. Typically, MVNOs have access network agreements with one or two MNOs, although a few MVNOs (such as TracFone in the US) have many such agreements.
12
Roger Entner, Entering the Wireless Market What you need to Know to Launch and Operate Your own Wireless Business, Ovuum 2006, p. 5.
The core network consists of switching and transmission, home location registers, intelligent network platforms, and so on. The service platform covers the activities and equipment used in the design and provision of services, while billing and pricing implement the pricing policies and billing options. Customer care encompasses all activities relating to after-sale customer care, and marketing and sales includes product and service marketing activities, sales operations, and customer acquisition activities. MVNOs can self-supply all the operational components listed above (with the exception of the access network). For instance, Frances MVNO Tele2 is known as an extended MVNO as it purchases only the access network from an MNO. The choice of the actual MVNO type depends on a variety of factors such as the amount of investment that MVNOs undertake and their possession of network elements from other telecommunications businesses, such as sites, leased lines, or billing systems, that could also be used in the provision of mobile services. Accordingly, MVNOs can be categorized as follows: Full or Extended MVNOs operate a core network that is comprised, at a minimum, of a mobile switching center and possibly a transmission network, home location registers, an intelligent network platform, and other network components. Full MVNOs have more independence and greater control over costs, traffic, and subscriber services. On the other hand, they are the most risky type of MVNO because they require relatively large amounts of investment. Full MVNOs cover the whole value chain except for the access network. Enhanced MVNOs are different from full MVNOs as they do not provide all the network elements and from basic MVNOs because they resell the MNOs services but provide additional value-added services. Basic MVNOs typically do not own network elements, but restrict themselves to customer-care operations, marketing and sales activities, and elements of pricing and billing. Pure Resellers do not offer customer care but simply provide marketing and sales functions to the MNO for a commission. As discussed above, we do not consider pure resellers to be MVNOs.
The network and operational components associated with the different types of MVNOs are summarized in Figure 1 below.
Figure 1 MVNO Operational Components
2.6. MVNO Business Models

MVNOs can expect to generate revenues and profits by targeting unserved market segments and adding value in the form of brand appeal, distribution channels, and other items to the resale of mobile services. In the US, in contrast to services offered by most MNOs, MVNOs focus mostly on the prepaid market segment, providing conveniently priced services that do not involve long-term contracts, credit checks, or early termination fees. Moreover, prepaid services are not subject to regulatory surcharges or additional fees for activation, E911, and mobile number portability.13 Most MVNOs also add innovative non-voice services, such as unique ring tones and games, to their service offerings. MVNOs in the US largely redefined the prepaid market segment and enjoyed initial success in the process. In Europe (in particular, the member states of the EU), the early success of prepaid services and the relatively quick adoption of mobile services caused mobile penetration rates to rise dramatically over the last decade.14 As a result, European MVNOs (unlike their US counterparts) faced a market where prepaid services were already well established and consequently tend to focus on adding value by offering data features, such as text messaging, ring tones, games, and music downloads.
13
14
Enhanced 911 (or E911) is an FCC program that requires wireless carriers to provide automatic number identification and location information for cell phone users that dial 911 emergency services. The fees associated with this program are recovered via a user charge. MVNOs are exempted from charging this mandatory recovery fee to their users. In October 2006, penetration rates exceeded 100 percent in 17 out the 25 EU member states.
10
Brand appeal is one of the major components of an MVNOs business model. As stated by Virgin Mobile CEO Tom Alexander: If you could put the spirit that exists throughout the whole of the Virgin Group into bottles to sell, youd be a multimillionaire . The brand is also our biggest asset because it symbolises the great customer experience that Virgin stands far [sic].15 In addition to capitalizing on brand appeal, MVNOs generate revenues and profits by selecting appropriate corporate structures and strategic partners. For instance, in its UK operations, Virgin Mobile formed a 50-percent-owned joint venture between the Virgin Group and MNO T-Mobile. Virgin finds it vital to have an agreement that is beneficial to both parties.16 As further elaborated by CEO Alexander: There is a network operator fear that MVNOs are going to steal a slice of their customer base and openly compete with them but what we actually do is provide T-Mobile with greater channels to market under another brand name, while the Virgin Group obtains an already built network.17 Other business models involve MVNOs having more control over their own fate. As described above, MVNOs can opt to become full or extended MVNOs. While this requires more start-up capital, it provides companies with more leverage to compete in the long run. Essentially, for each type of MVNO listed above, there exists a business model. In addition, each of these general types of business models can be customized in many different ways yielding a great many MVNO business models. For instance, in developing their business model, MVNOs must not ignore the importance of revenue management and billing systems. Whether to outsource such functions or keep them in-house along with determining which systems to select are strategic decisions that can affect an MVNOs growth and profitability. Other aspects of a sound business model include flexible pricing, multiple payment options, revenue sharing with MNOs, and the total cost of ownership. In addition, MVNOs frequently offer customers simplified prepaid service plans marked by the absence of long-term obligations or contracts, early termination fees, monthly bills, credit checks, deposits, age limits, activation costs, and other hidden costs. MVNOs offer lower prices to low-income customers, thus giving them more control over how much they spend on telecommunications. 2.6.1. Classification of MVNO business models In general, MVNO business models can be grouped into four (nonexclusive) categories: facilities based, target market, strategy based, and plan based. In the facilities-based grouping, MVNO business models range from partial facilities-based models (full or extended MVNO) to models where the MVNO does not own any network facilities and is a pure reseller. As discussed above, extended MVNOs enjoy the most control over their costs and have the largest competitive leverage in the long run. MVNOs operating as pure resellers, on the other hand, have no intrinsic advantages and are least likely to survive in the long run.
15 16 17
Matthew Secker, The Right MVNO Cocktail? View from the Top, Telecommunications International, May 2002. Ibid. Ibid.
11
The target-market-based grouping also has two general models: those that operate as discount MVNOs and those that serve lifestyle-based niche-market segments. For instance, US MVNO TracFone is a no-frills discount carrier, targeting lower-income households. On the other side of the spectrum is Voce Mobile, which targets high-end users with plenty of disposable income. Voce offers exclusive designs and services to a wealthy, image-conscious clientele. After an initial registration fee of $500, Voce charges $200 per month for unlimited national usage and provides a new handset every twelve months.18 The long-run sustainability of a pure, discount-based strategy is doubtful because, in the face of intensifying mobile (in particular, prepaid) competition, merely offering low prices without some form of product differentiation, brand value, or cost advantage could prove to be a losing proposition. In contrast, lifestyle MVNOs target specific niche-market demographics and differentiate their services from those offered by their competitors. By adding value for members of the niche-market segments that they serve, these MVNOs are more likely to remain viable in the long run. The strategy-based grouping categorizes MVNOs according to their specific strategies. This grouping is open ended, as MVNOs strategies can vary widely. For example, we have (1) low-cost international MVNOs, (2) brand-extension MVNOs, (3) subprime-credit MVNOs, (4) youth-focused MVNOs, and (5) network-carrier MVNOs.19 Again, the ultimate purpose of the different strategies is product differentiation and market segmentation. Finally, the plan-based grouping distinguishes MVNOs according to whether they offer prepaid or postpaid service plans. In the US, MVNOs that serve the prepaid market segment presently possess a comparative advantage, although their long-term prospects are uncertain. However, in countries where the prepaid segment is already at or near saturation and overall mobile market penetration rates are high that comparative advantage may not exist. 2.6.2. Challenges facing MVNOs Regardless of the business model selected, MVNOs must overcome challenges in order to survive in a competitive market. One challenge comes from their long-term dependency on, and relationship with, MNOs. Securing, negotiating, and maintaining wholesale contracts that are consistent with financial viability may not be easy. Even after negotiating such contracts, it can be challenging for MVNOs to capture and retain enough market share if their host MNOs or other MVNOs decide to focus on the same market segment as they do. A case in point is the market niche for tween mobile users. Historically, US MNOs did not specifically target the market for children between the ages of 8 and 12 (tweens). In March 2005, Firefly Mobile, an MVNO doing business with AT&T (formerly Cingular) on a wholesale basis, launched services and handsets geared toward tweens. Fireflys tweenfocused services, which are sold by AT&T, Alltel, and T-Mobile, are based on a simple pricing plan, $0.25 per minute for all domestic calls. Fireflys handset has speed dialing for parents, easy access to emergency services, and optionally rejects all phone calls from numbers not stored in the handsets phone book. Firefly prices its starter kit (handset,
18 19
Initially, Voce charged $1,500 and $500 per month for unlimited use and a free handset upgrade every four months. The prices above reflect Voces current offerings. See http://www.voce.com/main.html. Subprime credit refers to those with a credit history of late or missed payments. Frequently, MNOs check a new subscribers credit history before activating his account in an effort to make sure that the subscriber has the financial resources to pay for the plans monthly recurring costs. If a subscribers credit history is subprime, an MNO might refuse to provide service to him.
12
charger, clip, and 30 minutes of airtime) at $99.99. Only two years after Firefly went after this market segment, MNOs, MVNOs, and MNO/MVNO joint ventures, including LG Migo (offered by Verizon Wireless), TicTalk (offered by MVNO Enfora), and Kickflip (offered by MVNO Helio) are targeting the tween market segment. This is not surprising as economic theory suggests that in a competitive market short-term monopoly rents in a market segment will quickly be eroded. Another challenge for MVNOs is being able to capitalize on their brand reputations from outside telecommunications to identify or develop profitable niche-market segments within telecommunications. For instance, in autumn 2006, Walt Disney Company announced that it was shutting down its mobile ESPN wireless telephone service, an MVNO that had been launched with much fanfare and high expectations during the Super Bowl game just eight months earlier. ESPN Mobile is a prime example of an MVNO that was not able to capitalize quickly enough on its brand reputation from outside telecommunications (i.e., sports entertainment). A common problem for MVNOs, particularly in highly penetrated or saturated markets, is the lack of economies of scale, relative to the incumbent facilities-based MNOs. Long-term survival in these circumstances depends crucially on a value proposition that starts with being able to identify or develop the most profitable niche markets. Only then can an MVNO expect to produce sufficient revenues to remain in business until it achieves economies of scale. While not unique to MVNOs, mobile carriers that follow a niche strategy must constantly adapt their strategies to the changing needs of their target markets. MVNOs frequently introduce new and innovative services in order to differentiate themselves from their competitors, including the large, national MNOs. To do this, they must develop innovative business models and service offerings that are targeted to specific demographic groups and underserved consumer segments. However, as niche services become mainstream ones and the early advantages of product differentiation are dissipated, MVNOs must continually introduce new features and services in order to stay ahead of the game. Perhaps the greatest survival challenge for MVNOsand a source of new opportunities as wellis the ongoing process of convergence and intermodal competition in the expanded communications industry (involving fixed and mobile telecommunications, data transmission and storage, and multimedia and entertainment). The ability to offer a combination of voice, video, and data (or the triple play) over any one of several alternative platforms has become the aspiration of traditional networks and service providers and even MVNOs. With convergence, MVNOs are likely to face stiff competition from outside the mobile sector as well, particularly from cable and fixed providers that offer converged service offerings. However, convergence can also work to the MVNOs advantage in certain strategic relationshipsbest exemplified by the Bluephone product developed by BT (a fixedservice provider) and Vodafone (an MNO). Bluephone is a converged handset that connects via Bluetooth technology inside the home, but operates as a cellular device outside. In effect, outside the home, this makes BT an MVNO that uses Vodaphones backbone facilities. As similar converged services are introduced in communications markets, fixed- and cableservice providers will likely seek opportunities to become MVNOs themselves in order to further increase the demand for such services.
13
3.
REVIEW OF THE ACTIVITY AND IMPACT OF MVNO PLAYERS WORLDWIDE
In this section, we provide a comprehensive review of the introduction and activity of MVNO players in a number of benchmark countries. In particular, our review includes MVNO activity in the United States, the United Kingdom, Hong Kong, Ireland, France, Denmark, and Spain. We have also added anecdotal evidence from other countries where appropriate.
3.1. United States

The US was one of the first countries to experience the emergence of MVNOs. It is also one of the most fragmented markets in the world because of its technological infrastructure. Wireless technology has rapidly advanced from analog to digital (second generation or 2G), then to more advanced digital (third generation or 3G). While new and better technologies were emerging, wireless companies were merging with each other to form larger and larger companies. The FCC, chartered by Congress to conduct an annual review of the status of competition in the commercial mobile radio services industry, has repeatedly concluded that the marketplace is effectively competitive; therefore, they have declined to regulate it. 3.1.1. Available technological infrastructure With five different technological infrastructures available, the US is one of the most fragmented markets in the world. Currently, mobile service is primarily provided through four different 2G digital technology standardscode division multiple access (CDMA), global system for mobile communications (GSM), integrated digital enhanced network (iDen), and time division multiple access (TDMA). AT&T and T-Mobile have GSM networksthe leading technology in the US, while Verizon Wireless, Sprint Nextel (formerly Sprint), Alltel, and US Cellular have CDMA networks. Sprint Nextel (formerly Nextel) is the sole provider using iDen technology. iDEN is a proprietary Motorola technology, thus only Motorola manufacturers iDen handsets and equipment. TDMA is no longer used in new networks, and the few TDMA networks that remain are mainly for backup purposes.20 Estimates are that GSM has 39 percent of the US mobile market (in terms of subscribers), and CDMA has a 53 percent share and iDen 9 percent.21 Third generation services in the US are still relatively new. CDMA-technology users, such as Verizon Wireless and Sprint Nextel, have rolled out their services in a number of metropolitan areas using CDMA 1xEV-DO (evolution-data optimized), offering data speeds up to 400900 kbps for downloads and 100400 kbps for uploads.22 Due to the technical difficulties associated with upgrading and lacking the required spectrum, GSM providers still largely rely on EDGE (enhanced data rate for GSM evolution or 2.75G) networks, and 3G services (HSDPA or high-speed downlink packet access) are only available in selected cities
20 21 22
When referring to TDMA, we specifically refer to one application of TDMAthe IS-136 technology. We treat the other applications of TDMA (GSM and iDEN) separately. Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc., and NERA Research. Comrex Support, 3G Wireless Data Services, http://www.comrex.com/support/technotes/ 3GWirelessServices.html.
14
at speeds of 1.8 for downloads and 384 kbps for uploads (under optimal conditions).23 Currently, only about 2 percent of US mobile subscribers are on 3G networks.24 Unlike European and Asian markets, short message service (SMS) has not been popular in the US, where subscribers prefer to use their phones to talk or send emails rather than SMS messages. However, according to the CTIA-The Wireless Association, SMS messaging has been growing with the number of SMS messages sent doubling between 2005 and 2006. 3.1.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry In 2004 and 2005, the US wireless market experienced a number of MNO consolidations as Cingular Wireless merged with AT&T Wireless and Sprint PCS merged with Nextel, resulting in four remaining national MNOs. Furthermore, through the 2006 merger of AT&T and BellSouth, AT&T obtained the remaining 40 percent ownership share of Cingular Wireless, thereby becoming the sole owner. Consequently, Cingular Wireless was renamed AT&T Mobility. The mergers were largely in response to market conditions. Specifically, as illustrated in Figure 2, MNOs found their traditional market segments approaching saturation as wireless penetration rates grew modestly in recent years.
Figure 2 US Wireless Subscribers 19942006
250,000,000
200,000,000 Us Mobile Subscribers
150,000,000
100,000,000
50,000,000
0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Source: Derived from data in Federal Communications Commission, 11th CMRS Report.
23 24
NERA Research. Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. and NERA Research.
15
The expansion of traditional market segments was restricted by the limited availability of wireless spectrum. Moreover, MNOs were under pressure to recover the high costs of rolling out 3G services. As wireless service providers offered increasingly larger coverage areas, competition among them also increased. On the regulatory side, the FCC introduced mobile number portability (MNP) that lowered barriers to competition and, together with the recent convergence of communications technologies, greatly expanded customer alternatives for traditional voice services by enabling intermodal competition from cable, wireless, Internet, and broadband providers. Current monthly churn rates are 1.7 percent for AT&T, 1.1 percent for Verizon Wireless and 2.3 percent for Sprint Nextel (not including Sprint Nextels prepaid subsidiary, Boost Mobile).25 AT&T Mobility leads the market in terms of subscribers, with a share of 25.70 percent. Verizon Wireless is close second with 24.87 percent, followed by Sprint Nextel with 22.35 percent, and T-Mobile with 10.55 percent. Alltel Communications, the smallest US MNO holds 4.98 percent, while the largest two MVNOs, Tracfone and Virgin Mobile, hold 3.33 percent and 1.94 percent, respectively. The remaining 6.29 percent belong to other smaller MNOs and MVNOs.26 These shares are illustrated in the Figure 3 below:
Figure 3 US 2G Wireless Market Shares (by subscribers)
Others Tracfone Virgin T-Mobile Sprint Nextel Alltel
AT&T Mobility Verizon Wireless
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites
25 26
Amol Sharma, Sprint Nextel Swings to Loss Amid Market Share Decline, The Wall Street Journal, May 2, 2007, citing Gartner Dataquest 2006. Ibid
16
At 75.9 percent penetration, the US is not a highly penetrated market, lagging behind Europe and much of Asia. The number of wireless subscribers, however, continues to grow steadily as shown in Figure 4 below:
Figure 4 US Wireless Subscriber Growth
0.45 0.4 Annual Growth Margin 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Source: Derived from data in Federal Communications Commission, 11th CMRS Report.
In its annual reports to Congress, the FCC has repeatedly found the mobile market to be effectively competitive and, thus, has refrained from regulatory intervention.27 The FCC arrived at its finding using a framework that groups indicators of the status of competition into four categories: (1) market structure, (2) carrier conduct, (3) consumer behavior, and (4) market performance. With respect to market structure, the FCC found that by mid-2006 98 percent of the total U.S. population lives in counties with access to three or more different operators offering mobile telephone service, slightly higher than in the previous year, and up from 88 percent in 2000, the first year for which these statistics were kept.28 The FCC also analyzed the percentage of the US population living in counties with access to four or more different mobile carriers and found it to have increased as well during the second part of 2005 and the first part of 2006. While the percentage of the population covered by three or more operators has increased, providing US consumers with more choice, the population with an option of five or more carriers has decreased. The FCC concluded, that nevertheless none of the
27 28
See FCC Eleventh CMRS Report, 216. Ibid., 2.
17
remaining competitors has a dominant share of the market, and the market continues to behave and perform in a competitive manner.29 With respect to carrier conduct, the FCC found that innovative pricing plans and service offerings and service and price matching by other carriers are signs of a competitive market.30 Specifically, the FCC pointed out that new service plans geared towards historically untapped markets and innovative plan options are evidence of price rivalry among the carriers. The FCCs analysis also highlighted further metrics such as the deployment of nextgeneration networks (UMTS/HSDPA and EV-DO). With respect to consumer behavior, the FCC concluded that customers exert competitive pressure on price, contract terms and conditions, and quality of service by freely switching providers in response to differences in the cost and quality of service.31 The FCC found that monthly churn rates average between 1.5 percent and 3.0 percent, and cited the implementation of mobile number portability (MNP) as having lowered consumer switching costs.32 Finally, with respect to market performance, the FCC concluded that wireless competition continues to enhance consumer welfare. Specifically, the regulator cited increased subscribership, increased minutes-of-use (MOUs), increased use of SMS messaging, and reports of increased quality of service. In terms of retail prices, the FCC reported slightly ambiguous results. On one hand, it found revenue per minute and the cellular Consumer Price Index (CPI) declining, while a third indicator that measures the price of mobile service based on the consumption patterns of hypothetical users indicated a slight price increase. When benchmarking the per-minute price of mobile service in the US to Western Europe and Japan, the FCC found that US mobile voice calls are still far less expensive on a per minute basis.33 It is worthwhile to note that the FCC has repeatedly found the US mobile industry to be effectively competitive, starting with its market review in 20022003. At that point in time, the US had six mobile carriers (Verizon Wireless, Cingular Wireless, T-Mobile, AT&T Wireless, Nextel, and Sprint), a mobile penetration rate of 49 percent, 458 MOUs, and average revenue per minute of $0.12.34 Before 2002, the FCC referred to the US mobile market as increasingly competitive, but did not draw a firm conclusion on the level of competition. 3.1.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale, and impact) In the US, the case for regulation has traditionally rested on concerns about market power and monopoly leveraging opportunities (particularly in the presence of essential facilities) and the desire to capture beneficial network externalities. Because competition in the wireless sector
29 30 31 32 33 34
Ibid. Ibid., 3. Ibid., 4. Ibid. Ibid., 5. Before the Federal Communications Commission, 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, Eighth Report, 18 FCC Rcd 14783 (2003) Appendix D-14.
18
has long been considered robust, regulators believe that pro-MVNO regulatory intervention is not warranted. For example, with respect to market power in the wireless sector in the US, the FCC has noted: A mobile carrier can exercise market power only to the extent that mobile subscribers do not respond to price increases or other adverse competitive effects. If, to the contrary, enough consumers are sufficiently well-informed to take prices and other non-price factors into account when choosing their service provider, and likewise, if enough consumers have the ability and propensity to switch service providers in response to an increase in price or other harmful conduct, then the carriers will have an incentive to compete on price and non-price factors. Consumer behavior will be more effective in constraining market power when the transaction costs subscribers incur in choosing and switching carriers are low. Transaction costs depend on, among other factors, subscribers access to and ability to use information, and costs and barriers to switching carriers.35 Thus, US regulators seem not to be concerned about anticompetitive leveraging opportunities and forbear from imposing most regulatory constraints on wireless service providers, including pro-MVNO intervention. It is instructive in this regard to review the thinking of US regulators (in particular, the FCC) on the wireless sector and how that thinking may or may not have influenced the evolution of the US wireless sector (and MVNOs, in particular). The initial startup costs of facilities-based operations are high in both the wireless and wireline sectors. That is, to serve even a few customers, MNOs must make large investments to acquire spectrum rights, either at auction or in a secondary market, and to build an infrastructure of transmission antennas, base stations, and switches. Within a given metropolitan area, the average cost per customer of those investments is obviously much lower for the millionth customer than for the first. However, that fact alone does not make wireless telephony a natural monopoly, just as the high fixed costs and low unit costs of the semiconductor industry do not make semiconductor manufacturing a natural monopoly. That is because the long-run average cost of an MNO does not consistently decline with increases in market share and corresponding increases in output. Naturally, competition in the wireless sector can only thrive if certain economic conditions are present. For instance, regulators must make sufficient spectrum available in all economic markets for multiple MNOs to enter. Similarly, although not nearly as crucial as sufficient spectrum, mobile number portability (MNP) has a significant impact on competition as it enables customers to more freely switch among service providers without having to change their phone numbers. The FCC has made sufficient spectrum available and introduced competition-enhancing programs, such as MNP, to avoid the creation of local monopolies or duopolies and to ensure that entry barriers are minimal. While robust competition in wireless telephony has made most forms of regulatory intervention unnecessary, wireless providers still have benefited from a strong network effect. Just like a wireline network, or any other network for that matter, the value of the wireless network is strongly correlated with the number of people using it. This fact has given rise to
35
Before the Federal Communications Commission, 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, Ninth Report, 19 FCC Rcd 20597 (2004) 158.
19
much debate about whether and when the government should intervene in the wireless sector to force cooperation among different service providers. One such debate has been about whether facilities-based MNOs should be required to negotiate with resellers. Specifically, in 1981, the FCC concluded that restrictions on the resale of wireless services were contrary to the public interest.36 Consequently, the FCC decided to grant cellular licenses on the condition that a licensee does not restrict the resale of its services. This was done with the intention of promoting a highly competitive secondary market for distribution of cellular service.37 In 1992, the FCC amended this order and exempted fully operational facilities-based competitors from the resale rule.38 In 1996, with the implementation of the new Telecommunications Act, some parties argued that wireless competition was sufficiently strong to justify rescinding the resale rule. Other parties argued the opposite, claiming that barriers to entry were insurmountable and that wireless resale enhanced consumer welfare. In a subsequent order, the FCC concluded that, under the market conditions at that time, restrictions on resale by cellular, broadband personal communications services (PCS), and certain specialized mobile radio (SMR) providers will inhibit the development of competition in these services.39 Specifically, the rule prohibited cellular, broadband PCS, and certain SMR providers from unreasonably restricting the resale of their services during a transitional period.40 Importantly, the FCC concluded that once broadband PCS licensees have built their networks and are competing with wireless service providers, market forces should eliminate the need for explicit resale regulation. Thus, the FCC ordered that the resale rule be eliminated five years after the last group of initial licenses of broadband PCS spectrum was awarded. Therefore, the resale rule expired on November 24, 2002. It was the only temporary regulatory intervention that aided entry by MVNOs into wireless markets. While there were subsequent competition enhancement programs, such as MNP, none of those programs revisited the question of open access. Few, if any, MVNOs actually made use of the resale order before its expiration. TracFone was first in the US market with its launch in 1996. The only other MVNO that potentially benefited from the resale order was Virgin Mobile, which entered the market in July 2002. All other MVNOs entered the market after the expiration of the resale order. This fact indicates that at least in the US regulatory intervention was not necessary for MVNO entry to occur.
36
37 38
39
40
Before the Federal Communications Commission, In the Matter of Cellular Communications Systems, 86 FCC 2d 469, 511, 642 (1981) (FCC Cellular Order), modified, 89 FCC 2d 58 (1982), further modified, 90 FCC 2d 571 (1982), appeal dismissed sub nom. United States v. FCC, No. 82-1526 (D.C. Cir. Mar. 3, 1983). FCC Cellular Order, 511, 642 The Commission reasoned that the five years since the granting of the cellular licenses is sufficient time for a licensee to build its system and that permitting licensees to deny each other resale after this period would promote competition by encouraging each licensee to build out its network. See Federal Communications Commission, Petitions for Rule Making Concerning Proposed Changes to the Commissions Cellular Resale Policies, Notice of Proposed Rule Making and Order, CC Docket No. 91-33, 6 FCC Rcd 1719, 1724 (1991). Before the Federal Communications Commission, 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, First Report, 10 FCC Rcd 8844 (1995) 18455, 18456 (FCC First CMRS Report). Ibid.
20
While US regulators do not interfere in the MVNO-MNO relationship, MNOs and MVNOs are nevertheless regulated. For instance, all wireless carriers (including MVNOs) must register with the state concerned before providing service, must often provide E911 services, report annual figures to the regulators, and have universal service and customer complaint obligations. 3.1.4. MVNO entry (current and forecasted market shares of MVNOs) With increased competition in a progressively more saturated and spectrum-limited marketplace, wireless service providers have sought out new market segments to foster growth. MNOs are increasingly turning to MVNOs to differentiate themselves from competitors by targeting specific market segments such as young people, ethnic groups, businesses, and prepaid users. MVNOs have fewer overhead costs and are more closely attuned to these market segments. This trend is reflected in the large increase in US MVNOs as well as the exponential growth of prepaid subscribers. Particularly, following the success of Virgin Mobile and Boost Mobile, the MVNO market segment has grown rapidly as new players have joined. Currently, there are over 40 MVNOs operating in the US.41 An accurate count, however, becomes increasingly more difficult to obtain, as new MVNOs are entering, plan to enter, exit, and then reenter. Yet others enter with new, creative business models that make it difficult to conclusively label them as an MVNO. For instance, ESPN Mobile entered the MVNO market segment in February 2006, exited in October 2006, then consequently sold its brand name to Verizon Wireless and is planning to enter again through Verizon. Similarly, Apple, long rumored to enter the market, announced earlier this year that it will instead license its iPhone exclusively to AT&T Mobility and stay out of the MVNO market. To further confuse the issue, there is Sonopiaa recently launched provider that is a mix between an MVNE and an MVNO. Sonopia has a wholesale agreement with Verizon Wireless and resells MOUs along with turnkey MVNO packages, promising its customers anyone can create and start marketing a mobile service in 15 minutes or less.42 As of the writing of this report, there are 751 Sonopia members (up from 450 only a week prior).43 Per most MVNO definitions, these 751 Sonopia members should be counted as MVNOs. However, they differ significantly from regular MVNOs as they do not offer billing or customer service, and they are not required to offer any brand value; therefore, we would not classify them as MVNOs. Nevertheless, the new business model created by Sonopia is of much interest and will have a fundamental impact on this market segment. Moreover, in mid-April 2007, Japanese operator KDDI announced that it plans to enter the US MVNO market through a wholesale agreement with Sprint Nextel. The company plans to target Japanese expatriates with postpaid plans and offer prepaid plans to Japanese tourists. With innovative service plans, ready-to-use wireless service packages, low prices, and creative marketing techniques, existing and new MVNOs enjoy wide consumer demand. This, however, was not the case before the success of Virgin Mobile. While early data on MVNO subscribers are not available, in 1999, there were less than 500,000 MVNO
41 42 43
NERA Research. Sonopia, http://www.sonopia.com/mvc/network/main.html?html=/mvc/network/start.html#sonopias (accessed May 1, 2007). This number changes daily as new members join.
21
subscribers nationwide. This represented less than 1 percent of the wireless market. As illustrated in Table 1, the low MVNO success rate continued through 2002 when there were approximately two million MVNO subscribers, representing 1.5 percent of the entire wireless market. With the entry of Virgin Mobile, the number of MVNO subscribers doubled yearly, jumping to four-million subscribers in 2003 and eight-million subscribers in 2004. Currently, there are about 1620 million MVNO subscribers. The largest players in todays market are TracFone with over 7.5 million active subscribers, Virgin Mobile with over 4.5 million and Boost Mobile with 3.1 million.44 Industry analysts expect the success of MVNOs to continue through at least 2010, when the number of MVNO subscribers is expected to reach 25 million, or 1315 percent of the total wireless market.45
Table 1 US MVNO Subscribers as a Percent of Wireless Subscribers 19992010
Year (a)
MVNO Subscribers (b)
Total Wireless Subscribers (c)
Percentage
(d)
[(b)/(c)]*100
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Source: NERA research
344,188 437,912 513,498 2,106,824 4,285,493 8,014,176 12,190,200 19,286,750 20,878,000 21,158,500 23,541,000 25,000,000
86,047,003 109,478,031 128,374,512 140,454,918 158,721,981 182,140,362 184,700,000 187,250,000 189,800,000 192,350,000 196,175,000 200,000,000
0.40 % 0.40 0.40 1.50 2.70 4.40 6.60 10.30 11.00 11.00 12.00 12.50
Most US MVNOs offer prepaid services. The recent success of MVNOs, Virgin Mobile in particular, has resulted in a significant rise in the percentage of prepaid wireless subscribers. This, in turn, has caused most major MNOs to either introduce or reintroduce prepaid service plans under their own brand names. Currently, all major US MNOs offer prepaid service plans. In fact, T-Mobile reports that 11.4 percent of its subscribers are on prepaid plans,
44 45
See the TracFone homepage at http://www.tracfone.com/home_page.jsp?b=i&flash=YES&p=W and the Boost Mobile homepage at http://www.boostmobile.com/; see also FCC Eleventh CMRS Report, 2728. A.T. Kearney, MVNO Explosion: Paradigm Shift in the Making, http://www.atkearney.com/shared_res/ pdf/MVNO_Position_Paper_S.pdf.
22
AT&T reports 7 percent, and Verizon Wireless reports 5.5 percent.46 Sprint entered the prepaid market segment only after its merger with Nextel (Boost Mobile). The US MVNOs focus on the prepaid market segment has a direct impact on their financial performances. Specifically, the average monthly revenue per user (ARPU) for the prepaid wireless market segment is approximately $35, while the postpaid segment (mainly served by MNOs) reports ARPUs between $50 and $61.47 For instance, Sprint Nextel reports average monthly service revenue for postpaid customers of $61, while its prepaid services (offered through MVNO Boost Mobile) are obtaining monthly ARPUs of $31.48 US MVNOs report a wide range of ARPUs. For instance, TracFones ARPU is reported to be $14$17 per month.49 Bankrupt MVNO Ampd Mobile, on the other hand, reports to have an average content and data ARPU that exceeds $30 per month and a total ARPU well over $100/month.50 Research firm IDC estimates average MVNO ARPU for content and data at $6.74 per month and average total ARPU at $54 per month.51 The large variance in the ARPUs among MVNOs is due to the different usage patterns of the MVNOs target markets. In particular, Ampd Mobile targets teenagers who historically have the highest MOUs among all age groups and frequently download value-added services, such as ring tones, video clips, and music. In 2005, total MVNO revenue was $4.6 billion. Figure 5 shows US MVNO revenue forecasts through 2010. While the estimates for the latter years (20082010) may be on the high side, this study effectively demonstrates that MVNOs are expected to grow significantly.
46
47
48 49 50 51
See Implementation of Section 6002(b) of the Omnibus Budget Reconciliation Act of 1993, Annual Report and Analysis of Competitive Market Conditions with Respect to Commercial Mobile Services, Tenth Report, 20 FCC Rcd 15908 (2005) (FCC Tenth CMRS Report) 100, citing David Janazzo et al., US Wireless Matrix 4Q04, Merrill Lynch, Equity Research, March 4, 2005, p. 17. Before the Federal Communications Commission, In the Matter of Tenth Annual Report and Analysis of Competitive Market Conditions with Respect to Commercial Radio Services, WT Docket No. 05-71, PN Comments of TracFone Wireless, Inc., March 29, 2005, p. 3 (TracFone Comments). Sprint Nextel Corporation, 2006 Annual Report, March 1, 2007, p. 47. TracFone Comments, p. 3. Ampd Mobile Press Release, Ampd Mobiles Entertainment Focus Generates Strong 2006 Performance, http://get.ampd.com/About/Press/corpmetrics07.php (accessed May 1, 2007). Ampd Mobile Press Release, Leading the Charge as the Premiere Mobile Entertainment Company, Ampd Mobile Drives Continued Growth into 2007, http://get.ampd.com/About/Press/ (accessed May 1, 2007).
23
Figure 5 Projected MVNO Revenues in US$ 20022010

$35.00
$30.00
29.60
Projected MVNO Revenues in US (in billions)
$25.00
23.90
$20.00
18.90
$15.00
14.50
$10.00
9.60
$5.00 2.20 0.20 $0.00 2002 2003 2004 1.00
4.60
2005
2006 Year
2007
2008
2009
2010
Source: ARC Group
3.1.5. MVNO business models (positive and negative aspects and implications) While all the business models discussed above can be observed in the US market, clearly the dominant business model pursued by MVNOs is serving the historically untapped market of prepaid customers. Because the prepaid segment is known to largely attract lower-income classes, many MVNOs offer discounted prepaid services. In fact, 9 out of 10 US MVNOs offer prepaid services, while only a few offer exclusively postpaid plans (e.g., SK TelecomEarthlinkBlackBerry, EarthlinkTreo, and Voce). Prepaid rate plans range from $0.05 to $0.44 per minute.52 Starter packages, including a basic or refurbished handset and a few free minutes (e.g., 10 minutes), usually range from $50 $100. In contrast, Verizon Wireless charged up to $0.85 a minute for local prepaid calls in 1998 and $1.15 a minute for long distance.53 Now, Verizon charges $0.10 to $0.14 per minute for local and long distance calls.54 MVNOs pricing models differ significantly, depending on their target markets. Approximately 33 percent of the US MVNOs are discount carriers, targeting the budgetconscious, the credit challenged, or lower-income individuals. While this type of strategy may work in the short run, it is questionable whether it will work in the long run. As
52 53 54
At $8.33 per minute, OnStar is the most expensive MVNO in the US. However, because it offers its service mainly as an emergency feature in automobiles, its price level is of lesser importance. Andrew Backover, Customer-hungry cell firms seek prepaid callers, USA Today, February 6, 2002. See Verizon Wireless website, http://www.verizonwireless.com/b2c/index.jsp.
24
competition in the prepaid market continues to increase from both other MVNOs and MNOs, it can be expected that, in the long run, prices will approach marginal cost. With prices at or near marginal cost, future MVNO success stories will not be based on further discounting, but on value-added positioning. Some MVNOs can achieve this by relying on their brand value (e.g., Disney), innovative marketing (e.g., Helio), or expansive distribution channels (e.g., 7-Eleven). Other MVNOs or prepaid operators can succeed in this market by targeting the needs of substantial and well-defined groups, such as ethnic groups, travelers, and youths (e.g., Firefly). 3.1.6. Pricing Trends before and after MVNO entry In light of the continuing growth of US MVNOs (both in terms of numbers as well as financially), it is important to study their impact on competition and consumer welfare. In terms of consumer welfare, US MVNOs offer consumers more choice, particularly in the prepaid market segment. Where there were little to no offerings in prepaid services before the arrival of MVNOs, this market segment is now highly competitive and offers a wide selection of mobile carriers. MVNOs also have significantly contributed to meeting the mobile communication demands of niche markets, such as children (Firefly), teenagers (Boost Mobile), and university students (XE Mobile). Thus, MVNOs have widened and deepened the US market for mobile services. Determining the impact MVNOs had on prices and competition in general is more difficult. Given the tremendous technological advances as well as the fundamental restructuring of the US mobile market during the last ten years, the competitive impact of MVNOs can at best be approximated. Figure 6 below shows the monthly ARPU for an average US consumer from 1993 through 2005 (last year where data are available). As can be seen, rather than decreasing the ARPU, the ARPU actually increased in the years following MVNO entry.55
55
While we recognize the many shortcomings of using ARPU as a proxy for prices (such as its dependency on consumer price elasticity, change in consumption patterns, etc.), we nevertheless find it informative to illustrate that the impact of MVNOs is not readily observed in ARPUs and other metrics.
25
Figure 6 US Average Revenue Per User 19932005
$70.00 $60.00 Monthly ARPU $50.00 $40.00 $30.00 $20.00 $10.00 $1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Year
Source: Derived from data in Federal Communications Commission, CMRS Reports 19932005.
While we cannot conclude from this graph that MVNOs caused the ARPU to increase, we can conclude that MVNO entry did not have a sufficiently large negative impact on ARPUs. This finding, however, is not entirely surprising, given the previous conclusion that US MVNOs tend to serve previously untapped market segments, often with high ARPUs. As summarized by Virgin Mobile in its comments before the FCC: MVNOs that develop and market unique service offerings targeted to niche demographics traditionally ignored by larger carriers have a proven track record of competitive success.56 Pricing trends, or revenue per minute (RPM), offer another metric to assess the impact MVNOs had on the US mobile market.57 As illustrated in the graph below, US RPM reached its highest levels in 1994 with $0.47 per minute. From 19952001, RPM decreased at a steady annual rate of approximately $0.05 per minute. From 20022005, prices decreased at a slower rate of approximately $0.01 per minute. Thus, MVNO entry, which occurred largely after mid-2002 (with the entry of Virgin Mobile), does not appear to have reduced average retail prices.
56 57
FCC Eleventh CMRS Report, 28, citing Virgin Mobiles Reply Comments dated March 6, 2006. RPM, as described by the FCC, is calculated by dividing a carriers estimate of average monthly revenue per subscriber (often referred to as average revenue per unit, or ARPU) by its estimate of MOUs, yielding the revenue per minute that the carrier is receiving. (See FCC Eleventh CMRS Report, 154.)
26
Figure 7 US Average Revenue Per Minute 19932005
$0.50 $0.45 $0.40 Revenue Per Minute $0.35 $0.30 $0.25 $0.20 $0.15 $0.10 $0.05 $1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Year
Source: Derived from data in Federal Communications Commissions, CMRS Reports, 19932005.
The monthly CPI for wireless telephone service paints a similar picture. As shown in the graph below, the CPI for wireless has remained flat since spring 2001over a year before the entry of Virgin Mobile.
27
Figure 8 Monthly Consumer Price Index Wireless Telephony 19982005

120.0
100.0
80.0
60.0
40.0
20.0
0.0 Apr-03 Apr-04 Apr-01 Apr-02 Apr-00 Apr-99 Apr-98 Oct-98 Oct-99 Oct-00 Oct-01 Oct-02 Oct-03 Oct-04 Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-98 Jan-05 Jul-98 Jul-99 Jul-00 Jul-01 Jul-02 Jul-03 Jul-04
Source: Derived from data in Federal Communications Commission, CMRS Reports, 19982005.
We cannot conclude that retail competition was affected or unaffected by the entrance of MVNOs into the market. We simply find that if the MVNOs did increase competition, their impact cannot be observed in industry-wide averages, such as the ARPU, RPM, or the CPI.
3.2. United Kingdom

The UK provides an interesting benchmark study because with the entry of Virgin Mobile UK in 1999 it was the first European market to encounter MVNOs. However, eight years later, all of the MVNOs that entered after Virgin Mobile have had limited market success. 3.2.1. Available technological infrastructure As in most European countries, the dominant mobile technology in the UK is GSM. There are currently four MNO providers of GSM 2G/2.5G services, O2 UK, T-Mobile UK, Orange UK, and Vodafone UK, and five providers of 3G UMTS services (the four operators listed plus Hutchison 3G UK. Vodafone launched its services in 1992 and currently uses both 900 and 1800 MHz frequencies to provide GSM 2G/2.5G services and provides 3G UMTS services. T-Mobile entered the market in 1993 and currently offers 2G/2.5G services using
28
GSM 1800 and provides 3G UMTS. O2 and Orange entered the market in 1994.58 Both carriers also offer 3G UMTS. However, O2 operates its 2G/2.5G services using GSM 900 and 1800 MHz, while Orange operates only with GSM 1800 MHz. 3.2.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry The UK market for mobile services is one of the largest (in terms of subscribers) in Europe. Although wireless penetration surpassed the 100 percent mark in 2003, the market still is growing and has reached a penetration rate of 117 percent.
Figure 9 UK Wireless Subscribers 20002006
80,000,000 70,000,000 60,000,000 Mobile Subscribers 50,000,000 40,000,000 30,000,000 20,000,000 10,000,000 2000 2001 2002 2003 2004 2005 2006
Source: Data from Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc.
Currently, there are five MNOs and two large MVNOs serving UK mobile subscribers. O2 UK leads the MNO market in terms of total subscribers with a share of 23.02 percent. Vodafone UK is a close second with 22.12 percent, followed by T-Mobile with 22.08 percent, and Orange UK with 20.02 percent. Hutchison 3G UK is the smallest MNO with a share of 5.03 percent. MVNOs Virgin Mobile UK and Tesco Mobil hold 5.91 percent and 1.83 percent market share, respectively. These shares are illustrated in Figure 10, below. In terms of 3G services, Hutchison 3G UK serves 3.85 million (51 percent) of the 7.52 million 3G subscribers.
58
O2, like Vodafone, had been offering 1G (analog) service since 1985.
29
Figure 10 UK Total Wireless Market Shares (by subscribers)
Virgin Mobile UK Tesco Mobile
Hutchison 3G UK
O2 UK Vodafone UK
T-Mobile UK
Orange UK
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites.
The British regulator, Oftel, last conducted a competitive review of the relevant market for MVNOs in August 2003, which was four years after the entry of Virgin Mobile and about the same time when the second UK MVNO, Tesco Mobile, entered the market.59 This review addressed whether regulatory intervention regarding MVNOs was necessary. The review was stipulated by the ECs European Regulatory Framework, which specifies that regulation is only appropriate where the national regulatory authority finds that an operator has individual dominance or one or more operators have joint dominance in a given communications market. This framework calls for the scrutiny of 18 communications markets deemed most likely to be imperfectly competitive. 60 Specifically, the EC requires individual national regulatory authorities (NRAs) to conduct a formal assessment of the level of competition in each such market. If a market is found to be not competitive, appropriate regulatory measures (specified below) have to be drafted and implemented. Specifically, the NRAs must determine whether incumbent operators possess significant market power (SMP)
59
60
Oftel, Mobile access and call origination services market: identification and analysis of market and determination on market power, Explanatory statement and notification, August 4, 2003, http://www.ofcom.org.uk/static/archive/oftel/publications/eu_directives/2003/mobileaco0803.pdf (accessed May 2, 2007). As part of the new regulatory framework for the EU, the EC determined that there are 18 different economic markets for telecommunications services. This view is unique to the EC and not supported by the findings of either the FCC or the authors of this report.
30
in any given market. The EC finds SMP to be present if a party can raise prices by restricting output without incurring a significant loss of sales or revenues.61 The specific methodology to determine SMP is based on European competition law and can be found in the ECs guidelines for market analysis and assessment of significant market power.62 According to the ECs guidelines, a dominant position is found by reference to a number of criteria and its assessment is based on a forward-looking market analysis based on existing market conditions.63 The ECs guidelines pay particular attention to market share. Thus, a firm without significant market share is unlikely to be in a dominant position. However, the ECs guidelines correctly stress that a large market share alone is not sufficient to establish market dominance.64 The EC states that an operator with a large market share may be presumed to have SMP if its market share has remained stable over time.65 If its market share is gradually eroding, this may indicate that the market is becoming more competitive, although that fact alone does not necessarily preclude an SMP finding. On the other hand, fluctuating market shares over time may indicate a lack of SMP in the relevant market.66 In addition to market share analyses, the ECs guidelines direct NRAs to further analyze markets for other economic characteristics. For this, the EC has established the following criteria for testing whether an incumbent operator has the ability to attain a dominant market position:67 Overall size of the operator Control over network infrastructure that cannot be easily duplicated Technological advantage or superiority Absence of, or low, countervailing buying power Easy or privileged access to capital markets/financial resources Product/service diversification (e.g., bundled products or services) Economies of scale Economies of scope
61
62 63 64
65 66
67
European Commission, Commission guidelines on market analysis and the assessment of significant market power under the Community Regulatory Framework for electronic communications networks and services, (2002/C 165/03) July 11, 2002, 75 (EC Guidelines on Market Analysis). Ibid. Ibid. The limited use of market share analysis to determine market power has also been adopted by the US Federal Trade Commission and the US Department of Justice. Market share analyses can be useful screens to determine whether there is potential market power. However, they should not be seen as being dispositive on market power and can, at times, be highly misleading indicators of market power. EC Guidelines on Market Analysis, 75. Ibid. Fluctuating market share may be empirical evidence of a contestable market. Given the relentless lowering of barriers to entry and exit (through VoIP or other emerging technologies), certain communications markets have become contestable and, thus, lack market power. Ibid.
31
Vertical integration A highly developed distribution and sales network Absence of potential competition Barriers to expansion
Some combination of these criteria may suffice for supporting a finding that an incumbent operator has a dominant position, but no single criterion can be sufficient to support a finding of SMP. The ECs guidelines also note that a dominant position in the market can be held by more than one operator and that an operator can enjoy SMP either individually or jointly with others.68 Based on these criteria, any finding by an NRA that an incumbent operator possesses SMP in the study market is sufficient for that market to be declared not open to competition. Following such a declaration, the EC empowers the NRA to design and propose specific regulatory obligations for the operator found to possess SMP. Before imposing any such regulatory obligation, however, the NRA must submit its market SMP assessment to the EC for endorsement and approval.69 Upon completing its review, the EC has the authority to veto (and require withdrawal) of the proposed measure if it has serious doubts about the NRAs market assessment. Wholesale network access and origination services provided by MNOs to MVNOs appear to fall within the purview of access and call origination on public mobile telephone networks, one of the 18 markets identified by the EC for SMP assessments. 70 Thus, in determining the regulatory framework for MVNOs in the EU member states, Oftel had to assess the level of competition in that particular market and determine whether any particular MNO possessed SMP, which would provide economic justification for regulatory intervention on behalf of the MVNOs. In conducting its analysis, Oftel examined the market in terms of: Barriers to entry Market shares Maturity of the market Maturity of technology Pricing and profitability
68 69 70
Ibid. eCommunications Regulation, Electronic Communications, How the New Regulatory Package Works, Fact Sheet 14, September 2005. European Commission, Commission Recommendation on relevant product and service markets within the electronic communications sector susceptible to ex ante regulation in accordance with Directive 2002/21/EC of the European Parliament and of the Council on a common regulatory framework for electronic communication networks and services, (2003/311/EC) February 11, 2003. The market in question is sometimes referred to as wholesale market 15.
32
Elasticity of demand, awareness, and switching Countervailing buyer power Similarity of cost structures The extent of excess capacity Transparency of pricing Scope for operators to retaliate if an operator were to change its prices
Oftel first analyzed whether any carrier had individual SMP and concluded: no supplier has an appreciable market share lead over its competitors or any significant cost, technological, or other competitive advantage.71 Oftel then turned to collective dominance and concluded that there was no collective dominance due to significant historical fluctuations in relative market shares, significant asymmetries of relative profitability, continuing entry onto the retail market by service providers, fluctuating relative prices, and retailers countervailing buyer power.72 In addition, while Ofcom found significant barriers to entry at the wholesale level (because of limited spectrum and high sunk costs), it took note that a fifth major mobile network operator (Hutchison 3G UK) was about to enter the mobile market, which indicated low entry barriers to the retail market and countervailing buying power exercised by retailers.73 Finally, Oftel found that competition at both the wholesale and retail level was likely to persist into the future.74 Consequently, Oftel refrained from regulating Market 15 and with it the MNO-MVNO relationship. As noted, this latest review of Market 15 was conducted after the entry of the two most successful MVNOs in the UK. Hence, in order to examine whether Ofcoms view on competition was impacted by the presence of commercially negotiated wholesale agreements and the active competition of MVNOs, it is informative to examine the regulators view on competition prior to 2003. The competition review in the UK immediately preceding this last one was conducted in September 2001. The objective of that review was to ensure that regulation remains appropriate to the level of competition.75 With respect to regulation, Oftel noted that: Regulation is justified when competitive pressures are insufficient to enable consumers to get a good deal in terms of quality, choice and value for money. Under-regulation is not in consumers interests. Nor is over-regulation which can deter investment and hinder innovation.76
71 72 73 74 75 76
European Commission, Case No UK/2003/2001: Mobile network access and call origination, Comments pursuant to Article 7(3) of Directive 2002/21/EC, August 29, 2003, p. 2. Ibid. Ibid. Ibid. Ibid. Office of Telecommunications, Effective Competition Review Mobile, September 26, 2001, http://www.ofcom.org.uk/static/archive/oftel/publications/mobile/mmr0901.htm (accessed May 4, 2007).
33
In reviewing the level of competition in 2001 (two years after the entry of Virgin Mobile UK and two years prior to the entry of the second MVNO), Oftel examined the retail prices for mobile services in the UK and found them to be among the cheapest in Europe. Other facts Oftel uncovered included declining long-term pricing trends, high levels of consumer satisfaction, and evidence of quality competition. Oftel, however, did not conclude that the mobile market was effectively competitive because low consumer awareness of different tariffs, high prices for off-net calls, the recent implementation of MNP, and the practice of handset locking led Oftel to conclude that for services bought by mobile subscribers, the market is prospectively competitive.77 In terms of competition, Ofcom found that [i]n light of growing competitive pressures Oftel can start to reduce regulation.78 With respect to MVNOs, Oftel noted: When faced with complaints under the ICD, Oftel has carefully considered on a case-by-case basis what constitutes reasonable access and will continue to do so. Oftel would not, without the most careful investigation, conclude that any requests for new types of access under the ICD should be met. Oftel has no current intention of changing its policy that MVNOs should be negotiated commercially, rather than mandated under the ICD.79 It is worthwhile noting that in arriving at its decision, the regulator observed the costs associated with regulation, and, even though it did not find effectively competitive markets, it opted not to regulate wholesale access. Rather, it elected to continuously monitor the situation and carefully investigate each failed MNO-MVNO negotiation. 3.2.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact) Based on its most recent competitive review, there are currently no MVNO-related, Ofcomspecific regulatory requirements beyond those in the published General Conditions of Entitlement.80 The process for establishing an MVNO depends on what services the MNO is supplying to the MVNO and is a commercial matter between those two parties. 3.2.4. MVNO entry (current and forecasted market shares of MVNOs) There are currently six MVNOs in operation in the UKVirgin Mobile UK, Tesco Mobile, Value Telecom (Fresh), BT Mobile, OneTel Mobile, and Sainsburys.81 Virgin Mobile UK, a joint venture between the Virgin Group and MVNO T-Mobile, was the first MVNO in the UK and Europe. It started service in November 1999 and has over 4.5 million subscribers, thus serving approximately 6 percent of the UK wireless market. Virgin Mobile seems to benefit from its previous experience as an MVNO, its high brand recognition, and its reported 68 percent wholesale discount off retail prices. The second
77 78 79 80
81
Ibid., 4.6. Ibid., S.5. Ibid. 4.8. Ofcoms General Conditions on electronic communications network providers include obligations to negotiate interconnection, obligations concerning standardization and provision of specified interface data, obligations to ensure the proper and effective functioning of the network, and number portability. In addition to these consumer MVNOs, there are a number of resellers, reseller/MVNO hybrids, business MVNOs, and data MVNOs such as Kingston Communications, Wyless, and Dot-Dash/Prian Partners.
34
largest MVNO in the UK is Tesco Mobile, a 50/50 joint venture with O2, which started service in June 2003 and has approximately 1.4 million subscribers (1.83 percent market share). The success of the other five UK MVNOs has been limited thus far. For instance, the now defunct low-cost MVNO easyMobile started service in March 2005 and was able to attract only about 23,000 customers in its first six months in operation, and, while it eventually had around 80,000 customers, it closed operations in late 2006.82 Company officials blamed low customer growth, change in ownership, and internal management problems for easyMobiles failure.83 easyMobile also closed its operations in the Netherlands.84 Similarly, Fresh, which launched its service in 1999, only had 200,000 customers six years later.85 3.2.5. MVNO business model (positive and negative aspects and implications) Due to the UKs high mobile penetration rate and its high percentage (66 percent) of prepaid subscribers, the business models for MVNOs in the UK are largely limited to niche markets. Discounts offered on service prices are one of the biggest attractions of MVNOs in Europe. Those discounted prices attract low-income customers from untapped markets, including special interest groups ranging from football players to ethnic groups. From an economic perspective, these MVNOs seek out niche markets, essentially segmenting the wireless sector into subgroups with similar interests and purchasing propensities. For instance, Tesco Mobile offers low-cost, flat-fee prepaid plans; Sainsbury, the supermarket chain, offers discounted prepaid and postpaid plans; and Yes Telecom focuses on tailored plans for business customers. BT Mobile focuses on the business segment and high-usage residential customers. 3.2.6. Pricing Trends before and after MVNO entry While mobile prices in the UK and elsewhere have drastically decreased during the last five years, the decline cannot be entirely attributed to increased competition or MVNOs in particular. Isolating the specific price impact of MVNO competition is difficult as many other factors, such as technological development, different offerings and price bundles, changes in cost structures, consumer prices, and so on, will distort the accuracy of such an exercise. Some analysts have attempted to document the economic impact of MVNOs using metrics, such as RPM, profitability, and MOUs. However, as we pointed out in the case of the US, such trends, while informative, cannot be attributed solely to MVNOs. Figure 11 below shows the monthly ARPU for an average consumer in the UK from 2001 through 2005. The figure visually shows how the entry of Tesco in mid-2003 as well as the growing presence of Virgin Mobile did not have a large competitive impact on the mobile market with respect to the ARPU.86 While the data ARPU continues to grow, the voice ARPU dropped in 2005, when compared to 2004.
82
83 84 85 86
Nancy Gohring, IDG News Service, EasyMobile to close mobile phone service in UK, No stores and no phones means no clients for EasyMobile, InfoWorld, November 15, 2006, http://www.infoworld.com/archives/emailPrint.jsp?R=printThis&A=/article/06/11/15/HNeasymobileshuts_ 1.html (accessed May 1. 2007). Ibid. Ibid. Fresh has been relaunched twice, first in mid-2004 and then again in October 2005 (source: Telegeography) Note that this does not necessarily mean that MVNO entry did not have a competitive impact on ARPU. It simply implies that other factors, such as increased data and voice consumption have increased the average consumers monthly spending on mobile services. Moreover, mobile companies have frequently revised the
35
Figure 11 UK Average Revenue Per User 20012005
16 14 12 per month 10 8 6 4 2 0 2001 2002 2003 2004 2005 Voice Data
Source: Office of Communications, The Communications Market 2006, 3 Telecommunications, August 10, 1006, p. 149
Retail pricing trends, as measured by RPM, provide an alternative metric to analyze the impact of competition because this metric is independent of consumption patterns. As illustrated in Figure 12 below, 2005 mobile prices in the UK for on-net, off-net, international, and SMS calls have decreased since 2001. In particular, off-net calls have decreased from 26.2 pence a minutes to 11.3 pence per minute. On-net calls have also decreased, albeit only slightly from 5.9 pence per minute in 2001 to 4.2 pence per minute in 2005.
fashion by which they determine ARPU. For instance, in or around 2002, mobile carriers changed their ARPU calculation to include only active customers, which increases the ARPU if the amount of inactive customers is sufficiently large.
36
Figure 12 UK Average Revenue Per Minute 20012005
45 40 35 Revenue Per Minute 30 25 20 15 10 5 0 2001 2002 2003 2004 2005 On-net Off-net International SMS
Source: Office of Communications, The Communications Market 2006, 3 Telecommunications, August 10, 1006, p. 152
However, as before, it is unclear whether this pricing trend can be attributed to the entry of MVNOs. For instance, Italy, a country with no MVNOs, has an RPM of US$0.22, down 8 percent from the previous year. The UKs overall RPM is quite comparable at US$0.21, down 9 percent from the previous year.87 Thus, as was the case for the US, it is unclear whether MVNO entry had a competitive impact on retail prices in the UK. Additionally, if in fact MVNOs in the UK were to exert competitive pressure on MNOs, the MVNOs ARPUs and RPMs would similarly decrease if they became engaged in a pricing war. Interestingly, Virgin Mobile UKs ARPU has changed little in the last few years. In Q104, Virgin Mobile reported an ARPU of 147 per month.88 Almost three years later, Virgin Mobiles ARPU is at 142.89 This observation would suggest that rather than competing vigorously with MNOs, Virgin Mobile is serving a niche market in which it can maintain a rather stable ARPU.
87 88
89
Merrill Lynch, Global Wireless Matrix 3Q06, GEMs Continue to Shine, January 9, 2007. Virgin Mobile, News 2004, Virgin sees strong customer, revenue and margin growth in the first quarter to end June, July 27, 2004, http://about.virginmobile.com/aboutus/media/news/2004/2004-07-29/ (accessed May 7, 2007). Virgin Mobile, Annual Report 2006, April 13, 2007, p. 60.
37
3.3. Ireland
In assessing the need of regulatory intervention in the MVNO-MNO relationship, Ireland is a frequently cited case study. With no MVNOs, two players that serve over 80 percent of the market, and repeated attempts by the regulator to stimulate competition, Irelands MVNO debate drew domestic and international attention. 3.3.1. Available technological infrastructure Much like other European nations, 2G and 2.5G mobile services in Ireland are provided through one common digital technology standard, GSM. Similarly, all 3G and 3.5G services use the W-CDMA platform. There are three GSM providers (Vodafone Ireland, O2 Ireland, and Meteor Mobile) and three W-CDMA providers (Vodafone Ireland, O2 Ireland, and Hutchison 3G Ireland). In February 2007, Hutchison announced that it had completed its network upgrade to provide 3.5G services through HSPDA with a download speed of up to 3.6 Mbps. Vodafone has made similar announcements, although it is not clear whether the company has completed its network upgrades.90 The statistics of the available infrastructures are summarized in Table 2 below:
Table 2 Technological Infrastructure Ireland
Provider Hutchison 3G Ireland Hutchison 3G Ireland Meteor Mobile Meteor Mobile O2 Ireland O2 Ireland O2 Ireland Vodafone Ireland Vodafone Ireland Vodafone Ireland Vodafone Ireland
Service 3G 3.5G 2G 2.5G 2G 2.5G 3G 2G 2.5G 3G 3.5G
Platform W-CDMA W-CDMA (HSDPA) GSM GSM GSM GSM W-CDMA GSM GSM W-CDMA W-CDMA (HSDPA)
Frequency Band 2100 2100 900/1800 900/1800 900/1800 900/1800 2100 900/1800 900/1800 2100 2100
Launch Date July 2005 January 2007 February 2001 2003 1997 January 2002 December 2005 1993 March 2001 July 2004 Currently
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites
3.3.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry At year-end 2006, there were 4.76 million wireless subscribers with 90 percent (or 4.298 million) on 2G/2.5G GSM networks and 10 percent (or 462 thousand) on 3G/3.5G W-CDMA networks.91 Vodafone Ireland was the first carrier to offer 3G services in Ireland with its
90
91
Vodafone Press Release, Vodafone Ireland Joins Forces with Dell to Launch Irelands First Commercial Trial of 3G Broadband, May 24, 2006, http://www.vodafone.ie/aboutus/company/press/2006/ release240506.jsp (accessed April 16, 2007). Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc.
38
launch of a W-CDMA network in July 2004. Hutchison 3G Ireland, a newcomer to the Irish mobile market, entered in July 2005, and O2 Ireland launched its 3G services in December 2005. Vodafone Ireland and Hutchison 3G Ireland recently also launched 3.5G (HSDPA) services. Despite a penetration rate of 114.7 percent, the number of subscribers in Ireland continues to grow, as illustrated in Figure 13 below.
Figure 13 Wireless Subscribers Growth in Ireland 20002006
5,000,000 4,500,000 4,000,000 Wireless Subscribers 3,500,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 2000 2001 2002 2003 Year
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc.
2004
2005
2006
Vodafone Ireland has a market share of 45.76 percent, O2 Ireland has 34.29 percent, Meteor Mobile has 16.87 percent; and newcomer Hutchison has 3.09 percent.92 In the light of Vodafone and O2s large market shares, jointly over 80 percent, the Irish regulator ComReg has found the Irish mobile market to be insufficiently competitive.
92
Ibid.
39
Figure 14 Wireless Network Operator Market Shares Ireland
Hutchison 3G Ireland, 3.09% Meteor Mobile, 16.87% Vodafone Ireland, 45.76%
O2 Ireland, 34.29%
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc.
ComReg has repeatedly attempted to inject more competition in the market, with little to no success. For instance, in July 2003, ComReg introduced MNP. However, MNP had minimal impact on the churn rate, and the market structure remained unchanged. Consequently, ComReg attempted to break the near duopoly by introducing MVNO-related regulation. This effort failed as well. In 2005, ComReg provisionally awarded a fourth 3G license to newcomer Smart Telecom. However, in 2006, ComReg withdrew its license offer. After months of legal arguments, Smart Telecom ran into financial troubles, and it is currently unknown whether Smart Telecom will enter the Irish mobile market. 3.3.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale, and impact) In 2004, ComReg conducted a thorough analysis of the level of competition in the market for wholesale access and call origination on public mobile telephone networks. In doing so, the Irish regulator followed the Framework Directives set forth by the EC (and discussed above in the section for the UK). Under these directives: ComReg is required to impose ex ante regulatory obligations that are appropriate, based on the nature of the problem identified,
40
proportionate and justified in the light of the objectives set out in Article 8 of the Framework Directives.93 As a first step in its analysis, ComReg identified the relevant product and geographic markets for wholesale access and call origination. With respect to the relevant product market, ComReg found that it consists of all wholesale access and origination services provided by a Mobile Network Operator (MNO).94 The relevant geographic market was found to be limited at this point in time to the national territory of Ireland.95 However, because there was no activity in the wholesale mobile market (e.g., there existed no MVNOs or any other type of reseller), ComReg had to conduct its competition analysis on the retail market and then infer the level of competition in the wholesale market from evidence in the retail mobile market. Based on this fundamental assumption, the regulator then proceeded to examine the retail market positions of the individual players to determine if there was a dominant carrier or a duopoly consisting of two dominant carriers. ComReg analyzed: The existence of barriers to entry and expansion (beyond those strategic advantages which amount to a first-mover advantage) Economies of scale and/or scope The size and distribution of competitors, relative to the largest firm The existence of deep pockets and access to capital Significant advantages in terms of advertising spending and other issues relevant to brand image The ability to leverage key aspects of overall size or economic strength in the relevant market Historical conduct suggestive of the existence of market power Economic performance relative to other market operators, taking into account efficiencies The extent of vertical integration, or the monopolization of routes to market
The regulator concluded that: barriers to expansion are low for both O2 and Vodafone, and this factor, combined with a number of other factors as present (such as economies of scale and scope, access to capital, etc.), means that Vodafone and O2 could not act independently of each other.96 It thus concluded that there was no single dominant carrier. The above analysis, however, gave rise to concerns about a duopoly or joint dominance. Specifically, ComReg noted the following areas of regulatory concern:
93 94 95 96
Commission for Communications Regulation, Market Analysis Wholesale Broadband Access, Decision No: 03/05, Document No: 05/11r, Date Issued: 24 February 2005, p. 6. Commission for Communications Regulation, Market Analysis Wholesale Mobile Access and Call Origination, Document No: 04/118 and 04/118a, Date: 9 December 2004, p. 4. Ibid. Ibid., p. 5.
41
A highly concentrated market High market share of both Vodafone and O2 Decreasing difference in market share between 2001 and 2004 Stable market share held by O2 since the beginning of 2001 New competition has not disrupted the market in terms of key competitive characteristics Relatively high and stable prices in the postpaid market segment High entry barriers High profitability of Vodafone and O2 (measured in terms of rate of return on capital employedVodafone had a 39 percent ROCE, and O2 had a return of 38 percent) Evidence of parallel behavior by O2 and Vodafone in substantial segments in the market
Consequently, ComReg applied a three-pronged test, examining the market for: 1. The degree of market concentration: in particular, does a small number of undertakings account for a large share of the relevant market, and does any single undertaking create an individual dominant position? 2. The incentive to coordinate: in particular, do O2 and Vodafone have any incentive to align their conduct in the market in such a way as to elevate their joint profits by restricting production with a view to sustaining prices above those which would otherwise prevail? 3. The ability to coordinate: in particular, do O2 and Vodafone have the ability to coordinate by aligning their conduct to establish a consensus position? In conducting this test, ComReg noted that, in order for any coordination to be possible, a number of basic conditions must be met, including (1) the ability to detect cheating, (2) the enforceability of compliance, and (3) actual and/or potential market constraints. With respect to the degree of market concentration, ComReg was of the opinion that the market was highly concentrated and notes that O2 and Vodafone had a 94 percent share of subscribers in the retail market for mobile communications in September 2004.97 With respect to the incentive to coordinate, ComReg determined that there exist sufficient incentives on the part of Vodafone and O2 for them to coordinate behaviour in the market for wholesale mobile access and call origination in Ireland.98 In arriving at this particular conclusion, ComReg first examined the market shares of the various players, in particular, the symmetry of market shares and concluded:
97 98
Ibid., p. 41. Ibid., p. 44.
42
[t]he proximity in market shares between Vodafone and O2 relative to other competitors and in light of the maturity of the market, and especially given the existence of so many other factors facilitating coordinated effects, suggests that the conditions conducive to the alignment of commercial behavior are very much in evidence.99 ComReg notes, however, that the market shares alone do not determine the incentive to coordinate because they simply could reflect similar cost structures. Other factors ComReg considered were (1) the high level of interactions between O2 and Vodafone, (2) the nonnegative market growth, and (3) the lack of drastic innovations that would favor one carrier over the other. With respect to the third prong of the joint dominance test, ComReg found that O2 and Vodafone have the ability to coordinate because there are two distinct focal points in this market that provide the necessary transparency for coordination and denying access to independent entities to upstream elements, such as wholesale airtime.100 The regulator found that the two carriers were able to focus on price because (1) they offer homogenous products, (2) they have transparent pricing plans, and (3) their pricing plans can be, and have been, simplified.101 Of particular importance to this report is the fact that there were no MVNOs even though several entities [are] seeking to offer independent service.102 The regulator did note, however, that the absence of MVNOs might simply be because the MNOs are more efficient in retailing and prices are at competitive levels. As a consequence the margins available to independent service providers would be insufficient to generate a return to cover costs assuming access were made available on reasonable terms.103 After analyzing Vodafone and O2s profit levels, however, ComReg found that their high profitability suggests otherwise, and evidence of successful MVNO operations in other European nations indicates that MVNOs can be profitable. It thus concluded that the market is not sufficiently competitive, and Vodafone and O2 have the ability to coordinate their behavior in the Irish marketplace.104 ComReg also analyzed if O2 and Vodafone would have the ability to monitor whether or not they are adopting a common policy. For this, the regulator reviewed two forms of cheating prices and access to upstream wholesale elementsand concluded that both forms of cheating could readily be detected. The Irish regulator also reviewed the presence of switching cost as a countervailing aspect to cheating. In particular, switching costs could render any deviation from a common policy (i.e., cheating) ineffective, because customers could not readily switch carriers in response to a price decrease by one carrier. If so, the absence of cheating (or price differentiation) would not be a sign of coordinated behavior, but a sign of the high cost of switching carriers. ComReg, however, found that a sufficient number of customers are no longer under a
99 100 101 102 103 104
Ibid., p. 43. Ibid., p. 47. Ibid. Ibid., p. 48. Ibid. Ibid, p. 49.
43
contract (no early termination fees) or churn throughout the year, so that (undetected) cheating should still be highly profitable. ComReg reviewed why deviations have not occurred, despite the promise of high returns. It found that the most likely retaliatory response to a deviation along the price dimension, which would be immediately transparent, is via a reduction in price, as this can be effected swiftly. Finally, ComReg analyzed the impact of Meteor as a fringe competitor that could be considered a maverick firm. Due to its low market share in the lucrative postpaid market since its launch three years previously, its low revenue market share, and its lack of geographic coverage, ComReg concluded that Meteor has not been a strong competitive force in Ireland. Based on these considerations, ComReg concluded that the mobile market in Ireland has a structure that is conducive to coordinated effects. It even went further and stated that: ComReg believes that the evidence supports the view that O2 and Vodafone are tacitly colluding in this market.105 Consequently, ComReg proposed to impose on Vodafone and O2 the obligation to provide network access on nondiscriminatory terms, following a reasonable request. If this proved unsuccessful, ComReg proposed the imposition of three additional obligations: (1) price control by way of cost orientation, (2) preparation of separated accounts, and (3) implementation of appropriate cost-accounting systems. The EC agreed generally with ComReg, but noted that MNO performance at the retail level does not automatically mirror its potential performance in the wholesale market, as not necessarily all supply in the relevant market is provided internally by vertically integrated mobile network operators. Thus, retail market characteristics are only indicative, not conclusive, regarding wholesale market performance.106 The EC further noted that, having signed a national roaming agreement with O2, Meteor may be positioned to pose a competitive check on the two major MNOs. Even Hutchison 3G Irelands role cannot be discounted entirely. The EC asked ComReg to monitor these competitive impacts in order to revisit its original collective dominance finding. Finally, the EC reasoned that tacit coordination among the top two MNOs could be disrupted even by a relatively small competitorparticularly one that is able to sign up customers with contracts that compare favorably with the services of the top MNOs. Vodafone Ireland, O2, and Meteor separately appealed ComRegs conclusion regarding joint SMP to the Department of Communications, Marine and Natural Resources, which subsequently appointed the Electronic Communications Appeals Panel (ECAP) to hear the appeals. On December 13, 2005, the parties settled the appeals, at which time ECAP annulled ComRegs finding of joint dominance (SMP). ECAP has not provided details about its decision. In a related case on mobile call termination, however, ECAP also overturned a ComReg finding of SMP for all four of Irelands MNOs. That decision cited a lack of rigorous economic analysis in ComRegs SMP assessment. In the light of ECAPs annulment of the collective SMP finding, there is currently no MVNO-related regulation in Ireland, and MNOs may host MVNOs at their own discretion.
105 106
Ibid., p. 62. European Commission, Commission for Communications Regulation, Comments pursuant to Article 7(3) of Directive 2002/21/EC, Case IE/2004/0121: Access and call origination on public mobile telephone networks in Ireland, January 20, 2005, p. 6.
44
3.3.4. MVNO entry (current and forecasted market shares of MVNOs) Following the ECs approval of ComRegs proposed regulation regarding MVNO entry and before ECAPs annulment of that decision, there were a number of MVNOs that were at various stages of entering the mobile market in Ireland. For instance, in the first quarter of 2005, MinuteBuyer became the first company to officially announce its entry as an MVNO. Rumors were that other companies, such as Clever Communication, Eircom, Tele2, British Telecom, EasyMobile, and T-Mobile might serve Irish consumers through MVNO offerings. Upon ECAPs annulment, however, all plans were put on hold. Clever Communications, for instance, announced that it has put its plan into deep freeze, and Eircom acquired MNO Meteor Mobile. While ECAPs decision might have caused some potential MVNOs entry plans to be abandoned, it did not foreclose MVNO entry. In December 2006, Tesco announced that it will enter the Irish mobile market as an MVNO in 2007 through a joint venture with O2.107 Tesco Mobile Ireland will rely on O2s network and technical expertise, while providing brand recognition and retail outlets throughout Ireland. Rumors also exist that Clever Communications will attempt to enter the market again along with Dome Telecom, MinuteBuyer, Perlico, BT, Carphone Warehouse, and Ryanair.108 3.3.5. MVNO business model (positive and negative aspects and implications) While the business models in Ireland are no different than the ones identified at the onset of this report, it is worthwhile to note that MVNO entry seems to be occurring despite the annulment of the proposed regulatory measures. As observed in the US and other countries, however, the absence of regulation has caused some potential MVNOs to abort their business plans, while encouraging others with strong brand appeal and retail distribution channels to enter. For instance, Tesco Mobile Ireland brings brand recognition (supermarket chain and gasoline stations) as well as 79 stores throughout Ireland. 3.3.6. Pricing Trends before and after MVNO entry While it is too early to analyze the impact of MVNO entry on prices, Tony Keohane, Chief Executive of Tesco Ireland proclaimed: We are delighted to announce Tesco Mobile which is built from two of the strongest and most successful brands in Ireland, O2 and Tesco. The launch of Tesco mobile next year will bring new competition to the market and enable us to provide an excellent quality, value for money product for customers throughout the country. We believe that Tesco mobile will be widely welcomed when it launches.109
107 108 109
Tesco Ireland, Tesco Ireland and O2 Ireland Announce Plans for Irelands First MVNO, Press Release, December 19, 2006, http://www.tesco.ie/corporate_info/. Electronic News Net, Tesco Mobile goes virtual in Ireland, December 19, 2006, http://www.electricnews.net/frontpage/news-9859883.html. Ibid.
45
3.4. France
France provides an interesting benchmark study because the French regulators policy recommendations on MVNO entry attracted a lot of political debate especially since there were no MVNOs operating in France at the time. Moreover, after the EC struck down the French regulators proposed MVNO-related regulation, a number of MVNOs who negotiated voluntary MNO agreements have prospered in France. 3.4.1. Available technological infrastructure With only one 2G and one 3G infrastructure, Frances mobile market is not fragmented. All three MNOs (Bouygues Telecom, Orange France, and SFR) provide 2G and 2.5G services using a GSM platform. Orange France was the first to enter the country with 2G services in 1992, followed a year later by SFR. Bouygues Telecom entered in 1996. SFR paved the way for Frances 3G W-CDMA services with its launch in April 2004. Orange France followed suit eight months later, and Bouygues Telecom only now is rolling out its 3G network. 3.4.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry As of December 2006, Orange France leads the market with a share of 45.55 percent, followed by SFR with 36.60 percent, and Bouygues Telecom with 17.85 percent.110 The last competitive review of the French mobile market was conducted in December 2004 as part of the EC mandated review of Market 15. At that point in time, no MVNOs were operating in France. In its review of Market 15, the French regulator, ART, analyzed: Market shares Barriers to entry Countervailing buyer power The technological advantage the incumbent carriers have over entrants Product diversification Vertical integration Potential competition Price competition111
110 111
Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. Autorit de Rgulation des Tlcommunications, Dcision No. 04-937 de lAutorit de rgulation des tlcommunications en date du 9 dcembre 2004 portant sur linfluence significative de la socit Orange France sur le marche de gros de la terminaison dappel vocal sur son rseau et les obligations imposes a ce titre, December 9, 2004, p. 4.
46
ART divided its analysis into two parts: it first reviewed competition in the retail market and then followed with a review of the wholesale market. In terms of the retail market, ART found that the market was not sufficiently competitive due to high retail prices, insufficient innovation, and coordinated market behavior. ART found evidence of coordinated market behavior among the three incumbent firms with respect to handset subsidies, term contracts, the limited success of MNP, and the relatively constant market share distribution over the least few years.112 ART further concluded that competitive entry was unlikely due to the lack of spectrum, high barriers to entry, and regulatory (approval) barriers.113 In its analysis of the wholesale market, ART found that the nonexistence of such a market provided evidence that the three incumbents had a common interest in avoiding wholesale agreements with MVNOs.114 Based on this review, ART concluded that the incumbent carriers each had individual SMP in their respective markets. Consequently, ART submitted a draft decision to the EC with the intention of legislating mandated wholesale access as a way of stimulating competition. Specifically, ART proposed to impose obligations on: Wholesale access Transparency Nondiscrimination Accounting separation Price control115
After receiving the blessing of the French antitrust commission, ARCEP (the former ART) submitted its analysis of Market 15 to the EC. However, since the competition review in 2004, French MNOs had voluntarily entered into relationships with at least four MVNOs. In light of these agreements, ARCEP withdrew its decision as it is difficult to have a clear, forward-looking view of the degree of competition on the wholesale and retail markets for access and mobile call origination.116 Instead, it promised to monitor the mobile access and call origination market. In particular, ARCEP will check that the recently signed MVNO contracts offer the MVNOs an opportunity for financial and commercial autonomy and that they are in a position to offer an expanded range of services.117 The regulator also announced that it will evaluate the evolution of competition on the market, especially in terms of price and market share.118
112 113
Ibid. Ibid. 114 Ibid. 115 Ibid. 116 Autorit de Rgulation des Tlcommunications, ARCEP monitors the wholesale and retail access and call origination markets, May 31, 2005. 117 Ibid. 118 Ibid.
47
3.4.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact) In light of the French regulators withdrawal of its draft regulation that would have regulated mobile wholesale access, there is currently no MVNO-related regulation. Rather, ARCEP acts as a watchdog and has promised to reassess the market in due course. In 2006, the French competition commission indicated that additional powers may be needed to strengthen competition in the [mobile] sector, for instance by giving MVNOs access to the three major operators networks.119 We are not aware of any regulatory measures that have been implemented in this respect. 3.4.4. MVNO entry (current and forecasted market shares of MVNOs) Absent regulatory intervention, Frances MNOs began to voluntarily carry MVNO traffic. The first wholesale agreement in June 2004 (over one and one-half years before ARTs proposed regulation) was between the MNO SFR and the MVNO debitel for a period of nine years. debitel launched its services in July 2004, targeting young people and individuals who never had a cell phone with aggressively priced prepaid and postpaid plans.120 During the same time, the UKs The Phone House (an independent phone retailer) signed a wholesale agreement with Orange and subsequently launched service under the brand name Breizh Telecom.121 Breizh Telecom is a regional provider, focusing on service in Brittany with prepaid and postpaid plans.122 MNO Bouygues signed its first MVNO agreement in July 2004 with Universal Mobile.123 Targeting young and trendy consumers, Universal builds on its brand name as a movie studio and offers prepaid and postpaid plans, along with games, music clips, ring tones, and other vertical features. In 2005, two additional MVNOs entered Frances mobile market: Tele2 with a wholesale deal with Orange and Coriolis Telecom with a MVNO deal with SFR. It is worthwhile noting that the Tele2/Orange deal only became a reality after regulators threatened to intervene and drafted actual MVNO-related regulation when negotiations appeared to be faltering. In April 2006, Virgin Mobile entered the market via a negotiated agreement with Orange, and the supermarket chain Carrefour negotiated a wholesale deal with Orange in July 2006. In addition to these larger MVNOs, there are several smaller MVNOs that have entered the French mobile market, including the grocery chain Auchan in fall 2006, Mobisud in December 2006, NRJ Mobile in November 2005 all using SFRs network, and most recently TV station TV1 using Bouyues network.124 Currently, there are 16 MVNOs,125 serving 1.4 million customers (or 2.8 percent of the market).126 This figure has increased dramatically from previous years as illustrated in the figure below.
119 120 121 122 123 124 125
Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. debitel company website, www.debitel.fr (accessed May 8, 2007). Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. Breizh Telecom company website, http://www.breizhmobile.com/ (accessed May 8, 2007). Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. NERA research. LAutorit de Rgulation des Communications Electroniques et des Postes, Operateurs mobiles virtuels (MVNO) et accords the licence de marque, October 24, 2006, http://www.arcep.fr/index.php?id=8981 (accessed May 8, 2007).
48
Figure 15 France MVNO Subscribers June 2005December 2006

1,600,000 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 Jun-05 Sep-05 Dec-05 Mar-06 Jun-06 Sep-06 Dec-06
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites.
Market analysts expect the market share of MVNOs to remain relatively small in the coming years, but it should reach 17 percent by the year 2010.127 We understand that some MNOMVNO wholesale contracts restrict MVNOs in terms of target market and subscriber count. While we are unable to examine the details and rationale of these restrictions, they potentially limit the market shares MVNOs will achieve in France. 3.4.5. MVNO business model (positive and negative aspects and implications) Unlike other European countries that predominately focus on cheap, unsophisticated, prepaid business models, French MVNOs seem to rely on brand appeal (Universal, Virgin, NRJ), existing distribution channels (Carrefour, Auchan, Breizh), and traditional prepaid and postpaid plans. 3.4.6. Pricing Trends before and after MVNO entry At less than 3 percent market share, MVNO entry has not had an observable impact on retail prices or service offerings in France. Conversely, MNOs seem to benefit from MVNO entry as much of their annual growth can be attributed to the addition of wholesale (MVNO)
126 127
LAutorit de Rgulation des Communications Electroniques et des Postes, Survey of the Mobile Market, Statistic indicators for December 31, 2006, December 31, 2006. Paul Budde Communications Ltd., 2006-2007 Europe-Telecoms, Mobile and Broadband in France and Switzerland, December 13, 2006.
49
customers. As stated by France Telecom (owner of Orange France) in its most recent financial statements: The MVNO customer base in France went from 259,000 at 31 March 2006 to 1.001 million at 31 March 2007. The MVNO business acquired 160,000 customers in the first quarter of 2007, in line with the quarterly acquisition rate recorded in 2006.128 As the MVNO market in France develops, we expect that successful MVNO business models will be adopted/acquired by MNOs, and competition in these new market segments will ensue.
3.5. Denmark
Denmark serves as a unique example of MVNO-related regulation because it introduced mandatory wholesale access in 2000 and annulled the policy six years later as the new regulatory directives by the EU no longer supported the finding of collective dominance. 3.5.1. Available technological infrastructure Although Denmark did have an AMPS network from 1982 until 2002, all of the current networks are either 2G/2.5G GSM or 3G W-CDMA. The estimated 5.8 million subscribers in Denmark are served by four MNOs (TDC Mobile, Sonofon, Telia Denmark, and Hi3G Access Denmark) and a number of MVNOs. Sonofon and TDC Mobil launched the first GSM networks in July 1992, and in 1998 Telia Denmark launched the third GSM network. All three of these carriers eventually upgraded their networks to accommodate 2.5G services, and Sonofon and Telia Denmark also launched 3G and 3.5G networks. A fourth MNO, Hi3G Access Denmark, entered the Danish market in October 2003 with a 3G network. 3.5.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry As with many European nations, Denmarks wireless penetration rate is well over 100 percent, and wireless-subscriber growth continues to increase, despite a one-year negative growth rate in 2005. Figure 16 below shows wireless-subscriber growth from 2000 through 2006.
128
France Telecom, Financial Information for the First Quarter of 2007, April 26, 2007.
50
Figure 16 Denmark Wireless Subscribers 19962006
7,000,000 6,000,000 Mobile Subscribers 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Source: Globalcommms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites.
Among the four MNOs, TDC leads the market with a share of 50.65 percent, followed by Sonofon with 25.50 percent, and Telia with 20.48 percent. Hi3G Access Denmark is the smallest player with a share of 3.37 percent. However, it serves over 70 percent of Denmarks 3G subscribers.
51
Figure 17 Denmark MNO Market Shares (by subscribers)
Hi3G Access Denmark Telia Denmark
TDC Mobil
Sonofon
Source: Globalcommms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites.
In 2000, as part of the ECs 1999 telecommunications review, the National IT and Telecom Agency (NITA) concluded that TDC and Sonofon possessed single SMP. Consequently, it subjected the two carriers to a wholesale access obligation where they had to grant all reasonable requests for wholesale access. It is important to note that the finding of Sonofon and TDCs SMP was based on analyses conducted under the ECs old regulatory framework. Under this approach, operators with a market share in excess of 25 percent were automatically presumed to have SMP. Since at that time, TDC had a market share approaching 50 percent and Sonofons share was near 30 percent, both carriers were found to have SMP. The EC revised its old regulatory framework and replaced it with the current framework, which now defines SMP according to the economic concepts of market dominance described above. Based on the new definition of SMP, NITA reexamined Market 15 in mid-2005 and concluded that no carrier had SMP, either single or collective. In arriving at its revised decision, NITA considered the following criteria and market observations: Indicators of market concentration: With nine retail service providers, NITA computed a HHI index of 0.37 that it considered as pointing towards competition.129
129
European Commission, Case DK/2005/0243: Wholesale access and call origination on public mobile telephone networks in Denmark, October 21, 2005.
52
Barriers to entry: NITA found that, while the licensing requirement could pose a barrier to entry, there is one additional license available for a new entrant. Moreover, NITA found that because there are four networks in place MVNOs have a choice of four wholesale providers, which eliminated bottleneck concerns.130 Wholesale price trends: NITA found that falling wholesale prices and the increasing number of MVNOs also point to effectively competitive markets.131 Countervailing buying power. Information and transparency. Barriers to changing suppliers.
Consequently, NITA declared the market as effectively competitive and removed all mandatory wholesale access obligations.132 3.5.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact) Denmark was one of the first European countries to implement wholesale access regulation and to support and promote the MVNO concept with a mandatory wholesale access rule. Under the ECs new regulatory framework, however, this rule was withdrawn. Currently, there are no MVNO-related regulatory measures in place in Denmark. 3.5.4. MVNO entry (current and forecasted market shares of MVNOs) The first attempt by an MVNO to enter Denmark was Sense Communications (a Norwegian MNO) in 1998. Sense sought a wholesale access agreement from Sonofon that was subsequently denied. This led to a number of regulatory complaints in which NITA supported Sense and forced Sonofon to provide limited wholesale access. Specifically, NITA endorsed Sense to use SIM cards that contain Sonofon Mobile Network Codes (MNCs) and Senses own subscriber numbers. However, it decided that Sonofon should not be required to route calls received from mobile terminals with SIM cards containing Senses own MNC code, because the service desired by Sense constitutes the same functionality as roaming between mobile networks. In other words, NITA did not regard roaming as an extension of interconnection.133 The first MVNO to actually enter Denmark was Tele2 in August 2000 with a wholesale agreement with Sonofon. By 2006, there were four additional large MVNOsTelmore, debitel Denmark, CBB Mobil, and DLG Tele. Combined, these MVNOs have a subscriber market share of 26.8 percent as shown in the table below.
130 131 132 133
Ibid. Ibid. Ibid. See Office of Communications, Mobile Virtual Network Operators: Oftel inquiry into what MVNOs could offer to consumers, a consultative document issued by the Director General of Telecommunications, June 1999, http://www.ofcom.org.uk/static/archive/Oftel/publications/1999/consumer/mvno0699.htm (accessed May 15, 2007).
53
Table 3 Denmark Total Wireless Market Shares (by subscribers)
Mobile Carrier TDC Mobil Sonofon Telia Hi3G Tele2 Telmore CBB debitel DLG tele Others
Market Share 31.50% 19.00% 19.90% 2.70% 4.00% 10.10% 4.30% 4.70% 1.00% 2.70%
Source: National IT and Telecom Agency, Denmark, Telecom Statistics first half of 2006.
While all five MVNOs originally entered as independent companies, three were subsequently acquired by Danish MNOs. Specifically, in January 2004, Telmore was acquired by TDC.134 Three months later, CBB Mobil was acquired by Sonofon.135 In May 2007, Tele2 was acquired by Telenor, who owns MNO Sonofon.136 Currently, only debitel (owned by TeliaSonera) and DLG Tele are independent MVNOs. Combined with all the smaller (other) MVNOs in Denmark, these independent MVNOs have a market share of 8.4 percent, which is comparable to the US and other countries. 3.5.5. MVNO business model (positive and negative aspects and implications) Given the relatively high penetration rate in Denmark, MVNOs focus primarily on the no frills model, which is characterized by low per-minute pricing on voice services and cheap SMS. 3.5.6. Pricing Trends before and after MVNO entry Between 2000 and 2004, prices for mobile voice services declined on average by 36 percent, while SMS prices declined by 52 percent.137 Considering the fact that most MVNOs focus on offering lower prices for voice and SMS, this price decrease likely can be attributed, at least to some extent, to the entry of MVNOs. However, the exact impact of the MVNOs is unclear. Furthermore, in light of the three MVNO acquisitions by MNOs, it is probable that MVNOs will have little impact on retail prices in the future.
134 135 136 137
TDC Press Release, TDC acquires 100% of Telmore, January 27, 2004, http://tdc.com/publish.php?id=2502 (accessed May 15, 2007). Telenor Press Release, Sonofon acquires CBB Mobil, April 30, 2004, http://press.telenor.com/PR/200404/943682_5.html (accessed May 15, 2007). Tele2 Press Release, Tele2 divest Tele2 Denmark to Telenor for approximately MSEK 1,025, May 9, 2007, http://www.tele2.com/pages/Press.aspx?id=90&source=1125232 (accessed May 15, 2007). Ixis Securities (2005): Telecommunications Services
54
3.6. Spain
Spain is yet another country where prospective MVNO-related regulation has caused much political and public interest. Because of a lack of MVNOs in the Spanish wireless market, the Spanish NRA, Comisin de las Mercados de Telecomunicaciones (or the CMT), has recently implemented mandatory wholesale access. 3.6.1. Available technological infrastructure There are four MNOs in SpainFrance Telecom Espana (Orange), Telefonica Moviles Espana, Vodafone Spain, and Xfera Moviles (Yoigo). Telefonica and Vodafone have been serving customers in Spain since 1995 with a 2G GSM infrastructure. Orange entered in 1999, also with a 2G GSM infrastructure. All three carriers eventually upgraded their networks to accommodate 2.5G services and built 3G and 3.5G networks. The fourth MNO, Yoigo, entered the market in late 2006 after a lot of financial troubles and repeated delays. Third generation services in Spain are still relatively new and only approximately 8 percent of the over 47 million subscribers have signed up for them.138 3.6.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry With mobile penetration rates of over 108 percent, Spains wireless industry continues to grow approximately 10 percent per annum, as illustrated in the figure below. In 2005, mobile-telephony-retail-service revenue has surpassed the fixed-telephony-retail-service revenue.139
138 139
Globalcommms 3.0, Telegeography Research Product, PriMetrica, Inc., and company websites. CMT, The Industry in 2005, p. 63, table 48.
55
Figure 18 Spain Wireless Subscribers 20002006
50,000,000 45,000,000 40,000,000 Wireless Subscribers 35,000,000 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 2000
Source: CMT
2001
2002
2003
2004
2005
2006
Market shares have remained rather stable over the last three years. Telefonica has been leading the market with a share ranging between 4549 percent. Currently, Telefonicas market share is 45.58 percent. Similarly, Vodafone has consistently been the second largest carrier, with a market share ranging between 2831 percent. Its current market share is 30.74 percent. Orange is the third largest MNO with an almost constant market share of 23.62 percent. Yoigo has just recently launched its services and has less than 1 percent of the market. These statistics are summarized in the figure below.
56
Figure 19 Spain Total Wireless Shares (by subscribers)
Orange
TME
Vodafone
Source: Globalcommms 3.0, Telegeography Research Product, PriMetrica, Inc., CMT and company websites.
The CMT has been attempting to entice more mobile carriers to enter the wireless market in Spain for quite a while. In May 2002, it introduced legislation that would ease the entry of MVNOs by setting up specific licenses, the A2 license, for such carriers. The A2 license was subsequently retired with the introduction of the new regulatory framework. Currently, MVNOs are not required to submit a specific license to CMT. They simply have to communicate to CMT their intention to enter the market, along with some general information, such as company name, company ownership, type of services to be provided, technology employed, etc. We understand from CMT that permission to provide service is granted automatically. The only exceptions to this process are full MVNOs requesting numbering resources. In addition to the general information, these carriers have to submit their roaming agreements with network operators prior to being allowed to enter the market. By 2005, the CMT had licensed a number of carriers, including Tele2, BT Group, Jazztel, Grupalia Internet, Meflur Comunicarte, and TeleCable. However, none of these licensed carriers was able to successfully negotiate a wholesale agreement with an MNO. Consequently, in its competitive review of Market 15 in 2005, the CMT analyzed a number of the complaints that had been levied against the MNOs. Specifically, the CMT investigated: The stable wireless penetration rate The relatively constant market shares of the three MNOs
57
The complex MNO pricing plans (low fixed monthly fees and per-minute fees that vary with onnet/off-net and peak/off-peak hours and preferred number calls) The subsidized handset issue The similarities of the MNOs business models and pricing plans The high prices for mobile services in Spain when compared to the EU average The downward-trending ARPU of all three MNOs in 20022005 (that CMT attributed to sales promotions, customer contracts, etc., but not nominal price reductions) The high profitability (measured by accounting profit measures EBITDA and ROCE) for all three MNOs with the ROCE (return on capital employed) greatly exceeding the WACC (weighted average cost of capital)140
In addition to these specific investigations, the CMT also reviewed the market for: Market transparency Barriers to entry Lack of actual and potential competition Demand growth patterns Frequency of MNO interactions Empirical evidenceabsence of any major MVNOs Retaliation strategy141
Given that at the time of the analysis there were no transactions in the wholesale market, the CMT used retail market performance to determine the level of competition in the wholesale market. Furthermore, rather than first evaluating the market for single SMP, the CMT went directly to analyzing collective dominance. Based on this analysis, the CMT concluded that there was no evidence of collective dominance in the retail market. Rather curiously, however, the CMT found that certain characteristics of the retail market suggested a state of collective dominance in the wholesale market. Specifically, the CMT pointed to the following three findings: Barriers to entry: The CMT found that despite granting 10 MVNO licenses and 119 reseller licenses, no operator had been able to negotiate an access agreement with any of the three MNOs. Thus, it concluded that barriers to entry must be significant.
140 141
European Commission, Case ES/2005/0330: Access and Call Origination on Public Mobile Telephone Networks in Spain, Comments pursuant to Article 7(3) of Directive 2002/21/EC, January 30, 2006. Comisin Del Mercado De Las Telecomunicaciones, Notification of Draft Measures pursuant to 7(3) of the Directive 2002/21/EC for access and call origination on public mobile telephone networks, date unknown.
58
High transparency: Because pricing by each MNO was visible to the other MNOs, the CMT concluded that there was no possibility of any single MNO reducing prices without being detected. Sustainable tacit coordination: The CMT found that there were strong incentives for the three MNOs to coordinate their pricing actions. The CMT reasoned as follows. High and stable profitability reinforces the MNOs incentive to maintain high prices and deny access to MVNOs and resellers. Further, a high probability of retaliation deters any one MNO from cheating by lowering prices (because retaliation can be mutually destructive and reduce profits for all). The CMT also found that there are no actual or potential market constraints on any of the MNOs (Xfera is not yet a factor) and that the high customer acquisition costs in mature market deter entry/share growth of new entrants.142
Consequently, the CMT proposed to impose ex ante wholesale access obligation on MNOs. The EC generally agreed with the CMTs analysis, particularly the finding of collective dominance in the wholesale market. However, it noted that the CMT could have done a more rigorous analysis in the following areas: Identification of a focal point at the retail level regarding transparency that could have further supported the finding of sustainable coordination at the wholesale level. Presentation of further evidence on the retaliatory capability of MNOs (perhaps in relation to whether they can increase capacity in the short run). Evaluation of consumer responsiveness to price changes by looking at data on SIMlocking and the presence of operators in various market segments.143
3.6.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact) Based on its finding of wholesale joint dominance in Market 15, the CMT imposed mandatory wholesale access and required an MNO to offer reasonable prices for access to its network.144 Specifically, wholesale access required the three MNOs to: Offer free access to technical interfaces or other indispensable technologies to permit the interoperability of services Facilitate ways in which to share the use of installations Offer the services necessary to guarantee the interoperability of extreme services with those offered to users Give access to the operating support systems or to information systems with similar functions
142 143 144
European Commission, Case ES/2005/0330: Access and Call Origination on Public Mobile Telephone Networks in Spain, Comments pursuant to Article 7(3) of Directive 2002/21/EC, January 30, 2006. Ibid. Proyecto de medida relativo a la definicin del mercado de acceso y originacin de llamadas en las redes pblicas de telefona mvil, el anlisis del mismo, la designacin de operadores con poder significativo de mercado y la propuesta de obligaciones especficas, CMT, December 30, 2005.
59
Offer interconnection of networks and other resources145
The CMT stated that it would take the following factors into consideration when determining the reasonableness of a request for access: Whether the agreement would lead to long-term competition in the market How easy it would be to replicate the infrastructure required to provide the service The economic and technical feasibility of using or deploying the specific resources needed to offer MVNO services The existence of spare capacity in the MNOs network Any other objective reasons146
Finally, in terms of pricing, the CMT required the three operators to charge reasonable prices for wholesale access by negotiating with those seeking access in good faith. In defining what constitutes reasonable pricing, the CMT put the following two guidelines in place: Prices offered to third parties by SMP operators should not be excessive and should not result in a margin squeeze or deter the entry of efficient operators. The charges for access should allow the MNO to recoup its investments.147
Orange has appealed the CMTs decision on grounds that it does not have the authority to unilaterally impose conditions for MVNOs to gain wholesale access. The case is currently pending at the Supreme Court. 3.6.4. MVNO entry (current and forecasted market shares of MVNOs) In the months following the CMTs decision, four MVNOs entered the Spanish wireless market, and about 20 more have announced their intention of entering in the next three years.148 Offering services are: French hypermarket chain Carrefour with an Orange wholesale agreement British phone retailer The Phone House with an Orange wholesale agreement Spanish 3G MNO Xfera with a Vodafone wholesale agreement Spanish Internet provider Euskaltel with a Vodafone wholesale agreement UK mobile carrier Lebara Mobile with a Vodafone wholesale agreement149
Furthermore, the following agreements have been reached, although services have yet to be launched:
145 146 147 148 149
Ibid. Ibid. Ibid. Jesus Garcia Lozano, Spain: Mobile Virtual Network Operators (MVNO), US Commercial Service, United States of America Department of Commerce, April 2007. Ibid. and NERA research.
60
Spanish retailer El Corte Ingles with a Telefonica Moviles wholesale agreement150 Spanish cable provider Ono with a Telefonica Moviles wholesale agreement151 British telecommunications company BT with a Vodafone wholesale agreement152 Cable TV provider Telecable with a Vodafone wholesale agreement153 Galician cable provider R Cable y Telecomunicaciones de Galicia with a Vodafone wholesale agreement154 Tourism group Globalia with a Vodafone wholesale agreement155
Currently, Spanish MVNOs have attracted 450,000 subscribers (less than 1 percent of the market), but forecasts are that they will achieve a 10 percent market share by 2011. 3.6.5. MVNO business model (positive and negative aspects and implications) The two dominant business models in Spain are brand and up-selling (convincing the consumer to purchase more expensive items, upgrades, or other add-ons in an attempt to make a more profitable sale). El Corte Ingles and Carrefour, for instance, built their business model on brand and existing distribution channels. Most of the other MVNOs seem to be companies that already have a stake in the telecommunications industry and seek to expand their offerings to include mobile services. Not surprising, three cable providers have wholesale agreements. For these companies, the MVNO agreements mean that they will be able to offer one more segment in their quest to offer a quadruple play. Interestingly, thus far Spain has not seen the type of MVNOs that seem to be the standard in Europethe ones that offer low cost, no frill services, typically to niche markets. 3.6.6. Pricing Trends before and after MVNO entry Given the recent introduction of MVNOs to the Spanish wireless market, no change in prices can be detected yet.
3.7. Hong Kong

Hong Kong is an excellent case study for the examination of the impact of regulatory intervention in MNO-MVNO relationships. Despite a high mobile penetration rate (over 135 percent) and multiple market participants, in 2001 Hong Kongs regulator, OFTA, mandated regulatory MVNO access to 3G networks. While most other countries provide only examples of commercially negotiated MVNOs, Hong Kong provides the unique opportunity to examine the economic impact and success of regulatory-mandated MVNOs.
150 151 152 153 154 155
Ibid. Ibid. Ibid. Ibid. Ibid. Telecompaper.com, Tourism group Globalia signs MVNO deal with Vodafone, May 9, 2007.
61
3.7.1. Available technological infrastructure Having been a colony of the UK until 1997, Hong Kongs technological infrastructure largely mirrors the one in the UK and other European nations. That is, 2G and 2.5G services are provided mostly through GSM, while 3G and 3.5G use the W-CDMA platform. The exception to this is Hutchison Telephone, which runs four digital mobile networksGSM 900/1800, CDMA, and W-CDMA. After extensive industry consolidations in the last two years, Hong Kongs 6.9 million inhabitants have a choice of five MNO networks: China Mobile Peoples Telephone Company (Peoples), CSL New World Mobility (CSL), Hutchison Telephone (Hutchison), PCCW Mobile Sunday Communications (Sunday), and SmarTone-Vodafone (SmarTone). Peoples is the only carrier that does not offer any services beyond 2.5G. CSL, Hutchison, Sunday, and SmarTone all offer 2G, 2.5G, and 3G services. CSL and SmarTone also offer 3.5G (HSDPA) services, which became available in the second half of 2006. The available technological infrastructure in Hong Kong is summarized below:
Table 4 Technological Infrastructure Hong Kong
Provider Peoples Peoples CSL CSL CSL CSL CSL Hutchison Hutchison Hutchison Hutchison PCCW PCCW PCCW SmarTone SmarTone SmarTone SmarTone
Service 2G 2.5G 2G 2G 2.5G 3G 3.5G 2G 2G 2.5G 3G 2G 2.5G 3G 2G 2.5G 3G 3.5G
Platform GSM GSM TDMA GSM GSM W-CDMA W-CDMA CDMA GSM GSM W-CDMA GSM GSM W-CDMA GSM GSM W-CDMA W-CDMA
Frequency Band 900/1800 900/1800 800/900 900/1800 900/1800 2100 2100 900/1800 900/1800 2100 900/1800 900/1800 2100 900/1800 900/1800 2100 2100
Launch Date 1997 2001 1992 1993 2000 2004 2006 1995 1998 2001 2004 1997 2001 2005 1993 2001 2004 2006
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. and company websites.
3.7.2. General competitive conditions (competitive, oligopolistic, monopolistic) before and after MVNO entry As of December 2006, Hong Kong had 9,444,140 mobile subscribers. Eighty-five percent (or 8,112,489 subscribers) were on 2/2.5G networks, and 14 percent (or 1,331,651 subscribers) 62
were on 3/3.5G W-CDMA networks.156 With a penetration rate of over 135 percent, the number of Hong Kong subscribers still continues to grow, as illustrated in the figure below:
Figure 20 Wireless Subscribers Growth in Hong Kong 19902006
10,000,000 9,000,000 8,000,000 7,000,000 Subscribers 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Year
Source: OFTA
Of the current 9.5 million subscribers, CSL serves 31.41 percent. The second largest mobile carrier is Hutchison at 26.14 percent, followed by Peoples with 17.84 percent, and SmarTone with 13.36 percent. The smallest player is PCCW with 11.26 percent. The 2006 market shares are illustrated in the figure below and contrasted to the market shares in 2001 just before MVNO regulation.
156
Office of the Telecommunications Authority, Hong Kong, Key Statistics for Telecommunications in Hong Kong, Wireless Services, March 28, 2007, http://www.ofta.gov.hk/en/datastat/eng_wireless.pdf (accessed April 18, 2007).
63
Figure 21 Subscriber Market Shares 2001 vs. 2006
35.00% 30.00% 25.00% 20.00% 15.00% 10.00% 5.00% 0.00% CSL New World Mobility Hutchison Telephone Company China Mobile SmarTonePeoples Vodafone PCCW Mobile Sunday Jun-01 Dec-06
Source: Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc., company websites, and Datafile of Asia-Pacific Telecommunications, CIT Publications, 2001.
As the above graph indicates, since 2001 the two largest mobile carriers in Hong Kong (CSL and Hutchison) lost some market share to Peoples and PCCW.157 SmartTone also suffered a drop in market share. Notably, as we discuss below, this analysis does not include the approximately 7 percent of customers currently served by MVNOs. Hong Kongs mobile market has been found to be among the most competitive in the world.158 Specifically, with a Herfindahl-Hirschman Index (HHI) of 2,293, it has the leastconcentrated market amongst developed countries in the world. Moreover, Hong Kongs prices, measured as revenue per minute (RPM) is US$0.04, are the lowest in developed countries.159 Monthly wireless churn rates are estimated to be 3.4 percent; the highest reported number in developed countries.160
157 158 159 160
The 2001 market share for CSL New World Mobility reflects the combined shares of New World Mobility and CSL Hong Konga merger that took place in November 2005. Janusz A. Ordover and William Lehr, Mobile Service Relicensing in Hong Kong: Economic Considerations, June 19, 2004. Merrill Lynch, Global Wireless Matrix 4Q06, Shaky markets, solid fundamentals, March 28, 2007 Ibid.
64
3.7.3. MVNO policy (level and kind of pro-MVNO regulatory intervention, rationale and impact) Implemented in 2001, Hong Kong is one of the few countries in the world that has MVNOrelated regulation. The regulation, limited to 3G networks, requires 3G licensees to open up 30 percent of their network capacity to unaffiliated MVNOs and is part of the licensing conditions for 3G spectrum. The discussion about MVNO-related regulation was brought up as early as 2000 in the context of the licensing framework for 3G mobile services. At that point in time, Hong Kong had 11 mobile telecommunications networks in operation and a penetration rate of 55 percent.161 OFTA stated early in the licensing-framework-design process that it considers that the introduction of new entrants to the 3G market will be beneficial to market development and to consumers.162 It is important to note though that OFTA was primarily not looking for 3G entry from existing 2G licensees as new entrants would not be constrained by any legacy network elements.163 In addition, OFTA sought the industrys input as to whether any competitive safeguards should be introduced to preserve effective competition in the 3G-market segment.164 OFTA also specifically discussed MVNOs in the context of the 3G licensing framework. In particular, the Hong Kong regulator sought comment on whether wholesale-retail separation (and with it mandatory open wholesale access) should be put into place. Citing similar separation concepts in the UK, Denmark, Finland, Norway, and Sweden, the regulators objective was to introduce MVNO entry.165 OFTA expected that: Separating service provision from network operation in the 3G mobile services would enhance competition in the services market and provide customers with more choice and variety of service and price packages. Service providers would develop a wider range of value-added and multimedia services such as mobile access to the Internet, innovative data applications over mobile networks etc. to customers and customers would be able to enjoy wider geographical coverage if service providers team up with more than one mobile network operator.166 However, OFTA also was clear on the potential drawbacks from such regulation, stating: There are demerits with this concept. If the new 3G mobile network operators are obliged to open up their networks and systems and share use of the spectrum with the service providers, and such interconnection or sharing were based on cost-based charges, they would have less commercial incentive to invest and build up their own networks, infrastructure and service coverage. Because the 3G mobile network operators would also be service providers, there would be competition, charging, bill settlement and commercial issues to be tackled among 3G mobile network operators and the service providers.
161 162 163 164 165 166
OFTA, Licensing Framework for Third Generation Mobile Services, An Industry Consulting Paper, March 21, 2000, p. 12. Ibid. Ibid. Ibid, pp. 1718. Ibid, p. 20. Ibid.
65
Some regulators are of the view that such matters should best be resolved by commercial negotiations rather than by regulatory intervention (citing the UK regulator, OFTEL, in a statement issued in October 1999, indicating that it did not see the need for regulatory intervention to implement the concept of MVNO for the time being).167 Supporters of wholesale-retail separation urged OFTA to ensure that there would be nondiscriminatory access to the 3G MNOs networks and no regulatory barriers that might foreclose 3G MVNO entry. Because OFTA allowed wholesale providers to also operate in the retail market, open-network providers urged the regulator to minimize collusion among these players. Opponents of the proposed open-network requirement stressed that wholesale access should be voluntary and subject to market forces. They also raised concerns about potential technical and financial difficulties to meet the necessary interconnection requirements. After receiving and reviewing comments from the industry, OFTA issued preliminary conclusions and sought further comment from the industry. With respect to wholesale-retail separation, OFTA stressed again its intention to introduce open networks as an essential element in the approach for the selection of operators. The regulator found that in light of the limited spectrum availability, this policy was consistent with its goal of introducing more competition in the marketplace. OFTA maintained its position that such policy would enhance consumer welfare by increasing competition, and, at the same time, gave assurances that the 3G-license cost will not be passed on to consumers. The regulator concluded that to implement such an open network requirement, it is necessary to separate service provision from network operation and to institute some sort of regulatory intervention in the determination of wholesale prices if commercial negotiation fails.168 Consequently, OFTAs preliminary decision was to implement the open-network access policy and require MNOs to make 3050 percent of their network capacity available for MVNOs.169 OFTAs preliminary view on the wholesale price was to initially leave it to commercial negotiations and only intervene if such negotiations failed. If they had to intervene, OFTA noted that it would apply a retail minus and cost plus approach.170 Based on these and other industry consultations, in July 2001, OFTA issued its ruling on the 3G licensing framework. Essentially, the regulator confirmed its preliminary views on open-network access and required each licensee to open 30 percent of its network capacity to nonaffiliated service providers (NSPs).171 In order to assure nondiscriminatory access, the regulator put the following safeguards into place:
167 168 169 170
171
Ibid., pp. 2021. OFTA, Licensing Framework for Third Generation Mobile Services, Analysis of Comments Received, Preliminary Conclusions and Further Industry Consultation, October 3, 2000. Ibid, p. 12. The retail minus approach sets wholesale access prices based on the end-user or retail prices of the corresponding final services minus a discount, which is usually set as a fixed percentage of the retail price. The cost plus approach sets these prices by calculating the incremental costs of provisioning the wholesale service and then adding a profit margin to it. OFTA, Hong Kong Third Generation Mobile Services Licensing Information Memorandum, July 2001, p. 7.
66
NSPs should have access to the same transmission and supporting capabilities (e.g., data rates, bearer attributes and intelligent network functionality) as the host mobile network operator (MNO) has available to serve its own customers.172 Traffic associated with the NSPs is to be treated on a non-discriminatory basis compared to that of the Licensee, or its affiliated NSPs that use the Network, as the case may be, in all respects unless otherwise agreed between the parties concerned.173
OFTA limited the open-network requirement to NSPs, which meant that MNOs were not required to open their network to MVNOs or content providers that were affiliated with a licensee. It is important to note that OFTA made the open-network requirement part of the spectrumlicensing framework, allowing all potential bidders to consider this regulatory constraint on their use of the 3G licenses. Presumably, the open-network requirement limits the free use of 3G spectrum, resulting in a reduction in expected utility and thus maximum willingness to pay. In other words, economic theory would suggest that the prices for 3G licenses were lower in the presence of the open-network requirement, relative to the prices without such constraint. Similarly, with the open-network requirement, some bidders that might have considered entering as an MNO (i.e., bid for a license) now had the option to enter as an MVNO. The licensing framework also imposed a number of requirements on potential MVNOs. Specifically, in order to promote at least partially facilities-based MVNO entry, OFTA required MVNOs to: Provide, or intend to provide, mobile services to a customer base, including the general public Provide their own mobile switching and gateway infrastructure for circuit and/or packet switched traffic Enter into their own interconnection and roaming agreements Provide their own business support systems, such as billing and customers care Maintain their own Home Location Register of subscribers Satisfy requirements for call control Issue their own SIM cards174
OFTA also spelled out when and how it would interfere if commercial negotiations among MNOs and MVNOs fell apart175
172 173 174 175
Ibid. Ibid. Ibid, p. 9. Ibid. p. 18.
67
From 2001 through 2005, the above requirements for MNOs and MVNOs were only applied to 3G licensees. As 2G licenses were renewed in 2005 and 2006, OFTA extended its opennetwork access policy to also include the renewed 2G licenses. However, given the historically unconstrained 2G licenses, OFTA granted a five-year waiver from these requirements with a possibility for an extension thereafter. Thus, the earliest date for 2G open access requirements is 2010. OFTA has informed us that since the inception of the open-network access policy in 2001, it has not received any complaints with respect to the commercial negotiation of the MNOMVNO contracts and thus so far has never intervened beyond the initial policy. OFTA further has informed us that in the light of fixed-mobile convergence, the regulator is reviewing whether the open-network access policy is still appropriate. In fact, in a recent consultation exercise for the award of a CDMA200 mobile license (using 850 MHz spectrum), the regulator has proposed not imposing the open-network access requirement on the successful bidder. 3.7.4. MVNO entry (current and forecasted market shares of MVNOs) Following OFTAs 2001 decision, seven MVNOs applied and were granted an MVNO license: Trident Telecom Ventures (Trident), China Motion Telecom HK Ltd (China Motion), China Unicom International Limited (China Unicom), China-Hongkong Telecom Ltd. (China-Hongkong), CITIC Telecom 1616 Limited (CITIC), Telecom Digital Mobile Ltd (Digital), and IMC Networks Limited (IMC). While the initial MVNO response was strong, the commercial success of these MVNO is rather limited with only approximately 7 percent of all subscribers being served by MVNOs.176 Nevertheless, the Hong Kong MVNOs experienced over 30 percent growth in their subscriber base between 2005 and 2006. The number of MVNO customers increased from 529,554 in December 2005 to 689,667 in December 2006.177 Furthermore, all MVNO services in Hong Kong are 2G based, and 3G services are only now being rolled out. Because there are no open-network access requirements on 2G licenses, this development suggests that all seven MVNOs successfully negotiated wholesale agreements with MNOs, absent any form of regulatory intervention. On the other hand, because there are still no 3G services, this implies that, at least for now, OFTAs MVNO policy has had no impact on MVNO entry. It is worthwhile to note that Virgin Mobile (an MVNO) has not entered the Hong Kong market, despite initial announcements that it would, at the time it had operations in Singapore, and that several MNOs sought to have Virgin Mobile as its MVNO. It is also interesting to observe that once OFTA implemented the open-network access policy, MNOs were actively seeking out desirable MVNOs, possibly in a quest to share the 3G licensing costs and to mitigate their business risk.178
176 177 178
See Office of the Telecommunications Authority, Hong Kong, Key Statistics for Communications in Hong Kong, Wireless Services, 28.3.2007, http://www.ofta.gov.hk/en/datastat/eng_wireless.pdf. Ibid. China Peoples Daily, Virgin Expresses Interest in Hong Kong 3G Network, October 8, 2000, http://english.people.com.cn/english/200010/08/eng20001008_52013.html (accessed April 20, 2007).
68
3.7.5. MVNO business model (positive and negative aspects and implications) Similar to MVNO business models in other parts of the world, Hong Kong MVNOs largely target niche markets. China Unicom, China Motion, and China Hongkong offer international plans covering China, Singapore, and Hong Kong that allow business travelers to visit these countries without incurring international roaming fees. Trident, on the other hand, focuses on prepaid services, targeting business travelers from non-GSM nations, such as Japan, South Korea, and the US. 3.7.6. Pricing Trends before and after MVNO entry No data on pricing trends or public reports on such are available. Given the recent launch of MVNOs in Hong Kong, the limited subscribership they have attained, and their focus on previously untapped market segments, we do not expect the current MVNO offerings to have a strong competitive or pricing effect. Similarly, given the intensity of the MNO competition, we do not expect MVNOs to have a strong impact on competition, even with the introduction of 3G services.
3.8. Other Countries

In addition to the detailed review of the countries above, we think it is important to provide brief overviews of MVNO regulatory proceedings and debates in some other countries. 3.8.1. Japan In spring 2002, the Ministry of Internal Affairs and Communications (MIC) drafted the Guideline Concerning Applicable Scope of the Telecommunications Business Law (TBL) and the Radio Law (RL) Pertaining to Mobile Virtual network Operators (MVNO), in which the regulator outlined its approach with respect to MVNOs. Essentially, the MIC decided to intervene only if commercial negotiations failed.179 Due to the limited entry of MVNOs that followed this decision, the MIC has been reviewing its MVNO policy for the last two years, debating whether it should extend the current interconnection obligations in the wireline and wireless sectors to MVNOs. While the exact scope of such a proposed approach is still unclear, this would mean, at a minimum, that MNOs would be required to open their networks to MVNOs. In February 2006, NTT DoCoMo, Japans largest MNO, announced that it would voluntarily open part of its 2G network to MVNOs. This opening, however, did not include networks in major cities as this would not encourage facilities-based entry, and all wholesale agreements are subject to the successful negotiation of reasonable wholesale fees.180 Currently, there are two MVNOs in Japan, Jupiter Communications (J-Com) and Japan Communications Inc (JCI). J-Com, a Japanese cable TV operator, provides mobile services through a wholesale agreement with Willcom. Willcom, formerly DDI Pocket, is Japans largest wireless provider using a 2G technology known as personal handyphone system
179
180
Ministry of Public Management, Home Affairs, Posts and Telecommunications, Guideline Concerning Applicable Scope of the Telecommunications Business Law (TBL) and the Radio Law (RL) Pertaining to Mobile Virtual network Operators (MVNO), Spring 2002. Telegeography, Globalcomms 3.0, Japan Wireless Markets, last updated January 31, 2007.
69
(PHS), which covers a smaller area than standard mobile networks but at a lower cost. The company is owned by the US investment fund Carlyle Group and Japanese electronics conglomerate Kyocera Corp. Named J:Com Mobile, this MVNO mainly targets J-Coms cable subscribers through quadruple play or other bundled offerings. JCI, recognized as the first Japanese MVNO, commercially negotiated a wholesale agreement with Willcom in August 2001 and started offering services in December 2001.181 JCI also has partnerships with a number of Wi-Fi carriers in Japan and offers Japans most comprehensive wireless LAN broadband access network.182 JCI mainly offers plans for data services. 3.8.2. South Korea With a penetration rate of 83.2 percent, the South Korean mobile market has been stagnant (growing on average less than 5 percent per annum in the last four years). In March 2005, the Ministry of Information and Communications (MIC) licensed three operators (SK Telecom, KT, and Hanaro Telecom) to provide WiBro services.183 As part of this concession, the MIC stipulated that these MNOs must provide wholesale access to MVNOs three years after the launch of their WiBro service or when the total number of WiBro subscribers exceeded five million subscribers. When either occurs, the MNOs will be obliged to provide 30 percent of their network capacity to MVNOs.184 Until that time, MVNOs are not permitted in South Korea. In setting the MVNO policy, the committee advising the MIC (consisting of members from the MIC and the state-run Korea Information Strategy Development Institute), noted that: We dont envision MVNO as a measure to enhance competition and eventually relieve the market dominated by SK Telecom. Its rather to boost the overall use to networks and provide a wider variety of data-based services in the industrys transition to third-generation mobile telephony.185 WiBro was offered to the public in June 2006. According to the latest estimates, there are currently less than 1,000 WiBro subscribers in South Korea. Several parties have expressed an interest in becoming an MVNO in South Korea, including Kookmin Bank and longdistance, fixed-line operator Onse Telecom.186 3.8.3. Jordan The Telecommunications Regulatory Commission (TRC) of the Kingdom of Jordan has recently opened a proceeding regarding the implementation of MVNOs in Jordan. As an
181 182 183 184
185
186
Japan Communications, JCI Corporate History, http://www.j-com.co.jp/corp/en/history.html. Japan Communications, Company Vision, http://www.j-com.co.jp/corp/en/vision.html WiBro is a wireless broadband Internet technology developed in South Korea. It is essentially the Korean version of WiMAX. Ministry of Information and Communications, Korea to allow portable Internet service to start by June 2006, March 10, 2005, http://eng.mic.go.kr/eng/user.tdf?a=user.board.BoardApp&c=2002&board_id=E_03_01&seq=502. The Korea Herald, Ministry discusses adopting MVNO, June 21, 2004, posted on the website of the Ministry of Information and Communications, http://eng.mic.go.kr/eng/user.tdf?a=user.board.BoardApp&c=2002&board_id=E_05_06&seq=2820. Ibid.
70
initial step in this proceeding, it has requested comments from industry participants regarding a number of key questions. Among other things, the TRC sought input on: MVNO business models MVNO barriers to entry Consumer benefits from MVNO entry Reasons for the lack of MVNOs in Jordan Recommendations on the type of license issued to MVNOs Regulatory recommendations for the development of MVNOs in Jordan187
While the proceeding has just begun, most MNOs have rejected the idea of regulation aiding MVNO entry. The MNOs believe that competitive forces are sufficiently strong and regulation is not necessary. Further, they think that all MNO-MVNO agreements should be voluntarily negotiated. 3.8.4. Norway Like Denmark and Finland, Norway has been cited as an example of a country where regulation has enabled MVNOs to gain a relatively large market share.188 While we discuss Denmark in detail as part of our benchmark analysis, we find that brief discussions on the MVNO developments in Finland and Norway are also useful. As of September 2006, Norways two MNOs (Telenor and NetCom) serve approximately 78.5 percent of the mobile subscribers. The remaining 21.5 percent are served by MVNOs Chess (7.6 percent), Tele2 (7.4 percent), Ventelo (2.4 percent), and other smaller MVNOs (with a combined 4.1 percent).189 Compared to other countries, such as the US, the MVNO market share is high. In order to understand the cause of this high MVNO market share, one needs to review the market conditions and regulatory environment. In 2000, the regulator NPT awarded 3G licenses to Telenor Mobile, NetCom, and Tele2. In September 2002, Telenor and Tele2 reached a commercially negotiated agreement, which provided Tele2 with wholesale access to Telenors network.190 In light of this agreement, Tele2 returned its 3G license and became an MVNO. Thus, at least 7.4 percent of the 21.5 percent MVNO market share is not due to regulation, but to commercial negotiations. After much debate, in 2003, the regulators imposed an obligation on MNOs to provide access to MVNOs. In 2005, Telenor reached wholesale agreements with TDC and Ventelo. In early
187
188
189 190
See Telecommunications Regulatory Commission (TRC), Hashemite Kingdom of Jordan, Notice requesting comments on Implementation of Mobile Virtual Network Operator (MVNO) in Jordan, January 14, 2007. Telecommunications Reports, TR Daily, MVNOs Prosper More Quickly where Regulations Enable Them, April 23, 2007, citing a presentation by Analysys Consulting Ltd, which found that regulation is key for high MVNO market shares. BuddeComm, Norway Mobile Market Overview & Statistics, April 2007, p. 4. Norwegian Post and Telecommunications Authority, et al., Competition and regulation in the Nordic mobile markets, date unknown, p. 30.
71
2006, the regulator determined that a number of operators had SMP, under the guidelines imposed by the EC. Specifically, the NPT determined that Telenor, NetCom, Teletopia, and MVNO Tele2 had SMP and subjected them to individual regulation. For Telenor and NetCom, this finding required these MNOs, once more, to provide wholesale network access to MVNOs. The decision is in force, even though it has been appealed by Telenor. In 2005, TeliaSonera, owner of the NetCom brand, acquired Chess. Thus, the MNO bought a MVNO, although the MVNO is still marketed under its own name. With Tele2 being involved in a voluntary agreement and Chess being bought by an MNO, this left only Ventelo and some other (smaller) MVNOs in Norway as the subject of regulation. While it is unclear whether the MVNO agreements between Telenor and TDC and Telenor and Ventelo are a result of regulation, this would indicate that the effective MVNO market share in Norway is 6.5 percent, which is similar to the MVNO share in the US. 3.8.5. Finland Until recently, Finland had a number of MVNOs, serving in excess of 12 percent of mobile customers. For instance, in 2003, ACN, an MVNO on TeliaSoneras network, had approximately 300,000 customers, which was a market share of 6.3 percent.191 In the same year, MVNO Saunalahti had 144,000 subscribers, which was a market share of 3.03 percent.192 In 2004, however, ACN exited the market and its subscribers transferred to TeliaSonera. Upon ACNs exit, Saunalahtis subscribership grew significantly, totaling 419,351 subscribers at the end of 2004a market share of 8.5 percent. In 2005, however, Saunalahti was acquired by MNO Elisa Oyj. ACNs market exit in 2004 and the acquisition of Saunalahti reduced the MVNO market share in Finland to approximately 1 percent. With respect to regulation, the Finish regulator (FICORA) has traditionally encouraged MNOs to enter into wholesale agreements with MVNOs, but has refrained from actually intervening. Thus, all MVNO agreements in Finland have been the result of commercial negotiations. In 2004, as part of the ECs directive to review Market 15, the regulator concluded that none of the MNOs had SMP. FICORA based its decision on the following findings: Existence of commercially negotiated MNO-MVNO agreements Existence of incentives for MNOs to sell spare capacity to MVNOs Existence of competition in the wholesale market
It is worth noting that the Finish market differs from other markets in that handset subsidies are prohibited, thus switching barriers are relatively low. Similarly, MNP, introduced in 2003, requires number porting within five days of a request, which led to higher porting rates than in other Nordic countries.193
191 192 193
Ibid., p. 23. Saunalathi Group Oyj, Annual Report 2003, http://www.saunalahtigroup.com/annualreport2003/saunalahti.html (accessed April 27, 2007). Norwegian Post and Telecommunications Authority, et al., Competition and regulation in the Nordic mobile markets, date unknown, p. 24.
72
Thus, the relatively large market share of MVNOs observed in Finland was short-lived and not due to regulation but voluntary agreements. Moreover, with FICORAs decision not to regulate MVNOs and the recent (2006) unchallenged introduction of term contracts with lengths up to 24 months, including SIM-locking, the regulator seems to be increasingly adopting a hands-off approach to regulating the mobile sector. 3.8.6. Canada Much like the US, Canadas wireless market has repeatedly been found highly competitive. Consequently, the Canadian regulator, the Canadian Radio-Television and Telecoms Commission (CRTC), has refrained from regulating the MNO-MVNO relationship. Moreover, in March 2007, MNP was introduced, which is expected to further increase competition. Interestingly, while the six MNOs have grown only slightly, the MVNOs seem to continue to grow strongly. Between 2005 and 2006, the MNOs grew as follows: Bell Aliant: 2 percent Bell Mobility: 3 percent Manitoba Telecom: 3.5 percent Rogers Wireless: 3.4 percent SaskTel Mobility: 3.8 percent Telus Mobility: 3.7 percent194
In contrast, Telus Mobility informs its shareholders that it expects MVNO markets to continue to expand in 2007, as Virgin Mobile, using Bell Mobilitys network, grew its presence in 2006, and was joined by Videotron, partnering with Rogers to offer an MVNO wireless phone service.195 Telus also points out that the number of MVNOs continues to grow as [r]etailer brands such as Presidents Choice and 7-Eleven stores also launched MVNO offerings.196 In fact, in 2006, Telus launched its own MVNO under the Ampd Mobile brand. Bell Mobility and Rogers Wireless also entered the MVNO business by launching MVNO Solo and Fido, respectively. Thus, while MVNOs appear to be growing in Canada, they serve as a way for MNOs to further expand their market reach.
3.9. Lessons Learned from the Benchmark Countries

As demonstrated above, competitive conditions, consumer demand, technological progress, and approaches to regulatory intervention in each country are different. Hence, there is no one correct way of approaching the question of how to encourage MVNO entry into a countrys wireless sector. Nevertheless, there are a number of commonalities among the benchmark countries that provide valuable lessons for Israel.
194 195 196
Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc. and NERA Research. Telus Mobility, Growing Together, 2006 Financial Review, p. 45. Telus 2006 Online Annual Report, looking forward to 2007, http://about.telus.com/investors/annualreport2006/en/financials/looking_forward.html
73
Lesson 1: Lesson 2: Lesson 3: Lesson 4: Lesson 5: Lesson 6: Lesson 7:
All benchmark countries have at least one MVNO providing service, although it is too early to conclude that they all will be able to survive. MVNOs tend to serve previously unserved market segments rather than competing with MNOs, and they deepen and widen the market. MVNOs appear in both fragmented and unfragmented wireless markets. Typically two to three MVNOs serve the majority of MVNO subscribers. Independent MVNOs serve less than 10 percent of the wireless market, and forecasts place combined MVNO market shares at less than 20 percent. There are no established 3G MVNOs because all larger MVNOs are 2G providers. Generally, national regulatory agencies have not intervened in MNO-MVNO relationships with the exception of Spain and Hong Kong (in the case of 3G). The decision in Spain is currently under appeal; moreover, in Hong Kong, regulation produced no MVNO entry and the regulator is currently considering lifting all MVNO-related regulation. Most regulators have maintained a watchdog position; that is, they are carefully continuing to monitor the interactions between MNOs and MVNOs. Regulators decisions to intervene are typically based on a competitive review of the relevant market. The guiding principle is market failure, as observed through market power or anticompetitive behavior. Generally, regulation has not directly led to MVNO entry, although threats of regulation might have encouraged MNOs to negotiate with potential MVNOs, with the exception being Spain. Most MVNO business models are built on brand appeal, niche market targeting, existing distribution channels, discount offerings, and prepaid plans and are complementing, rather than competing with, the MNOs offerings. Most MVNOs are competitive tools used by MNOs rather than competitors of MNOs. The consumer-welfare impact of MVNOs is in extended and innovative service offerings and not in lower prices. The most successful MVNOs are frequently acquired by MNOs.
Lesson 8: Lesson 9:
Lesson 10:
Lesson 11:
Lesson 12: Lesson 13: Lesson 14:
74
4.
ASSESSING THE NEED FOR MVNO-RELATED REGULATION IN ISRAEL
The objective of this section is to assess whether MVNO-related regulation in Israel can be economically justified. In order to answer this question, we examined indicators of the level of competition, potential competition, and entry conditions in the Israeli wireless market. Specifically, we examined the market for signs of market failure with respect to MVNO entry. Market failure prevents competitive forces from working properly. Therefore, if the economic analysis reveals sufficiently competitive market conditions to alleviate concerns of market failure in the face of entry by MVNOs, then regulation becomes at best superfluous and at worst counterproductive. If not, then regulation might be required. We stress that the focus of our review of competition in the Israeli mobile market is to identify the implications for MVNO entry. It is not and is not intended to be an assessment of the overall level of competition in the market.197 Hence, we do not provide an opinion as to whether the overall market is competitive, effectively competitive, non-competitive. We also caution against using the present assessment to draw conclusions on the level of competition in areas other than MVNO entry, as competitive problems that do not relate to MVNOs might still exist. We also evaluate whether conditions in Israel are likely to be conducive to a viable MVNO business case. This analysis, together with the competitive review, will help us answer the question of why there are currently no MVNOs in Israel and what the competitive effects of mandated MVNO entry would be.
4.1. Competitive Conditions in the Israeli Mobile Market

MVNOs interact with MNOs at the wholesale level. Therefore, ideally one would examine competition in the provision of wholesale network access. However, since there are currently no MNO-MVNO wholesale agreements in Israel, it is necessary to rely on retail competition instead. Retail competition is generally regarded as a suitable proxy for wholesale competition, as it is unlikely that there will be no wholesale competition if the carriers compete in the retail market. Similarly, it is unlikely that there will be wholesale competition if there is no retail competition. Moreover, this methodology is consistent with the approach adopted in a number of countries, such as Spain and Ireland, where the need for MVNOrelated regulation, absent the presence of MVNOs, was assessed by the regulators. 4.1.1. Market definition Competition occurs within a market; supply and demand conditions within a market dictate prices; and the public interest is easier to evaluate when viewed as the welfare of all consumers within a market. Hence, we conduct our analyses in the context of a market. In assessing the extent of retail competition, we began our analysis with the most conservative (and typically undisputed) market definition. That is, we only include wireless
197
For instance, we do not analyze the interactions between mobile and fixed networks.
75
services. We are well aware that this definition does not take into account any competition Israeli MNOs might be facing from intermodal competitors, such as wireline or cable networks, and emerging services such as Wi-Fi and WiMAX. As we explain below, however, expanding the size of the market to include these potential competitors is not necessary as it is unlikely to have a material impact on our findings and conclusions. Furthermore, to the extent that emerging services provide for potential competition, we do consider them by analyzing the entry conditions in the Israeli wireless market. We also note that defining the market to include wireless only is consistent with the practice of FCC, the EC, and most of the EU NRAs when defining the market. 4.1.2. Analysis of the level of competitiveness of the Israel retail market with respect to MVNOs Consistent with economic theory and the practices employed by the NRAs in the benchmark countries, we evaluate the level of competition of the Israeli wireless retail market by assessing the risk of market failure with respect to MVNO entry.198 Specifically, we examine the Israeli retail market for signs of market failure in the form of market power or anticompetitive behavior. The European Court of Justice defines market power, or market dominance, as: a position of economic strength enjoyed by a undertaking which enables it to prevent effective competition being maintained on the relevant market by affording it the power to behave to an appreciable extent independently of its competitors, customers and ultimately of its consumers.199 As outlined above, the ECs guidelines state that a dominant position is found by reference to a number of criteria and its assessment is based on a forward-looking market analysis based on existing market conditions.200 Hence, we assess the competitive conditions in Israel by analyzing: Market shares and market-share trends Barriers to entry and expansion Pricing and profitability Coordination incentives and abilities Historical behavior Market contestability
When applicable, we benchmark our findings in Israel against those in the countries discussed above.
198 199 200
We emphasize, once more, that our analysis only addresses the risk of market failure with respect to MVNO entry. It is not an all-inclusive competitive review of the wireless market in Israel. European Court of Justice, United Brands v. Commission, Case 27/76, 1978, ECR 207. Ibid.
76
4.1.3. Market shares and market-share trends Market shares and market-share trends are an integral part of a competitive assessment as concerns regarding potential market failure are more pronounced where there are only a few firms in the market, or where one or some firms have very large market shares, thus making it difficult for smaller firms to compete. For instance, as held by the European Court of Justice: very large market shares are in themselves, and save in exceptional circumstances, evidence of the existence of a dominant position. An undertaking which has a very large market share and holds it for some time is by virtue of that share in a position of strength.201 While we do not entirely agree with the absoluteness of this statement, particularly for regulated industries, we do agree that market share and market-share trends are a relevant part of a rigorous competition review. This view is consistent with the ECs guidelines, which pay particular attention to market share, yet correctly stress that a large market share alone is not sufficient to establish market dominance.202 4.1.3.1. Israeli Market Shares Cellcom leads the Israeli wireless market in terms of total subscribers with a share of 34.5 percent, followed by Partner with 31.9 percent, and Pelephone with 29.0 percent. MIRS is the smallest MNO with a total market share of 4.6 percent.
201 202
European Court of Justice, Hoffman-La Roche v. Commission, Case 85/76 1979, ECR 461, 41. The limited use of market-share analysis to determine market power has also been adopted by the US Federal Trade Commission and the US Department of Justice. Market share analyses can be useful screens to determine whether there is potential market power. However, they should not be seen as being dispositive on market power and can, at times, be highly misleading indicators of market power.
77
Figure 22 Israel Total Wireless Market Shares 2006
MIRS, 4.55%
Partner, 31.91% Pelephone, 29.02%
Cellcom, 34.52%
Source: Companies financial reports, data request responses
Pelephone, which has been wholly owned by Israelis fixed-line incumbent Bezeq since 2004, was the first MNO to enter the Israeli mobile market in 1986. It enjoyed an eight-year monopoly until December 1994 when Cellcom entered the market. Partner launched its services in October 1998 after paying reportedly US$400 million in licensing fees and investing an additional US$700 million in building its 2G GSM network.203 With its roots in providing trunking services to business and the government, MIRS entered the wireless market in 2001, focusing primarily on the business sector. The incumbent, Pelephone, has lost much of its market share and currently is the smallest of the three large MNOs. The two largest firms, Cellcom and Partner, had in 2006 a combined market share of 66.4 percent. In 45 of the 53 countries (including Israel) for which such data are tracked, the combined market share of the two largest MNOs is higher than that of Cellcom and Partner.204 In seven countries (Russia, Hong Kong, UK, US, India, Taiwan, and Brazil) the combined share of the two largest MNOs is smaller than in Israel.205 Under the EUs old regulatory framework all carriers with a market share exceeding 25 percent were
203 204 205
Partner Communications Company Ltd., Response to NERAs Mobile Network Operators Data Inquiry MOCs Letter dated 15.04.07", May 21, 2007, p. 3. (Partner Data Request Response). Merrill Lynch, Global Wireless Matrix 4Q06, Shaky markets, solid fundamentals, March 28, 2007, tables 5 and 6. Ibid. The average combined share of the two largest MNOs in the 53 benchmark countries is 77.3 percent.
78
automatically assumed to have SMP. Hence, under this old standard, Cellcom, Partner and Pelephone would all be considered as having SMP. However, under the new EU framework the size of the Israeli MNOs market shares alone would not be sufficient for a finding of dominance (SMP) and do not in themselves point to market failure. Similarly, as shown below, market concentration in Israel, as measured by the HerfindahlHirschman Index (HHI) is 3,073, which is lower (hence less concentrated) than an average HHI of 3,706 in the 53 countries mentioned above. Ten countries have lower concentration levels than Israel: Pakistan, Taiwan, Germany, Iraq, Russia, Hong Kong, UK, Brazil, US, and India. The remaining countries have more concentrated mobile markets.
Figure 23 Global Comparison of HHIs 2006
7,0 0 0
6 ,0 0 0
5,0 0 0
4 ,0 0 0
3 ,0 0 0
2 ,0 0 0
1,0 0 0
Source: Companies financial reports, data request responses, Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc.
Given MIRS limited role in the private segment, it is also informative to calculate separate HHIs for the private sector and the business sector in Israel. The Israeli HHI for the private wireless sector is 3,197, while the concentration in the business sector is 2,757. These numbers indicate that the more concentrated private sector is below the overall average HHI for the benchmark countries, with 15 countries having lower overall concentration levels. We conclude that 2006 market shares in Israel, overall and by customer segment, are relatively favorable in terms of the potential for competition and by themselves do not provide evidence of the existence of a dominant position by any MNO or a group of MNOs.
Mexico China Morocco Norway Colombia New Egypt Peru South Africa Philippines Switzerland Turkey Indonesia Japan Algeria Korea Portugal Venezuela Ukraine France Chile Finland Thailand Denmark Czech Hungary Spain Ireland Bangladesh Belgium Nigeria Sweden Malaysia Netherlands Singapore Australia Poland Austria Argentina Canada Italy Greece Israel Pakistan Taiwan Germany Iraq Russia Hong Kong UK Brazil US India
79
4.1.3.2. Israeli Market Share Trends The EC states that an operator with a large market share may be presumed to have SMP if its market share has remained stable over time.206 If its market share is gradually eroding, this may indicate that the market is becoming more competitive, although that fact alone does not necessarily preclude an SMP finding. On the other hand, fluctuating market shares over time may indicate a lack of SMP in the relevant market.207 As illustrated in the figure below, the market shares of the four Israeli MNOs (Partner, Cellcom, Pelephone, and MIRS) fluctuated between 1999 and 2003 but have remained relatively stable thereafter. Figure 24 Israel Total Wireless Market Shares 19992006
60% 50% Wireless Market Shares 40% 30% 20% 10% 0% 1999 2000 2001 2002 2003 2004 2005 2006
Partner Cellcom Pelephone MIRS
Specifically, Partner, which entered the wireless market in 1998, gained market share quickly, eroding the shares of the two incumbent MNOs (Cellcom and Pelephone). Partner exceeded Pelephones subscriber count in 2002, four years after its launch, and reached near parity with Cellcom a year later. The fact that market shares have not changed significantly in the last four years is not necessarily a source of concern regarding market failure. First, as markets mature, shares tend
206 207
EC Guidelines on Market Analysis, 75. Ibid. Fluctuating market share may be empirical evidence of a contestable market. Given the relentless lowering of barriers to entry and exit (through VoIP or other emerging technologies), certain communications markets have become contestable and, thus, lack market power.
80
to become more stable. In fact, similar market-share patterns can be observed in the US and the UK where regulators have declared these markets as effectively competitive. Second, while market shares might have been more stable between 2003 and 2006, the MNOs individual patterns of growth in the number of subscribers fluctuated over the same period, as summarized in the table below. Table 5 Israel MNO Subscriber Growth 20032006
Partner Cellcom Pelephone MIRS Total Subscribers 2003 14% -7% 11% 5% 5% 2004 11% 7% 9% 9% 9% 2005 8% 6% 7% 7% 7% 2006 5% 11% 6% 17% 8%
Third, while total market shares have remained relatively stable over the last four years, the market shares for the various market segments (private, business, and prepaid) have been less stable. Specifically, as, illustrated in the figures below, in the private sector, since 2003 [].208
208
We note that part of the fluctuations in market shares might be due to changes in the MNOs definitions of active and inactive subscribers.
81
Figure 25 Israel Wireless Market Shares Private Sector 20022006
Similarly, in the business sector, market shares have shown some variation over the past four years. Most notably, [].
82
Figure 26 Israel Wireless Market Shares Business Sector 20022006
Finally, in the prepaid sector, market positions are rather different than in the private and business sector, [].
83
Figure 27 Israel Wireless Market Shares Prepaid Sector 20022006
The market share trends in Israel, although demonstrating greater stability in recent years, are by themselves not inconsistent with the behavior of a competitive market, nor do they necessarily indicate market failure. We do recommend, however, that MOC continues to monitor the market positions in the private sector, as we would expect the market shares of the largest operators to change over the next few years, rather than converging to equality. 4.1.4. Pricing and Profitability Pricing and profitability levels are commonly used in evaluating the level of competition in an industry. In the case of MVNOs, the UK regulator, Ofcom, decided against regulation due in part to significant asymmetries in profitability. Prices and profitability levels also played an important role in Ireland and Spain, where regulators justified their regulation proposals in part due to the high profitability of MNOs. The rationale for examining prices and profitability is simple: high prices and high profitability might be indicative of supracompetitive prices and excessive returns (i.e., monopoly rents). This in turn might be evidence of market failure.
84
4.1.4.1. Prices in Israel To examine retail prices in Israel, we first benchmarked the average voice revenue-perminute (RPM) against 42 other countries for which 2006 RPM data were available.209 As shown in the figure below, Israels PPI adjusted RPM is lower than in 35 other countries and higher than in 13 countries. It is also lower than the average (on both a weighted and unweighted basis) and lower than the countries where regulators proposed (at least initially) wireless access regulationSpain, France, and Ireland.
Figure 28 PPI Adjusted RPM of Outgoing and Incoming Voice Traffic 2006
0.6 0.5 0.4 0.3 0.2 0.1 0 Bangladesh Pakistan US Hong Kong Ukraine Singapore India Thailand Russia China Algeria Finland Canada Israel Sweden Norway Denmark France Australia Korea Turkey Venezuela New Zealand Ireland Malaysia UK Mexico Austria Netherlands Chile Colombia Belgium Italy Spain Taiwan Germany Switzerland Portugal Greece Hungary Japan Czech Poland Argentina Egypt Brazil Peru Philippines South Africa
Source: Derived from Merrill Lynch, Global Wireless Matrix 4Q06.
While it provides some information, the above comparison suffers from a number of shortcomings which make it difficult to draw a solid conclusion. First, the source data for the RPM includes interconnection revenues. In many countries, interconnection revenues are regulated. Hence, RPM data which include regulated revenues might not be a good indicator of competitive prices. Second, it is unclear how the RPMs were calculated for each country. Particularly, the denominator, MOU, by which voice revenue is divided, might differ by country. In some countries MOU reflect actual conversation minutes, while in others MOU might be billed minutes. Depending on the billing increments (60 second, 30 seconds, 12 seconds, etc), actual MOU and billed MOU might differ and thus distort the above analysis. Third, in countries such as Israel where the size of the handset subsidy is a function of the MOU, prices might appear artificially low, unless handset revenues are part of the voice revenues.210 Fourth, the comparison does not reflect the ratio between prepaid and
209 210
The RPM is calculated by dividing monthly voice-only ARPU by monthly minutes of use (MOU). To illustrate this point, consider a hypothetical customer whose monthly voice charges amount to $100 and whose MOUs are 2000 minutes per month, resulting in an RPM of $0.05. Assume a handset subsidy of $50
85
postpaid customers. Specifically, at less than 30 percent, the proportion of prepaid subscribers in Israel is approximately half of that in EU countries.211 Given the limitations of a simple RPM comparison, it is also useful to examine pricing trends. The following figure compares the 4Q06 year-to-year percentage change in RPM for Israel to 49 countries for which data are available. For example, in 4Q06 Israels RPM decreased 3.6 percent from 1Q2006. As the figure shows, RPM in Israel has dropped at a slower rate than in most other countries. However, in many countries, the RPM at the beginning of 2006 was significantly higher than in Israel and thus arguably there was greater scope for a reduction in prices (although the precise position is again clouded by the comparability problems mentioned above).
Figure 29 Change in Revenue Per Minute from 1Q06 to 4Q06
Iraq Venezuela Russia Brazil Algeria South Africa Argentina
20.00% Percentage Change in RPM 10.00% 0.00% -10.00% -20.00% -30.00% -40.00% -50.00%
Source: Derived from Merrill Lynch, Global Wireless Matrix 4Q06, Table 31.
Furthermore, data for cumulative changes in RPM between 2002 and 2006 indicate that there was a proportionally greater RPM decrease in Israel than in developed countries in the AsiaPacific region and in Europe.212 Israels decrease, however, was much smaller than the ones in
211
212
was given, but was independent of the extent to which the consumer used his phone. Now consider an identical scenario, except now the handset subsidy is a function of the extent to which the consumer used his phone. If he obtained the handset subsidy only if he reached 2000 minutes, then the voice revenue is $100-$50=$50. Hence, the RPM is $0.025. Using subscriber, ARPU and MOU data provided by the operators for business, postpaid and prepaid we estimated the impact on RPM if pre-paid subscribers represented 60 percent of the market while business subscribers retained their current share. The impact was an increase in total market RPM (outgoing and incoming) of less than 0.2 US cents per minute. We note that part of this decrease might be due to regulator mandated decreases in interconnection fees.
Korea Hong Kong Canada Italy Israel Taiwan Portugal Denmark Chile Finland Spain Japan Singapore Bangladesh Turkey Belgium Switzerland Netherlands Norway Australia US UK Peru New Zealand Czech France Colombia Malaysia Philippines Greece Hungary Egypt Sweden Ireland China Germany Mexico India Austria Pakistan Poland Ukraine Thailand
86
North America and emerging countries in Asia and Europe. This is illustrated in the following table.
Table 6 Cumulative Change in Revenue Per Minute 20022006
Countries Developed Asia-Pacific Developed Europe North America Emerging Asia Emerging Europe Latin America Israel
Percent Change 20022006 -7% -4% -50% -47% -54% 0% -17%
Source: Derived from data in Merrill Lynch, Global Wireless Matrix Q4 2006, Table 30
In order to examine the Israeli MNOs pricing behavior more closely, we disaggregated the total RPM figures and examined the RPMs of Pelephone, Partner, and Cellcom by type of customer, accounting only for revenues for outgoing services. Specifically, we started with average revenue per minute for each type of call (on-net, off-net, and fixed) and for each type of customer (see Appendix A). We then calculated an overall weighted average outgoing call RPM for each customer type, using traffic volumes as weights. Finally, for private postpaid and business customers, we added the average subscription charges per minute. The comparative position of the different companies in 2006 is shown below:
Table 7 Outgoing RPMs (US$) by Sector and Company 2006
Prepaid Pelephone 0.17 Partner 0.15 Cellcom 0.17
Postpaid 0.11 0.10 0.11
Business 0.09 0.07 0.09
Source: Derived from data request responses. This analysis includes subscription charges and all outgoing calls apart from international calls, for which data could not be obtained for all companies. International calls account for approximately 1 percent of total call minutes and their exclusion thus has little impact on the results.
RPMs for prepaid private services appear to be substantially higher than RPMs for postpaid private services which in turn exceed RPMs for business services by between [] percent ([]) and [] percent ([]). So at least from an average revenue perspective, prices for private services appear to be substantially higher than the prices for business services. The following three figures depict the pricing trends for the business, postpaid private, and prepaid sectors by company over time. []
87
Figure 30 Outgoing RPM (US$) Trends Postpaid Private Sector Israel 2004-2006
Source: Derived from data request responses by the operators Figure 31 Outgoing RPM (US$) Trends Business Sector Israel 2004-2006
Source: Derived from data request responses by the operators
Figure 32 Outgoing RPM (US$) Trends Prepaid Sector Israel 2004-2006
88
Source: Derived from data request responses by the operators In addition, the RPM information summarized in Appendix A reveals the following: Israeli MNOs on average charge different prices: [] Israeli MNOs appear to follow different pricing strategies: Examining the pricing trends for outgoing RPMs suggests that the three MNOs are pursuing different pricing strategies. [] Subscription charges differ: []. Per minute prices in the private sector are higher than the prices in the business sector: Consistent with our findings above, the prices for each type of private call are higher than the corresponding prices for business calls, for each year since 2003 and for all carriers, although subscription charges in the business sector are substantially higher than in the private sector.213
Based on a review of the available data, the evidence on pricing (i.e. the overall decrease in prices and differences between operators in price levels and price changes) does not appear in itself to indicate market failure. It does, however, seem to indicate that prices in the private sector exceed prices in the business sector. Given, however, that call volumes per subscriber are much higher in the business sector, lower prices in the latter sector need not necessarily imply lower profitability. The evidence on sectoral profitability is examined in the next section.
213
The overall difference in the extent to which business and private subscribers receive handset subsidies would also need to be taken into account. We do not have any data on this.
89
4.1.4.2. Profitability of Israel Mobile Operators The UK regulator, Ofcom, cited significant differences in the profitability of different operators when reaching its conclusion that the UK retail wireless market was competitive. Similarly, when declaring Vodafone and O2 as having joint SMP, Irish regulator ComReg cited the high profitability of the two MNOs and pointed to the companies respective 39 percent and 38 percent rates of return on capital employed (ROCE). In Spain, profitability also played a key role in the decision to implement wireless access regulation, as CMT found high profitability, measured by EBITDA and ROCE, for all three MNOs, with ROCE greatly exceeding the WACC.214 The WACC is generally used to calculate the overall cost of a companys financing. It is calculated as the average of the interest rate on the companys debt and the required rate of return on the companys equity. The WACC is calculated taking into account the respective proportion of debt and equity in the companys capital structure. The WACC, therefore, is somewhere within the range between the companys cost of debt and its cost of equity. It represents the capital markets overall assessment of the rate of return that should be earned by the company to cover the time value of money and the risk premium. The WACC is predominantly used by economists and financial experts when undertaking evaluations of a projects profitability and is frequently discussed in the economic and financial literature.215 The reason for focusing on profitability is straightforwardhigh and stable profitability could be indicative of a noncompetitive market and could be an incentive for MNOs to maintain high prices and deny access to MVNOs and resellers. Return on Capital Employed In light of the pivotal role of profitability in assessing the level of competition, we have examined the profitability of the Israeli MNOs in some detail. A commonly accepted way of measuring a firms profitability is ROCE. ROCE is calculated by dividing the profit before interest payments and tax (PBIT) by the average capital employed (MCE), where capital employed includes the shareholders funds and long term liabilities (i.e., debt) and is equal to total assets (tangible fixed assets + goodwill + current assets) minus current liabilities. In the table below, we summarize ROCEs for the four Israeli MNOs from 20032006. The detailed calculations of ROCE are attached as Appendix B.
214 215
European Commission, Case ES/2005/0330: Access and Call Origination on Public Mobile Telephone Networks in Spain, Comments pursuant to Article 7(3) of Directive 2002/21/EC, January 30, 2006. For example, a leading economic finance textbook, Principles of Corporate Finance, by Richard A. Brealey and Steward C. Myers (McGraw-Hill Companies; 5th edition, July 1996), dedicates an entire chapter to explaining this use of the WACC.
90
Table 8 ROCE of Israeli MNOs 20032006
Source: Company accounts and financial statements.
As is readily seen, the profitability of the Israeli MNOs has fluctuated over the years and differs substantially between operators.216 Furthermore, Israeli ROCEs are significantly lower than the levels reached in Ireland where the regulator decided to intervene. This fact alone, however, is not sufficient to conclude that the Israeli ROCEs are not excessive.217 Moreover, it does not address the concern of the Spanish regulator CMT about ROCE exceeding the WACC. In those jurisdictions, for example the UK, where profitability measures are used in the assessment of monopoly and other competition law cases, the starting point is to compare the ROCE with the WACC. While WACC varies over time and between operators, a figure of 15 percent in pretax nominal terms would not be atypical for a mobile operator.218 Using this rate, the profitability of the Israeli MNOs over the past five years has been marginally higher than the WACC. For instance, Partners profitability averages about 13 percentage points above a 15 percent WACC. However, as we discuss in Appendix C, the fact that the ROCE is above the WACC does not by itself indicate an excessive rate of return. Most importantly, the combined rate of return of the operators is on average [] percentage points higher than a typical 15 percent WACC and, while Partners rate of return is substantially higher than the 15 percent figure, the difference between it and the ROCE is too small to warrant a finding of excessive profits given: The substantial margin of error surrounding any estimate of WACC The likelihood of there being investments in brands and in process improvements that are not reflected in the asset base as measured in the accounts The likely existence of real options in the case of 3G investment that warrant a margin between the ROCE and the WACC
216 217 218
The decrease in profitability from 2004 to 2005 might be partially due to the regulator mandate decrease in interconnection fees. By excessive we mean that profitability is above the level that could reasonably be expected in a properly functioning competitive market. In the Analysys LRIC model built for the MOC in 2004, the real pretax WACC was 13.4 percent (Analysys, Response to Issues Raised Concerning the Analysys Cost Model, 15 December 2004). If the long-run inflation rate is 1.5 to 2 percent p.a., this would be consistent with a pretax nominal WACC of around 15 percent.
91
The ROCE for an individual year or even for a period of several years may exceed the WACC even when the overall IRR of the investment in the service is not in excess of the WACC
While it is likely that the use of historic cost accounting by the MNOs means that asset values are overstated and the ROCE is understated, we would expect the impact to be small compared to the combined effect of the factors listed above.219 Moreover, the Israeli MNOs ROCE levels are significantly below the levels of a recent study that examined what constitutes excessive profits.220 Specifically, the study focused on the findings of the UK Competition Commission (CC) over the period 1973 to 2000. The UK is one of the few countries where the competition authorities have made substantial use of profitability measures in their application of competition law and where a great deal of effort goes in to trying to measure profitability. The table below summarizes the findings of this study and lists the average ROCE for which the UK Competition Commission (CC) determined that the profitability levels were excessive, as well as the average ROCE for which it concluded that profits were not excessive.
Table 9 Profitability and CC Findings
All cases All companies investigated in a given case Average ROCE found excessive Average ROCE found not excessive Only largest company investigated in a given case Average ROCE found excessive Average ROCE found not excessive 51.0% 30.5%
Monopoly pricing only 102.6% 28.3%
62.6% 39.9%
119.0% 50.4%
Source: Paul A. Grout and Anna Zalewska, Profitability Measures and Competition Law, in W.D. Collins (ed) Issues in Competition law and Policy, American Bar Association, 2006.
The all cases column covers investigations into all types of potential abuses of monopoly power including price discrimination, predatory pricing, exclusive distribution, and monopoly pricing. The monopoly pricing cases only column includes only those cases where monopoly pricing was a major concern. As the table illustrates, the average profitability level in those cases where profitability was found not be excessive was 30.5 percent. In monopoly pricing cases, the equivalent figure was 28.3 percent. Given that these are average figure, ROCEs higher than 30 percent frequently were also found not to be excessive. Similarly, in cases where only the largest
219
220
In a study for Ofcom (The Profitability and Efficiency of the UK Mobile Network Operators, NERA, August 2001) NERA found that, under reasonable assumptions about the rate of decline of asset prices, the impact on the ROCE of moving from historic cost to current cost accounting was likely to be small. Paul A. Grout and Anna Zalewska, Profitability Measures and Competition Law, in W.D. Collins (ed) Issues in Competition law and Policy, American Bar Association, 2006.
92
company was investigated, and profits were found not to be excessive, the average ROCE was 39.9 percent in all cases and 50.4 percent in monopoly pricing cases. The findings of this study have at least two implications for Israel. First, they demonstrate that profit levels far exceeding a typical (15 percent) WACC are frequently deemed not to be excessive. Second, when profits were found to be excessive, ROCE typically substantially exceeded 50 percent. None of the Israeli MNOs has exceeded this level in recent years, or even approached it. To summarize the position for aggregate profitability: the average ROCE of mobile operators in Israel is only slightly above WACC; there is substantial variation between the operators; and, even for the most profitable operator, Partner, ROCE has averaged 28 percent over the past four years, which is well below the level of profitability in cases in the UK where the Competition Commission found profitability to be excessive. Company profitability at the aggregate level is not therefore suggestive of market failure. During our meetings in Israel with various interested parties, it was brought to our attention that while the business sector might be competitive, or even highly competitive, the private sector might not be competitive. It was alleged that the lack of competition was due to the absence of MIRS from the private sector as well as possible concerns of coordinated behavior. To examine these claims, we attempted to allocate the revenue, costs and average capital employed for Cellcom, Partner, and Pelephone between their business and private sector operations. Based on these allocations, we then calculated ROCE for the business and private segments separately. Certain costs, for example general overheads, are difficult to allocate to the two segments as they are essentially common costs. Appendix D details the underlying assumptions we employed in allocating these costs for the purposes of identifying relative profitability. The results of this analysis are summarized in the table below:
Table 10 ROCE by Sector Largest Israeli MNOs 2003-2006
Business Private Total
2003 12% 23% 18%
2004 17% 30% 24%
2005 15% 23% 19%
2006 19% 33% 26%
Source: Financial data from Cellcom, Partner, and Pelephone; published company accounts and financial statements, and NERA calculations. Industry average does not incorporate MIRS.
Table 11 ROCE by Sector Partner Communications 2003-2006
Prepaid Pelephone 0.17 Partner 0.15 Cellcom 0.17
Postpaid 0.11 0.10 0.11
Business 0.09 0.07 0.09
93
Source: Financial data from Partner; published company accounts and financial statements, and NERA calculations.
Table 12 ROCE by Sector Cellcom 2003-2006
Pelephone Partner Cellcom
Prepaid 0.17 0.15 0.17
Postpaid 0.11 0.10 0.11
Business 0.09 0.07 0.09
Source: Financial data from Cellcom; published company accounts and financial statements, and NERA calculations.
Table 13 ROCE by Sector Pelephone 2003-2006
Pelephone Partner Cellcom
Prepaid 0.17 0.15 0.17
Postpaid 0.11 0.10 0.11
Business 0.09 0.07 0.09
Source: Financial data from Pelephone; published company accounts and financial statements, and NERA calculations.
We emphasize that the analysis above relies on a number of assumptions and thus carries a significant margin of error.221 Nevertheless, the sector-specific ROCE approximation for Cellcom, Partner, and Pelephone indicates higher profitability in the private sector than in the business sector. This, in turn, could indicate less competition in the private sector, relative to the business sector. However, given the margin of error associated with the sectoral profitability analysis and given the fact that even the MNOs private sector ROCE numbers are not excessive by international standards, a firm conclusion on this evidence alone cannot be drawn. Earnings before Interest, Taxes, Depreciation, and Amortization In addition to analyzing ROCE for the mobile operators in Israel, we also benchmarked their EBITDA margins against the same measure for companies in other countries. Our findings also suggest that Israeli MNOs aggregate profitability is not excessive.222 As the figure below
221
222
Specifically, we assumed that the ratio of variable to fixed cost was 80:20. While this is consistent with international experience, we were unable to confirm this ratio for Israel. Furthermore, we assumed that the peak demand occurs between 9:00 am and 8:00 pm. While this is consistent with the data provided by Pelephone, other data suggests that peak demand occurs between 10:00 am and 7:00 pm. Finally, due to a lack of data, we assumed that sales and marketing costs per subscriber are the same for business, prepaid and postpaid private customers. We note that ROCE is superior to EBITDA margins in measuring a firms profitability. This is because the EBITDA margin fails to take account of the relationship between revenue and capital employed. If the ratio
94
suggests, Israel has a low EBITDA margin by international standards, exceeding only those in Australia, Denmark, Japan, UK, Argentina, Finland, Brazil, and Hong Kong.
Figure 33 EBITDA as a Percentage of Revenue 4Q06
80 70 EBITDA as % of Revenue 60 50 40 30 20 10 0
Philippines Indonesia Algeria Morocco Iraq Nigeria Taiwan Russia Egypt Malasia Czech Mexico Singapore Canada Turke Bangladesh New Zealand Belgium Norway Italy China Pakistan India Chile Switzerland Sweden Greece Portugal Colombia Austria Spain Thailand France Hungary Ukraine Korea Germany Poland Ireland Peru Venezuela US Netherlands South Africa Israel Australia Denmark Japan UK Argentina Finland Brazil Hong Kong
Source: Derived from data in Merrill Lynch, Global Wireless Matrix Q4 2006, Table 32.
As shown in the figure below, more concentrated markets tend to have better margins, and Israels EBITDA margin falls below the trend line. In fact, it has the second lowest margin for countries with comparable HHIs.
of revenue to capital employed is high, a low EBITDA margin could still be consistent with a high ROCE. Hence, where EBITDA and ROCE analyses yield inconsistent results, ROCE analyses tend to be more accurate.
95
Figure 34 EBITDA Margin vs. HHI 2006
Source: Companies financial reports, data request responses, Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc, Merrill Lynch, Global Wireless Matrix 4Q06
A similar picture emerges when benchmarking Israels EBIDTA per minute of voice traffic to the levels found in other countries.223 Countries with higher GDP per capita tend to have higher profits per minute. The relationship is shown below. It can be seen that Israels profits per minute appear to be among the lowest compared to countries with similar GDP per capita.
223
EBIDTA per minute of voice traffic is calculated by subtracting data revenues from service revenues, multiplying the result by the EBIDTA margin and dividing the product by the total voice minutes. Formulaically: EBITDA per minute = [(Service Revenue Data Revenue)*(EBITDA/Revenue)] / (MOU*12*Number of Subscribers).
96
Figure 35 Profit per Minute vs. GDP 4Q06
Source: Companies financial reports, data request responses, Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc, Merrill Lynch, Global Wireless Matrix 4Q06.
Benchmarking Israels EBIDTA margin from 20022006 against those in different regions reveals that, with the exception of Latin America, EBIDTA margins have remained relatively stable. Israels margin was the lowest in 2002 and is the third lowest today (after Latin America and Asia Pacific).224
224
We note that the implications of this finding are not fully clear absence an analysis of the factors that determine the level of depreciation the various countries.
97
Figure 36 EBITDA Margins by Region 20022006
50.00% EBITDA Margins US$ 45.00% 40.00% 35.00% 30.00% 25.00% 20.00% 2002 2003 2004 Europe Emerging Europe Israel 2005 2006
Asia Pacific Emerging Asia Weighted Average

Source: Merrill Lynch, Global Wireless Matrix 4Q06.
North America Latin America
Given the relatively high cell phone usage in Israel, as measured by MOU, profit per minute may give an unduly favorable impression of the performance of Israeli MNOs. We therefore also calculated the voice profit per user in different countries and looked at the relationship with GDP per capita. This is summarized in the figure below:
98
Figure 37 EBITDA per User vs. GDP 4Q06
Source: Companies financial reports, data request responses, Globalcomms 3.0, Telegeography Research Product, PriMetrica, Inc, Merrill Lynch, Global Wireless Matrix 4Q06.
On a per subscriber basis, Israels MNO appear to make a higher profit than the trend line. However, they still appear no more profitable than countries with similar GDP per capita. Overall, therefore, as in the case of ROCE, we did not find evidence that supports the conclusion that the operators in Israel are making an excessive overall rate of return. The above figures do not of course address the situation in different sectors. 4.1.5. Barriers to entry and expansion Market entry barriers refer to asymmetries between incumbent firms and new entrants that make it more costly or harder for the latter to compete. A frequently cited entry barrier to the wireless market is the lack of spectrum. For instance, as described above, one of the reasons OFTA required wholesale 3G services to be offered separately from retail 3G services in Hong Kong was the limited availability of spectrum. OFTA found that the lack of spectrum was a barrier to entry for potential competitors. Hence, it implemented regulatory measures to reduce potential market entry barriers with the hope that increased entry would lead to increased competition. Barriers to expansion refer to situations where a company has already entered the market but faces obstacles in expanding its market share. The costs of customers switching providers (switching costs) is a frequently cited barrier to expansion and refers to situations where subscribers are tied, financially or otherwise, to their existing provider and cannot take advantage of special service offerings or price promotions by competitive service providers.
99
While the existence of entry and expansion barriers does not necessarily preclude competition, if the barriers are low competition is enhanced. This, in turn, can benefit consumers as the incumbent operator(s) cannot sustain supra-competitive prices in the long run when competitors can easily enter the market. On the other hand, if entry or expansion barriers are high, firms with a large market share are more likely to be able to maintain their share over time and possibly sustain prices above competitive levels. As we discuss below, however, high entry and/or expansion barriers alone do not guarantee that incumbent players can charge supra-competitive prices. Rather, other factors, such as the number of incumbent players, countervailing buying power, and regulatory barriers contribute to the level of competition in a given market. Below, we discuss a variety of entry barriers and assess their level in Israel. Spectrum availability Wireless spectrum, in particular spectrum for 2G and 3G services, is a scarce resource. In most countries, the number of parties seeking a license for spectrum far exceeds the number of operators that can be supported by the available blocks of spectrum. Governments around the world have dealt with this issue using different allocation methods. Some countries resorted to auctions, allocating the available spectrum blocks to the highest bidders. For instance, Germanys 3G auction raised approximately US$46 billion, while the UK regulator sold its five 3G licenses at approximately US$36 billion.225 Other countries allocated their spectrum blocks using the beauty contest method where only companies that met the regulators definition of beauty (typically an appropriate business plan, attractive service offerings, financial soundness, etc.) received a spectrum license. Some countries sold their spectrum to interested parties without any particular allocation method, and others used some sort of hybrid version. Regardless of the allocation method, companies that did not obtain a spectrum license face a significant barrier to entry. That is, without a spectrum license, they cannot enter as a facilities-based competitor. Ofcom, in its competitive assessment of the UK wireless market, found that there were significant barriers to entry at the wholesale level due to limited spectrum and high sunk costs. In France, ART concluded that competitive entry was unlikely due to the lack of spectrum, and in Denmark the regulator concluded that such a barrier did not exist as there was one additional license available for a new entrant. The MOC have informed spectrum is available for one additional MNO (to provide both 2G and 3G services).. Hence, we conclude that the availability of spectrum does not pose a barrier to entry in Israel. Regulatory approval Many regulators require mobile carriers to be licensed as wireless service providers. Depending on the requirements and process specified by the regulator, obtaining regulatory approval can be lengthy and expensive. Some governments, such as Spain, have attempted to remedy this potential entry barrier by creating special licenses for MVNOs (the A2 license) and then ultimately entirely waiving such a requirement.
225
Sag Harbor Group, Reflex Rules, The Role of Pervasive Low-Cost Networks and Devices in the Future of Mobile Data Messaging, 2001, p. 42.
100
We have been informed that the MOCs licensing process for MNO entry has historically differed by MNO. We understand that Pelephone obtained its right to operate through its affiliation with Israels wireline incumbent, Bezeq. Cellcom and Partner entered through a spectrum auction, while MIRS evolved from a trunking provider to an MNO and did not participate in a spectrum auction. We understand that the licensing fees differed significantly among MNOs. An overall conclusion on whether regulatory approval presents a barrier to entry cannot readily be reached as the process seems to have varied over time. If future licenses require significant upfront payments, such as the license for Partner, then regulatory approval represents a significant entry barrier. On the other hand, if MOC were to license new MNOs with lower or no fees, then this barrier would be reduced significantly. Another source of potential regulatory barriers stems from local government jurisdiction over rights-of-way and the location of cellular antennas and towers. Given the increasing health concern associated with wireless infrastructure, obtaining site and right-of-way approvals from local governments can represent a significant barrier of entry. We understand that local governments in Israel are reluctant to grant permits for the construction of new cell sites. Given that this is a widespread phenomenon, it is a significant barrier to entry for new Israeli MNOs. Access to capital markets Access to capital markets and financial resources can present a barrier to entry and expansion, particularly for new MNOs that cannot secure sufficient funding. NERA has no information about the potential entry and expansion barriers of MNOs on this matter, as no facilitiesbased entrant is available to provide such information. Economies of scale and scope A common problem faced by new MNOs, particularly in highly penetrated or saturated markets where it is not possible to pick up a share of an expanding market, is the initial lack of scale that puts them at a cost disadvantage relative to the incumbent MNOs. The problem is especially pronounced for MNOs who have to incur significant sunk costs upfront and cannot enjoy any economies of scale in the short run. For instance, per Partners financial records, it invested U$100 million in licensing fees and $300 million to build its 3G network, and this was after spending US$1.1 billion on its 2G network. A new entrant might focus more heavily, or even exclusively, on a 3G or WiMAX network, thus incurring different costs than Partner. Nevertheless, in Israels highly saturated wireless market, it will likely take years to recoup the investment in a new network. The lack of economies of scale creates a barrier to entry or expansion as it might prevent firms from entering the wireless market or at least hinder their development. This, however, is not unique to Israel and applies to most countries. Economies of scope are a slightly different matter because companies, such as HOT Cable, could attain economies of scope through the offering of quadruple play (TV, telephony, and Internet services, as at present, plus mobile). For example, if HOT were to enter the market as an MNO it could sell its wireless services at or near cost and make its profits on the overall quadruple play bundle. Consequently, while for some firms, the necessary economies are not present to justify entry, for others the synergy with other business operations could make it worthwhile.
101
Switching barriers If customers are reluctant or unable to switch wireless providers due to certain characteristics of the incumbent MNOs offerings, this would make it harder for competing carriers to enter the market or expand their market share. Consequently, switching barriers (costs) can be a competitive constraint. A switching barrier is particularly troublesome for competitors if the customers that are less likely to switch are those customers who represent low default risk. Switching costs are prevalent throughout most economies (e.g., health clubs, airline and hotel loyalty programs, credit cards, etc.) and are typically not subject to regulation. Accordingly, there is an abundance of economic literature that addresses whether the resulting lock-ins represent a market failure that requires government intervention.226 During the US regulators proceeding on MNP in 2002, switching costs and their relation to market failure was a central issue in the discussion. In this proceeding, economist Dr. Hal R. Varian stated: Switching costs in a competitive market do not inevitably harm customers. In fact, the higher the switching cost, the greater the inducement a competitor will likely offer to acquire a new customer. For example, in the wireless context, such inducements could take the form of free or subsidized handsets. If switching costs are eliminated, competitors will have less incentive to offer up-front sweeteners to acquire customers.227 Dr. Varian based his findings on the economic argument that lock-in effects can be managed by consumers prior to being locked in because they predict and extract upfront payments (sweeteners) whose value equals the switching costs. We agree with Dr. Varians statement that the mere existence of a switching barrier does not mean that there is market failure. As long as the market is competitive, a switching barrier follows the classic bargain then rip-off cycle. Thus, in determining if switching barriers are present in the Israeli wireless market, it is important to recognize that if they exist, they should be evaluated in the context of the overall level of competition. In wireless telephony, switching barriers can take various forms, including: Term contracts SIM/handset locking Lack of mobile number portability Opaque pricing plans Product bundles
We address each of these potential barriers below. Term contracts require customers to remain with a particular carrier for a certain period of time, typically in exchange for a substantial handset subsidy and/or lower prices. We
226 227
See, generally, Carl Shapiro and Hal R. Varian, Information Rules: a Strategic Guide to the Network Economy (Harvard Business School Press: 1998). Letter from Hal R. Varian to Magalie Roman Salas, January 25, 2002, FCC Docket WT 01-184, p. 2.
102
understand that in Israel MNOs offer subsidized handsets to customers willing to sign a 36month contract, subject to an early termination fee (ETF) that can amount to the sum of remaining monthly recurring charges or higher. Several parties, particularly the Consumer Trust and the Israel Consumer Council, have argued that term agreements and ETFs represent high switching barriers and thus limit the level of wireless retail competition in Israel. Specifically, the Israel Consumer Council stated that 36-month tariff contracts tie consumers to a particular tariff plan, particularly as they involve an ETF of up to NIS 1,500. Information provided by [] indicates that, until November 2005, all private postpaid customers were on pricing plans with no term obligation or ETF. After that date and the introduction of MNP in April 2005, 36-month contracts were offered, which required a guaranteed minimum number of minutes of use in exchange for a large (sometimes full) handset subsidy. The ETFs for these contracts were prorated over the duration of the contract and amounted to NIS 700, NIS 500, and NIS 300, depending on whether a consumer terminated the contract in the first, second, or third year. [] pointed out that these contracts allow customers to switch among its available price plans at no additional cost. []s data indicate that approximately [] percent of its postpaid residential subscribers are now on a 36-month contract.228 For non-corporate business customers, 36-month contracts have long been the norm for [] and [] percent of non-corporate business customers are on such contracts. Payments for early termination are NIS 1400 per contract in the first year of the contract, NIS 1100 in the second year, and NIS 800 in the final year. If the customer wishes to stay with [] but change his price plan before the end of the 36-month original contract period, the termination fees are NIS 170, NIS 150, and NIS 140, respectively, in the first, second, and third years of the original contract. All contracts offered to corporate customers (defined as customers with 100 or more handsets) are for 36 months. The fees for early termination (per handset) during the customers first contract period are the same as those for non-corporate business customers. It is possible for corporate customers to renew their existing contracts before the end of the 36month period without incurring an early termination fee. For contracts after the initial contract, the early termination fees per handset are lower at NIS 170, NIS 150, and NIS 140, respectively, in the first, second, and third years of the additional contract. Similarly, data provided by [] reveals that, at the end of 2006, [] percent of its private customers and [] percent of its business customers were on 36-month contracts. [] states that [] percent of its private customers and [] percent of its business customers are under 36-month contracts. [] numbers are [] percent and [] percent for private and business customers, respectively. Based on the data provided by the MNOs, a variety of aspects of term contracts can be identified that are important to the examination of switching barriers. First, the large majority of private customers are not yet under a term contract which means that they can still readily switch providers without incurring an ETF and do not face term switching barriers. However, given that these 36 months term contracts have only been introduced recently, this number is likely to decrease in the near future.
228
[...]
103
Second, one operator, MIRS, offers 36-month term contracts with handset subsidies alongside plans that do not require any term commitments and no handset subsidy. Third, 36-month contracts are currently mostly observed in the business segment (noncorporate business customers and corporate business customers). The various parties we interviewed in Israel seem to agree that the business segment is competitive, indicating that these might be competitive responses rather than signs or sources of market failure. Finally, the percentages of [] customers under 36-month contracts suggests that, at least with respect to business customer contracts and probably also residential contracts, its behavior has differed in the past from that of [] , indicating no apparent parallel behavior. Notwithstanding this, various concerns remain regarding the switching barriers associated with term contracts in Israel: First, 36-month contracts have been all but phased out in Europe and the US. Most term contracts are 12, 18, or less frequently 24 months. A problem with term contracts of 36 months is that handsets typically have a shelf life of approximately two years and consumers on average change their handsets every 12-24 months. Thus, if a consumer is locked into a 36 months contract, then he/she will have to renew the contract (adding up to three years to the current tenure) simply to update his/her handset. Second, we understand that MNOs can raise the price throughout the length of the term. Unless these price increases are clearly indicated in the contract and presented to the consumer prior to committing, such changes appear not to be consistent with the behavior of a competitive market. For instance, in the US, any changes in price, terms, and conditions of the contract, regardless of size, allow subscribers to terminate their contract at anytime without incurring an ETF. Third, ETFs should serve as a liquidated-damages clause, meaning that they are not a penalty but make the company financially whole in case of a breach of contract. We do not have the necessary data to conduct an analysis of whether the size of the ETF properly corresponds to the economic damage caused by a contract breach, but believe it would be a worthwhile exercise for this to be verified in order to ensure the maximum level of competition.
At a minimum, consumers should be clearly informed of the terms, conditions, and pricing when they sign up for a contract. Only then, can a consumer properly trade off sweeteners against switching costs. The view of the consumer advocacy groups in Israel was that typically consumers were not aware of what they were signing up for and the contracts were long, full of small print, and complicated. Sim/handset locking is the name given to software locking that prevents a handset from being ported to another providers network or, in the case of GSM, prevents a handset from accepting SIM cards from another provider. Thus, similar to term contracts, several parties have argued that handset locking presents a switching barrier to consumers. Since currently only Cellcom and Partner share the same network technology (GSM), this issue would only appear to be relevant for switching between these two providers.229 We understand that Pelephone is in the process of migrating its 3G network from CDMA based technology to UMTS. Once this has happened, the issue of handset locking will also apply to Pelephone.
229
It is not possible to used CDMA handsets on GSM networks or vice versa.
104
[]. The SIM can be unlocked at any time at a Cellcom service center for a fee of NIS 73. Cellcoms practice is less restrictive than the practices followed in the US and in the UK two wireless markets that have been declared effectively competitive by their regulators (and hence free of market failure). In the US, all GSM and most CDMA handsets are locked, and, while CDMA handsets remain unlocked for product integrity reasons, GSM handsets are unlocked if a consumer has remained with the carrier for a certain period of time (typically at least three months) and if his/her account is in good standing. In the UK and much of the EU, GSM handsets are also locked. In 2002, the UK regulator Oftel (now Ofcom) concluded that SIM locking was a barrier to consumers switching suppliers, and that the removal of locking, or easing of its terms, would benefit competition in an number of respects, including price, consumer choice and entry barriers.230 Despite this finding, the UK regulator found that it is inappropriate to use more formal regulatory power to reduce SIM-locking or change the terms for unlocking, certainly in the absence of a clearer understanding of how far SIMlocking matters to consumers.231 Instead, it stressed that consumers are best served by improved awareness. Consequently, it allowed SIM locking. Like Israeli MNOs, UK MNOs charge a fee, which varies by provider, to unlock a handset. The UK practices are more restrictive than in Israel as they require customers to subscribe to their services for a certain number of months (not exceeding one year) before offering to unlock a handset. The fees and minimum lock-in period are summarized in the table below.
Table 12 UK SIM Locking Policies November 2002
MNO Vodafone O2 T-Mobile Orange Virgin Mobile Lock-in period Unlock at any time, but only pay the fee if unlock within first year 12 months Pre-pay 3 months, contract 1 month Unlock at any time Unlock once customer has spent 30 on Virgin Mobile airtime Fee 19.99 15 35.25* 20 Free Whats locked? Pre-pay & a minority of contract phones Pre-pay & a minority of contract phones Pre-pay & contract phones Pre-pay &contract phones Pre-pay & pay monthly phones
Source: Company websites * T-Mobile reduced its unlocking fees to 15 in early 2003.
Furthermore, contract customers typically must also pay the outstanding subscriptions for the remaining period of the contract before the handset is unlocked. Danish regulator NITA ruled that the private consumer is entitled to have his telephone unlocked free of charge after six months, so that the telephone may be used under another company.232 We are not aware of other EC countries where SIM/handset locking was an issue of dispute. However, in 1996, the EC issued a set of guidelines that loosely govern the practice of SIM locking. Specifically, the EC guidelines state:
230 231 232
Oftel, Review of SIM-Locking Policy, 26 November 2002, 1.4 Ibid. National IT and Telecom Agency, Annual Report 2004, Chapter 4.
105
End users should be made aware at the time of purchase whether their handsets are locked. Network operators/service providers should tell end users that unlocking is possible or provide upon request the information necessary to unlock the handset. The existence and amount of the handset subsidy (and any conditions for repaying monies due under the contract) should be made clear to customers at the time of purchase. Handsets need not to be unlocked until the subsidy has been repaid.233
As stated above, Cellcom unlocks its handsets at anytime for a fee of NIS 73. We understand that previously this fee was as high as NIS 148. Similarly, Partner charges an unlocking fee of NIS 90 (previously this fee was195 NIS) in addition to any outstanding rate plan or handset subsidy payments. We understand that Cellcom and Partner offer handset subsidies to their customers.234 If this is the case, unlocking fee and/or a minimum lock-in period may be economically justifiable provided that the unlocking fee is not excessive in relation to the costs involved. Handset locking generally benefits consumers as it enables them to purchase their handsets at a reduced price. Moreover, even though unlocked GSM handsets typically work on other GSM networks for voice services, lack of firmware or specific phone characteristics limits the use of ported handsets. Given Cellcoms and Partners willingness to unlock and the fact that their practices (including the size of unlocking fees relative to GDP per head) are generally in line with the practices followed in the UK, which have been accepted by Ofcom, it would be difficult to conclude that handset locking in Israel was anticompetitive or a source of market failure. However, as with term contracts and consistent with the ECs guidelines, we recommend that wireless subscribers be clearly informed about the carriers locking policies. Absence of Mobile Number Portability is a frequently cited entry or expansion barrier as it requires a subscriber to change his/her telephone number when changing providers. MNP has been implemented in most developed nations, including the US and the UK. We understand that MNP in Israel is expected to be implemented during the second part of 2007.235 With a strong MNP program being implemented in Israel by the end of the year, this switching barrier is soon to be removed. We do caution though that MNP has had mixed success in terms of its impact on customer churn. For instance, in Ireland and France, MNP had little impact on churn, while in the US, Hong Kong and Denmark it proved to be rather successful. Opaque pricing plans often are viewed as barriers to expansion in that they prevent subscribers from comparing product and service offerings among competing carriers. During our interviews in Israel, various parties argued that opaque pricing plans offered by the MNOs were a deterrent to competition. This is because, if customers are not able to compare the offerings of different operators, they will tend to stay put rather than risk making a choice that leaves them less well off.
233 234 235
Oftel, Review of SIM-Locking Policy, 26 November 2002, 1.6, citing the European Commission. We understand these subsidies to be discounts off the market prices of the phone, not necessarily indicating that the MNOs sell the handsets at a loss. We are informed that MNP was to be implemented by September 1, 2006. However, MNOs in Israel, along with other telecommunications operators, have not yet complied with the law (section 5A (e)(1) of the Israeli Communications Law, Telecom and Broadband, 5742-1992) obliges them to do so.
106
It is important to note that the existence of a plethora of differing pricing plans could also be the result of competition in which the MNOs are attempting to differentiate their products and services from those of their competitors. Moreover, if in fact the pricing plans were intentionally confusing, then in the absence of some type of collusion competing firms could use this to their advantage and offer competitive products with clearly defined pricing plans. Also of particular concern to various parties was the introduction of billing plans that offered 1 minute billing increments. While other plans are available that offer billing increments of 12 seconds, questions arose about whether and how customers were informed about the existence of alternative plans. In particular, customers need to be equally aware of the different alternatives in order to be in a position to decide which option would be more beneficial to them. While a thorough investigation of this issue is outside the scope of this report, clear pricing plans, full disclosure and comparable promotion of alternative options are important to consumers and to competition and their absence can create a barrier to switching. Product bundles can act as a barrier to entry and expansion in that they tie consumers to a range of products and services, making switching between different operators more costly. For instance, for a consumer of a quadruple play package to react to the price change of a competing carriers wireless plan, he has to port his wireline and wireless numbers, change his email address, possibly reconfigure his Internet and browser settings, and change his TV subscriptionassuming that the new provider offers a quadruple play. If it does not, the consumer would then have to search for a synthetic package and sign up for and source the previously bundled components individually from different operators. Similar to term contracts, however, this type of barrier does not necessarily harm the consumer as quadruple play bundles are often associated with a volume discount. Hence, per Varians argument, the consumer can predict these upfront discounts (sweeteners) and offset them against the increased switching costs. Churn Rates Churn rates measure customer turnover and are typically expressed as the rate of disconnections per month. For instance, a 3 percent monthly churn rate implies that 3 percent of the average customer base for the specific month disconnected its services. Churn rates are most useful as business metrics, providing individual companies with information on the impact of price changes (price elasticity) and quality of service changes, the intensity of competition, and the average lifespan of a customer. High average churn rates in a market are sometimes also associated with strong competition, while low average churn rates usually indicate less competition. From an economic viewpoint, the correlation between churn rates and competition is not necessarily clear. For instance, high average churn rates for the wireless market might be caused by one carriers inferior quality of service and thus might not be a testament to the level of competition. At best, churn rates indicate the level of competition on the basis that, if there is more intense competition, it is more likely, other things being equal, that consumers will switch networks. This treatment of churn rates is consistent with the view adopted by UK regulator Ofcom, who referred to the high level of switching between providers or tariff plans as a positive indicator of competition.236
236
Office of Communications, Effective Competition Review: Mobile, Statement issued by the Director General of Telecommunications, September 26, 2001, 2.28.
107
According to data provided by Merrill Lynch, the average monthly wireless churn rate in Israel is 1.4 percent. As illustrated in the figure below, this rate has increased in the last few years and currently falls between the average rates observed in the Asia-Pacific region and Europe. However, it remains low compared with most countries.
Figure 38 Average Monthly Churn Rates by Region 20022006
4.00% Monthly Average Churn Rate 3.50% 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00% 2002 2003 2004 2005 2006 Asia Pacific Europe North America Emerging Asia Emerging Europe Latin America Weighted Average Israel
Comparing overall average churn rates, however, might be misleading for a variety of reasons. First, churn rates can be expected to be higher for prepaid mobile subscribers as there is no contract period. However, as illustrated in the figure below, data for EU countries indicate that while there is a positive correlation between the overall churn and the share of prepaid subscribers, it is not statistically significant.237
237
This remains true if the churn rate (for EU countries) is regressed against the percentage of subscribers that are prepaid and the market share of the two largest firms.
108
Figure 39 Average Churn Rates vs. Share of Prepaid Subscribers

Churn Rate v % Pre-Paid
3.0%
2.5%
2.0% Churn Rate
1.5%
1.0%
0.5%
0.0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% % of Subscribers that are Pre-Paid
Source: NERA analysis based on data from Merrill Lynch, Global Wireless Matrix Q4 2006, Table 22 and EC 12th Annual Implementation report, 29 March 2007.
Based on information received from the MNOs, the churn rate for prepaid subscribers in Israel is much higher than for postpaid or business subscribers. At the same time, the proportion of subscribers that are prepaid is around half of that in EU countries. Specifically, on average, 60 percent of mobile subscribers in the EU use prepaid services. If that were the case in Israel (and the churn rates for different types of customer were unchanged), then based on the disaggregated churn rate data provided by [], the overall churn rate for Israel would range between 1.6 percent and 1.7 percent. We explain this calculation in Appendix E. Hence, this analysis suggests that a substantial proportion of the difference between the overall churn rates in Israel and the EU is explained by the significantly lower proportion of prepaid subscribers in Israel. Second, the presence or absence of MNP can further influence churn rates. All EU countries have implemented MNP, whereas Israel has not, although we understand that MNP will be implemented by the end of this year. While the effect MNP has on Israeli mobile subscribers remains to be seen, the current lack of MNP might explain much, if not all, of the remaining gap between churn rates in Israel and the EU. In light of the considerations above, we find that low churn rates in Israel do not in themselves indicate the existence of market failure. Once account is taken of factors such as the low incidence of prepaid subscribers in Israel, there is only a small remaining difference between churn rates in Israel and the EU, many of whose countries have been found to have effectively competitive mobile markets. At the same time, one cannot say that churn rates similar to those in Europe would necessarily rule out the possibility of market failure. In short, churn rates by themselves do not tell us a great deal.
109
4.1.6. Historical behavior In analyzing the level of competition in the Irish wireless sector, ComReg specifically focused on historical conduct by one or more MNOs that was suggestive of the existence of market power. During our interviews in Israel, the view was frequently expressed that the MNOs had exhibited coordinated behavior in various aspects of their operations. Particularly, several parties (i.e., the consumer protection agencies, the antitrust commission, and the MOC) were concerned about 1) the MNOs simultaneous price increases, 2) intentionally confusing pricing plans (fogging practices), and 3) stable market shares among MNOs. In addition, responding to one of NERAs data requests, MOC added a number of additional concerns, including: The MNOs all introduced new pricing plans with one minute billing increments for the first minute and 12 or 60 seconds increments thereafter. While other plans with lower billing increments were offered, MOCs investigation revealed that consumers were routinely not made properly aware of their availability and thus were not able to make an informed choice as to whether a move to one minute charging would benefit them. In some cases, consumers were allegedly informed that no other options (other than the plan with the one minute initial billing increment) were available. In a competitive market such behavior would not be sustainable over long run, as other MNOs would promote their 12 second plans and inform customers if the one minute plans were inferior choices. Apparently, no such market correction took place. MNOs have introduced 36 month term contracts in the residential sector and continue to promote them heavily. Previously, the residential customers typically had no term contract. As indicated above, the lack of competitive alternatives (i.e., contracts shorter than 36 months) raises concerns regarding the competitiveness of the Israeli mobile market. MNOs apparently reserve the right to raise tariffs throughout the term period, allegedly without properly notifying customers at the time of signing. This, allegedly, is counter to the MNOs licensing terms. Again, MOC finds that in a competitive market such issues would be corrected. In January 2007, MOC implemented a rule which requires MNOs to inform the customer if his call is answered by voice mail prior to charging the customer for the call. In response, the MNOs allegedly all increased their rates to recoup the revenue lost due to this required announcement. MOC finds that the simultaneous rate increase is inconsistent with the behavior of a competitive marketplace.
Investigating each of these claims is beyond the scope of this report as it does not directly relate to MVNO entry. However, as we have discussed above, our analysis of some of these concerns (mainly the excessively long term contracts and the MNOs ability to raise prices throughout the term period) does raise some concerns about the way the market is currently operating. 4.1.7. Likelihood of collusion Regulators frequently examine the likelihood of tacit collusion by examining market conditions that are conducive to such behavior. In cases where there is a high likelihood of tacit collusion, regulators might intervene. Conversely if market conditions make collusion
110
unlikely, then regulators would tend to refrain from intervention. For instance, even though it did not find any evidence of single or collective SMP, the Spanish regulator, CMT, found conditions in the wholesale market to be conducive to sustainable tacit collusion. Specifically, CMT found that there were strong incentives for the three MNOs to coordinate their pricing actions. High and stable profitability, in turn, would reinforce the MNOs incentive to maintain high prices and deny access to MVNOs and resellers. Consequently, CMT opted to intervene in the wireless wholesale market. In this section, therefore, we examine the likelihood of collusion in the Israeli retail wireless market. For tacit collusion to be feasible, there are two conditions that need to hold. That is, MNOs have to (1) have the incentive to coordinate, and (2) have the ability to coordinate. We will discuss each of these necessary conditions for tacit collusion below. Incentive to coordinate As noted by ComReg, the incentive to coordinate is largely based on assessing the merits of going it alone versus aligning behavior. Incentives for coordination are likely to be stronger when there are: (1) few firms in an oligopoly, and (2) the key competitiveness indicators of those firms are relatively similar. Few firms in an oligopoly: There are four competitors in the Israel mobile business market and effectively three in the private market, with MIRS having the opportunity to develop business in the private market. While US authorities typically consider two or more in a market as sufficient, many European nations have considered four MNOs as few. While we do not agree or disagree, we conservatively opt to consider the four MNOs in Israel as few. Thus, at least in this respect, the market structure in Israel is conducive to coordination. Similarity of key competitiveness indicators of the firms: in the sections above, we have shown that the MNOs differ in a number of important respects, including market shares, pricing (RPM), pricing plans, technology, profitability (ROCE), costs, and target markets. We summarize the most important of these factors below: Similarity in market shares: Although there has been significant variation in market shares over time, currently the market shares for the three largest MNOs are relatively similar. In markets where the largest operators have similar market shares, collusive behavior is more likely to be sustainable since no company has more to gain and less to lose from breaking ranks. Hence, the similar market shares tend to contribute to collusive behavior.
Similarity of costs: Where operators have differences in costs or cost structures, collusive behavior is less likely. One reason is that a high cost firm is in a less strong position to use price cuts as a means to discipline non-collusive behavior by a low cost firm. This, in turn, means that collusion is unlikely to be enforceable. [] . Providing access to an MVNO would be one way of doing this. Consequently, it does not seem that costs are sufficiently similar to be conducive to coordinated behavior. Similarity of capacity constraints: If there are differences in the degree of spare capacity, this might be expected to reduce the likelihood of collusive behavior. This is because a firm with spare capacity would stand to gain more from attracting business away from its rivals than a firm with no spare capacity. [], although MOC has indicated that they might possess spare spectrum within their spectrum allocations. [] While it is difficult to reach a
111
categorical conclusion, it appears that if an MVNO were to approach it with a commercially viable proposition, it is likely that it would be in a position to increase the volume of traffic over its network at relatively low cost. Similarity in ownership structure: The four MNOs have different ownership structures, which in turn affects their respective abilities to offer triple and quadruple play product offerings. For instance, Cellcom owns both wireless and fixed line licenses, and because of its sister companies (NetVision, Barak and Globcall) strong presence in the international long-distance market and the Internet segment, Cellcom has a competitive advantage in triple and quadruple plays. Pelephone, on the other hand, is wholly owned by Bezeq. While this could also serve as an advantage over other MNOs, we understand that current regulations restrict Pelephone from offering a quadruple play. Partner also recently acquired a fixed-line license. The different ownership structures and opportunities to offer triple and quadruple plays tend to reduce the incentives for collusion.
Accordingly, we believe that, based on the data we examined and the information we have received, there are factors in Israel that are likely to restrict the incentives that the four Israeli MNOs have to coordinate behavior in the wireless retail market. Ability to coordinate In order to be able to coordinate, a number of basic conditions must be met, including (1) the ability to detect cheating, (2) the enforceability of compliance, and (3) actual and/or potential market constraints. Even if there was an incentive to align behavior, the MNOs would need to be able to detect cheating. If cheating cannot be detected, then the ability (and incentive) to coordinate is limited. For cheating to be detectable requires a high level of transparency in each MNOs product and service offerings, as well as their terms and conditions. Transparency can facilitate reaching a consensus position and ensures that deviations from this position can be detected. Fundamentally, the products and services offered by the MNOs are publicly available, and the terms and conditions are well known. Moreover, the mobile market is characterized by continuous and repeated interactions, and, if an MNO were to change prices or offer access to an MVNO, this would be immediately known to its competitors, prompting retaliation. Thus, MNOs appear to be able to detect cheating. Some doubt, however, remains as many parties have informed us that a comparison of price plans is almost impossible as the plans are structured differently, contain different variables, and are rather confusing. Nevertheless, we believe that, given the MNOs sophistication, they would be able to detect cheating. Furthermore, in order to reach a consensus position, there needs to be a simple way to coordinate. In Ireland, ComReg found that the two MNOs had a common focal point with two dimensions. One dimension was price and the other was denying access to independent entities to upstream elements, such as in the case of wholesale airtime. We agree with ComReg that price and wholesale access are two important dimensions of a possible focal point. In addition terms and conditions offer a third dimension. For the MNOs to coordinate on price, their products must be generally homogenous. Given the fact that the products and services offered by the MNOs are substitutes (i.e., in the same 112
product market), it is not surprising to find that they broadly offer the same portfolio. That is, they all offer voice, SMS, email, and Internet access, in addition to some attachable features, such as ringtones, news, and so on. However, there are also some fundamental differences between the Israeli MNOs. For instance: MIRS mainly focuses on the business sector, while all others serve both the business and the private sector. MIRS offers push-to-talk, while none of the other MNOs offer this feature. Partner and Cellcom are based on the GSM and W-CDMA technologies, MIRS uses iDen and Pelephone uses CDMA and CDMA2000. The MNOs differ in their provision of 3G services and offer different features, such as traffic cameras, music download, and access to news channels.
Furthermore, while similar overall, the MNOs service offerings do differ in terms of the monthly recurring costs and/or number of free minutes included in the price plans. As regards, terms and conditions, they appear to be largely similar across MNOs. Most important, as we have pointed out above, the existence of 36-month contracts is longer than what we observe in the US and the EU. However, there are differences in the cost of unlocking handsets and ETFs. With respect to denying access to independent entities (e.g., MVNOs) to upstream elements, such coordination would be simple and transparent. It is simple as there are only two possible resultsan MNO either permits or denies wholesale access. It is transparent because cheating (i.e., allowing wholesale access) is easily detectable, as an MVNO would enter the market. We note that this situation is not unique to Israel and is likely to apply to all wireless markets. As we elaborate further below, however, it is important to note that none of the Israeli MNOs has formally been approached to offer wholesale access. Therefore, there does not exist any historic behavior that would indicate that such a focal point has been tacitly or otherwise agreed. Thus, although Israeli MNOs offer generally the same portfolio, we do not believe that coordination would be easy, as the Israeli MNOs differ in some key competitive respects. The enforceability of compliance is a necessary condition for coordinated behavior as, without it, the members of the oligopoly would deviate from the focal point and the coordinated behavior could not be sustained in the long run. Each member must be aware that cheating (in order to gain market share) would result in identical action by the other members, thus terminating the oligopoly. As noted in the case law and cited by ComReg: [t]he Commission must not necessarily prove that there is a specific retaliation mechanism involving a degree of severity, but it must none the less establish that deterrents exist, which are such that it is not worth the while
113
of any member of the dominant oligopoly to depart from the common course of conduct to the detriment of the other oligopolists.238 The economic literature diverges on what constitutes a sufficient deterrent mechanism. Some economists find that the mere breakdown of collusion is sufficient to deter cheating. If so, then the condition of enforceability of compliance is presentin the Israel wireless market and likely many other markets. As noted by ComReg, actual and/or potential market constraints can discipline prices and exert significant competitive pressure on the members of the oligopoly.239 Market constraints can come in many forms. First, a smaller maverick MNO can upset collusion by undercutting other MNOs through lower prices or increased quality of service. While we cannot say with certainty whether such a maverick exists in Israel, we find that MIRS could potentially undercut prices in the residential sector, if coordination existed in that sector. MIRS compares to Nextel (the US MNO that uses iDen technology and is now part of Sprint). Nextels history mirrors MIRSs evolution, with the company commencing operation exclusively as a business MNO. Nextel, however, reached one of the first MVNO agreements in the US, forming Boost Mobile. Boost Mobile (now the third largest MVNO in the US) allowed Nextel to enter the residential market. Similarly, we are of the view that, if coordination on MVNO entry or price existed in the retail wireless market in Israel, MIRS would have the ability to destabilize it by entering the residential market, much like Nextel did in the US. Second, given the availability of additional spectrum, an existing communications firm (such as HOT, or Xfone) could potentially enter the market and force the market price down. This, however, would require that profits at the time of entry be above normal levels, as otherwise the profits will be too low to overcome the initial lack of economies of scale. In contrast, based on our review of the Israeli market, our view is that intermodal competition, such as WiFi and WiMAX, is not strong enough yet to pose a competitive threat. This, however, might change in the coming years. Other considerations In addition to the considerations above, there are a number of other attributes that contribute to or subtract from the likelihood of collusive behavior occurring in the Israeli retail market. Specifically: Rapidity of market growth: if the market is growing rapidly, then the short term gains from deviating from collusive behavior are more likely to be outweighed by the losses that result from future retaliation by competitors, thereby making collusion more likely. In the Israel mobile market, growth has recently slowed down as market saturation is approached. This means that collusive behavior is now less likely than in the past. The fact that Pelephone introduced a pricing plan for the government that involved heavily discounted calls indicates the potential for non-collusive behavior in the corporate sector. Moreover, no apparent retaliation seems to have taken place.
238
239
Commission for Communications Regulation, Market Analysis-Wholesale Mobile Access and Call Origination, Document No: 04/118 and 04/118a, December 9, 2004, 4.89, citing the European Court First Instance in the Airtours judgment, 62. Ibid., 4.98.
114
Stability of demand: Similar to the point above, in markets with large demand fluctuations, there is a greater incentive for non-collusive behavior at the peak of the cycle, as any future losses from retaliation will be reduced as demand falls. Thus, noncollusive behavior could appear at the peak even though it would not do so at other points of the cycle. In the Israel mobile market, demand does not appear to be subject to cyclical variation. Hence, it is not likely to be a potential source of non-collusive behavior, although it is also not likely to be a source of collusive behavior. Market price elasticity of demand: If the market elasticity of demand is high, the cost of retaliatory price reductions will be mitigated by the resulting increase in market demand. In such circumstances, collusive behavior is less likely to be sustainable. Recent studies (outside Israel) that have attempted to estimate mobile-market price elasticities of demand suggest that the latter are relatively low. Assuming that Israel is similar, this would suggest that elasticities of demand are unlikely to be a source of non-collusive behavior. Entry barriers: Generally, there are substantial barriers to entry into the mobile market, especially the requirement for a spectrum license, approval from MOC and local municipalities, and the high upfront costs required to commence operations. Entry makes collusive behavior less likely as it is potentially unsustainable. Therefore, high barriers to entry make coordinated behavior more likely. Thus, the entry barriers in Israel tend to be conducive to collusion. As we have noted though, while barriers to entry are high in Israel, they are lower than in many other countries, as there is spare spectrum available. Furthermore, MIRS could develop business in the private market with relatively low entry costs. As mentioned above, one way of doing this would be to team up with an MVNO. Buyer power: Large (countervailing) buyers can make noncollusive behavior more attractive because the gains from breaking ranks and winning the business are potentially substantial and like for like retaliation is more difficult. There are groups with substantial buying power in Israel, such as the armed forces and the government (or its employees). Hence, this tends to reduce the attractiveness of coordination in the business sector. Customer switching costs: These reduce the potential benefits of non-collusive behavior, but also reduce the impact of retaliation. They therefore probably have a broadly neutral effect.
In the table below, we summarize our conclusions regarding the possibility of tacit coordination.
115
Table 13 Considerations of Collusive Behavior in Israel Summary

Conducive to collusive behavior X X X X X X X X X X X X X X X X X X X X X X X X X X X X X Conducive to noncollusive behavior
Consideration Number of MNOs Symmetries Marketshares Costs RPM ROCE Pricing plans Target markets Technology Capacity Detect cheating Known products Frequency of interactions Transparency Product comparability Price coordination General portfolio Product features Network technology 3G offerings Ownership structure Contract terms Enforcability of compliance Breakdown of collusion Market constraints Market contestability Potential entry or expansion Intermodal competition Other considerations Market growth Stability of demand Price elasticity Entry barriers Buyer power Switching costs
Indeterminable
As this summary reveals, there are various market conditions that are conducive to collusion, while others seem to undermine such a possibility. It is important to note that, in order for firms to collude successfully for a significant period of time, it is necessary that most, if not all, conditions are conducive to such behavior. This is not the case in Israel. Hence, while we cannot rule out the possibility of tacit collusion, there are factors that would make it difficult to sustain in the long run.
116
4.1.8. Summary of competitive conditions in the Israeli wireless market In light of the above analyses and theoretical considerations, we conclude that there is no compelling evidence of market failure in the Israel wireless market with regards MVNO entry. Hence, we do not need to further expand our narrow market definition, even if it were warranted. At the same time, while we find that in some respects the market shows signs of behavior consistent with competitive markets, there are other aspects were it does not and improvements can be made. Specifically, our findings regarding the competitive conditions in the wireless retail market in Israel are: Overall market shares: The market shares of the largest operators are relatively low and do not suggest the existence of a dominant position by any MNO. Specifically, the two largest firms, Cellcom and Partner, have a smaller combined market share than is the case of most other countries. Similarly, the HHI index in Israel is lower than in most other countries. Sector-specific market shares: The limited role of MIRS in the private sector results in a higher concentration in this sector relative to the overall market. However, even at this elevated level, the concentration is still low by international standards. Thus, the near absence of MIRS in the private sector is not necessarily of competitive concern. Overall market share trends: While market shares have changed less in recent years, this might be a sign of a maturing market. Similar patterns can be observed in the US and the UK. More importantly though, the MNOs growth patterns from 2003-2006 differ, which is consistent with competition, rather than market failure. Sector-specific market share trends: Israeli market shares by customer segment do differ and have fluctuated over the last five years, which is not consistent with the existence of market power or collusion. Given the increasing convergence in market shares, however, we do recommend that MOC continues to monitor developments in sector specific market shares as in a competitive market we would expect market shares to continue to change over the next few years, rather than converge to the same level. Prices: Israeli prices, measured by RPM, appear low by international standards. While we have a number of concerns surrounding the conclusions that can be drawn from such a comparison, which we have set out above, it can be concluded that the available evidence on pricing does not of itself indicate market failure. As regards the evidence on sectoral price differences, it is not clear whether it supports the allegation that there is less competition in the private sector. It certainly appears to be the case that prices are higher for private customers than for business customers. However, call volumes per subscriber are also much higher in the business sector, which means that relative prices are not in themselves an indicator of relative profitability. At the same time, it also necessary to take account of the comparative level of handset subsidization in the two sectors, for which we did not have the necessary data. Pricing trends: Recently Israeli prices overall have fallen less than in other nations. Prices for postpaid services and business services have decreased while prices for prepaid services have increased. The large variance in the MNOs price changes is an indicator that they have not acted in a concerted manner. Hence, neither pricing trends nor current price levels by themselves suggest market failure.
117
Overall profitability: Profitability levels of Israeli MNOs, measured by ROCE and EBITDA margins, (1) have fluctuated significantly over time, (2) are asymmetric, (3) are lower than the levels reached in Ireland where the regulator decided to intervene, (3) are well below what UK regulators have historically considered as excessive, and (4) are lower than in most countries. Furthermore, profits per minute are low relative to countries with similar GDP per capita. Finally, profits per user are comparable or on the low side relative to countries with similar GDP per capita. Hence, the overall profitability of Israeli MNOs by itself provides no evidence of market failure. Sector-specific profitability: A sector-specific ROCE approximation for Cellcom, Partner, and Pelephone indicates higher profitability in the private sector than in the business sector. This, in turn, could indicate less competition in the private sector, relative to the business sector. However, given the margin of error associated with this type of analysis and the fact that the three MNOs private ROCE numbers do not appear clearly excessive by international standards, a firm conclusion cannot be drawn. Nonetheless, we believe that MOC should keep the situation under review. Barrier to entry and expansion: Spectrum in Israel appears not to be a barrier as spare spectrum is available for one additional MNO (to provide both 2G and 3G services). Regulatory approval, both for MNOs and MVNOs, might amount to a barrier to entry. Access to capital markets should not be an entry barrier, at least for large carriers already in the communication industry. Economies of scale are a barrier to entry, although they appear to be the result of first mover advantage and market saturation. Another source of potential regulatory barriers stems from local government jurisdiction over rights-of-way and the location of cellular antennas and towers. Switching barriers do exist in the Israeli market and, while in many cases they may result from competitive behavior rather than attempts to restrict competition, we have concerns about the uniform offering of a single term contract that binds customers for 36 months. Typically term contracts in other countries are of between 12 and 24 months in length. Hence, we conclude that some of the existing barriers to entry and expansion are not signs of market failure. However, the current contract term length, possibly some of the terms and conditions, and possibly the promotion of seemingly inferior contracts, appear not to be consistent with competitive markets. Hence, we recommend MOC to carefully review these instances and take any necessary steps in order to minimize barriers to entry and expansion. Churn rates: Customer churn in Israel is low by international standards, although we do question the correlation of customer churn and competition. Given the impending introduction of MNP, we dont find current churn rates to be evidence of market failure. Historic behavior: In addition to seemingly excessively long term contracts, various parties, including MOC, have alleged historic behavior by the Israeli MNOs, which is suggestive of the existence of market power. It is not within our remit to review the evidence of such alleged behavior. If substantiated, it would not necessarily prevent market forces from working (i.e., it is not necessarily a sign of market failure). It would, however, suggest that the level of competition in Israel could still be improved and that specific issues such as contract opacity and the absence of information available to customers regarding the existence and content of alternative service offerings should be addressed directly by the MOC.
118
Collusive behavior: Some conditions in Israel are conducive to non collusive behavior, yet others make such behavior possible. While we find that collusion would be difficult at best, we cannot entirely rule out the possibility of it occurring, particularly with respect to MVNOs.
4.2. Business Models for Israeli MVNOs

The term business model refers to the economics of a companys revenue, cost, and profit strategy. Its primary objective is to demonstrate the long-term viability of the company, or in the present instance, an MVNO. Most generally, there are four components to a business model: 1) access to infrastructure, 2) value proposition, 3) customer relations, and 4) finances. We will discuss each of these components below and address how they apply to MVNOs and Israeli MVNOs in particular. 4.2.1. Access to Infrastructure Much like any company, a new MVNO will have to have an able and experienced leadership that can efficiently execute an MVNOs business model. We have no reason to believe that such leadership is not available in Israel. Particularly, we find that at least HOT is well experienced in the communications industry and should have the necessary qualifications to cover this core requirement. Critical to any MVNO is access to network infrastructure and possibly other aspects of the business which complement the business model. For instance, HOT would need access to a network (either by self supply or through a wholesale agreement with an MNO). In addition, its existing media network will provide it with valuable synergies, which in turn will impact its cost structure. There are only two ways by which an aspiring MVNO in Israel would gain access to another network; either it commercially negotiates a wholesale agreement with an MNO or it obtains access on terms mandated by MOC. As we will discuss below, mandating wholesale access and the terms on which it takes place can be costly and comes with significant economic risk. If market conditions are such that competitive forces cannot discipline prices, service offerings and quality (i.e., there is market failure), then regulation might be economically justified. If market conditions are not conducive to market failure, then commercially negotiated agreements are economically superior to regulatory intervention. A commercially negotiated agreement will only exist if both parties stand to benefit from it. For the MNO, this means that at a minimum the wholesale revenue obtained from an MVNO must cover its wholesale operational expenses, customer capital expense (see below), and its expected rate of return. Alternatively stated, wholesale revenues must at a minimum equal the residual of total service cost minus the avoided costs (from only providing whole service rather than retail service). The actual willingness to enter an agreement with an MVNO will also depend on a number of other aspects. For instance, if an MVNO were to compete directly with the MNO, rather than widening or deepening the market, then the MNO might also consider the lost profits due to the wholesale agreement. Spare capacity, reputation effects, and costs of integrating an MVNO are examples of other factors influencing an MNOs willingness to provide an MVNO with wholesale access. For the aspiring MVNO to enter a commercially negotiated agreement, its ARPU must at a minimum equal the wholesale access cost, its operational expenses, and its rate of return.
119
Over the lifetime of the customer, it will also need to recover customer acquisition fees, such as handset subsidies and marketing. The common component of these two equations is the wholesale discount, measured as a percentage off the retail price. That is, if a discount or range of wholesale discounts exist where both the MNO and MVNO cover their minimum requirements, as outlined above, then commercially negotiated agreements are possible, albeit not guaranteed. Conversely, if such a point or range does not exist, such agreements are unlikely to occur and there might be no MVNO business model. To examine whether the commercial conditions exist for aspiring MVNOs in Israel to negotiate successfully to gain access to the MNOs infrastructure, we developed an illustrative discounted cash flow (DCF) model. As is the case with most models, simplifications need to be made. Thus, the existence or absence of a commonly accepted bargaining range is only an indicator as to whether or not a wholesale agreement will occur. The analysis begins by discounting back to the present time the current and all future cash flows the MNO expects to earn when serving (1) a postpaid customer, (2) a prepaid customer, (3) a wholesale postpaid customer, and (4) a wholesale prepaid customer. While there exist a number of additional variations, we limit our analyses to these scenarios in order to determine whether the MNO has a financial incentive to reach a wholesale agreement with an MVNO. In order to determine whether the MVNO will benefit from this relationship, we examine the net present value of (1) a postpaid MVNO customer and (2) a prepaid MVNO customer. The DCF calculation is:
1+ g DCF = CF0 t =0 1 + r
n
where DCF is the discounted cash flow or NPV, CF0 is the current cash flow (in time period 0), g is the constant rate at which the cash flow is expected to grow in each future period, r is an expected rate of return that is used to discount future cash flows, t is an index for the time period, and n is the customer lifetime over which the DCF analysis is conducted. Below, we detail how each component of the DCF model is selected. Expected rate of return (r) The expected rate of return is the opportunity cost of the MNOs investment. The weighted average cost of capital (WACC) for any investment, whether for the entire firm or simply one project conducted by the firm, is the rate of return that capital providers typically expect to receive from investing their capital elsewhere. Thus, for r, we use the WACC of a typical wireless carrier. While WACC varies over time and between operators, a figure of 15 percent in pretax nominal terms would not be atypical for a mobile operator. Customer lifetime (n) The customer lifetime is the inverse of the customer churn rate. The churn rate is the percentage of customers who end their relationship with a company within a given time period. For instance, a typical monthly churn rate for a postpaid wireless service provider in
120
Israel is 1 percent. This means that 1 percent of the postpaid customers of a wireless service provider switch each month to another wireless service provider. Alternatively, a churn rate of 1 percent means that, not counting additions, all current customers would be completely replaced within 1/0.01, or 100, months. Thus, the implied customer lifetime in this example is 100 months. Based on the data received from the MNOs in Israel, on average, 1.03 percent of Israeli postpaid customers churn in a given month, yielding a postpaid customer lifetime of 97 months. Prepaid customers churn more frequently and based on the data received, 1.86 percent of Israeli prepaid customers churn in a given month, resulting into a lifetime of 54 months. We assume that an MVNO would have identical churn rates. This is a rather conservative assumption as MNOs have typically much higher brand value and are more experienced in serving the wireless market. Nevertheless, based on our expectation that the most likely entry will occur from current communications providers, such as Xfone and HOT, or possibly MIRS moving into the private market in a comprehensive way, we believe this assumption to be reasonable. Cash flow per user (CF) The series of current and future CF generated by each wireless (MNO or MVNO) customer is calculated as the gross profit for that customer in each month over the customers lifetime. Specifically: CF = ARPU CCPU CPGA CCapEx, where CF = cash flow generated each month by the customer ARPU = Average revenue per user (customer) CCPU = Cash cost per user CPGA = Cost per gross addition CCapEx = Customer capital expenditures Average revenue per user (ARPU) ARPU measures the average monthly revenue generated per customer. It is frequently used to assess a wireless service providers operating performance. The ARPU of a typical private postpaid wireless service provider in Israel is NIS 139 (approximately US$33) per month; that is, the average customer adds US$33 to the service providers monthly total revenue. Similarly, the ARPU of a typical prepaid wireless service provider is NIS 58 (approximately US$14) per month. We assumed that an MVNO will earn identical ARPUs to those of the MNOs. This is a reasonable assumption if the MVNO were to compete directly with the MNO. However, if the MVNO were to pursue a niche strategy, it might obtain lower ARPUs (e.g., discount carriers) or higher ARPUs (e.g., music download providers) than the MNOs. Cash cost per user (CCPU)
121
CCPU is the wireless (MNO or MVNO) service providers average monthly cost to serve its customers. Stated differently, it is the amount of spending in a given time period necessary for the wireless service provider to retain an existing customer. Retention costs include customer support, billing, promotional incentives, and so on. In economic terms, CCPUs are monthly operational expenses. Based on data contained in Partners and Cellcoms financial records, we estimate postpaid CCPU at NIS 11 (approximately US$2.57) per month.240 No data exists on the CCPU for prepaid customers. However, given that prepaid services do not require billing and possibly involve fewer promotional incentives, we assume CCPU for prepaid services at 40 percent of the CCPU for postpaid services, i.e., NIS 4 (approximately US$1.03). This ratio is consistent with data we have previously analyzed in the US. We assumed wholesale CCPU to be 50 percent of retail CCPU. This assumption takes into account that the MNO will not incur any retention costs for wholesale customers but instead will have additional operational expenses due to the wholesale interaction with the MVNOs. Finally, the MVNOs CCPU will again be a function of the type of MVNO that enters the Israeli market. An MVNO equally efficient as the MNO will incur a grossly identical CCPU, while MVNOs with high brand value and/or superior distribution channels will incur lower costs. For the purpose of this calculation, we have assumed that the MVNO is as efficient as the MNO. Cost per gross addition (CPGA) CPGA is the wireless service providers average cost to acquire a new customer. Typically, the largest cost items in this category include the handset subsidy provided to the customer and the marketing and distribution costs of supporting the service providers service plans. Based on financial data received from Cellcom, Partner and MIRS, a postpaid wireless operator in Israel spends on average NIS [] (approximately US$ []) per gross addition. No data exists on the CPGA for prepaid services. However, based on [] s data, we know that of the NIS [] spent on postpaid net additions, it paid NIS [] on handset subsidy and NIS [] on commissions. Prepaid customers typically do not enjoy any handset subsidy. While we have no information as to the specific commission arrangement for the sale of prepaid services, at a minimum, for [] we can deduct the handset subsidy to derive the CPGA for prepaid services, i.e., CPGA for prepaid service is NIS [] (approximately US$ []). Partners CPGA appears to be an overall average, hence it remains unchanged. For MIRS, we reduce its postpaid CPGA by the same percentage as []. Averaging the resulting three figures yields an approximate prepaid CPGA of NIS [] (US$ []). Again, for simplicity, we assume an MVNO would incur the same CPGAs as the MNOs. While this is reasonable for established communication and media providers (such HOT and Xfone), brand new MVNOs with no prior exposure to the communications industry would be likely to incur significantly higher CPGAs. Customer capital expenditures (CCapEx)
240
To arrive at this number, we subtracted commission expenses from selling and marketing expenses and divided the remaining amount by the number of subscribers in the same year and then by 12 to arrive at monthly figures.
122
CCapEx is the wireless service providers cost for network capacity upgrades needed to add customers to its existing network. When an existing network has significant spare capacity, this cost may be assumed to be zero. We understand from the MNOs that there is not sufficient spare network capacity to accommodate MVNOs. Hence, in our illustrative example we apply the LRIC estimate for call termination plus a markup for common fixed costs, as estimated by Analysys in its report to MOC in 2004. Allowing for inflation of 1 percent per annum (which is the average annual increase between 2003 and 2006) yields a 2007 figure of NIS 0.148. Typically fixed costs account for approximately 20 percent of the total. Hence, subtracting fixed costs, results in a long run marginal cost estimate of NIS 0.118 per minute for call termination, including network maintenance. We assume the cost for call origination to be the same. Based on the same report, we calculate off-net termination at NIS 0.142 per minute. These figures indicate that an on-net call costs an average MNO approximately NIS 0.236 per minute, while an off-net call costs NIS 0.260 per minute. We assume 66 percent on-net and 33 percent off-net calls. The table below summarizes the input values of this model.
123
Table 14 Generic DCF Model Assumptions Israel
Source: NERA research, MNO financial data
Model results Based on the data and assumptions described above, the DCF analysis reveals the following: The NPV of a postpaid customer to an MNO is approximately NIS [] or US$ [] The NPV of a prepaid customer to an MNO is approximately NIS [] or US$ []
124
The NPV of a wholesale postpaid consumer to an MNO is approximately NIS [] or US$ [] This assumes that an illustrative 35 percent wholesale discount off the retail price is given. The NPV of a wholesale prepaid consumer to an MNO is approximately NIS [] or US$ [] This assumes that an illustrative 35 percent wholesale discount off the retail price is given. The NPV of a postpaid customer to an MVNO is approximately NIS [] or US$ []. This assumes that a 35 percent wholesale discount is received. The breakeven point discount is approximately [] percent, meaning that for the MVNO to be profitable under the assumptions described, it needs to obtain at a minimum a wholesale discount of [] percent or higher. The actual required wholesale discount will be likely to be significantly higher to cover also fixed and upfront costs, not reflected in the calculation above. The NPV of a prepaid customer to an MVNO is approximately NIS [] or US$ []. This assumes that a 35 percent wholesale discount is received. The breakeven point discount is approximately [] percent, meaning that for the MVNO to be profitable under the assumptions described, it needs to obtain at a minimum a wholesale discount of [] percent or higher. The actual wholesale discount will again be likely to be significantly higher to cover also fixed and upfront costs, not reflected in the calculation above. Appendix F shows the assumed input values, calculations, and NPVs in the six scenarios. The table below summarizes the results of the DCF model. In addition, the DCF analysis leads to the following conclusions:

Postpaid MNO customers provide the highest value. This conclusion is evidenced in the marketplace where MNOs thus far have mainly focused on the postpaid market segment. Prepaid MNO customers can be profitable. In addition to its other advantages (such as continued subscriber growth, reduction of bad debts, optimized network utilization, etc.), prepaid service produces positive NPV. Wholesale postpaid and prepaid customers can be profitable. MNOs can profitably supply wholesale services to resale-based service providers like MVNOs. This suggests that, under the right circumstances, providing wholesale access to its wireless network is indeed in the MNOs economic interest. For instance, if an MNO is unsuccessful in expanding its operations into the prepaid market segment, it would make economic sense for it to open its network to wholesale access.
Aspiring MVNOs that propose to compete directly with their host MNO are unlikely to reach a wholesale agreement. As discussed above, however, the prerequisite for a wholesale profit is that the MVNO does not cannibalize the MNOs customer base. If the MVNO competes directly with the MNO, such relationship is less likely to happen as the MNO would offset the NPV of a retail customer with the NPV of a wholesale customer. In the example above, with a 35 percent wholesale discount, the MNO would forgo an NPV of $ [] for a wholesale NPV of $ [] incurring a net loss of NPV of $ [] per customer ($[] -$ []). Only a reduction in the wholesale discount would make the MNO indifferent between supplying an MVNO and selling to the retail customer. In the example above, the MNO
125
would want to reduce the wholesale discount to approximately [] percent. At that level of discount, however, the MVNO would not make any profit for each customer addition to cover upfront costs and thus unlikely to be sustainable in the long run.
MVNOs that offer identical products as the MNOs are unlikely to be socially beneficial. As the point above has illustrated, an MVNO that is purely and completely a substitute for an MNO (and is roughly equally efficient) is unlikely to be profitable. Hence, it entry is not beneficial from a social point of view, as nothing is gained and there is an additional set of upfront costs incurred. If, however, the MVNO adds value, by deepening or widening the market, or by operating more efficiently, then it is more likely to be profitable and can increase social welfare (through lower prices and/or new services) at the same time. Hence the key to MVNO entry is that it is economically efficient.
MVNOs targeting customers with higher ARPUs are more likely to survive in the long run. For Cellcom, in 2006, handset subsidies plus all selling and marketing costs accounted for [] percent of service revenues, excluding equipment revenues. We have no data on the cost of customer care. However, if we assume that they accounted for 100 percent of rent and related costs, salaries and related costs, and the residual other cost category within the cost of services, then the resulting retail costs (handset subsidy, plus selling and marketing plus customer care) amounts to [] percent of service revenues. In practice they will account for much less than 100 percent of these items, as these items include salaries, rent, and other costs associated with operating and maintaining the network. Based on these figures and assumptions, it is difficult to see how the retail discount could exceed [] percent without Cellcom making a loss on their wholesale operations. At a wholesale discount of [] percent, the MVNO is only marginally sustainable, making $ [] for each postpaid net addition and $ [] for each prepaid net addition.
Successful MVNOs will likely be facilities based. By becoming facilities-basedeven if partiallyan MVNO can put itself in a position to add more value than may be possible merely on the strength of its brand appeal. First, the facilities-based MVNO can reduce its dependency on the MNOs network and, in particular, on the wholesale discount. Second, the facilities-based MVNO is likely to incur lower incremental costs to expand short run supply (particularly if it has excess capacity available to it) than an MVNO that has no facilities at all. Even a partially facilities-based MVNO may, therefore, have better prospects for long-term survival. Successful MVNOs will have a comparative advantage in customer management. Whether or not they are facilities-based, MVNOs can further reduce their risk and network dependency by bringing value to their customers through superior customer management. MVNOs that need to outsource billing, customer service, and other customer management functions are likely to have higher operational expenses and, therefore, require greater wholesale discounts to stay viable. In contrast, MVNOs like Virgin Mobile that perform their own customer management functions stand a better chance of long-term survival. The most risk-diversified MVNO is one that is facilitiesbased, owns its own customer management operations, and has sufficient distribution channels which enable it to enjoy economies of scope in retailing.
Successful MVNOs will need to cover its fixed costs by targeting higher ARPU customers or by quickly establishing a presence in the market. At NPVs of US$ [] and US$ [] respectively, for prepaid and postpaid customers, MVNOs benefit from each additional
126
customer. However, at this level, it is far from clear whether they have a successful business case, as this rate might not be sufficient to cover upfront costs and fixed costs. Hence, the successful MVNO will be one that either targets customers with high ARPUs (such as US MVNO Helio) or attracts a large customer base by using a discount strategy (such as US MVNO Virgin Mobile USA). MVNOs that simply stand to replicate the MNO model and are not more cost efficient in doing so are unlikely to survive. The above analysis indicates that access to infrastructure depends crucially on the type of MVNO and its value proposition. While no firm conclusion can be drawn as to whether MVNO will gain access to the necessary infrastructure, the successful business case will rest on a model that adds value to the MNO by deepening and widening its target market and therefore becoming a more effective competitor.
4.2.2. Value Proposition
The MVNOs value proposition is crucial in at least two aspects. First, by adding value to the MNO, the aspiring MVNO is more likely to obtain a wholesale agreement as the engagement is more likely to be mutually beneficial. Second, in a saturated market, such as Israel, an MVNO needs to provide an innovative service or service bundle, which will increase its likelihood of long terms survival. Given the high level of saturation in the Israeli market, the value proposition will be a key challenge to the successful business plan. For example quadruple play would be a value proposition for the Israeli wireless market.
4.2.3. Customer relations
Clearly defining, targeting and supplying its customer is the third prong in a sound business model. While the level of saturation, again, works against achieving this requirement, companies that have existing customer relationships and distribution channels, such as HOT, are more likely to be successful.
4.2.4. Finances
As discussed above, aspiring MVNOs in Israel will either need to reduce their costs or target high ARPU customers in order to be profitable. Given Israels low prices (relative to international standards) and advanced networks and service features, identifying customers the meet these requirements will be challenging. Based on these considerations, we find the MVNO business case in Israel to be challenging at best. Of particular concern is the fact that the saturated market might not provide an MVNO with sufficient opportunities to obtain sufficient revenues to remain viable over the long run.
4.3. Understanding the Absence of MVNOs in Israel

With over 40 MVNOs in the US, six in the UK, and emerging MVNO entry in many other nations, the question arises as to why there are no MVNOs in Israel. In order to answer this question, it is important to emphasize that the mere presence of MVNOs does not guarantee an increase in consumer welfare. As we have shown above, for Israeli MVNOs to be profitable, they would need to serve new market segments, rather than simply compete in existing market segments (unless they were much more efficient that the MNOs when it came to acquiring and retaining customers). As was revealed by the benchmark analysis, however,
127
most current MVNOs do not compete directly with the MNOs. Rather, they find long-term survival in deepening and widening the market. That is, rather than offering exactly the same products as the MNOs, they offer innovative products and services. Generally, there are three possible answers to why MVNOs have not sought to enter the Israeli wireless market. First, MVNOs could have decided not to enter the Israeli wireless market because its competitive conditions would not support a viable business case. Second, MVNOs might want to enter the Israeli wireless market, but find MNOs not receptive to opening up their networks for wholesale access. Third, MVNOs might simply be a question of time. We will discuss each of these possible answers below. First, our review of the market conditions in Israel has revealed a number of market characteristics that do not appear to be conducive to MVNO entry, as well as others that are potentially favorable. Among the characteristics that are not conducive to entry are the following:
The existing penetration rate is 119 percent. Even though this is partly explained by some users having more than one handset, it suggests that there are few, if any, areas of the market that have not been opened up yet. Typically, MVNOs are most successful in countries with limited penetration rates, such as Denmark in 2000, or the prepaid market segment in the US, which was largely ignored by US MNOs prior to MVNO entry. Motivated by the success of Virgin Mobile, many MVNOs enter the market in the prepaid sector. The few MVNOs that have entered in the postpaid sector have had limited success. For instance, US MVNO Ampd mobile is one of the few US carriers that offers both prepaid and postpaid plans. Ampd recently filed for Chapter 11 bankruptcy, stating that bad debts from postpaid customers, among others, force them to seek bankruptcy protection. On the other hand, US MVNOs Virgin Mobile and Boost Mobile seemed to be growing steadily (although Virgin Mobile has yet to generate profits). The market in Israel is predominantly a postpaid market. Israeli prices for wireless services, measured by RPM, appear to be relatively low by international standards after allowing for a difference in GDP per capita. As indicated by the figure below (which includes all countries with a GDP per capita of US$10,000 or more), the average RPM in Israel is low, even after taking differences in GDP per capita into account. Notwithstanding the points made earlier about the difficulties in interpreting and using RPM comparisons, this does suggest that Israel may be a less attractive market for MVNO entry than others where the level of prices appears to be higher.
128
Figure 39 Revenue per Minute vs. GDP per Capital Israel 2006
0.4
0.35
0.3
Revenue per Minute $US
0.25
0.2
0.15
0.1
Israel
0.05
0 0 10000 20000 30000 40000 GDP per Capita $US 50000 60000 70000 80000
Taking all segments of the market together, Israeli MNOs are not collectively, or individually, making an excessive rate of return on capital. Hence, from an overall market perspective, an entrant MVNO could therefore not anticipate large profit margins to exploit, unless it had some particular advantage in terms of low retailing costs. There exist a number of customer switching barriers, such as opaque contracts, the current absence of MNP and the increasing prevalence of 36-month contracts, coupled with early termination fees. These tend to reduce customer churn which will make MVNO entry more difficult. As we were informed during our interviews, the perception of prepaid wireless service in Israel is bad, signaling low societal status. Since most MVNOs rely either exclusively or at least partially on prepaid service offerings, MVNOs might not view Israel as an attractive market. Further, we understand that the MOC currently has no licensing process for MVNO entry. Naturally, this is an entry barrier, as without regulatory approval process, there will be no MVNOs in Israel. In order to minimize the regulatory barriers for MVNO entry, we recommend implementing a simple regulatory approval process.
129
There are at least three market characteristics that are potentially favorable to MVNO entry, including:
While customer churn in the private and business segments seems low by international standards, churn is higher in the prepaid sector, which is where MVNO typically enter the market. Depending on costs, low prices could still be consistent with profitable opportunities for MVNOs to exploit, particularly as profitability in the private sector appears higher than in the business sector. An existing operator, such as HOT, could bundle mobile services along with fixed telephony, broadband and video. The economies of scope involved would offer it a potential advantage over the Israeli MNOs. As indicated above, both Partner and Cellcom could offer wireless and fixed services, but would still be missing video for a quadruple play. The low percentage of prepaid subscribers could serve as an opportunity for an MVNO to increase the attractiveness of prepaid services, much like Virgin Mobile did in the US.
Based on these considerations, the analysis above regarding a potential business model for Israeli MVNOs, and the findings of the DCF model, we find that entry conditions for MVNOs are difficult at best. Much like in other countries, it is unclear whether MVNOs would be able to cover their upfront and fixed costs, given the current market conditions and the expected NPV per customer. Hence, it is possible that thus far MVNOs could have decided not to enter the Israeli wireless market because its competitive conditions would not support a viable business case. For the second theory (collusion) to hold, MNOs would need to collude tacitly to prevent wholesale access. As we discussed above, while we cannot entirely rule out the possibility of collusion on wholesale access, it does not seem likely that successful collusion could be maintained under the current market conditions. Hence, our conclusion is that collusion is not the reason for the absence of MVNOs in Israel. Finally, to address the third theory, based on the information received from Partner, Cellcom and Pelephone, none of the Israeli MNOs have recently been approached by a potential MVNO. []: [] [] While, according to these statements, Israeli MNOs appear not to have been approached by any potential MVNOs in the recent past, our research indicates that in December 2006 Xfone approached MOC regarding an MVNO license. No information, however, is available as to whether Xfone has since approached an MNO for wholesale agreement negotiations. Furthermore, Xfone informed us during our interviews that no such negotiations have taken place. Similarly, in January 2007, HOT approached the MOC regarding an MVNO license,
130
and the Israeli media reported that HOT was in talks with MNOs.241 Based on our interviews, this does not seem to be correct, although HOT has expressed an interest in becoming an MVNO as part of their quadruple play strategy. Given the previous interest in Israel, NERA contacted Virgin Mobile to obtain information regarding possible reasons for the failed negotiations with Pelephone. While we were unable to obtain specific information from Virgin regarding the Israeli negotiations, Virgin UK informed us that as a company they dont support wholesale tariff regulation as [it] does not tend to lead to successful, high growth MVNOs and that they believe markets should be open to MVNOs (i.e., permitted), but then market participants should be free to negotiate as they wish.242 When asked about their attitude to MVNOs, the MNOs claim to be open for discussion, stating that they would be willing to negotiate with any MVNOs. However, we do note that the general attitude towards MVNOs (possibly only regulator mandated MVNOs) was negative. As to the MVNOs potential for success in the Israeli wireless market, the MNOs seem to be skeptical, citing, amongst other things, high penetration rates, high and increasing customer service levels, high customer satisfaction levels, and the large necessary advertising budgets as reasons why the prospects for successful MVNO entry were limited. While MNOs do not seem to believe that there is a viable business model for MVNOs, the possibility of such entry remains. In our view a successful MVNO entrant would require either to have some network facilities itself or to have substantial potential economies of scope. We would therefore expect that HOT cable would be one of the companies that would consider launching an MVNO. Similarly, MIRS would have an incentive offer access to its network, particularly where such entry focused on the private sector of the market where it does not currently operate to any meaningful extent. That, of course, assumes that MVNO entry is potentially profitable, an assumption that may be open to question, as we discussed above. In evaluating the three possible explanations as to why there are currently no MVNOs in Israel, we believe that it is reasonable to rule out collusion by MNOs not to offer access. Only time will prove whether there is a real business case for MVNOs in Israel. The chances are not, in our view, that high, but for a few selected MNOs, such as MIRS, and a few selected potential MVNOs, such as HOT, striking a wholesale agreement could be economically beneficial and thus remains a possibility.
4.4. Effects of MOC-Mandated MVNO Players in Israel

The effects of MOC mandated MVNO entry in Israel can have far reaching consequences if such regulation is not warranted. Primarily, such intervention can impact on consumers, the industry, competition and hence the general economy. The potential harm to economies where regulation is not warranted is multifold and well documented in the economic literature. At the heart of this literature is the fact that regulation
241 242
Globes Online, HOT in talks to set up virtual cellular operator, A virtual operator buys wholesale air-time from an existing operator, November 29, 2006. Email correspondence from Kim Baker, Virgin UK, on behalf of Robert Samuelson, Director of Corporate Development, Virgin UK, with C. Dippon, April 5, 2007.
131
is not free; it comes with substantial costs. The costs materialize in the form of regulatory costs (e.g., setting of rates, access terms and conditions) and market risk. As found by Kahn (1994), regulation and market forces compete in disciplining market players.243 Kahn notes that there is no rational half-way house between thorough regulation and free competition.244 In addition, there is a high likelihood that, because regulators do not have full visibility of the costs involved, regulated terms of access will either be too high, therefore precluding entry, or too low, thereby encouraging inefficient entry. The latter means that costs in the mobile telecommunications industry will actually increase as the result of regulation. Hence, unless there is market failure, imposing regulation does not guarantee an increase in competition and hence social welfare. Specific to the MVNO question in Israel, unless there are clear signs of market failure, imposing MOC mandated MVNOs may well be detrimental to competition and innovation in the wireless market, as it might also lead to inefficient entry.
4.5. Regulatory Principles of Competition Policy

In order to assess the economic necessity, or lack thereof, for regulatory intervention, it is necessary to start with the need for regulation itself. The main justification for regulatory intervention in an industry is market failure, a major contributory factor to which is the possession of market power by one or more firms. Economic theory defines market power as the ability to profitably raise and sustain the price of a product above the level that would prevail in a competitive market. Usually, this power over price is made possible by a firms ability to significantly affect the market supply of the product by restricting its own output. The corollary of this definition is that, in an effectively competitive or contestable market, no single firm, regardless of its relative size, can expect to charge a supracompetitive price and earn greater than normal profits. In such a market, countervailing forces are likely to defeat an attempt by any one firm to permanently raise the market price above the competitive level. What are the sources of market power? Even in circumstances in which the possession or the very acquisition of monopoly control is prohibited by law, incumbent firms can still acquire market power (to various degrees) through either industry or product characteristics or various forms of market conduct. Historically, one source of market power in telecommunications networks around the world (primarily fixed networks) has been a condition most frequently referred to as natural monopoly; that is, a monopoly that emerges not because of illegal or nefarious attempts by one firm to monopolize the market, but because operational and market structure characteristics are such that the product or service in question can most economically and efficiently be produced or provided by a single firm. Those characteristics typically include economies of scale and scope, commonly observed to arise in industries in which to operate at the minimum efficient scale, a firm must incur high initial fixed costs and produce multiple products or services using shared or common plant and equipment. Because of this, the provision of telecommunications services in countries around the world had historically been entrusted to a single entity, whether a privately owned firm or a public
243 244
Alfred E. Kahn, The Uneasy Marriage of Regulation and Competition, Telematics Vol. 1, Number 5, pp. 1-17. Ibid., p. 8.
132
or government undertaking. Where the danger of impermissible monopolistic conduct arose (e.g., excessive earnings and abuse of customers through exploitative prices), such as with private natural monopoly service providers, governments imposed various forms of regulation and franchise responsibilities (such as universal service obligations) in exchange for preserving their single-provider status. Incumbent firms can also acquire market power through vertical control of essential facilities. An essential facility has four main characteristics: 1. It is wholly owned and controlled by a dominant vertically integrated firm. 2. It is a critical input into the production or provision of a product in an adjacent market, typically a downstream retail market. 3. Competitors can only acquire access to the facility from the firm that owns and controls it (i.e., there are no existing alternatives). 4. Competitors cannot economically or technologically duplicate the facility. A vertically integrated incumbent that has full control over an essential facility (usually a wholesale product or service) has the power to prevent access by its competitors or to exert a price squeeze by setting prices for access that are above costs and which make it difficult or impossible for even equally efficient competitors in the retail market to earn sufficient profits to survive. While the mitigation of incumbent market power is the primary focus of regulation, there is another source of market failure in network industries such as telecommunications. This stems from the existence of network externalities. However, there is no reason why regulatory measures to deal with externalities should involve mandatory access by MVNOs to MNO networks. In telecommunications, such externalities arise as a result of network effects; that is, the value of the network to existing subscribers increases as new subscribers join and are served by that network (or other interconnecting networks). As a result, the marginal social benefit of a new subscriber joining a network exceeds the marginal private benefit, a situation that leads to a sub-optimally low number of subscribers because the decision to join the network is based on the marginal private benefit. Because of this, public policies that help to increase the number of network subscribers (through subsidies, if necessary) are often implemented. Examples include measures to promote universal service and special prices for low-income groups. Such measures generally require some element of regulation because telephone companies acting purely in their own commercial interest do not have the incentive to implement them (i.e., the prices that subscribers pay do not reflect the value of externalities). Requiring mandatory access by MVNOs to MNO networks is neither a means of persuading network operators to extend their network coverage nor a way of introducing special prices for low income groups (unless in the latter case it can be demonstrated that MVNOs will specifically target such groups when acting in their own commercial interests). In sum, any justification for the regulation of MNOs, in particular, mandating the opening of MNO networks to MVNOs through wholesale access at regulated rates must depend on governments or regulators being able first to establish that MNOs exercise market power and prevent the normal operation of competitive markets from which consumers, and society at
133
large, can benefit. Specifically, they need to demonstrate that MNOs possess natural monopoly characteristics and, most importantly, retain vertical control over wholesale wireless services (thus rendering them essential facilities).
4.6. Is Regulatory Intervention Necessary for MVNO Entry into the Mobile Market in Israel?
Our analysis of the wireless retail market in Israel has not revealed clear signs of market failure. Furthermore, we have not been provided with evidence that the possibility of having commercial agreements with any MVNO has been rejected by an Israeli MNO. Hence, at this point in time, we recommend against regulatory intervention to mandate the terms of the MNO-MVNO relationship. However, given that we identified some aspects of behavior that appear to be inconsistent with competitive markets, we do recommend that the MOC proceeds in stages and reconsiders this recommendation after potential MVNOs have been licensed and have actually negotiated with MNOs. If negotiations between MVNOs and MNOs have been unsuccessful and MOC can establish that the reason for the absence of such agreements lies with the refusal of the MNOs to reach an economically justified agreement with the MVNOs, then there would be cause for intervention.
4.7. Are There Policies or Regulatory Measures for MVNO Activity to Have a Positive Effect or in Furthering the MOC Mandate?
We recommend that MOC implements a licensing process for MVNOs. In addition, to further improve the competitive conditions with respect to MVNO entry in Israel, we recommend:

Implementing MNP Encouraging or possibly even mandating shorter term contracts Ensure that ETFs are commensurate with the economic damages caused by early termination Ensure that pricing plans are transparent and that consumers understand the different plans and options available to them Encourage intermodal competition from emerging technologies, such as WiFi and WiMax
In addition, we encourage MOC to review their own list of issues to ensure that conditions in the mobile market in Israel are consistent with those in competitive markets and thus open to potential MVNO entry, if such entry is economically viable.
4.8. MVNO Policy Recommendation for Israel

As noted above, at present we recommend against the adoption of a policy of mandating the terms of MVNO access. Rather, we recommend allowing market forces the opportunity to
134
work by licensing MVNOs and encouraging them to negotiate with the MNOs. Should these negotiations be successful within a timeframe sufficient to allow such agreements to be reached (i.e., 6 months after the first licensed MVNO has contacted an MNO in writing expressing its interest in negotiating a wholesale access contract or 9 months after the MOC gives public notification that it intends to issue MVNO licenses, whichever is longer), then we recommend MOC to continue refraining from intervening in the wholesale market. However, should no wholesale agreements be reached within this time period, we recommend MOC to carefully review the reasons behind each of these failed negotiations. Should this investigation determine that the failed negotiations are not the result of anticompetitive behavior but rather the lack of a viable MVNO business plan under reasonable access terms, we recommend MOC to continue refraining from regulation. If this analysis reveals that the failed negotiations are due to anticompetitive tactics by the MNOs, we recommend MOC to consider implementing wholesale access regulation, according to the guidelines detailed below. Our recommendation is summarized in the figure below:
Figure 40 MVNO Policy Recommendation for Israel
Guidelines for intervention and pricing principles Regulatory intervention can take several forms, ranging from mandatory wholesale access only (and reliance on commercially negotiated prices and terms and conditions) to a full fledged intervention, regulating wholesale access rates, as well as the non-price aspects of the relationship, such as quality of service. If upon thorough investigation, the MOC finds that the absence of MVNOs in Israel is due to market failure (rather than the absence of a viable business case at reasonable access charges), it must select the proper level of intervention. The objective of the MOCs intervention should be to mandate the MNOs to offer reasonable
135
prices and terms and conditions for access to their networks. Specifically, MOC might consider, requiring MNOs to:
Offer access to technical interfaces or other key technologies, necessary for MVNOs to interoperate with the MNOs. This access can be offered at cost, at cost plus a reasonable return, or for free Offer joint use of facilities, such as switching centers and cell sites, at cost, cost plus a reasonable return, or for free Offer the services necessary to guarantee the interoperability of the MVNOs products and services and the MNOs underlying network Provide access to the operating support systems or to information systems with similar functions Set the wholesale quality of service metrics at comparable levels as those offered to the MNOs own customers Generally offer interconnection of networks and other resources
An important aspect of potential intervention is the setting of wholesale prices. In this respect, the MOCs objective should be to set prices that would mimic the prices in a competitive market. This can be achieved by either requiring the operators to charge reasonable prices for wholesale access by negotiating with potential MVNOs in good faith, or by regulating these prices. Either way, the prices set must neither be excessive, nor should they be too low. Regulators frequently use LRIC, with a mark up to recover common fixed costs, to set wholesale access prices. It is important that the wholesale prices must allow the MNOs at a minimum to recoup their investments, as otherwise regulation runs the risk of promoting inefficient entry.
4.9. Recommended Business Models for MVNO Success and Furtherance of the MOC Mandate
We conclude that the successful MVNO business model in Israel is likely to be limited mainly to convergence communications companies seeking to offer triple or quadruple play. These MVNOs should ideally come into being as the result of commercial negotiations and be subject to market forces, rather than regulatory intervention. In our view this will further the MOC mandate to increase competition and improve customer welfare.
136
APPENDIX A OUTGOING RPM ANALYSIS
Partner
Pre-paid On-net Off-net Fixed International Simple average of peak and off-peak Private post-paid On-net Off-net Fixed International Peak and off-peak are the same Business On-net Off-net Fixed International Peak and off-peak are the same Subscription Private post-paid Business
Average revenue per minute (NIS) 2004 0.61 1.08 0.67 1.99 2005 0.60 0.89 0.65 1.94 2006 0.57 0.80 0.63 1.78
0.30 0.68 0.52 1.39
0.31 0.58 0.53 1.39
0.31 0.49 0.50 1.22
0.24 0.49 0.43 1.33
0.18 0.45 0.40 1.21
0.13 0.34 0.32 0.91
NIS per month 4.40 38.48 4.74 36.89 6.26 35.79
Source: Data requests responses from Partner, NERA calculations.
137
Cellcom
Pre-paid On-net Off-net Fixed International
Private post-paid On-net Off-net Fixed International
0.37 0.80 0.52 1.21
0.37 0.59 0.50 1.03
0.36 0.51 0.49 1.05
Business On-net Off-net Fixed International
0.26 0.76 0.38 0.98
0.27 0.55 0.37 0.87
0.27 0.43 0.34 0.80
Subscription Private post-paid Business
NIS per month 2.76 2.56 15.68 13.50
5.09 17.06
Source: Data requests responses from Cellcom, NERA calculations.
138
Pelephone
Pre-paid On-net Off-net Fixed International
Private post-paid On-net Off-net Fixed International
0.32 0.77 0.36 1.85
0.29 0.63 0.33 1.89
0.26 0.56 0.30 1.95
Business On-net Off-net Fixed International
0.22 0.67 0.26 1.75
0.20 0.54 0.24 1.81
0.19 0.49 0.23 1.87
Subscription Private post-paid Business
NIS per month 4.00 5.00 21.05 28.86
16.00 36.92
Source: Data requests responses from Cellcom, NERA calculations.
139
APPENDIX B - PROFITABILITY OF ISRAELI MOBILE OPERATORS
140
Source: Company accounts and financial statements
141
APPENDIX C ROCE RATES IN EXCESS OF WACC DO NOT NECESSARILY INDICATE EXCESSIVE RETURNS
The fact that the ROCE is above the WACC does not by itself indicate an excessive rate of return. When comparing the ROCE with the WACC, for the purposes of assessing whether profitability is excessive, it is important to take account of a variety of factors, including:

Margin of error associated with estimates of WACC Fact that return on equity (which is part of WACC) measures the return on all assets in which shareholders have an interest as a result of owning the shares of a company Valuation of capital employed Value of real options Profitability over the life cycle of a product or service
Each of these is discussed below. Margin of Error Associated with WACC When applying the capital asset pricing model to measure the WACC, what is measured is a best estimate of the WACC, around which there exists a substantial margin of error. Specifically, the WACC is calculated as the average of the interest rate on the companys debt and the required rate of return on the companys equity and taking into account the respective proportion of debt and equity in the companys capital structure. The WACC, therefore, is somewhere within the range between the companys cost of debt and its cost of equity. It represents the capital markets overall assessment of the rate of return that should be earned by the company to cover the time value of money and the risk premium. Although there may be doubts concerning the precise figures for the upper bounds of the confidence intervals, the implication of this recalculation is that ROCE needs to be substantially greater than WACC before one can be confident that profitability is excessive. Taking Account of the Total Asset Base Shareholders own a share of all the assets in the companies in which they invest. Such assets can include more than just the physical and financial assets that appear in a companys balance sheet. For example, a company may invest in establishing a brand or in improving its processes. Such investments do not appear in the balance sheet and hence, in order to earn a return on them, it is necessary for the companys accounting rate of return, which is measured as profit relative to physical and financial assets only, to exceed its overall rate of return (profit relative to all assets including investment in brand and better processes). WACC measures the return required on all investment (assets) and hence the accounting rate of return needs to exceed WACC if there are material investments in brand and better processes etc which are not reflected in the balance sheet.
142
Valuation of Capital Employed Fixed assets in company accounts are normally valued on a historic cost basis and this is the case for the mobile operators in Israel. If the price of new assets is changing over time, either as a result of general inflation or because of technological progress, valuing fixed assets on a historic cost basis will not correctly measure their true value. For example, if asset prices are tending to fall over time, historic cost accounting will exaggerate the value of the asset base and hence capital employed. If asset prices are rising, the opposite will be true and capital employed will be understated. Whether asset prices are falling or rising in Israel is an empirical question. In Analysyss LRIC model developed for MOC in 2004, the prices of virtually all assets were assumed to be falling, with only site acquisition, preparation and infrastructure costs remaining constant. As a result, the use of historic costs to value assets will exaggerate capital employed and understate the true value of ROCE. Consequently, ignoring other factors, the companies would not need to earn a ROCE (based on historic cost accounting) that is as high as WACC. Real Options If there is uncertainty and investment is irreversible (i.e., there are sunk costs), then there is value in having the option to delay investment if, for example, greater information about the market or about costs can be gleaned during the period when investment is delayed. Put another way, there is cost associated with investing now and giving up the option to delay. In such circumstances, the return required from investing in physical assets will, other things being equal, be greater than the return from investing in the shares of a company, which can be sold at any time. This in turn means that ROCE should be expected to be larger than WACC. The fact that real options can exist in particular circumstances, for example where investment is being made in next generation technologies in the fixed network and there is uncertainty about demand and costs and the option to delay exists, has been recognized by Ofcom, the UK regulator.245 The question is whether such circumstances exist in the case of mobile operators in Israel. The answer is not clear, although 3G would appear to share at least some of the risk and uncertainty characteristics of next generation technologies in the fixed network. How important such effects are is not known. However, it seems likely that they would justify some gap between ROCE and WACC. Service Life Cycle During the life cycle of a service, it is to be expected that there will be variations in the level of profitability. For example, in a competitive market, a typical profile would have losses being made in the early years, followed by a period of time when profitability is above WACC as the firm concerned is able to benefit from first mover or other advantages that stem from its investment, followed finally by a period when profitability falls back as a result of the impact of competition.
245
See, for example, Ofcoms Approach to Risk in the Assessment of the Cost of Capital, Ofcom, 26 January 2005, Chapter 6.
143
Hence, even if it were appropriate for the internal rate of return of an investment over its lifetime to equal WACC, there will be substantial variation in ROCE on an annual basis. In some years ROCE will be above WACC and in others it will be below it. Care therefore has to be taken that false conclusions regarding the level of profitability are not drawn as a result of focusing on individual years or too limited a period of time. In the analysis presented above, data has been derived showing annual ROCE and for each mobile operator and a five year average. In our view is that, for the reasons just given, it is more appropriate to use the five year average figure than the annual figures when comparing ROCE with WACC.
144
APPENDIX D SECTOR-SPECIFIC ROCE METHODOLOGY
[]
145
Appendix E Normalized Israeli Churn Rates

Partner 2006 Year ave Monthly Subs ('000)Churn rate Business 556 0.50% Post-Paid 1275 0.90% Pre-Paid 768 2.60% Total 2599 1.32% Year ave Monthly Subs ('000)Churn rate Business 556 0.50% Post-Paid 483 0.90% Pre-Paid 1559 2.60% Total 2599 1.83%
If 60% prepaid
Cellcom
2006 Year ave Monthly Subs ('000)Churn rate Business 713 0.41% Post-Paid 1062 0.65% Pre-Paid 942 2.08% Total 2717 1.08%
If 60% prepaid
Year ave Monthly Subs ('000)Churn rate Business 713 0.41% Post-Paid 374 0.65% Pre-Paid 1630 2.08% Total 2717 1.45%
Weighted average for Cellcom and Partner
1.64%
146
APPENDIX F DISCOUNTED CASH FLOW ANALYSIS
Table F1: Value to an MNO of Adding a Single Customer Using Postpaid Service [] Table F2: Value to an MNO of Adding a Single Customer Using Prepaid Service [] Table F3: Value to an MNO of Adding a Single Customer Using Wholesale Postpaid Service [] Table F4: Value to an MNO of Adding a Single Customer Using Wholesale Prepaid Service [] Table F5: Value to an MVNO of Adding a Single Customer Using Postpaid Service [] Table F6: Value to an MVNO of Adding a Single Customer Using Prepaid Service []
147

1328

Enviado por

Dados do documento

Descrição original:

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

1328

Enviado por

Direitos autorais:

Formatos disponíveis

MOBILE VIRTUAL NETWORK OPERATORS (MVNOS) IN ISRAEL

ECONOMIC ASSESSMENT AND POLICY RECOMMENDATION

Final Report: 1 August 2007

Note: "[]" indicates confidential content that has been removed

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

2.1. MVNO Definitions

NERA Economic Consulting

2.2. Historical Overview

2.3. Technical Considerations of MVNOs

NERA Economic Consulting

2.4. MVNO Integration

2.5. MVNO Operational Components

NERA Economic Consulting

NERA Economic Consulting

Figure 1 MVNO Operational Components

2.6. MVNO Business Models

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

REVIEW OF THE ACTIVITY AND IMPACT OF MVNO PLAYERS WORLDWIDE

3.1. United States

NERA Economic Consulting

200,000,000 Us Mobile Subscribers

NERA Economic Consulting

Others Tracfone Virgin T-Mobile Sprint Nextel Alltel

AT&T Mobility Verizon Wireless

NERA Economic Consulting

See FCC Eleventh CMRS Report, 216. Ibid., 2.

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

NERA Economic Consulting

MVNO Subscribers (b)

Total Wireless Subscribers (c)

NERA Economic Consulting

NERA Economic Consulting

Figure 5 Projected MVNO Revenues in US$ 20022010

Projected MVNO Revenues in US (in billions)

$5.00 2.20 0.20 $0.00 2002 2003 2004 1.00

Source: ARC Group

NERA Economic Consulting

NERA Economic Consulting

Figure 6 US Average Revenue Per User 19932005

NERA Economic Consulting

Figure 7 US Average Revenue Per Minute 19932005

NERA Economic Consulting

Figure 8 Monthly Consumer Price Index Wireless Telephony 19982005

3.2. United Kingdom

NERA Economic Consulting

NERA Economic Consulting

Figure 10 UK Total Wireless Market Shares (by subscribers)

Virgin Mobile UK Tesco Mobile

NERA Economic Consulting