Canadian Radio-television and Telecommunications Commission

Proceeding Nº 2006-14

Review of regulatory framework for wholesale services and definition of essential service

Report of Michael Osborne
6 September 2007

Affleck
Greene
Orr  LLP

Affleck Greene Orr LLP created the original version of the document, which has been translated by a third party.

Contents

I Executive summary

II Introduction

III Legislative and regulatory framework

IV Essential facilities in Canadian competition law

V Defining "essential facilities" for purposes of telecommunications regulation

VI What facilities are essential?

VII How should essential facilities be priced?

VIII When should future reviews be conducted?

IX What regulatory régime should apply to non-essential services?

I Executive summary

1 A review of the law in Canada as it pertains to abuse of dominance suggest that it is not certain whether denial of access to an essential facility can constitute an anti-competitive act, or what the test would be.

2 Nevertheless, I propose that the CRTC adopt a test mandating access to essential facilities and services that is based on the test derived from US jurisprudence, with modifications to make it consistent with Canadian competition law.

3 As all parties agree, it is clearly in the public interest that the CRTC should continue to mandate access to interconnection services.

4 So far as wholesale access services are concerned, residential and business end-user markets are different. In residential markets, access should no longer be mandated for areas where facilities-based competition from cablecos is present.

5 In business markets, there appears to be less facilities-based competition. The CRTC should adopt an ex post regulatory approach that allows wholesale access-dependent competitors to demonstrate that continued is necessary in particular markets.

6 There is also broad agreement among the parties that access to support structures should continue to be mandated.

7 To set terms of access to mandated facilities, the CRTC should adopt a "negotiate first" approach with strict timelines and speedy final offer arbitration.

8 In theory, periodic reviews should not be necessary, as the parties could apply for mandatory access, or to terminate mandatory access, as the case may be. Nevertheless, the CRTC should conduct a review in five years to ensure that the principles governing mandatory access continue to be respected.

9 The CRTC should phase out mandatory access to non-essential facilities.

II Introduction

A. Scope of this report

10 I have been asked to prepare provide comments and recommendations from the perspective of a practising competition lawyer relating to the questions posed at paragraph 26 of Telecom Public Notice CRTC 2006-14. The issues, in brief, are:

a) What definition should the CRTC adopt for "essential services"?

b) What facilities, functions and services should be considered essential?

c) What pricing principles should apply to essential services?

d) When should future reviews be conducted?

e) What regulatory regime should apply to non-essential services?

11 The CRTC has also asked me to comment on the competition law régime applicable to services or markets the CRTC forbears from regulating.

B. My qualifications

12 This report represents my views on the above issues, and not those of the CRTC or any party to the proceeding.

13 I am a barrister and solicitor called to the bar in Ontario in 1998. I practise primarily competition law and commercial litigation, but also some constitutional law (division of powers, taxation powers, etc), and administrative law. I am the chair of the Reviewable Matters / Unilateral Conduct Committee of the National Competition Law Section of the Canadian Bar Association. I am also an active member of the American Bar Association, particularly the Sherman Act Section 2 committee. I write frequently on competition law and commercial litigation topics in our firm newsletter, The Litigator,¹ and, less frequently, for other publications. Most recently, I assisted in editing the CBA's first competition law book, tentatively titled Fundamentals of Competition Law. I attach a copy of my curriculum vitae.

14 I have been following developments in the telecommunications industry since Rogers announced its acquisition of Microcell. At that time I subscribed to press releases from several telecos. Last year, I was responsible for leading the team that drafted the CBA Competition Law Section's comments on the Competition Bureau's draft Information Bulletin on the Abuse of Dominance Provisions as Applied to the Telecommunications Industry ("TAB").

15 I have been involved in four major contested proceedings in the Competition Tribunal and numerous matters that did not result in proceedings. I will mention a few of the more relevant cases that I have worked on.

16 I acted for the Commissioner (as junior) in the merger of Air Canada and Canadian Airlines in 1999. This case is of some significance because the Commissioner obtained a number of undertakings from Air Canada that suggest an essential facilities approach. Air Canada agreed to grant access to Aeroplan to other carriers; surrender certain airport facilities, including gates, ticket counters, and slots at Pearson; and enter into interlining and joint fare agreements with other carriers.²

17 I also acted, again as second chair, in the Commissioner's abuse of dominance case against Air Canada, which alleged that Air Canada had cut fares and increased capacity in order to deter entry in the eastern Canada market by WestJet and CanJet. Only the first phase of this case was heard by the Competition Tribunal. It concerned the definition of Air Canada's avoidable costs for purposes of determining whether Air Canada was operating below its avoidable costs.

18 Commission of Competition v. Canadian Waste Services involved a merger between two landfill companies. I acted for the Commissioner, as second chair, in a challenge to CWS' acquisition of the Ridge Landfill in southwestern Ontario. We argued, and the Tribunal accepted, the merger would substantially prevent competition, which makes it of some relevance to the present proceeding.

19 Finally, and most recently, I acted for B-Filer Inc. in its application under section 75 (refusal to deal) to force Bank of Nova Scotia to provide B-Filer with certain banking services that B-Filer said were essential to its business. This was the first private application under section 75 to proceed to a full hearing.

20 I do not have any formal training in economics. However, competition law is a mixture of law and economics. The language and concepts employed by economists have entered into the vocabulary of competition lawyers such as me because these concepts have been adopted in the caselaw. Where I use and discuss economic concepts, it is because they have been adopted by the cases, and not because I pretend to any expertise in the underlying economics.

C. Terminology

(1) "Essential facilities"

21 In its notice and past decisions, the CRTC has used the term "essential services", which includes facilities, functions and services. The Telecommunications Act distinguishes between "telecommunications facilities" and "telecommunications services".³

22 As the term usually used by competition lawyers is "essential facilities", I shall use that term, but with the same meaning as "essential services" as the CRTC uses that term. When referring to the "essential services" as defined by Telus, I use the term "essential retail services".

(2) "Facilities-based"

23 As discussed below, the government re-defined "facilities-based" to include carriers that lease facilities, in its order varying the CRTC's Local Forbearance decision.

24 In this report I use "facilities-based" in the same way as the Companies use "end-to-end facilities-based"; that is, a carrier that uses its own facilities and does not lease them from another carrier..

III Legislative and regulatory framework

25 This proceeding is subject to a patchwork of legislation and subordinate legislation whose provisions are not wholly internally consistent. I will attempt, in this section, to make some sense of them.

A. Telecommunications Act 1993

26 Canada's telecommunications policy is set out at section 7 of the Telecommunications Act 1993:

Objectives

7. It is hereby affirmed that telecommunications performs an essential role in the maintenance of Canada's identity and sovereignty and that the Canadian telecommunications policy has as its objectives

(a) to facilitate the orderly development throughout Canada of a telecommunications system that serves to safeguard, enrich and strengthen the social and economic fabric of Canada and its regions;

(b) to render reliable and affordable telecommunications services of high quality accessible to Canadians in both urban and rural areas in all regions of Canada;

(c) to enhance the efficiency and competitiveness, at the national and international levels, of Canadian telecommunications;

(d) to promote the ownership and control of Canadian carriers by Canadians;

(e) to promote the use of Canadian transmission facilities for telecommunications within Canada and between Canada and points outside Canada;

(f) to foster increased reliance on market forces for the provision of telecommunications;

(g) to stimulate research and development in Canada in the field of telecommunications and to encourage innovation in the provision of telecommunications services;

(h) to respond to the economic and social requirements of users of telecommunications services; and

(i) to contribute to the protection of the privacy of persons.[Emphasis added.]⁴

27 As is often the case with statutes, the purpose clause of the Telecommunications Act 1993 contains something for everyone. References to "efficiency", "increased reliance on market forces", "encourage innovations", and the like, highlighted above, dovetail with a presumption against regulation and in favour of market forces, with reliance on mandated wholesale unbundling only where absolutely necessary.

28 That being said, the Act does not itself mandate this approach. As well, certain other of the Act's purposes suggest less reliance on market forces, such as "facilitate the orderly development", "promote the ownership and control of Canadian carriers by Canadians", and "respond to the economic and social requirements of users of telecommunications services". These purposes do not negate increased reliance on market forces, however.

29 Moreover, the Act retains a presumption that the CRTC will regulate tariffs for telecommunications services (including any services that are incidental to the business of providing telecommunications services). Sections 23-33 deal generally with regulation of services, particularly tariff regulation.

30 Section 34, however, permits, and in some cases, requires, the CRTC to forbear from regulating telecommunications services. Forbearance can be partial or total, and conditional or unconditional. The basic test for forbearance is whether forbearance would be consistent with the Canadian telecommunications policy objectives set out in section 7.

31 Subsection 34(2) requires the CRTC to forbear from regulating

Where the Commission finds as a question of fact that a telecommunications service or class of services provided by a Canadian carrier is or will be subject to competition sufficient to protect the interests of users...

32 However, the CRTC retains considerable leeway to forbear to the extent that it considers appropriate; and forbearance under this provision can be conditional or unconditional.

33 It is also noteworthy that forbearance consists of the CRTC refraining from exercising its powers and duties under sections 24, 25, 27, 29, and 31. It does not relieve carriers themselves of any duties or prohibitions. Thus forbearance would not relieve a carrier of the obligation to file tariffs pursuant to section 25, nor of the obligation to charge rates that are just and reasonable (s. 27(1)) and that do not unjustly discriminate between customers (s. 27(2)). Indeed, the forbearance provisions show signs of poor legislative drafting. For instance, were the CRTC to refrain from exercising its duty to approve agreements respecting interchange of telecommunications under section 29, carriers would theoretically be prohibited from entering into such agreements.

34 There is no special statutory régime relating to mandatory access. Rather, the CRTC has a general power to impose conditions on TSPs, principally under section 24, that would appear to be the statutory basis for mandatory access. Thus, while tariffs are presumptively regulated, with the possibility of forbearance, mandatory access is not presumed in the Telecommunications Act. So far as mandatory access is concerned, therefore, it is arguable that the Telecommunications Act already contains a presumption of reliance on market forces.

35 The Telecommunications Act gives the government considerable power to direct the CRTC's decision-making. The government can issue directions to the CRTC under section 8:

Directions

8. The Governor in Council may, by order, issue to the Commission directions of general application on broad policy matters with respect to the Canadian telecommunications policy objectives.

36 Pursuant to section 11, these directions are binding on the CRTC.

37 Section 12 gives the government the power to vary or rescind decisions of the CRTC.

38 On December 14, 2006, the Governor in Council made an Order Issuing a Direction to the CRTC on Implementing the Canadian Telecommunications Policy Objectives⁵ (the "Policy Direction"). This order was made under section 8.

39 The Policy Direction relates directly to this proceeding, as, among other things, it instructs the CRTC to review its framework for mandatory access to wholesale services.

40 The Policy Direction is divided into three parts. The first part articulates two general principles:

1. In exercising its powers and performing its duties under the Telecommunications Act, the Canadian Radio-television and Telecommunications Commission (the "Commission") shall implement the Canadian telecommunications policy objectives set out in section 7 of that Act, in accordance with the following:

(a) the Commission should

(i) rely on market forces to the maximum extent feasible as the means of achieving the telecommunications policy objectives, and

(ii) when relying on regulation, use measures that are efficient and proportionate to their purpose and that interfere with the operation of competitive market forces to the minimum extent necessary to meet the policy objectives;

41 Both of these principles originate from recommendations contained in the Telecommunications Review Panel Report, which recommended that Canada move from a presumption of regulation to a presumption against regulation.

42 It is odd that the Policy Direction uses the word "should" rather than "shall" in the opening phrase of subsection 1(a). The Interpretation Act provides that "shall" is imperative and "may" is permissive.⁶ "Should" is the past tense of "shall", but is commonly used to express that one "ought" to do something rather than one "must". While this might be taken to suggest that the CRTC has some discretion to depart from these principles, I would not place much weight on this distinction. The Policy Direction clearly instructs the CRTC to rely on market forces as much as possible, and on regulation, as little as possible.

43 The second part of the Policy Direction articulates criteria to be met when the CRTC does rely on regulation. These criteria should be interpreted as fleshing out the broad principle in paragraph 1(a)(ii) that regulation must be efficient, proportionate, and minimally intrusive.

44 The criteria in subsection 1(b) reflect two important concerns:

a) The effect of regulation on incentives. The CRTC is thus instructed to walk the fine line between deterring economically efficient entry and promoting inefficient entry.

b) Non-economic regulation (such as regulated access to networks) must be symmetrical and competitively neutral as between market participants and technologies.

45 The third part of the Policy Direction contains some specific instructions to the CRTC, including conducting the present review

(ii) with a view to increasing incentives for innovation and investment in and construction of competing telecommunications network facilities, to complete a review of its regulatory framework regarding mandated access to wholesale services, to determine the extent to which mandated access to wholesale services that are not essential services should be phased out and to determine the appropriate pricing of mandated services, which review should take into account the principles of technological and competitive neutrality, the potential for incumbents to exercise market power in the wholesale and retail markets for the service in the absence of mandated access to wholesale services, and the impediments faced by new and existing carriers seeking to develop competing network facilities.⁷

46 As I read this instruction, the regulatory goal is to increase incentives for innovation and development of new facilities. The means is a review of the regulatory framework for mandated wholesale access.

47 In addition, this review must take into account three factors:

a) "the principles of technological and competitive neutrality";

b) "the potential for incumbents to exercise market power in the wholesale and retail markets for the service in the absence of mandated access to wholesale services"; and

c) "the impediments faced by new and existing carriers seeking to develop competing network facilities"

48 If the overall regulatory approach articulated in subsection 1(a) seems clear enough, the particulars set out in subsections (b) and (c) introduce a certain degree of ambiguity into the Policy Direction. For instance, the Policy Direction appears to require phasing out mandatory access to non-essential services, yet prefaces this with "the extent to which", which suggests that the CRTC might continue to require access to non-essential services.

49 In my opinion, any ambiguity in the particulars in subsections 1(b) and (c) should be resolved in favour of the general principles articulated in subsection 1(a). It would, in my view, be inconsistent with the principle of maximum reliance on market forces to continue to mandate access to non-essential services.

50 I would note, parenthetically, that two of the factors listed in paragraph 1(c)(ii) may have a bearing on the approach to essential facilities to be taken by the CRTC. The direction to consider the extent to which incumbents might exercise market power in the absence of wholesale access suggests that this should form part of the analysis of what services are essential. The reference to the "impediments faced by new and existing carriers seeking to develop competing network facilities" is consistent with an approach that would prioritize the development of competing network facilities over resale arrangements, as is the statement of the regulatory goal of increasing incentives for innovation and investments in facilities. It is also worth noting the inclusion of existing carriers in this instruction. This points to the importance of not disincenting the development of new facilities by existing carriers.

51 On April 4, 2007, the Governor in Council made the Order Varying Telecom Decision CRTC 2006-15⁸ (the "Local Forbearance Variation Order"). This order was made under section 12 to vary the CRTC's decision on Forbearance from the regulation of retail local exchange services.

52 Strictly speaking, the Local Forbearance Variation Order does not govern this proceeding. It is relevant in two ways, however. First, it provides additional guidance in implementing the spirit of the Policy Direction. Second, forbearance from rate regulation in local markets may have an impact on mandatory wholesale access.

53 The Local Forbearance Variation Order has an extensive preamble, which explains the variations made in the order and further articulates the policy direction that the Governor in Council wishes the CRTC to take. These include:

a) An express statement of preference for facilities-based competition:

Whereas the Governor in Council considers that facilities-based competition is a durable form of competition that delivers the greatest benefits to consumers, imposes competitive market discipline on incumbents and strengthens investment in telecommunications infrastructure;

b) A distinction between business and residential markets:

Whereas the Governor in Council considers that local business markets and local residential markets should be considered separately;

c) A definition of geographic markets as local exchanges:

Whereas the Governor in Council considers that local exchanges are the appropriate geographic component of a relevant market, as they often reflect a social and economic community of interest and are less likely than LFRs to contain pockets of uncontested customers;

d) A suggestion that mobile wireless technology should be considered as part of the same relevant product market as landline telephony:

Whereas the Governor in Council considers that the use of mobile wireless technology by consumers is increasing and will likely continue to increase, and that for many consumers the exclusive use of mobile wireless services is an increasingly attractive alternative to wireline local exchange services;

54 The operative part of the Local Forbearance Variation Order varies the CRTC's Local Forbearance Decision in a number of important respects. First, consistent with the preamble, it defines the relevant geographic market as a local exchange.

55 Second, it changed the test for local forbearance. The original decision required ILECs to demonstrate a 25% market share loss, plus certain quality of service requirements.⁹ The new test replaces the market share loss test with three alternative branches, which are worth quoting in full:

a) the ILEC demonstrates that one of the following circumstances exists in the relevant market:

(i) that the ILEC does not have market power, based on the criteria set out in paragraph 213,

(ii) that, if the ILEC offers residential local exchange services, there are, in addition to the ILEC, at least 2 independent facilities-based telecommunications service providers, including providers of mobile wireless services, each of which offers local exchange services in the market and is capable of serving at least 75% of the number of residential local exchange service lines that the ILEC is capable of serving, and at least one of which, in addition to the ILEC, is a facilities-based, fixed-line telecommunications service provider, or

(iii) that, if the ILEC offers business local exchange services, there is, in addition to the ILEC, at least one other independent facilities-based, fixed-line telecommunications service provider that offers local exchange services in the market and is capable of serving at least 75% of the number of business local exchange service lines that the ILEC is capable of serving;¹⁰

56 The criteria in paragraph 213 are the criteria proposed by the Competition Bureau:

213. The Competition Bureau proposed that the Commission should adopt a structured rule of reason (SROR) approach that could serve as the basis for streamlined analysis of ILEC requests for local exchange service forbearance, once the relevant product market had been identified. In the Competition Bureau's view, this approach used the following set of conditions that, if satisfied, should be sufficient for the Commission to conclude that an ILEC did not possess market power in the provision of local exchange services:

At least two independent facilities-based service providers must exist, the ILEC and a facilities-based entrant, capable of offering local service that has been determined to fall within the relevant product market for ILEC local service;

• The entrant was able to obtain and retain a customer base;

• The entrant's variable costs of providing local service are similar to or lower than the ILEC's variable costs of providing local service;

• Neither the ILEC nor the entrant was capacity-constrained;

• There was evidence of vigorous rivalry between the ILEC and the entrant in the provision of local service; and

• Industry characteristics are such that the ILECs are unlikely to engage in anti-competitive behaviour.

57 It should be emphasized that the Competition Bureau's test is expressly stated in both the Local Forbearance Decision and the Local Forbearance Variation Order as a test for lack of market power. That is, if an ILEC can demonstrate the existence of these criteria, the ILEC is deemed not to have market power in the relevant geographic market.

58 The definition of "independent facilities-based telecommunications service provider" is important. "Facilities-based" is not defined in the Local Forbearance Decision. From the way the term is used, it seems to connote a telecommunications service provider that owns its own facilities, or at any rate, does not lease them from an ILEC. The term appears to have the same meaning in the CRTC's Local Competition Decision.¹¹ The Competition Bureau appears to use this term in the same sense in its submission to the CRTC in the local forbearance proceeding.¹² From the CRTC's description of submissions by other parties, this appears to have been the common understanding of the term.

59 The Local Forbearance Variation Order expressly defines "facilities-based" for purposes of two of the three branches of the test. This definition considers CLECs that lease facilities from other providers, presumably ILECs, as facilities-based TSPs:

243. For the purposes of subparagraphs 242a)(ii) and (iii) and paragraph 523, the Commission considers that a telecommunications service provider is independent if it does not have the same owner as, and is not affiliated with, any other service provider referred to in the respective subparagraph. Further, for the purpose of those provisions, the Commission considers that a facilities-based telecommunications service provider is one that provides services in the relevant market either by using its own facilities and services or by using a combination of its own facilities and services together with those leased from other service providers.

60 This expanded definition of "facilities-based" does not apply to the first branch of the test for forbearance, that is, the Competition Bureau's test.

61 The revised definition of "facilities-based" contained in the Local Forbearance Order has the potential to undermine the goal of reducing reliance on wholesale access, as the Competition Bureau notes in its Supplementary Evidence.¹³ If the CRTC forbears from regulating a market because of the presence of competitors using leased lines, and there are no true facilities-based competitors present, it may be necessary to continue to mandate access at the wholesale level.

IV Essential facilities in Canadian competition law

62 There are three provisions in the Competition Act that could provide the locus for the essential facilities doctrine in Canadian competition law: the abuse of dominance provision (s. 79), the refusal to deal provision (s. 75), and the special remedies for misuse of intellectual property (s. 32).

63 As discussed in greater detail below, denial of access to an essential facility has yet to be accepted by the Competition Tribunal as an "anti-competitive act" for purposes of the abuse of dominance provision, although there is some indication that refusals to deal by dominant firms may in some circumstances be an anti-competitive act. However, the refusal to deal provision could be characterized as an essential facilities provision of a sort, as could the misuse of intellectual property provision. The remedies for misuse of intellectual property are not relevant to telecom facilities; thus I do not address them.

A. Abuse of dominance

64 Sections 78 and 79 deal with situations where a firm that is dominant in a particular market in Canada engages in anti-competitive conduct in that market, or another market, in order to extend or maintain that dominant position.¹⁴

65 Section 79(1) defines the elements of abuse of dominant position. Section 78 lists examples of anti-competitive acts.

66 The elements of abuse of dominance are three-fold:

a) Dominance;

b) Practice of anti-competitive acts; and

c) Substantial lessening or prevention of competition

67 In the Canada Pipe appeal, the Federal Court of Appeal held that each of the three elements set out in section 79 must give rise to discrete legal tests.¹⁵

68 The Act defines dominance by the phrase "substantially or completely control". This indicates that to be dominant, a firm must have market power.¹⁶ Testing for market power necessarily involves defining the relevant market. The phrase "throughout Canada or any area thereof" expresses the geographic aspect of market definition, and the phrase "class or species of business", the product aspect.¹⁷

69 The definition of dominance contemplates that two or more firms might enjoy joint dominance over a market as a result of co-ordinated behaviour. However, most abuse of dominance cases involve one dominant firm.

70 The dominance element thus involves two questions:

a) What is the relevant market (both product and geographic)?

b) Does the firm have a substantial amount of market power over the relevant market ("substantially or completely control")?

(i) Relevant market

71 Often the outcome of abuse of dominance as well as other competition cases turns on market definition. Paradoxically, market definition, whether product or geographic, is not a precise exercise, as the Tribunal has cautioned.¹⁸

72 The relevant market is simply the products and places that compete with the product and place in question. It is the "universe of effective competition"¹⁹ in both product and geographic space.

73 The purpose of defining the relevant market is to identify the possibility for the exercise of market power.²⁰ Market power is, in turn, defined as the ability of a firm or firms to maintain prices above the competitive level over a sustained period of time.²¹

(ii) Product market

74 "Substitutability" is the "fundamental test or ‘touchstone' for determining the boundaries of the relevant product market".²² Products that are close substitutes for one another are in the same market. Close substitutes are identified by looking at evidence that buyers are willing to switch from one product to the other in response to a relative change in price.

75 Evidence of substitutability can be both direct and indirect. Direct evidence of substitutability includes statistical evidence of whether buyers switch to another product in response to small changes in the price of the product at issue²³ and anecdotal evidence of switching behaviour.²⁴ Lack of evidence of buyers actually switching between products is a strong indication that they are not in the same market.²⁵

76 Often direct evidence of substitutability is not available and indirect evidence must be used. Indirect evidence of substitutability consists, broadly speaking, of evidence about the characteristics of the product and its buyers. Exactly which characteristics must be considered varies from case to case. Types of indirect evidence that have been considered include:

a) End use / functional interchangeability: the question is whether the two products can be put to the same end use by a purchaser. If they can, they might be in the same market. Functional interchangeability has thus been described as a "preliminary filter"²⁶ or a "necessary but not sufficient condition"²⁷ for finding that two products are in the same market. Even defining the end use can raise issues. Defining it too broadly results in over-inclusion. In Tele-Direct, the Tribunal noted the danger of placing automobiles and bicycles in the same market just because both provide a means of transport.²⁸

b) Physical and technical characteristics: a comparison of the physical and technical characteristics of the products can show whether one product is better or more valuable than another, or more suited to a particular end use.²⁹

c) Views, strategies, behaviour and identity of buyers: do buyers of the products consider them to be close substitutes?³⁰

d) Trade views, strategies and behaviour (inter-industry competition): do industry participants, including the respondent, consider the products to be close substitutes or competitors of one another?

e) Switching costs: the higher the cost of switching between two products, the less likely it is that they are in the same market.³¹

f) Price relationships and relative price levels: a lack of correlation between the price movements of two products over time suggests that they are not in the same market. Relative price levels between two products in two separate geographic markets can also indicate whether they are in the same market or not.³²

77 The Tribunal has been willing to define product markets narrowly even where there are products that can be substituted for the product under examination for some, but not all purposes. For instance, in Canada Pipe, the Tribunal found that the demand for cast iron pipe was falling over time as plastic pipe gradually became accepted for applications previously reserved to cast iron pipe. Nevertheless, because there remained applications for which only cast iron pipe was acceptable, it constituted a separate product market, the Tribunal held.³³ Similarly, in NutraSweet, other sweeteners were substitutable for aspartame in some, but not all applications.³⁴

(iii) The hypothetical monopolist test

78 The test for market definition typically used in merger cases is the "hypothetical monopolist" test. This test works by provisionally adopting a market delineation and asking whether a hypothetical monopolist over that market would impose a small but significant price increase (usually assumed to be 5%) and sustain it for one year (hence, a small but significant non-transitory increase in price, or SSNIP). If the answer is yes, that is the market. If the answer is no, this means that other products would constrain this price increase; thus the market is expanded to include those other products.³⁵

79 The Tribunal has suggested that hypothetical monopolist test may not be appropriate in abuse of dominance cases. The question of whether a hypothetical monopolist can raise prices depends on the assumption that prices are at competitive levels. By contrast, abuse cases necessarily involve an allegation that the respondent firm has a dominant position already.³⁶ It is unsafe to assume prices are at competitive levels if the monopolist is not hypothetical, but real. The Tribunal has, however, applied the hypothetical monopolist test in a similar situation, where it found that pre-merger prices were not competitive, by defining the market with references to anticipated future, competitive prices.³⁷

(iv) The cellophane fallacy

80 A related problem is the so-called "cellophane fallacy". Where a firm with a dominant position has already raised prices above competitive levels, evidence of apparent substitutes can be misleading. This is because supra-competitive prices can make purchasers switch to products they would not in fact choose if prices were at competitive levels, making the products appear to be substitutes for one another when they are not. In those cases it is not appropriate to use the supra-competitive prices to define the market. The "cellophane fallacy" is named after a case where the US Supreme Court is said to have fallen into this trap by finding that cellophane wrap and other flexible wraps such as wax paper were in the same market.³⁸

81 One way to avoid the cellophane fallacy is to use the likely future, or competitive, price (where future competition is expected to lower prices), as the Tribunal noted in CWS.³⁹ Another is to use other factors for delineating the market than price, as the Tribunal did in three abuse of dominance cases, Laidlaw, Nielsen and Tele-Direct.⁴⁰

(v) Telecommunications product market definition

82 In the TAB, the Bureau cautions that market definition in telecommunications cases may be more difficult because of a lack of data:

In recently deregulated telecommunications markets, there may be added difficulty in assessing evidence and data. For example, in an abuse of dominance analysis, the Bureau may consider evidence, such as business plans, strategic documents and data, during both the regulated and forborne time periods but interpreting this evidence in the context of both time periods may prove difficult.⁴¹

83 The TAB suggests that the Bureau would use the indicia set out above, that is, the evidence of switching, the SSNIP test, and characteristics of the products:

For example, in considering an allegation of abuse of dominance by a firm in its provision of local residential telephone service, the Bureau would first assess the willingness and ability of consumers to substitute to other local residential telephone services provided by different technologies (e.g., circuit switched, IP, and wireless technologies) to determine whether the services provided by these technologies are in the same product market. As part of this exercise, the Bureau would consider whether there are characteristics of a particular local residential telephone service (e.g., reliability and clarity) that sufficiently differentiate it from other local residential telephone services such that customers are unlikely to switch in response to a SSNIP. The Bureau would also consider whether there are costs involved in switching, which would make switching a less likely response to a SSNIP.⁴²

84 The Bureau also suggests that bundling may be relevant to product market definition in telecommunications markets, and that markets might be defined around bundles.

85 To date, there has been only one example of market definition for competition law purposes in telecommunications markets: the Rogers-Microcell merger. In allowing the merger, the Bureau commented that it had concluded that wireline and wireless telephone services were not in the same product market, as there had been little substitution between them. The Bureau concluded, however, that GSM and CDMA wireless telephone services were in the same product market.⁴³

86 As discussed below, identifying substitutes for wholesale inputs involves an exercise in product market definition. However, the existing "market" for wholesale access is defined and priced by regulation. Consequently, we cannot say with confidence whether prices are above, below, or at competitive levels. This suggests that applying the SSNIP test to existing pricing would be inappropriate.

(vi) Geographic market

87 The phrase "throughout Canada or any part thereof" in s. 79(1)(a) refers to geographic market.⁴⁴ The process of delineating the geographic market is in principle very similar to that of delineating the product market. The factors are essentially the same, except that they are applied in a geographic sense. Specifically geographic factors include evidence of foreign competition and imports, and transport costs.⁴⁵

88 Telecommunications services obviously present special geographic market definition issues. As the Bureau points out,

For many telecommunications services, the provision of the service is tied to a location and the number of competitive alternatives that are available to consumers can differ depending on where they live or carry on business.⁴⁶

89 Put another way, telecommunications services must be delivered through some medium, either wireline (including cable and fibre) or wireless. The choices of telecommunications service providers available to any person necessarily depend on the ability of those providers to reach that person's residence, either through facilities they own or lease, or by delivering applications through a broadband internet connection.

90 This fact led to the Bureau's statement in the TAB that every household could theoretically be defined as a relevant geographic market. The impracticality of this leads to aggregation of locations based on the footprints of competing networks.

91 The Local Forbearance Variation Order defined local exchanges as geographic markets. To the extent that the same networks cover the whole of any given local exchange, there is no harm in using local exchanges as relevant geographic markets, even if larger geographic markets could be drawn using network footprints. To the extent, however, that the networks available in a local exchange vary according to locations within the exchange, it may be inappropriate to use that local exchange as a relevant geographic market.

(vii) Market power

92 Market power is the ability to set prices above competitive levels and to maintain them at that level for a significant period of time.⁴⁷ There are both direct and indirect indicators of market power.

93 Performance results and conduct are direct indicators of market power. If firm's performance results (such as profits) or conduct (such as pricing policy) are more likely to be associated with firms that have market power, then they directly indicate that the firm has market power. Conversely, results or conduct similar to that of firms without market power indicates a lack of market power.⁴⁸ Thus evidence of high profits or high prices can be evidence that prices are above competitive levels, and thus, that a firm has market power. The Tribunal relied on evidence of high profits or high prices in Tele-Direct,⁴⁹ CWS,⁵⁰ NutraSweet,⁵¹ Canada Pipe,⁵² and Laidlaw.⁵³ Similarly, price discrimination can be evidence of market power. The Tribunal relied on evidence of price discrimination in Tele-Direct⁵⁴ and CWS.⁵⁵

94 Direct evidence of market power is frequently not available; normally one must consider indirect indicators of market power: market share and barriers to entry.⁵⁶ The Tribunal follows a two-step approach. First, it examines the market share of the respondent. Second, it considers other factors, principally whether there are barriers to entry, but also how many competitors there are in a market and their respective market shares, and how much excess capacity firms in the market have.

(viii) Market shares

95 Very high market shares raise a presumption, or prima facie case, of dominance, absent evidence of easy entry into the market.⁵⁷ Exactly what market share is required to trigger this presumption is uncertain. The respondents' market shares were near monopoly levels in NutraSweet (95%), Laidlaw (87-100%), Nielsen (100%), Tele-Direct (80-96%) and Canada Pipe (80-90%). The Tribunal commented in Laidlaw that a market share below 50% could not give rise to the presumption of dominance.⁵⁸

96 In the Abuse Guidelines and the TAB, the Bureau adopts a 35% threshold: market shares above 35% "prompt further examination"; those below do not.⁵⁹ The Bureau defends this on the basis that the Tribunal did not exclude the possibility that a firm with a market share below 50% could be dominant. However, paragraph 79(1)(a) requires a showing that the respondent substantially or completely controls a class or species of business, not just a showing of some market power. In my view, the possibility that the Tribunal would conclude that a firm with a market share of less than 50% is dominant is remote.

97 The appropriate way to measure market shares is not always obvious. While the Tribunal most frequently compares revenues earned by each market participant, there are other possible measures. Most ways of measuring market share involve comparing respective revenue, output or capacity; but there can be a plethora of industry-specific statistics that could be used.⁶⁰ It is easy to imagine scenarios where using different measures would yield very different results. For instance, where an allegedly dominant firm's competitors have more surplus capacity than it does, the dominant firm's share of capacity will be less than its share of revenues.⁶¹

98 In the TAB, the Bureau indicates that capacity, including network coverage, may be the appropriate measure of market shares in telecommunications markets. Where there are two or more facilities-based providers with substantial excess capacity, the Bureau suggests, the market shares of each would be equal. Economists refer to this as 1/n ("one over n") market shares, where n is the number of such market participants. The Bureau explains the point thus:

When substantial excess capacity remains in a market, allowing firms to easily increase supply in response to an increase in price, the ability to raise price above competitive levels may be considerably lower than what a simple concentration measure might suggest.⁴²

⁴² For example, in geographic markets where there are two independent facilities-based service providers (a facility-based service provider is one that owns and operates its own networks) with sunk costs that are not capacity constrained and are equally capable of offering the relevant product, the capacity market share of each would be 50%.⁶² [Emphasis added.]

99 It follows that, where there are three or more independent facilities-based service providers that are not capacity constrained serving a particular geographic market, none of them can be dominant, no matter that one of them may have a high share of customers or revenues. Where there are two such service providers, it is unlikely that the Tribunal would find one of them to be dominant. How could the Tribunal find one service provider with a 50% market share to be dominant when its competitor also has a 50% market share?

100 In the result, on the analysis proposed in the TAB, the dominance element of the abuse of dominance test would not be met in any market where there is more than one facilities-based service provider that is not capacity constrained.

(ix) Barriers to entry

101 Without the existence of barriers to entry, there can be no dominance, because any attempt to exercise market power would attract immediate entry:

In the absence of barriers to entry, even a single seller cannot exercise market power. Any attempt by the incumbent to price above the competitive level will attract immediate entry by competing sellers.⁶³

102 The question is: how easy – or hard – is it for a firm to start doing business in the relevant market and establish itself as a viable competitor on a sustainable basis.⁶⁴ So-called "hit-and-run" entry,⁶⁵ where firms quickly enter a market and are forced out, does not count.

103 Factors considered by the Tribunal in assessing barriers to entry in abuse of dominance cases include:

a) Observed entry and exit: evidence of sustained (or viable) entry of new competitors tends to demonstrate that entry is easy. However, the nature of the entry is important: limited entry, or entry into niche sub-markets does not rebut a presumption of dominance of the larger market.⁶⁶ Similarly, lack of entry, or evidence that entrants quickly leave, is evidence of high barriers to entry.

b) Sunk costs: costs that cannot be recovered if entry is unsuccessful can be a significant barrier to entry.⁶⁷ In the TAB, the Bureau notes that the telecommunications industry is characterized by large sunk costs and significant economies of scale, density and scope.⁶⁸ This would support a finding that there are barriers to entry into this industry.

c) Cost of entry: even costs that are potentially recoverable may be a barrier to entry.⁶⁹

d) Economies of scale: large economies of scale mean that an entrant must enter on a large scale in order to achieve costs similar to an incumbent. This increases sunk costs and raises barriers to entry.⁷⁰

e) Patents or other intellectual property: entry often requires access to technology protected by patents or other intellectual property; difficulties in obtaining this access raise barriers to entry.⁷¹

f) Incumbent advantages: the incumbent may have certain advantages derived from its position in the market.⁷²

g) Conduct by the incumbent: the incumbent's anti-competitive conduct can itself create barriers to entry.⁷³

h) Industry-specific barriers: there are often industry specific barriers in any given industry, including regulatory barriers.

i) State of the market: a market that is mature and has little potential for growth is unlikely to attract entry, because it is unattractive to investors.⁷⁴

(x) Market concentration and excess capacity

104 In Laidlaw, the Tribunal noted that the number of other competitors in the market and their respective market shares is relevant. In merger cases, evidence that a market is highly concentrated, that is, that a few large competitors control the bulk of the market, suggests that the merger will lessen competition.⁷⁵ The Tribunal has not discussed market concentration extensively in the four contested abuse of dominance cases. In each of those cases, the market was concentrated in the hands of the one dominant firm; the few competitors had low market shares.

105 Excess capacity in the hands of competitors of the allegedly dominant firm weighs against a finding of dominance. These competitors could in theory expand their production and thus their market share. It is easier for an existing competitor with excess capacity to expand production than for a new competitor to enter the market. However, these competitors might have excess capacity because the anti-competitive conduct of the dominant firm has prevented them from expanding output.⁷⁶

(i) The test for identifying anti-competitive acts

106 Section 78 provides a non-exhaustive list of anti-competitive acts. Other conduct can be – and has been – found to be anti-competitive because it has characteristics in common with the conduct listed in s. 78. In NutraSweet, the Tribunal discerned that the common thread running through the list in s.78 was that anti-competitive acts are performed for a predatory, exclusionary or disciplinary purpose:

A number of the acts share common features but, as recognized by the Director and the respondent, only one feature is common to all: an anti-competitive act must be performed for a purpose, and evidence of this purpose is a necessary ingredient. The purpose common to all acts, save that found in paragraph 78(f), is an intended negative effect on a competitor that is predatory, exclusionary or disciplinary.⁷⁷

107 In the Canada Pipe Appeal, Desjardins J.A. adopted this gloss on section 78. She distinguished two aspects to "purpose". The first relates to whether it must be subjective or objective.

[67] First, the meaning of the term purpose deserves some comment. As the Tribunal observed in Tele-Direct, supra at p. 180, "‘purpose' is used in this context in a broader sense than merely subjective intent on the part of the respondent. . . it might be more apt to speak of the overall character of the act in question" (emphasis added). In order to apply paragraph 79(1)(b), the purpose or character of the impugned conduct must therefore be determined. Relevant factors to be considered and weighed to determine this overarching "purpose" include the reasonably foreseeable or expected objective effects of the act (from which intention may be deemed, as I explain further below), any business justification, and any evidence of subjective intent, if available (see Tele-Direct, supra at p. 180).

108 The second aspect is the type of purpose. Here, Desjardins J.A. approached the Tribunal's gloss on section 78 as though it were itself the statutory text. She emphasized that the purpose must relate not to effects on competition, but on competitors:

[68] The second aspect describes the type of purpose required in the context of paragraph 79(1)(b): to be considered "anti-competitive" under paragraph 79(1)(b), an act must have an intended predatory, exclusionary or disciplinary negative effect on a competitor. The paragraph 79(1)(b) inquiry is thus focused upon the intended effects of the act on a competitor. As a result, some types of effects on competition in the market might be irrelevant for the purposes of paragraph 79(1)(b), if these effects do not manifest through a negative effect on a competitor. It is important to recognize that "anti-competitive" therefore has a restricted meaning within the context of paragraph 79(1)(b). While, for the Act as a whole, "competition" has many facets as enumerated in section 1.1, for the particular purposes of paragraph 79(1)(b), "anti-competitive" refers to an act whose purpose is a negative effect on a competitor. [Emphasis in original.]

109 Desjardins J.A. went on to emphasize that evidence of impacts of the impugned acts on competition or on consumers is irrelevant. Only the effects on competitors count for purposes of paragraph 79(1)(b).⁷⁸

110 Desjardins J.A. did however confirm that it is not necessary to show subjective intent to show that an act is anti-competitive.⁷⁹ The respondent is deemed to intend the effects of its actions.⁸⁰ Evidence of a subjective anti-competitive intent is certainly helpful however, and has been used by the Tribunal.⁸¹ However, it is clear from Desjardins J.A.'s reasoning that a subjective anti-competitive intent must exist, and must be the over-riding purpose for the conduct. That being said, a subjective anti-competitive intent would not make an otherwise appropriate competitive response anti-competitive.⁸²

111 The Federal Court of Appeal also changed the law on business justification. Desjardins J.A. held that a business justification is not a defence; it only goes to rebut evidence of anti-competitive purpose, by providing an alternative explanation for the conduct. The range of business justifications that can be accepted is limited to those that "provide a credible efficiency or pro-competitive explanation, unrelated to an anti-competitive purpose, for why the dominant firm engaged in the conduct alleged to be anti-competitive".⁸³

112 The Tribunal had accepted Canada Pipe's business justification that its rebate program enabled it to sell sufficient volumes of standard products that it could maintain a full line of product, including specialty products that were in less demand but were still important to have available in the market. This, the Tribunal accepted, benefited consumers. Desjardins J.A. rejected this as irrelevant: "Simply stated, improved consumer welfare is on its own insufficient to establish a valid business justification for the purposes of paragraph 79(1)(b)".⁸⁴ She held that

the Tribunal's reasons do not establish the requisite efficiency-related link between the SDP and the respondent, and hence do not supply a legitimate explanation for the latter's choice to engage in the impugned conduct, unrelated to an anti-competitive purpose.⁸⁵

113 As will be discussed below, this narrowing of the business justification defence could have serious consequences in any essential facilities case brought under section 79.

114 This is not the place to embark on a detailed criticism of the Federal Court of Appeal's decision in Canada Pipe. Suffice it to say that I think that Desjardins J.A.'s approach to paragraph 79(1)(b) is inconsistent with the fundamental principle that competition law protects competition, not competitors,⁸⁶ inconsistent with past Tribunal jurisprudence, and will likely lead to serious analytical problems.

115 Two of these analytical problems are, however, worth mentioning, as they may crop up in telecommunications cases. First, the emphasis on "a competitor" or "competitors" strongly suggests that there must be actual competitors in the market that experience these effects. Where there are no competitors, a monopolist will be able to engage in anti-competitive conduct designed to prevent entry scot-free. The same result would obtain where the dominant firm's conduct has no impact on existing competitors, but prevents entry by as yet non-existent competitors. In this, the court's reasoning is inconsistent with paragraph 79(1)(c), which refers to "substantial prevention of competition" in addition to "substantial lessening". It is tempting to solve the problem by saying that "competitors" includes "potential competitors", but that would in effect be to allow effects on "competition" into paragraph 79(1)(b), which the Federal Court of Appeal has stated is not permissible.

116 Thus, if a dominant owner of a telecommunications facility refused to grant access to new entrants, but did not cut off or otherwise affect its existing competitors, that conduct likely could not be considered an anti-competitive act under the test as modified by the Federal Court of Appeal.

117 The second problem is that looking at effects on actual competitors may lead to over-identification of anti-competitive acts. A competitor may fair poorly not because of the impugned acts, but because its costs are too high, it is undercapitalized, has a poor business plan, is poorly managed, or has an inferior product. Unless the Tribunal is careful to isolate these factors from the effects of the anti-competitive acts, a vigorous but appropriate competitive response by an incumbent may be mislabelled as anti-competitive as a result of focussing on the competitor. But isolating these factors effectively means considering the impact of the impugned acts, not on actual competitors, but on a hypothetical efficient, well-funded competitor with a good business plan, etc. In short, we are driven to looking at the effects on competition, not on a competitor.

118 This problem may be magnified in telecommunications markets, for two reasons. First, the markets are in transition from monopoly supply by ILECs to vigorous facilities-based competition, and from traditional telephone technology to IP based technology. (Indeed, for approximately 82% of Canadian homes that can obtain VoIP telephony over broadband cable internet, that transition may be nearly complete.) In this environment, however, there remains considerable suspicion about the once-dominant ILECs. This suspicion is apparent in many of the submissions in this proceeding.

119 Second, a market that is in transition and characterized by rapid innovation, such as the present telecommunications markets, might well attract more entrants than a mature market, some of whom may be weak. Indeed, both the US and Canada have recently experienced a raft of failures and consolidation of CLECs with resale business models. When these entrants fail, the suspicion may fall on the ILECs.

(ii) Examples of anti-competitive acts

120 As noted above, section 78 contains a non-exhaustive list of anti-competitive acts. In fact, most of the conduct found by the Tribunal to be anti-competitive conduct was not on the list in section 78.

121 Conduct found to be anti-competitive in past cases can be divided into contracting practices, and non-contracting practices. Anti-competitive contracting practices consist of practices that lock up customers, such as exclusive contracts, particularly those with evergreen clauses and/or punitive termination provisions.⁸⁷ Meet or release, right of first refusal, right to compete, and most-favoured nation clauses can also be anti-competitive.⁸⁸ Finally, financial inducements to exclusivity can also be anti-competitive.⁸⁹

122 Non-contracting practices identified as anti-competitive acts to date include sham litigation,⁹⁰ a pattern of anti-competitive acquisitions,⁹¹ predatory pricing,⁹² refusal to deal,⁹³ leveraging intellectual property to foreclose competition in other markets,⁹⁴ excluding competitors from networks, marketplaces or associations,⁹⁵ and leveraging dominance in one market to achieve market power in another.⁹⁶

123 Potential anti-competitive acts that may be relevant to the telecommunications industry, include:

124 Raising rivals' costs. In the TAB, the Bureau gives a number of examples of potential conduct by a dominant TSP that might raise rivals' costs.⁹⁷

125 Margin squeezing (s. 78(1)(a). De-regulation of wholesale access will remove regulatory barriers to margin squeezes. This is not necessarily a reason to de-regulate, as margin squeezing can be challenged under the abuse of dominance provisions. As the Bureau discusses margin squeezing in the TAB, I do not propose to deal with it here.⁹⁸ The European Union has brought two major margin squeezing cases against ILECs.⁹⁹ Price squeezes can be conceptualized as an indirect way of cutting off supplies to a competitor; they are thus in principle analogous to refusal to deal or essential facilities cases.¹⁰⁰

126 Pre-emption of scarce facilities or resources required by a competitor, with the object of withholding the facilities or resources from the market (s. 78(1)(e)). This might occur where a dominant telecommunications service provider locks up all available support structures or rights-of-way in a particular area.

127 Predatory pricing.¹⁰¹ Predatory pricing is one of the most problematic and controversial areas in competition law. The theory is deceptively simple: a dominant firm lowers prices in order to drive out its competitors. It then raises prices again to recoup the lost profits. The problem is that vigorous competition also involves firms lowering prices. A number of important limitations have been developed to constrain the scope of predatory pricing theories and distinguish between vigorous and abusive competition. Thus, for instance, in order to be predatory, prices must be below an appropriate measure of costs. Just what that measure is has been the subject of great debate. The usual candidates are average variable costs and avoidable costs.¹⁰² But determining what costs are avoidable is extremely complex. Moreover, it is well accepted in criminal predatory pricing cases that the dominant firm is entitled to match its competitor's price, even if that price is below the dominant firm's costs (this is called the "meeting the competition defence"). As well, in the US, there must be a likelihood that the predator will recoup lost profits after forcing out or disciplining competitors. It is not clear whether recoupment is an essential element in Canada.

128 The TAB contains the Bureau's most comprehensive statement to date on predatory pricing. Among other things, the Bureau adopts the avoidable cost standard; it appears to treat likelihood of recoupment as an essential element; and it recognizes the meeting the competition defence.

129 The European Union fined Wanadoo, a subsidiary of France Télécom, for predatory pricing of its broadband internet services in July 2003. The Commission's decision was upheld by the Court of First Instance in January 2007.¹⁰³ The case may be of little assistance, however, as the test for predatory pricing in the EU is very different from the test in Canada. In the EU, prices that are below average variable costs are presumed to be predatory; prices that are above average variable costs but below total costs are considered predatory if they form part of a plan to eliminate a competitor.¹⁰⁴ Nor does the EU recognize the meeting the competition defence¹⁰⁵ or require a showing of a likelihood of recoupment.¹⁰⁶

130 Targeting. The TAB suggests that targeted responses by an incumbent might be an anti-competitive act. However, in my view, the Tribunal's decision in Tele-Direct all but categorically excludes "targeting" as constituting a stand-alone anti-competitive act.¹⁰⁷

(iii) "Practice"

131 The word "practice" suggests a pattern of conduct. The Tribunal has held that its meaning is broad: "a practice may exist where there is more than an ‘isolated act or acts'".¹⁰⁸ Similarly, several different anti-competitive acts taken together can be a "practice".¹⁰⁹ In the result, the "practice" requirement will only exclude cases involving a few isolated anti-competitive acts.

132 Despite the breadth of this definition, it is difficult to see how a single denial of access by an owner of a facility could constitute a "practice". Indeed, absent a policy of denying access by the owner, it would be hard to infer that the owner had the requisite anti-competitive purpose.

133 The final element is whether the practice of anti-competitive acts substantially lessens or prevents competition. Substantial lessening refers to a reduction in the level of competition from what existed before the anti-competitive conduct started. Put another way, if the anti-competitive conduct increases the dominant firm's market power, competition is lessened. Substantial prevention generally refers to a situation where the market was not competitive (or less competitive) at the beginning, but would have become competitive (or, more competitive) had the dominant firm not engaged in the anti-competitive conduct. Put another way, if the dominant firm's market power would have decreased but for the anti-competitive conduct, competition is prevented.

134 In the Canada Pipe Appeal, Desjardins J.A. held that a relative, or "but for" assessment must be carried out. That is, the level of competition in the market with the impugned practice in place must be compared with the competition without the impugned practice. The comparison is carried out over three time frames, past, present, future:

...the question that must be assessed for the purposes of paragraph 79(1)(c) is, would the relevant markets – in the past, present or future – be substantially more competitive but for the impugned practice of anti-competitive acts?¹¹⁰

135 As well, in the Canada Pipe Appeal, Desjardins J.A. held that the absolute level of competition is irrelevant, nor is the fact of entry in the absolute sense.¹¹¹

136 Subsection 79(4) directs the Tribunal to consider whether the practice is a result of superior competitive performance in determining whether a practice of anti-competitive effects substantially lessens or prevents competition.

137 It is unclear what, exactly, subsection 79(4) means. The Abuse Guidelines suggest that subsection 79(4) might be engaged where a firm exploits legitimate competitive advantages such as lower costs, better distribution or production techniques, or a broader array of product offerings.¹¹² This provides little assistance: exploiting these advantages is unlikely to constitute a practice of anti-competitive acts in the first place. As well, subsection 79(4) does not appear to create a defence; it merely instructs the Tribunal to consider this factor. In Canada Pipe, the Tribunal accepted that the advantages that Canada Pipe's Canada-wide distribution network and broad product range gave it met this test.¹¹³ An ILEC facing an abuse of dominance proceeding likely would argue that this reasoning applies to its existing network and product range, such that its conduct falls within subsection 79(4).

138 Under section 79, the Tribunal can make orders designed to remedy the substantial lessening or prevention of competition caused by the practice of anti-competitive acts. The Tribunal cannot award damages, and cannot make punitive orders.¹¹⁴ The Tribunal's approach in choosing a remedy is "only to go as far as it considers necessary in order to restore competition in the relevant markets".¹¹⁵ In Southam, the Supreme Court enunciated a similar principle, noting that in a merger case, the test is whether the remedy eliminates the substantial lessening of competition, not whether it restores the pre-merger competitive situation. The least intrusive of the possible effective remedies should be preferred.¹¹⁶

139 The Tribunal can make three kinds of remedial orders under s. 79:

a) It can order the respondents to stop the practice of anti-competitive acts;

b) It can order the respondents to take positive action to overcome the effects of the practice of anti-competitive acts;

c) It can order the sale of assets or shares.

140 Section 79 envisages that the main remedy will be an order that the respondent stop the anti-competitive conduct (option a). The Tribunal can only make a mandatory or structural order (options b and c) if an order prohibiting the practice of anti-competitive acts is insufficient. There is an additional limitation on the mandatory or structural orders in s. 79(3): the Tribunal can interfere with the rights of the person against whom the order is made (or the rights of any other person affected by the order) only to the extent necessary to achieve the purpose of the order. While an order breaking up a dominant company is possible, the Tribunal has never ordered the sale of assets or shares in an abuse of dominance case.

141 The distinction between an order prohibiting a person from engaging in an anti-competitive act, and an order mandating positive acts, is not always clear. For example, in Tele-Direct, the anti-competitive act was discriminating against consultants and customers who used consultants. The Tribunal prohibited Tele-Direct from rejecting orders from customers who used consultants, and from "not processing" such orders. This, of course, amounts to a requirement that Tele-Direct accept and process such orders.¹¹⁷ Any order "prohibiting" a refusal to grant access to an essential facility would likely involve convoluted mandatory provisions disguised as prohibitions, and/or a mandatory order.

142 For purposes of this proceeding, the key question is whether the refusal of a dominant firm to grant access to a competitor to an essential facility that it owns could be an anti-competitive act under section 79. The CRTC, as an industry regulator, can impose a duty on TSPs to grant access to essential facilities, regardless of whether the doctrine is recognized under section 79. However, to the extent that parties are urging the CRTC to adopt competition law essential facilities principles in defining when it will mandate access to wholesale services, it is important to consider what the competition law in Canada on essential facilities might be. As well, to the extent that the CRTC forbears from mandating access to wholesale facilities, the Competition Act will apply. Thus it is relevant to consider whether the Tribunal might mandate access even where the CRTC has not.

143 As discussed in detail below, in my view neither the statutory language of sections 78 and 79, nor the case law, permits one to say with certainty whether, or not, denial of access to an essential facility can be an anti-competitive act.

(i) Sections 78 and 79 are inconclusive

144 First of all, denial of access to an essential facility by an airline is listed among the anti-competitive acts, at paragraph 78(1)(k). This provision was added in 2000, shortly after the Air Canada-Canadian merger left Canada with one dominant airline. Otherwise, denial of access to an essential facility is not listed in section 78. One could argue from this that denial of access to an essential facility in other industries is therefore not an anti-competitive act using the principle of statutory interpretation expressio unius, exclusio alterius. This argument is unlikely to be decisive, however. The list in section 78 was deliberately left open-ended, and paragraph 78(1)(k) was added long after the enactment of section 78. Assuming it was open to the Tribunal to find that denial of an essential facility was an anti-competitive act before the addition of paragraph 78(1)(k), there is nothing in the 2000 amendments to change that.

145 An argument in favour of recognition of essential facilities can be made from the inclusion of paragraph 78(1)(a), margin squeezing. If a dominant firm cannot be made to supply a competitor, then it would be inconsistent to require that if the dominant firm chooses to supply a competitor, it must not squeeze the competitor's margins by charging the competitor too much in the upstream market. Thus recognition that denial of access to an essential facility can be an anti-competitive act may be implicit in the recognition of margin squeezing as an anti-competitive act. I think this is a strong argument, but I am not sure that it would be determinative.

(ii) The case law is inconclusive

146 The Tele-Direct case is the only contested case to consider refusals to deal under the abuse of dominance provisions.

147 Tele-Direct was an application under the tied selling (s. 77) and abuse of dominant position provisions of the Competition Act. Tele-Direct (Publications) Inc., a subsidiary of Bell Canada and BCE Inc., published Yellow Pages directories. The Director of Investigation and Research¹¹⁸ accused Tele-Direct of a variety of anti-competitive conduct directed at independent advertising service suppliers. The Director alleged that Tele-Direct engaged in tied selling by forcing or inducing customers seeking advertising space in its directories to buy telephone directory advertising services as well. The Director also accused Tele-Direct of anti-competitive acts in two markets, the market for advertising space, and the market for advertising services.

148 The Tribunal held that telephone directory advertising was the relevant product market, and that Tele-Direct had control or market power over this market. The Tribunal held that Tele-Direct had tied the sale of advertising space to advertising services, and prohibited it from continuing this practice.

149 The abuse of dominance case consisted of two separate cases. The first accused Tele-Direct of abusing its dominant position in the market for telephone directory advertising space. The alleged anti-competitive act consisted of various competitive responses that the Director claimed constituted "targeting". The Tribunal rejected this theory.

150 The second accused Tele-Direct of leveraging its market power over the advertising space market into the advertising services market. This was a leveraging case, as Tele-Direct did not have market power in the services market. The Director alleged that Tele-Direct engaged in anti-competitive acts in its dealings both with agents and consultants. With respect to agents, the Director alleged that Tele-Direct squeezed their returns through various tactics, and discriminated against them in providing space on less favourable terms than those available to Tele-Direct's internal sales force. The Tribunal rejected these allegations on the ground that the lack of erosion of the agents' market share showed that Tele-Direct's actions had not given rise to a substantial lessening of competition.

151 With respect to consultants, the Director attacked Tele-Direct's policy of not dealing with consultants as agents for a customer, and various tactics employed by Tele-Direct against consultants. Consultants operated by suggesting to Yellow Pages advertisers that they could save money by using smaller ads or removing colour. They were paid out of the savings achieved by the customer. Tele-Direct had a policy of not accepting orders placed by consultants for customers. Tele-Direct also used various tactics against consultants.

152 Tele-Direct argued that, as a matter of law, it was not obliged to help its competitors, and thus refusal to do so could not be an anti-competitive act. Tele-Direct also argued that anti-competitive acts involved doing something, as opposed to not doing something:

Each of the anti-competitive acts listed in section 78 require the dominant firm to actively initiate some action. . . . None of the listed acts are triggered simply by the dominant firm not doing something or refusing to assist. . . .¹¹⁹

153 The Tribunal stated that it agreed with the general proposition that competitors do not have to help one another, but that this might not apply in a case under section 79:

As stated above, as a general proposition, competitors should not be required to assist one another. But, this general proposition may be shown to be inapplicable in a given section 79 case by the Director proving that the "act" of the respondent meets the elements of that section and is an anti-competitive act leading to a substantial lessening of competition. Then, any order of the Tribunal which may issue is, by definition, not an order to "assist" a competitor but rather, in the case of subsection 79(1), an order to cease and desist from anti-competitive conduct.

¶ 589 It is, therefore, not sufficient, in circumstances such as these, to argue the general proposition. Nothing can be determined by simply labelling the alleged anti-competitive "act" as "doing something" (active) or "not doing something" (passive). The anti-competitive effect of the conduct of the respondents, whether "active" or "passive", must be weighed against any business justification in order to conclude whether there has or has not been a substantial lessening of competition. That can only be done by reference to the evidence.¹²⁰

154 The Tribunal thus proceeded to weigh the anti-competitive effects of the acts against the business justification proffered by Tele-Direct. Tele-Direct's refusal to deal might have two types of adverse effects: it might raise consultants' costs and damage their reputations, the Tribunal held. However, Tele-Direct's refusal to deal did not impose any significant cost increases, or even increase consultants' costs at all. Similarly, the Tribunal could not attribute any damage to consultants' reputation to the refusal to deal (as opposed to Tele-Direct's other practices).¹²¹

155 By contrast, Tele-Direct would face additional costs to deal with consultants, because it would have to set up an interface to deal with them and a system to deal with the fact that consultants do not pay up front when submitting the order (unlike agents). The Tribunal accepted that this was a valid justification for Tele-Direct's not dealing with consultants:

In the circumstances, we think that the additional costs that Tele-Direct would incur if it were forced to deal with consultants directly on behalf of advertisers is a valid justification for not doing so, given that no adverse cost effects on agents were proven and that any negative reputational effects that are attributable to the refusal to deal directly are, at best, weak.¹²²

156 The Tribunal also addressed Tele-Direct's refusal to supply its specifications to consultants. On this point the panel split. The majority, consisting of the judicial member, Justice Rothstein, and a lay member, Christine Lloyd, held that this refusal had not adversely affected the consultants. The potential for customers confusing consultants with Tele-Direct representatives if consultants used the specifications from Tele-Direct to suggest they had a relationship with Tele-Direct provided a valid business justification, the majority held.

157 The Tribunal's full-time economist, Dr. Frank Roseman, disagreed. He considered that Tele-Direct's refusal to supply specifications would likely "significantly hamper the consultants' ability to compete", and that there was no business justification for this refusal.¹²³

158 Unfortunately, the Tribunal did not articulate any test in Tele-Direct for the degree of impact that must be felt by competitors before a refusal to deal would become an anti-competitive act. The Tribunal did consider whether the refusal would increase consultant's costs, and at one point referred to the fact that they were not "significantly" increased. However, there is no exploration of what "significantly" might mean in the circumstances. Nor is there any discussion of whether there must be no alternative feasible sources of supply (presumably because, here, there obviously were none). Dr. Roseman, in his dissent, used the phrase "significantly hamper", which may be an improvement, but there is no further indication of what this might mean.

159 As well, the Tribunal's approach of balancing the impact of competitors against the cost or inconvenience to the respondent could lead to an overbroad approach to refusals to deal. By using this analysis, the Tribunal could end up ordering respondents to supply competitors even though the impact of the refusal to deal was minimal, simply because the inconvenience to the respondent was less than the impact on competitors. This balance of inconvenience approach would not, in my view, be an appropriate basis for forcing dominant firms to supply their competitors.

160 The Tribunal's approach to business justifications has in any event been over-ruled by the Federal Court of Appeal. Based on the Federal Court of Appeal's reasoning, it seems unlikely that the costs to the dominant firm of supplying a competitor could be a factor at all in the analysis, except to the extent that they could be included in an efficiency argument (the dominant firm might, for example, argue that it is inefficient to supply the competitor because of the costs involved).

161 The Tribunal's suggestion that an order to cease an anti-competitive act, where the anti-competitive act is a refusal to supply or assist a competitor, is not an order to supply or assist the competitor, makes no sense to me. This is exactly the same kind of semantic reasoning the Tribunal accused Tele-Direct of. Indeed, for an order to "cease" refusing to deal to be meaningful and enforceable, it must specify the terms under which the respondent will supply its competitor.

162 The fact is, there is a big difference between doing something and not doing something. In criminal law, it can mean the difference between going to jail and remaining free. Tele-Direct was correct in its submission that each of the acts enumerated in section 78 involved the respondent "doing something".¹²⁴ Moreover, all of the acts found by the Tribunal in the contested abuse of dominance cases involve "doing something". There is, in my view, a big difference between engaging in the anti-competitive acts listed in section 78 or in the cases, and refusing to supply a competitor.

163 It is interesting to contrast the Tribunal's approach to refusal to deal with its approach to refusal to license intellectual property in Tele-Direct. The Tribunal refused to qualify Tele-Direct's refusal to license its trademarks (such as the well-known "walking fingers" logo) to competitors as an anti-competitive act, holding that the right to determine whether or not, and to whom, to license a trade mark, was inherent in the nature of the right to license the trademark.¹²⁵ This finding was not made under section 79(5), which exempts acts engaged in pursuant only to the exercise of certain intellectual property rights.

164 All property, not just IP, is a legal construct. Property is usually defined as a "bundle of rights".¹²⁶ The right to control "whether or not, and to whom, to grant a licence" is just as inherent in the legal concept of property itself as it is in intellectual property.¹²⁷ Thus the Tribunal's reasoning in relation to trade marks could be applied to any kind of property, not just IP.

165 In summary, while the Tribunal has shown a willingness to treat refusals to deal as potential anti-competitive acts, there is insufficient jurisprudence to say with certainty whether refusal to grant access to an essential facility would be an anti-competitive act, nor what the test would be in Canada.

(iii) Essential facilities in the airline industry

166 After Air Canada took over Canadian Airlines in December 1999, leaving Canada with one dominant airline, airline industry-specific statutory and regulatory provisions were added to the Competition Act. In February 2001, the Bureau published draft Enforcement Guidelines on the Abuse of Dominance in the Airline Industry (the "Airline Guidelines") to explain these provisions. The airline industry-specific provisions remain on the books, although the Airline Guidelines were never finalized. Interestingly, although the Abuse Guidelines were finalized in July 2001, they only mention denial of access to an essential facility as a possible anti-competitive act, without discussion.

167 Paragraph 78(1)(k) codifies the essential facilities doctrine as it applies to airlines:

(k) the denial by a person operating a domestic service, as defined in subsection 55(1) of the Canada Transportation Act, of access on reasonable commercial terms to facilities or services that are essential to the operation in a market of an air service, as defined in that subsection, or refusal by such a person to supply such facilities or services on such terms.

168 It should be noted that paragraph 78(1)(k) does not contain an express purpose requirement. Purpose is thus not relevant to essential facilities in the airline industry.

169 In August 2000, the government made the Regulations Respecting Anti-Competitive Acts of Persons Operating a Domestic Service (the "Airline Regulations"). Section 2 defined essential facilities for purposes of paragraph 78(1)(k):

ESSENTIAL FACILITIES AND SERVICES

2. (1) For the purposes of paragraph 78(1)(k) of the Competition Act, facilities and services that are essential to the operation in a market of an air service, as defined in subsection 55(1) of the Canada Transportation Act, are those

(a) that are required in order to provide a competitive air service;

(b) that cannot reasonably or practicably be purchased, acquired, provided or replicated by another air carrier on its own behalf;

(c) that are effectively controlled by the air carrier who denies access to them or refuses supply of them; and

(d) that can be feasibly provided to another air carrier, having regard to operational or safety considerations, or legitimate business justifications of the air carrier referred to in paragraph (c).¹²⁸

170 The similarity of these criteria to those derived from the US cases by Professor Robinson is obvious.

171 The Bureau unpacks these elements somewhat in the Airline Guidelines. The guidelines explain that the first element means that "it must not be possible to offer the competitive air service without the facility or service for which access is being denied".¹²⁹

172 The Bureau also explains the term "competitive air service". It is basically meant to close a loophole, and prevent the dominant airline from arguing that the facility in question was not required to operate an air service generally, as opposed to one that competes with the particular route in question.¹³⁰

173 The Bureau cautions, however, that "competitive air service" does not mean identical air service.¹³¹ While the Bureau does not expand on this, given the competition law context "competitive" must mean that the air service is in the same product market as the dominant airline's air service.

174 The Bureau does not expand much on the second element. It suggests it will consider whether the competitor could have provided the facility for itself on reasonable terms.¹³²

175 Similarly, the Bureau merely points out that the third element, controlled by a dominant carrier, makes it clear that the dominant carrier must be in a position to grant access because it controls the facility.¹³³

176 The fourth element, feasible to provide, permits the introduction of legitimate business reasons for refusing access. The Bureau suggests as an example that the dominant carrier might need the facility itself and not have spare capacity to provide access, without investing in new facilities. The Bureau appears to accept that it is not reasonable to expect a dominant firm to invest in new facilities to supply a competitor.¹³⁴

177 There have been no cases involving denial of an essential facility in the airline industry.

(iv) Essential facilities in the TAB

178 The Bureau states the doctrine in the TAB in general, non-industry specific terms. It is clear, therefore, that the Bureau regards essential facilities as being of general application in competition law.

179 The Bureau's statement of the essential facilities doctrine is also not based on any Canadian precedent – there is none, apart from the incomplete analysis in Tele-Direct, and the Airline Regulations, as discussed above. The Bureau's test has therefore not been tested in the crucible of a contested case, nor against facts other than those considered by the Bureau in its drafting. This is an important point, because real cases tend to throw into vivid relief any defects in proposed legal tests.

180 The Bureau begins by defining as essential facility in economic terms:

For the purposes of section 79 of the Act, an "essential" facility is an input that provides the firm controlling it with the power to lessen or prevent competition in a relevant downstream market.¹³⁵

181 The Bureau did not add the modifier "substantially" to "lessen or prevent". This may be problematic. Competition law recognizes that in theory there can be a lessening or prevention of competition that is less than "substantial". Without the "substantial" modifier, a facility that provides the firm controlling it with the power to lessen or prevent competition downstream by a trivial degree could nonetheless be considered "essential" under this definition. Of course, no remedy would be available under the Competition Act, nor should the CRTC mandate access under such circumstances. It seems wrong in principle, however, to define "essential facilities" in such a way that facilities that are not really essential meet the test. I am inclined to assume that this definition was meant to include the modifier "substantial".

182 The Bureau cautions, however, that

Section 79 of the Act is concerned with abuse of this market power that has resulted, results or is likely to result in a substantial lessening or prevention of competition, not the exercise of any market power inherent in the facility.

183 This follows, as the Bureau notes in a footnote, from the principle that the Competition Act abuse of dominance provision is not concerned with the existence of dominance, but with its abuse. The distinction between exercise of market power inherent in the facility, and abuse of that market power, is difficult to draw in practice. I have difficulty seeing how a mere denial of access could constitute abuse of dominance, except in limited circumstances.¹³⁶

184 The Bureau then notes that in an essential facilities case, the actual conduct at issue would be "an actual or constructive denial of access to the facility". A constructive denial of access would be conduct that is equivalent to denying access, such as terms of access that are prohibitively expensive.

185 I set out the Bureau's test below, for convenience:

(i) A vertically integrated firm that is dominant in two markets. The first relevant market is the upstream market (or wholesale market) for the facility. The second relevant market is the downstream market (or retail market) in which the facility is an input. A necessary condition for concluding that there is dominance in the upstream market is that it is not practical or feasible for competitors to duplicate the facility in question.

(ii) A denial of access to the facility for the purpose of excluding competitors from entering or expanding in the downstream market or otherwise negatively affecting their ability to compete.

(iii) The denial has had, is having or is likely to have the effect of substantially lessening or preventing competition in the downstream market. [Footnotes omitted.] ¹³⁷

186 This test has already been the subject of extensive comment by economists retained by the parties. I will comment briefly on each element.

187 The first element is really two elements: dominance in the upstream market, and dominance in the downstream market (I refer to this as the "double dominance requirement"). A qualification is added to the first element, namely, that the vertically integrated firm is not considered dominant unless it is not practical or feasible for competitors to duplicate the facility.

188 The double dominance requirement is problematic. I agree that on a practical level, it is a useful screen, because if the vertically integrated firm is not dominant in the downstream market, one must question whether the purported essential facility really is essential. However, maintaining this requirement as a matter of law is inconsistent with the possibility of a firm that is dominant in an upstream market to leverage that dominance into the downstream market by cutting off its competitors. Thus I would regard lack of dominance in the downstream market as strongly indicating, and perhaps even raising a rebuttable presumption, that the facility is not essential and that the refusal to grant access will not substantially lessen or prevent competition.¹³⁸ It should be noted that in the telecom industry, only downstream competitors that are not dependent on the upstream dominant firm should be counted when considering this factor.

189 The second difficulty that I have with the first element is that the definition of "dominance" in the upstream market is different from the definition of dominance in the downstream market, where, presumably, dominance would be defined in the usual way (as discussed above). I agree that the "not practical or feasible for competitors to duplicate the facility" requirement is an essential element of any essential facilities theory. However, it is potentially problematic to play around with the definition of "dominance" in order to fit this in. The statutory language setting out the dominance requirement is common to all anti-competitive acts, which gives rise to the presumption that the requirements for establishing dominance will not vary depending on the anti-competitive act being alleged. Consequently, this element would be better placed under paragraph 79(1)(b). That being said, it must be admitted that the Bureau's approach in the TAB is consistent with its approach to margin squeezing in the Abuse Guidelines, where the Bureau equated dominance in the upstream market with there being no alternative sources of supply.¹³⁹

190 As discussed above, the Bureau's approach to market shares in the telecom industry results in 1/n market shares where there are competing networks that are not capacity constrained. It is unlikely that a network owner could be considered dominant in these circumstances.

191 The first element stated by the Bureau maps to paragraph 79(1)(a) ("...substantially or completely control... a class or species of business"), and the third, to paragraph 79(1)(c) (substantial lessening or prevention of competition). The second element stated by the Bureau should therefore map to paragraph 79(1)(b), and define the nature of the anti-competitive act.

192 The Bureau's second element focuses exclusively on intention. While the point is arguable, given the heavy emphasis of the Federal Court of Appeal in Canada Pipe on intention, earlier case law in the Tribunal focussed on whether the character of the act was predatory, exclusionary, or disciplinary. Intention was inferred from the character of the act, in the absence of evidence of a subjective intention.

193 As stated above, the "not practical or feasible to duplicate" element should form part of the inquiry into the character of the act. A similar concept is found in one statutorily enacted essential facilities provision, paragraph 78(1)(k), relating to airlines.

194 Another important issue is whether the dominant firm can reasonably be expected to supply the facility. This is expressly required by the Airline Regulations, and the Bureau touches on it in its discussion of the second element. In this discussion, the Bureau also repeats the "not practical or feasible to duplicate" requirement. It also adverts to the capability of the supplier to supply the facility:

With a finding of market power, a denial of access is an anti-competitive act when its purpose is to exclude or impede actual or potential competitors. To infer such a purpose, it must be difficult or impossible for those competitors to substitute other inputs or to practically or reasonably duplicate the facility. The requirement that it is not practical or feasible for a competitor to duplicate the facility means that such an entrant would not find it feasible to enter or compete effectively if it had to self-supply the facility. At the same time, for the purpose to be anti-competitive, the supplier must have the necessary capacity, or have the willingness and ability to build the necessary capacity, to supply those competitors.¹⁴⁰

195 The third element of the test in the TAB is simply paragraph 79(1)(c): substantial lessening or prevention of competition. As held in the Canada Pipe Appeal, this involves a comparative assessment of the market with, and without, the impugned refusal to deal.

196 The Bureau has indicated it will take a narrow approach to this element:

Before the Tribunal is able to issue any remedial order under section 79, it must be shown that the practice of anti-competitive acts must actually, or be likely to, substantially lessen or prevent competition in the downstream market. Accordingly, if control of the facility is a source of dominance in both an upstream and downstream market and the denial of access has been for an anti-competitive purpose, the Bureau's "but for" analysis would then focus on whether the denial of access leads to materially higher prices downstream than would occur if the dominant firm charged the profit-maximizing access price to downstream competitors.⁵⁵

⁵⁵ The "profit-maximizing access price" would be that at the time of the denial (i.e., assuming the downstream market structure absent the denial).¹⁴¹

197 As I understand this paragraph, the comparison is made between the following two situations. In the first, "with" the denial of access, the incumbent vertically integrated firm may, or may not, have market power in the downstream market. If it has market power downstream as a result of its denial of access, then it is able to – and likely will – charge monopoly prices in that market. If it does not have market power downstream, it will not.

198 In the second situation, "without" denial of access, the vertically-integrated firm grants access to a competitor. Now if the vertically-integrated firm had market power downstream, and decided to grant access voluntarily, it would be able to price access up to the point where its competitor could not beat the monopoly prices it was charging downstream. So the competitor ends up charging essentially the same price as the incumbent in the downstream market. It follows that competition has not been increased by granting access; put another way, denial of access has not prevented competition.

199 Thus the Bureau appears to have recognized the "single monopoly profit critique", which holds that there is only one monopoly profit to be earned.¹⁴² It does not matter whether the monopoly profit is earned upstream or downstream.

200 The July 2005 report by the EU's Economic Advisory Group on Competition Policy, An economic approach to Article 82"("EAGCP Report") explains:

However, the monopolization of a downstream market need not have anticompetitive effects per se. After all, there is only one final market and only one monopoly profit to be reaped. If the dominant firm is able to capture the monopoly profit of the final market even if there is downstream competition, then monopolization of the downstream market cannot have anticompetitive effects because there is no competition anyway.¹⁴³

201 It follows that the situations where the Bureau's test would be met are very limited.¹⁴⁴

202 The fact that the Bureau's test appears to tolerate monopoly pricing by the incumbent may strike some as odd. It is, however, correct from a competition policy standpoint. The abuse of dominance provisions do not exist to provide a remedy for existing market power. They are only engaged when firms that have market power engage in anti-competitive conduct in order to preserve or extend their market power. The Tribunal stated this point in Tele-Direct:

section 79 is not intended to condemn a firm merely for having market power. Instead, it is directed at ensuring that dominant firms compete with other firms on merit and not through abusing their market power.¹⁴⁵

B. Refusal to deal

203 Section 75 allows the Tribunal to intervene in certain cases where a person's business is suffering because of an inability to obtain a supply of a product needed for that business. There have been only three applications under s. 75 by the Commissioner, the last being in 1997. However, now that private applications are allowed under s. 75, this provision is assuming greater importance. The first application to go to a full hearing, B-Filer, was heard and decided in 2006. Fifteen applications for leave to commence a private application under s. 75 have been filed; leave has been granted in five, and denied in nine. One application was withdrawn.¹⁴⁶

204 The elements that must be shown before the Tribunal can intervene, as set out in subsection 75(1), are:

(a) a person is substantially affected in his business or is precluded from carrying on business due to his inability to obtain adequate supplies of a product anywhere in a market on usual trade terms,

(b) the person referred to in paragraph (a) is unable to obtain adequate supplies of the product because of insufficient competition among suppliers of the product in the market,

(c) the person referred to in paragraph (a) is willing and able to meet the usual trade terms of the supplier or suppliers of the product,

(d) the product is in ample supply, and

(e) the refusal to deal is having or is likely to have an adverse effect on competition in a market [Emphasis added.]

205 It is immediately apparent that the test under section 75 is similar to the essential facilities doctrine as developed in the US, although there are differences.

206 The first step, as always, is to define the market surrounding the product whose supply has been refused, that is, the upstream market. Market definition will essentially determine the question of whether the complainant is able to obtain an adequate supply of a product in a market. The broader the product or geographic market, the more likely there are alternative sources of supply. In the two successful cases to date, the product market was limited to the particular product that the complainant was unable to obtain: genuine Chrysler parts in Chrysler¹⁴⁷ and Xerox copier parts in Xerox.¹⁴⁸

207 In B-Filer,¹⁴⁹ the Tribunal picked up on its statement in Chrysler that "In the case of para. 75(1)(a), the ultimate test concerns the effect on the business of the person refused¹⁵⁰ The Tribunal then articulated what it called the "Chrysler test" for market definition under paragraph 75(1)(a):

For the purposes of 75(l)(a), products are substitutes, and so are included in the same market, if a person is not substantially affected in his business (or if the person is not precluded from carrying on business) as result of switching to these other products.¹⁵¹

208 This requirement raises two questions: first, what is the relevant business – does the business that is "substantially affected" refer to the "specific line or product" whose supply has been refused; or does it refer to the entire business of the complainant? Second, how substantially must the complainant's business be affected?

209 In Chrysler, the Tribunal answered the first question as follows: "the effect on the entire activity of which the refused supplies are a part should be used" and elaborated criteria for making this determination.¹⁵²

210 The Tribunal has held that "‘substantial' should be given its ordinary meaning, which means more than something just beyond de minimis".¹⁵³

211 If the complainant substitutes the input with another input that is in the same product market as the refused input, then its business will not have been substantially affected.¹⁵⁴ The Tribunal has dismissed a number of applications for leave on the basis that the lost sales did not constitute a sufficiently large percentage of overall sales.¹⁵⁵

212 There are no cases interpreting paragraph 75(1)(a) in the telecom industry, nor has the Bureau provided any guidance on section 75 as it applies in this sector. Suppose that, after the CRTC ends mandatory unbundling in a given geographic market, an ILEC cuts off CLECs that obtain wholesale services and facilities from it. Suppose also that no other facilities-based TSP in the market is willing to provide wholesale services and facilities to that CLEC. If that market represents a sufficiently large percentage of the CLEC's overall business, then the CLEC will likely be able to establish this element. If that market only represents 10% or so of the CLEC's sales, it is unlikely to be "substantially affected" under the analysis employed by the Tribunal to date. If the CLEC is able to point to national accounts it has lost, is likely to lose, or perhaps is unlikely to gain, as a result of not being able to provide services in all Canadian markets, that might get the CLEC over the threshold. Also, the CLEC might try arguing being forced out of business in a particular geographic market meets the test, regardless of the percentage of its revenue involved.

213 Section 75(1)(b) requires that the complainant's inability to obtain adequate supplies of the product be because of insufficient competition among suppliers. This imports two requirements: first, at the time of the refusal to supply, a "particular market situation must exist",¹⁵⁶ namely, "insufficient competition among suppliers". A market of many suppliers acting independently would not involve insufficient competition; a market with a monopoly supplier would.¹⁵⁷

214 The second requirement is that the insufficient competition must be the cause of the complainant's inability to obtain adequate supplies; that is, "the overriding reason that adequate supplies are unavailable must be the competitive conditions in the product market".¹⁵⁸

215 The presence of an objectively justifiable business justification may rebut an inference that insufficient competition is the cause of the complainant's inability to obtain adequate supplies.¹⁵⁹

216 An ILEC cutting off a CLEC might have difficulty showing an objectively justifiable business justification. While the ILEC might reasonably say that the price at which it had been forced to sell wholesale services was insufficiently remunerative, the response would be that the ILEC should increase the price rather than cut off the CLEC. If the ILEC could show that it needed the facilities, however, and would have to build new ones if it did not cut off the CLEC, that may be a legitimate business justification.

217 Sections 75(1)(c) and (d) establish two preconditions to the complainant's obtaining any relief. The complainant must be "willing and able to meet the usual trade terms of the supplier or suppliers of the product", and the product must be in ample supply.

218 "Trade terms" is restrictively defined in subsection 75(3) as "terms in respect of payment, units of purchase and reasonable technical and servicing requirements". "Servicing requirements" relates to upkeep or maintenance requirements. Thus contractual conditions of supply, apart from those relating to payment and units of purchase, are not "trade terms".¹⁶⁰

219 The phrase "usual trade terms of the supplier" necessarily implies that the supplier is in fact engaged in the business of supplying the product. Thus, section 75 will not apply if the vertically integrated firm does not supply the upstream product to any third party. There would be no basis upon which the Tribunal could find that the applicant was willing to meet the supplier's usual trade terms, as the supplier would not have any usual trade terms. Thus, if a CLEC decided not to supply wholesale services to anyone, section 75 could have no application in respect of that CLEC.

220 Further, for the concept of "usual trade terms of the supplier" to make sense, those terms must not have been imposed by regulation. The purpose of the requirement must be to ensure that the supplier is only required to supply the applicant in accordance with the same terms under which it has freely chosen to supply others. Such terms, presumably, do not impose an unreasonable hardship on the supplier. Terms imposed by regulation are not really the suppliers' terms.

221 The telecom industry presents a transition problem here. It would be inappropriate to treat the regulated prices set by the CRTC as "usual trade terms" in a deregulated environment. The lack of a history of sales giving rise to "usual trade terms" will give rise to difficulties in any application by a CLEC under section 75.

222 Section 75(1)(e) requires that the refusal to deal "is having or is likely to have an adverse effect on competition in a market". This requirement was enacted in 2002, after Chrysler and Xerox were decided. B-Filer is the only case to consider this provision so far.

223 The Tribunal held that, like the substantial lessening requirement in the abuse of dominance provision, the adverse effect requirement mandates a comparative assessment of the competitiveness of the market with, and without, the refusal to deal. The competitiveness of the market refers to the degree of market power that prevails in the market, the Tribunal held. Thus,

for a refusal to deal to have an adverse effect on a market, the remaining market participants must be placed in a position, as result of the refusal, of created, enhanced or preserved market power.¹⁶¹

224 However, section 75 sets a lower threshold than the "substantial" threshold required by section 79 (and several other provisions in the Competition Act).¹⁶² The adverse effect need not be in the same product market as the refused product.¹⁶³

225 The remedy available under s. 75 is narrow: the Tribunal can order any supplier of the product (not necessarily the one that cut the complainant off) to accept the complainant as a customer on usual trade terms.

226 Moreover, the Tribunal is unable to impose a time limit on its order under s. 75.¹⁶⁴ An application under s. 106 is the only way to vary or terminate an order under s. 75.

227 The Tribunal has a residual discretion to refuse an order even though all of the statutory requirements are met.¹⁶⁵

228 It should be apparent from the foregoing that section 75 shares many similarities with the essential facilities doctrine as developed by US cases. For instance, section 75 has requirements that are equivalent to essentiality and non-duplicability. That is, if the complainant is able to substitute another input for the refused product without serious impact on its business, no remedy will be available under section 75.

229 However, as mentioned above, section 75 probably is not available where the owner of the input does not supply it to anyone. As well, the competitive effects test is lower than the test applicable in an abuse of dominance case.

C. Forum shopping

230 To the extent that the CRTC forbears from mandating access to wholesale services, CLECs who consider that they have been denied access to an essential facility will have three options:

a) They can apply to the CRTC for a determination of the matter.

b) They can complain to the Competition Bureau under the abuse of dominance provisions.

c) They can apply to the Tribunal for leave to commence a private application.

231 Forbearance thus unavoidably creates forum shopping opportunities for aggrieved CLECs. CLECs' choice of forum will naturally be based on where they think they have the best chance of winning. This in turn will depend on a combination of the legal test applicable in each forum and the particular facts.

232 To the extent that the legal test is appreciably easier to meet in one forum than another, CLECs will naturally gravitate to that forum. Thus if the CRTC establishes a test for essential facilities that is much harder to meet than the tests applicable under the Competition Act, CLECs will be more likely to apply to the Commissioner or the Tribunal than the CRTC. One complicating factor is that because the test applicable to denial of access to an essential facility under section 79 is uncertain, it will be difficult for CLECs to evaluate which forum will give them a better shot.

233 A related issue that will likely arise is whether decisions by the CRTC will be regarded as res judicata by the Tribunal, and vice versa. I do not intend to attempt to resolve this issue.

V Defining "essential facilities" for purposes of telecommunications regulation

A. Past CRTC definitions

234 Through a series of decisions, the CRTC has produced a test for when it will mandate wholesale access of an essential facility. A brief survey of those decisions is therefore an appropriate starting point.

235 In September 1994, the CRTC issued its Review of Regulatory Framework decision.¹⁶⁶ As its name suggests, this decision announces a new regulatory framework intended to increase competition and reliance on market forces in the telecommunications industry, and assist in the development of the "information highway", which was then a term much in use.

236 The CRTC wanted to promote local competition, but recognized that local exchange networks were monopoly products owned by vertically integrated incumbents. The CRTC rejected structural separation of wholesale and retail services. Instead, it announced a regulatory framework that included the following:

a) New rate bases and tariffs designed to increase equity between incumbents and competitors;

b) Local price floors to prevent anti-competitive pricing;

c) Removal of barriers to entry through open access, unbundling and co-location;

d) Price caps in the "utility" segment (bottleneck services).

237 The CRTC thus held that telephone companies should provide co-location where requested and directed them to file tariffs for access to local switches that is comparable in terms of price and quality to the access they provide to their own long distance operations. The CRTC also held that telephone companies must provide unbundled services. The CRTC appears to have adopted a form of "negotiate first" approach, directing that competitors should advise incumbents of their unbundling requirements, and the incumbents should file tariffs. The CRTC also adopted the principle of reciprocal access and interoperability.

238 Overall, the discussion of wholesale access in the Review of Regulatory Framework decision is stated in broad terms. The CRTC considered that bottleneck services, as well as those subject to dominant supply, should be unbundled. While the decision does not elaborate an "essential facilities" test, the decision clearly rests on an appreciation of the essentiality of certain facilities:

In telecommunications, the monopoly product (the local exchange network) is, as a practical matter, the sole source of ubiquitous switched distribution to end users, and is, therefore, an essential input, one on which both the vertically integrated firm and its competitive market rivals depend.

The Commission agrees that the telephone companies have both the incentive and the opportunity to use their vertically integrated structures to lessen competition by exploiting control over bottleneck facilities.¹⁶⁷

239 In May 1997, the CRTC issued its decision in the Local Competition proceeding.¹⁶⁸ The CRTC adopted a test for determining when a facility is essential derived from competition policy:

73. The Commission considers that either too narrow or too broad a definition of an essential facility may impair the development of competition. If it is too narrow, competitors may not be able to enter the market because of an inability to obtain the necessary network components. If it is too broad, giving overly generous access to ILEC inputs, CLECs may not have sufficient incentives to invest in their own facilities, and would enter and remain in the market primarily as resellers. The Commission is of the view that efficient and effective competition will be best achieved through facilities-based service providers; otherwise, competition will only develop at the retail level, with the ILECs retaining monopoly control of wholesale level distribution.

74. In light of the above, the Commission concludes that ILECs should generally not be required to make available facilities for which there are alternative sources of supply or which CLECs can reasonably supply on their own. Accordingly, the Commission considers it inappropriate to define an essential facility as a facility that is provided by a dominant firm with market power because it would require facilities to be treated as essential even in the face of the demonstrated feasibility of alternative provision, including self-supply. The Commission concludes that to be essential, a facility, function, or service must meet all three of the following criteria: (1) it is monopoly controlled; (2) a CLEC requires it as an input to provide services; and (3) a CLEC cannot duplicate it economically or technically. Facilities that meet this definition shall be subject to mandatory unbundling and mandated pricing. As well, the tariffed rates for these facilities shall be treated as costs in the imputation test. [Emphasis added.]

240 A number of points in this reasoning are noteworthy. First, the CRTC announces a clear goal of achieving facilities-based competition. Resellers are not expected to remain resellers; they should have incentives to develop their own facilities.

241 Second, the CRTC recognized that if unbundling was too generous, CLECs would not be incented to develop their own facilities.

242 Third, the CRTC initially made it clear that ILECs should not have to unbundle facilities for which there are alternative sources of supply or that CLECs can self-supply; that is, facilities that are not essential. Of course, the CRTC departed from this principle later in the same decision.

243 Fourth, the test for an essential facility in the Local Competition Decision is similar to that derived from US case law by Professor Robinson. The main difference is that the Local Competition test refers to the facility as being essential to "a CLEC", whereas Professor Robinson says "essential to competition". This is in fact a big difference, if what is intended by "a CLEC" is that if one CLEC requires the facility where others do not, the facility will be "essential". The same difference exists in the non-duplicability element. If the focus is on "a CLEC" as opposed to competition/competitors generally, the danger is that an underfunded or poorly managed CLEC will be coddled.

244 The CRTC then defined three essential facilities: central office codes, subscriber listings, and local loops in certain bands (small urban and rural areas). The CRTC also ordered unbundling of certain other facilities for a period of five years, even though it found that these facilities did not meet the requirements for an essential facility. There were, generally, local loops in urban areas, and various kinds of transiting. The CRTC did not make a finding on 911 service, but accepted Stentor's proposal to make it available to CLECs. Similarly, although it found that Message Relay Service was not essential, the CRTC accepted Stentor's offer to unbundle it and price it as an essential facility.

245 Interestingly, the CRTC held that rights-of-way and support structures were not essential facilities, because of the framework in the Telecommunications Act 1993. The CRTC also found that inside wire and local switching were not essential facilities. Nor were directory assistance and white pages directories, but this was because the CRTC ordered exchanges of subscriber listings.

246 The CRTC ordered that essential services be priced at Phase II costs plus a 25% markup, in order to recognize common costs in addition to Phase II costs.

247 The CRTC imposed certain obligations on all LECs, including CLECs, such as interconnection, 911 service, privacy protection, and subscriber listings.

248 As noted in the Public Notice for this proceeding, in the Price Cap Regulation and Related Issues decision, the CRTC began using the term "Competitor Services" to describe services used by competitors, including interconnection services used by wireless carriers. The CRTC stated that it intended to designate services designated as competitor services in a follow-up proceeding.¹⁶⁹

249 The follow up proceeding was Final Rates for Unbundled Local Network Components, wherein the CRTC established rates for wholesale access to unbundled services. This decision is illustrative of the complexity of establishing cost-based pricing for wholesale access.¹⁷⁰

250 In Regulation Under the Telecommunications Act of Certain Telecommunications Services Offered by "Broadcast Carriers", the CRTC referred to previous findings that the retail internet service market was highly competitive. There were numerous providers and switching was relative easy. Cable companies were not yet a major factor and did not have market power. The CRTC noted that

...the retail level ISs market is characterized by rapid and significant innovation, minimal barriers to entry and minimal restrictions on competitors' ability to increase supply in response to a non-transitory price increase.¹⁷¹

251 The CRTC then reached the opposite finding in the market for high speed access services (that is, wholesale access to underlying facilities for high speed internet access), finding that the incumbent cablecos and telcos had substantial market power and that it would therefore regulate rates and terms on which they provide access.

252 There does not appear to have been an analysis or determination as to whether high speed access services from cablecos were an "essential" or "near-essential" facility in this decision.

253 In September 1999, the CRTC ordered incumbent cablecos that offer high speed internet access to make these services available for resale at a discount of 25% from the lowest retail price they charge. The cablecos had not, at this point, provided access as provided for in decision 98-9, for technical reasons. The CRTC declined to order that they provide immediate access.

254 As noted in the Public Notice for this proceeding, in 2001, the CRTC's order Local competition: Sunset clause for near-essential facilities extended indefinitely mandated wholesale access to "near-essential" facilities. The CRTC did so on the basis that these facilities are "critical inputs required by entrants" and that ILECs are the only source of these facilities in virtually all cases. Despite this, the CRTC declined to re-categorize near-essential local loops as essential.¹⁷²

255 In its 2002 decision, Regulatory framework for second price cap period,¹⁷³ the CRTC announced the price cap régime that would apply to ILECs for the next four year period. Among the objectives of the pricing régime was fostering facilities-based competition. The CRTC divided Competitor Services into Category I, essential and near-essential services, and Category II, services developed for use by TSPs.¹⁷⁴

256 The CRTC noted that, by the end of 2000, competitors served 10.3% of the business market (but up to 16% in large urban centres) and only 0.2% of the residential market. It concluded that local competition is in its early stages, and almost non-existent in residential markets.¹⁷⁵ The CRTC noted that quality of service provided by ILECs to CLECs was substandard in an unacceptably high number of months.¹⁷⁶

257 The CRTC was alive to the concern pricing wholesale access to ILECs' facilities were priced too low would introduce disincentives to the construction of facilities and harm an important source of revenue for existing facilities-based CLECs. It thus rejected proposals for steeply discounted pricing of wholesale access.¹⁷⁷

258 The CRTC considered proposals to designate new services as Competitor Services. It designated Digital Network Access services ("DNA") as a Category I Competitor Services, although it recognized that was some self-supply of some components of DNA facilities. It rejected Distributel's submission that Centrex tie trunk terminations should also be designated.

259 The CRTC reduced the markup on Category I services to 15% above Phase II costs (from 25%), and addressed certain specific markup issues.

260 The CRTC reached the conclusion that the monitoring régime then applied to ensure quality of service was not sufficient. It decided to introduce incentives, in the form of rate adjustments, including rate adjustments for competitors.

261 In June 2002, the CRTC dismissed a complaint by some internet service providers. The complaint was basically in the nature of a margin squeezing complaint. The IPSs said that Bell offered wholesale DSL for up to $228 per month while charging retail customers $34.95 for DSL. The CRTC relied on earlier findings that low and high speed internet access were in the same product market and that there was a high degree of competition between telcos and cablecos for high speed residential internet access.¹⁷⁸

262 In its 2004 decision, Ethernet Services, the CRTC approved, on an interim basis, Bell Canada's application to offer Ethernet services. At the same time, the CRTC ordered Bell Canada and other ILECs to make Ethernet services available to competitors on an interim basis, until the CRTC could determine the adequacy of Ethernet substitutes.

263 In its 2005 decision, Competitor Digital Network Services,¹⁷⁹ the CRTC re-affirmed its finding in decision 2002-34 that ILECs should develop and make available competitor DNA ("CDN") services. The CRTC found that competitors relied heavily on ILEC provided DNA; non-ILEC supply (self-supply and third party supply) of DNA was 23% of the total, but that certain adjustments would reduce this number.¹⁸⁰ Competitors faced greater constraints on facilities construction than ILECs, and competition in the local wireline market continued to be very limited, the CRTC found.

264 The CRTC examined each of the various services that make up CDN services. The CRTC seems to have applied two slightly different tests in determining whether to mandate access to a particular service. In some cases, current reliance by CLECs led to mandating access. In others, the feasibility of self-supply from a cost perspective was considered.

265 The CRTC then considered whether these services should be assigned to Category II or II, and, within Category I, whether they were essential services. The CRTC held that only two services were essential (optical CO co-location link and other CO connecting link). DS-0 and DS-1 access were classified as Category I near essential services. The remaining services were classified as Category II services.

266 The CRTC determined that competitors could use any of the CDN services in conjunction with other ILEC services or non-ILEC services.¹⁸¹

267 In January 2007, the CRTC issued an order, Ethernet Services, containing its final determination on Ethernet services.¹⁸² The CRTC did not expressly apply the test for essential services it first elaborated in Decision 97-9. It did, however, find that the same constraints apply to Ethernet services as apply to fibre-based CDN services.¹⁸³ The CRTC noted its finding in Decision 2005-6 that third-party fibre-based suppliers have the potential to increase their supply to competitors,¹⁸⁴ and held that incentives to build fibre-based facilities would be unduly diminished if it designated Ethernet access services as Category I, and it classified them as Category II. The CRTC also classified Ethernet transport services as Category II because competitors have other options. It classified Ethernet CO link services as Category I because only ILECs can supply them.

268 Taken as a whole, the CRTC's decisions after the Local Competition decision show an increasing departure from the principled approach articulated in that decision toward what comes close to a presumption in favour of wholesale access. The CRTC went from the view that "ILECs should generally not be required to make available facilities for which there are alternative sources of supply or which CLECs can reasonably supply on their own" to doing exactly that.

B. The Commissioner of Competition's definition

269 The Commissioner of Competition has proposed a variation on the definition of an essential facility in the TAB for purposes of determining whether to mandate wholesale access:

• The firm controlling the facility in question is vertically integrated and dominant in two markets. The first relevant market is the upstream market (or wholesale market) for the facility. The second relevant market is the downstream market (or retail market) in which the facility is an input. A necessary condition for concluding that there is dominance in the upstream market is that it is not practical or feasible for competitors to duplicate the facility in question;
• Mandating access to the facility is likely to result in competitors entering or expanding in the downstream market; and
• Such entry or expansion is likely to result in competition sufficient to allow for removal of economic regulation in the downstream market within a reasonable period of time.

270 The first element is virtually identical to the first element in the test in the TAB. It retains the double-dominance requirement and retains the non-duplicability requirement as part of the test for dominance.

271 The comments I made about these requirements above apply here as well. However, my concern about having a different standard for dominance may not arise here, given that this test is not intended to be a test under section 79.

272 The Bureau explains why it would require a showing of downstream dominance in its Supplementary Evidence:

The requirement for dominance downstream is to recognize that if the owner of the facility is not dominant downstream, it does not have market power that is durable and significant. In these circumstances, there is already facilities-based competition in the downstream market, and thus the benefits of mandated wholesale access will be limited. Moreover, mandating wholesale access could be a disincentive for further investment by competitors in their own facilities so the costs of mandating access could be high. Consequently, the benefits from any increase in competition resulting from mandated access is unlikely to warrant the costs of mandated access. [Emphasis added]¹⁸⁵

273 It is clear from that the Bureau is limiting its consideration to facilities-based competitors downstream. This is obviously sensible: otherwise an ILEC could point to downstream competitors that depend on it for access under the current mandatory regime to argue that the compulsion to provide them with access should be removed.

274 Given the Bureau's 1/n market shares for facilities-based competitors that are not capacity constrained, the presence of one facilities-based competitor means that the owner of the facility is not dominant downstream. Thus if there is one facilities-based competitor, in addition to the ILEC, the Bureau's test is not met and access should not be mandated to the facility.

275 I agree that the Bureau's downstream dominance requirement is a useful screen, although it may, technically, be redundant. There are two reasons for this. First, what we care about is competition in the downstream market. If the owner of the facility does not have market power in this market, then there is no problem to fix. Second, if there is facilities-based competition in the downstream market, it follows that is has been practical and feasible for at least one competitor to duplicate the facility in question. Unless we adopt some arbitrary target for the number of facilities-based competitors we want in the downstream market, that is the end of the matter, and an element of the test cannot be made out and the facility is not essential.

276 I find the Bureau's discussion of the "strict" and "weaker" interpretations of the non-duplicability requirement confusing. As I understand it, the difference stems from effectiveness of a competitor that has duplicated the facility. One is looking at the costs of duplicating the facility, as opposed to actual competitors.

277 Thus I am not sure that this is not just another way of stressing that the requirement is not that it must be impossible to duplicate the facility, but that it must not be practical or feasible for a reasonably efficient competitor to duplicate the facility. Professor Robinson makes this very point in his first report:

Non-duplicability. Non-duplicability, of course, refers to the absence of any source of supply for the facility in question other than the monopolist owner. Practical non-duplicability can be a product of a number of different constraints that make self-provision impractical. For example, there might be technical or legal constraints (such as a patent) on self-provisioning. However, the usual ground for finding that a particular functionality cannot be duplicated is that it is not feasible to do so because market conditions do not make it economical to do so (the demand for the end services enabled by the facility will not support the cost of duplicating a particular facility). Consistent with the definition of essentiality above, non-duplicability is an objective condition of the market, not the financial status of a particular rival. If it is practical for a reasonably efficient competitor to self supply the facility in question, it is not relevant whether a particular entrant can do so.¹⁸⁶

278 That being said, Professor Robinson also stresses that "‘essential' is not the same as ‘convenient'".¹⁸⁷ That is, the fact that the alternative is more expensive is not a reason to mandate access – unless, presumably, the cost difference makes duplicating the facility uneconomic.

279 The second and third elements in this definition are functionally equivalent to the substantial prevention part of the third element in the TAB. If the first element in the Bureau's definition is satisfied, then neither the upstream nor the downstream markets are competitive. The second and third elements ask the question: will competition be increased, and by how much, if we mandate access? The question could be turned around to ask: if we do not mandate access, will competition be prevented, and to what degree?

280 In its Supplementary Evidence, the Bureau explains the second and third elements by pointing to a distinction between "substantial lessening or prevention of competition" and "competition sufficient to allow for the removal of economic regulation". I do not agree with this distinction. In principle, "competition sufficient to allow for the removal of economic regulation" would be a substantial increase in competition, which would imply that any denial of the owner to grant access to the facility would lead to a substantial prevention of competition.

281 The Bureau explains in its Supplementary Evidence what it means by "competition sufficient to allow for the removal of economic regulation":

The Bureau's test assumes that retail forbearance occurs only when there is competition in downstream markets that is sufficient to replace regulation as a means of controlling the exercise of market power. If retail forbearance is granted prior to the establishment of such competition, there will be pressure to inefficiently expand the wholesale regulation regime as an imperfect substitute for retail regulation.¹⁸⁸

282 In other words, the Bureau's third element requires an improvement in the downstream market from a situation where the owner of the facility is dominant, to one where the facility owned cannot exercise market power (or, at any rate, has less scope to exercise market power). I would have thought that if that test were met, the substantial prevention of competition test would also be met.

283 However, the test for local forbearance assumed by the Bureau above is not the test. When viewed against the test for forbearance from local rate regulation as set out in Decision 2006-15 as modified by the Local Forbearance Variation Order, "competition sufficient to allow for the removal of economic regulation" turns out to be a very low threshold. That test (which was set out above) consists of three alternative tests, any one of which is sufficient to impel forbearance (provided quality of service requirements are met).

284 The first is the test for a finding that the ILEC does not have market power proposed by the Bureau. Its main requirement is the presence of two facilities-based providers, where the competitor has similar or lower costs and neither is capacity constrained.

285 The second test is based simply on the presence of at least two independent facilities-based TSPs able to serve at least 75% of the market, including mobile wireless services. Given that there are at least two facilities-based mobile wireless providers in most (if not every) major Canadian city, this test comes down to a mere requirement for one facilities-based fixed-line provider.

286 The third test relates to business customers and is also based on the presence of a competitive, fixed-line provider that serves business customers.

287 It is important to note that "facilities-based", for purposes of the second and third tests, includes a carrier that leases its lines, presumably even those whose lines are leased from the ILEC.

288 It follows from this test that if mandating access would be likely to cause entry or expansion of a CLEC relying on wholesale access, with the ability to serve 75% of the market, the second forbearance test would be met (assuming the QoS requirements are also met) and thus also the second and third elements of the Bureau's test.

289 In its Supplementary Evidence, the Bureau cautions that the definition of "facilities-based" imposed by the Local Forbearance Variation Order must be interpreted narrowly.¹⁸⁹

290 The problem with the Bureau's advice is that it is not open to the CRTC to adopt a different test for local forbearance than that set out in the Local Forbearance Variation Order. As well, I see nothing in the definition of "facilities-based" in that order that supports a "narrow interpretation" that would exclude competitors that rely on facilities leased from an ILEC. On the contrary, the order says that TSPs that use a combination of their own facilities and facilities leased from "other services providers" are "facilities-based" service providers. "Other services providers" must mean what is says, and include the local ILEC.

291 Finally, and perhaps most importantly, events have overtaken the Bureau's test. In July, the CRTC announced that it was forbearing from local rate regulation in most major Canadian cities, and even some smaller ones. The Bureau's forbearance-based test has thus become virtually meaningless.

C. Definitions proposed by the parties

292 The Companies (Bell Canada and affiliates, and SaskTel) support the Bureau's approach as set out in the TAB.

293 There is one apparent difference between the Companies' approach and the Bureau's. The Companies' test requires significant market power. The Tribunal has long equated the requirement in section 79 of the Competition Act of showing substantial or complete control of a market as requiring proof of market power. However, merely having a little bit of market power would not, in my view, meet the statutory requirement of showing substantial or complete control. This issue has never come up in an abuse of dominance case as in every case, the respondent did have significant market power.

294 I note that the Companies, relying on Dr. Taylor's report, argue that the TAB test is substantively the same as the test in the Local Competition decision, which is in turn based on the same test as that proposed by Telus.

295 Telus is an ILEC in Alberta and British Columbia and a CLEC primarily serving business markets in other provinces. Telus also offers wireless telephony.

296 Relying on Dr. Weisman, Telus proposes the following the three part test for determining whether a facility is essential:

An essential facility is a facility, function, or service that at a minimum satisfies all three of the following criteria: (1) it is "monopoly" controlled; (2) CLECs require it as an input to provide services; and (3) CLECs cannot duplicate it economically or technically.¹⁹⁰

297 Professor Robinson's phrasing is slightly different, but not substantively so, except that he adds a requirement that it be feasible for the owner of the facility to provide access:

(1) control of the essential facility by a monopolist; (2) a competitor's inability practically or reasonably to duplicate the essential facility; (3) the denial of the use of the facility to a competitor; and (4) the feasibility of providing the facility.¹⁹¹

298 This test is based, as Professor Robinson says, on decisions of the circuit courts in the US (federal appeal courts). As Professor Robinson concedes, the status of the doctrine has not been finally determined in the US. Indeed, the Trinko case, the US Supreme Court pointed out that it had never recognized the doctrine:

This conclusion would be unchanged even if we considered to be established law the "essential facilities" doctrine crafted by some lower courts, under which the Court of Appeals concluded respondent's allegations might state a claim. ... We have never recognized such a doctrine, ... and we find no need either to recognize it or to repudiate it here. [Emphasis added; citations omitted]¹⁹²

(i) Monopoly control

299 Dr. Weisman and Professor Robinson make it clear that the first requirement, monopoly control, means just that: the facility must be a natural monopoly controlled by one owner.¹⁹³

300 Whatever the merits of the monopoly requirement from a theoretical standpoint, it is a difficult fit with Canadian competition law. The first element of section 79, dominance ("substantially or completely control"), requires a high degree of market power, but not monopoly.¹⁹⁴

(ii) Essentiality

301 Professor Robinson points out that the essentiality requirement "should be defined by reference to the functionality not the form of the facility".¹⁹⁵ I agree, so long as "functionality" is not taken to imply that functional interchangeability is the only factor to be considered. A product market needs to be defined upstream as well as downstream. Any upstream products that are good substitutes for the supposedly essential facility must be included. This exercise will, of course, affect not only the determination of whether the facility is essential, but whether the owner of it is dominant.

302 An important question is, essential for whom? Dr. Weisman points to his use of "CLECs" as opposed to "a CLEC". What he means is all CLECs, or CLECs in general:

It is important to recognize that the second necessary condition states that "CLECs require it as an input to provide services" rather than "a CLEC requires it as an input to provide services" as provided in the CRTC's current definition of an essential facility. Indeed, a single CLEC that does not require the input to provide the downstream service is sufficient to establish that the facility, function or service is not essential. This is necessarily the case because the CLEC that does not require the input has served to establish that the functionality can be duplicated and if it can be duplicated it is not essential.¹⁹⁶

303 I agree, up to a point. Where a CLEC has duplicated the facility, it follows that the facility can be duplicated. This may not follow, however, where an incumbent cableco has upgraded its network to handle high speed internet, and thus also VoIP. Incumbent cablecos have advantages similar to those of ILECs when it comes to facilities in residential markets. Thus the fact that a cableco has duplicated the facility may not mean that another competitor also can.

304 Another consideration with Dr. Weisman's use of "CLECs" is that not all CLECs are offering the same products. Residential local and long distance telephone service can now be offered through access-independent VoIP services. Thus it would be difficult for a local and long distance provider to claim that local loops are an essential facility for CLECs generally. However, this does not answer a claim by a CLEC that offers high speed internet service that it needs access to facilities. Thus the essentiality of a facility may turn on the products offered in the downstream market.

(iii) Non-duplicability

305 This essentiality and non-duplicability elements are related. If the supposedly essential facility or its functionality can be duplicated economically, then it cannot be essential.

306 Professor Robinson also points out that we do not look at what actual competitors think they need; rather, we ask whether hypothetical reasonably efficient competitors need the facility.¹⁹⁷

307 Professor Robinson uses the phrase "practical non-duplicability", and observes:

the usual ground for finding that a particular functionality cannot be duplicated is that it is not feasible to do so because market conditions do not make it economical to do so (the demand for the end services enabled by the facility will not support the cost of duplicating a particular facility).¹⁹⁸

308 I think another way of saying this might be: if the cost of duplicating the facility is so high that the competitor's service could not be priced to compete with the incumbent's, then it is not feasible to duplicate the facility.¹⁹⁹ As discussed above, I think this is less at odds with the Bureau's approach than Professor Robinson asserts.

309 Professor Robinson also points out that, like essentiality, duplicability refers to the function, not the facility. He stresses that this is a particularly important consideration in an industry where technology is changing:

As with the element of essentiality the non-duplicability inquiry is an instrumental one and should not be tied to any particular means of obtaining it. Especially in a field of changing technology – telecommunications is a particular apt example -- it will often be possible to provide a particular product or service with alternative inputs even though conventionally the product or service was provided by an asset that cannot be replicated.²⁰⁰

310 I agree.

(iv) Feasibility of providing access

311 Professor Robinson points out that the facilities owner should not have to incur unusual costs or suffer distortions in its business, such as disadvantaging its own customers in order to accommodate its competitors' customers, in order to share.²⁰¹

312 I agree in principle. In particular, ILECs should not generally have to invest to build or expand a facility in order to share it. Generally speaking, if the ILEC can build it, so can the CLEC. There may be exceptions. For instance, if the ILEC owns the physical space or support structures required for the facility, it may be appropriate either to require the ILEC to grant access to the space or support structures, or to build the facility for the CLEC if it is not prepared to grant access. Presumably in those cases the CLEC should be granted access to build its own facilities, however.

(v) Implied competitive effects test

313 Telus' approach contains an implied competitive effects test. Professor Robinson states:

If a facility is truly essential, its control must allow the owner to prevent or foreclose competition. In fact this seems to be implied as well by the Bureau's definition of non-duplicability, below.²⁰²

314 While Professor Robinson's statement may be a correct statement of the law of the US, it is not consistent with the Competition Act. Here it is important to recognize the difference between the US monopolization provision in the Sherman Act, and the abuse of dominance provision in the Competition Act. Sherman Act §2 reads as follows:

Every person who shall monopolize, or attempt to monopolize, or combine or conspire with any other person or persons, to monopolize any part of the trade or commerce among the several States, or with foreign nations, shall be deemed guilty of a felony, ...²⁰³

315 Section 2 contents itself with stating the general principle – thou shalt not monopolize – and leave the details to be worked out by the courts. There is thus considerably more scope for US courts to adopt different tests depending on the particular theory of monopolization under consideration.

316 By contrast, sections 78 and 79 of the Competition Act contain a detailed code, including an express competitive effects test, substantial lessening or prevention of competition. It would be inconsistent with this provision for the Tribunal to say that complete foreclosure of competition – total prevention – is the standard to be applied in abuse of dominance cases.

317 Telus takes up Professor Robinson's point and asserts that "the essential facilities doctrine is concerned exclusively with the prevention of competition in a downstream market and not a mere lessening of that competition".

318 This is not quite right. First, in paragraph 79(1)(c), and elsewhere in the Competition Act, "substantial" modifies both "lessening" and "preventing".²⁰⁴ The Competition Act does not, anywhere, rely on a total prevention standard.

319 Second, it is quite conceivable that an essential facilities case might involve a substantial lessening of competition. Suppose an owner of an essential facility cuts off a downstream competitor that relies on it for access to the facility. Assuming all the other elements of the test are met, the denial of access would have caused a substantial lessening of competition as well as a substantial prevention of competition.

320 Rogers is a cableco and both a facilities-based and wholesale-access based competitor, primarily in residential markets. Rogers offers wireless telephony. Rogers also offers portable wireless broadband internet, through a joint venture with Bell.

321 Rogers proposes the following definition of "essential facility":

An essential facility means:

(i) an input used to provide services in downstream markets;

(ii) where the input is controlled by a supplier that possesses market power in respect of its supply, such that, absent mandated supply of the input on regulated terms, the supplier could use its market power in the relevant upstream market to prevent or lessen competition in the relevant downstream markets; and

(iii) it is not feasible to duplicate the input, having regard to economic and/or technical factors.²⁰⁵

322 This test sets a lower threshold than either the Competition Bureau's test or Telus'. This is because it only requires that the owner of the upstream facility be able to prevent or lessen competition downstream through the exercise of market power upstream, as opposed to requiring a substantial prevention or lessening of competition.

323 There is no express "essentiality" requirement. Essentiality may be implicit in the requirement that the owner of the facility could use its upstream market power to lessen or prevent competition downstream. If there are substitutes for the upstream facility, then any attempt to exercise market power upstream should in theory cause downstream competitors to switch, and the downstream market should remain relatively unaffected.

324 Similarly, this test does not directly answer the "essential for whom" question. Given the low competitive effects threshold – mere lessening or prevention, instead of substantial lessening or prevention, it might be inferred that the answer is "a competitor" rather than "competitors" generally. This is because knocking out one competitor might lessen competition a bit, but not substantially.²⁰⁶

325 The third element of Rogers' definition appears to be substantively the same as the equivalent requirement in Telus' and the Bureau's test. The debate lies in the application of this element.

326 In its submission, Rogers places considerable emphasis on economic and technical barriers to construction of facilities. As discussed above, the Tribunal has accepted that a combination of economies of scale and sunk costs give rise to a barrier to entry. The Tribunal has recognized that incumbents may have certain advantages derived from their incumbency that constitute a barrier to entry, although I do not believe that "first mover cost advantages" has been accepted as a barrier to entry. I am not at all sure that "absolute cost advantages" would be accepted as a barrier to entry. Competition law encourages efficient entry. My understanding is that an entrant with higher costs is not generally considered efficient in comparison to the incumbent.

327 That being said, to the extent that building up facilities results in a cost structure that is too high for the entrant to be able to compete, it is not feasible or practical to build the facility, as discussed above.

328 In any event, Rogers' evidence shows that even under the current régime, which, arguably, under-incents facilities building by both incumbents and competitors, Rogers and the predecessor to its CLEC operations, Call-Net, found it economical to engage in a massive facilities building and upgrading project.

329 Rogers proposes various criteria for applying its definition of essential facilities. Generally speaking, for business markets, Rogers proposes that facilities be considered essential if the number of business lines is less than a certain number (the threshold varies depending on the facility) and there are fewer than four fibre-based collocated competitors in the market.²⁰⁷ The basis for this four-competitor rule is not because that is what is required for retail competition, but because, in Rogers' estimation, if there are fewer than four such competitors, other competitors will not be able to obtain access to the facility without regulatory compulsion.²⁰⁸ In short, Rogers' test appears to be aimed at ensuring that there are incentives for facilities owners to grant wholesale access to their facilities.

330 So far as local loops are concerned, Rogers says they will always be essential.²⁰⁹

331 MTS Allstream is the ILEC in Manitoba (MTS) and a CLEC in other provinces (Allstream, formerly AT&T Canada).

332 MTS Allstream's believes that the CRTC's access régime has been too narrow. It proposes expanding it.

333 MTS Allstream proposes adopting the Competition Bureau's definition of an essential facility in the TAB, but not the Bureau's test.²¹⁰ My comments about the missing modifier, "substantial", therefore apply here as well.

334 MTS Allstream comments that

Under the proposed definition, essential services would include any facility or service that is critical to a competitor's ability to compete effectively in one or more downstream retail telecommunications service markets and is controlled by a dominant, vertically integrated carrier (in this instance, a former monopoly carrier). More specifically, essential services would include any facility or service that in large part is only available from a former monopoly carrier and for which it is "difficult or impossible for downstream competitors to substitute toward other inputs or to practically or reasonably duplicate."²¹¹ [Emphasis added.]

335 MTS Allstream then proposes the following test for identifying an essential facility:

a. The facility or service provided by the former monopoly is required as an input by a competitor or competitors to provide downstream retail services; and

b. The former monopoly dominates the wholesale supply of the facility or service.²¹²

336 Although MTS Allstream incorporates a form of essentiality requirement (using the terms "critical" and "required"), when MTS Allstream applies this test, it looks at reliance in fact on the facility. For example, in the case of unbundled local loops, MTS Allstream points to the fact that in 2005, 40% of competitors' residential local lines were provided using leased loops.²¹³ In other words, MTS Allstream is effectively relying on a presumption that if the facility is being used, it is "critical" or "required".

337 Whereas Telus' answer to the question "essential for whom" was all competitors, MTS Allstream's answer appears to be any competitor. MTS Allstream defines an essential facility as a facility that is critical to or required by a competitor. This is consistent with MTS Allstream's application of its test, where unbundled local loops are said to be essential even though the majority – 60% – of competitors' local lines do not use leased loops.

338 MTS Allstream insists on the principle of "competitive neutrality". It takes issue with the CRTC's stated preference for facilities-based competition.²¹⁴ MTS Allstream also cites the Policy Direction's direction "not to artificially favour either Canadian carriers or resellers". However, under paragraphs 1(b)(iii) and (iv), these principles only apply where the CRTC is relying on non-economic regulation, including arrangements for network interconnection and access. As the decision whether to regulate has already been made before these principles come into play, competitive neutrality has no application at the stage of deciding whether to regulate. The fact that paragraph 1(c)(ii) states that this review must take competitive neutrality into account does not change at what stage of the decision making it applies.

339 Further, as I argue above, the over-riding instruction in the Policy Direction are the two principles of maximum reliance on market forces, and minimally intrusive regulation, when regulation is used.

340 The Policy Direction also states the regulatory goal for this proceeding: "increasing incentives for innovation and investment in and construction of competing telecommunications network facilities".²¹⁵

341 Finally, the Local Forbearance Variation Order expressly states the government's preference for facilities-based competition:

Whereas the Governor in Council considers that facilities-based competition is a durable form of competition that delivers the greatest benefits to consumers, imposes competitive market discipline on incumbents and strengthens investment in telecommunications infrastructure

342 Consequently, MTS Allstream' submission that the CRTC cannot favour facilities-based competition over wholesale access based competition is inconsistent with the directions to the CRTC and must therefore be rejected.

343 Primus is a CLEC offering local and long distance telephony, resold cellphone service, and broadband internet. Primus also offers an access-independent VoIP service.

344 Like MTS Allstream, Primus supports the Bureau's definition but proposes a test for identifying essential facilities that is different from the Bureau's test. Primus' test is as follows:

(a) It is an input important to a competitor or new entrant in order to compete effectively or efficiently in the provision of retail services; and

(b) The firm controlling the input is assumed to possess the power to lessen or prevent competition in downstream markets provided the CRTC has not determined, as a matter of fact, that there is sufficient evidence of an effective substitute for the input available at a competitive price as part of a sustainable, vigorous wholesale market (i.e. a feasible alternate source of supply of the input). [Emphasis added.]²¹⁶

345 The first criterion replaces the essentiality requirement with a requirement that the facility be "important" to "a competitor". Primus makes it clear that "importance" is a relatively low threshold: "it will be necessary to make a determination that the facility or service is of some importance" [emphasis added]. Primus is thus proposing, not a test for essential facilities, but a test for important facilities.

346 The second criterion begins with a presumption that every owner of a facility has the power to lessen or prevent competition downstream. While that presumption may have been valid before the advent of VoIP telephone service over broadband cable internet, it is no longer a valid presumption. In my view, market power must be proven, not assumed.

347 The proof required to displace the presumption of market power is that there is a vigorous wholesale market for the facility. In other words, under this test it does not matter whether the retail market is competitive or not. What matters is whether a competitor can get into the retail market by obtaining the upstream input. This reasoning could lead to absurd results. Suppose there were five facilities-based competitors offering broadband internet and telephony (traditional or VoIP) and competing vigorously in a particular market. But none of the five offers wholesale access. Under Primus' test, those upstream facilities would be considered essential.

348 It is apparent that Primus' test is intended to ensure that any and every wholesale access-based competitor can enter any market it wants to. It is, in short, a test designed to protect competitors, not competition, and, as such, fails the satisfy one of the fundamental principles of competition law.

349 Cogeco is a cableco and a facilities-based competitor. It takes the position that mandatory access to non-essential services should be phased out:

In many markets competition has evolved to a stage that no longer requires the ILECs to provide some wholesale services, including unbundled local loops and co-location, on a mandated basis.²¹⁷

350 Cogeco proposes the following definition:

A telecommunications facility, function, or service is essential if:

(1) it is either an interconnection service required to permit the interchange of traffic or a network access service that cannot be economically or technically duplicated or substituted and is required to provide retail telecommunications services; and,

(2) it provides the firm controlling it with the power to prevent or lessen competition in a relevant downstream telecommunications market.²¹⁸

351 The first branch of the definition gives rise to a presumption that the facility is essential. Interconnection is presumed to be essential without any inquiry. Network access services are subject to a non-duplicability requirement. Cogeco defines network access services as

...all services provided by ILECs or other carriers, namely CLECs or cable companies, by which they make parts of their network available and provide ancillary services to other telecommunications carriers enabling them to provide retail telecommunications services.²¹⁹

352 The definition incorporates a non-duplicability test, but not an express "essentiality" test. Perhaps Cogeco is assuming, not illogically, that some form of network access is essential to providing retail telecommunications services. The question is whether a competitor can duplicate it. In addressing this point, Cogeco makes it clear that it is the functionality that matters, not the precise service. As well, Cogeco's answer to the "essential for whom" question appears to be "competitors" not "a competitor":

In assessing whether the input can be economically or technically duplicated or whether the same functionality could be obtained elsewhere, the Commission should consider other geographic markets where facilities-based competition exists. For example, if similar exchanges exhibit competition in the supply of local exchange services by facilities-based competitors, the Commission should conclude that the input can be economically or technically duplicated and that a functional alternative for the input exists. [Emphasis added.]²²⁰

353 Cogeco does not support the "weaker" interpretation of the non-duplicability element referred to by the Bureau, because of the effect on incentives for investing in competing facilities that this interpretation may have. Cogeco favours "a clearer test, with less ambivalence than proposed by the Competition Bureau", and proposes that the test be that the facility "cannot be economically or technically duplicated".²²¹

354 The second branch of the test is basically the definition of an essential facility stated by the Bureau in the TAB. Cogeco adopts, with reservations, the Bureau's requirement of showing dominance in the downstream market, and proposes two alternative criteria for determining when the second branch of its test is met:

• The telecommunications carrier controlling the input is found dominant in the relevant downstream markets, meaning that the dominant firm has the power to prevent the entry by a potential competitor by denying access to the essential service required to offer competitive retail services in a relevant downstream market; or,
• The telecommunications carrier controlling the input is found non dominant in any relevant downstream markets, but has the power to significantly preserve or enhance barriers to entry and/or [prevent entry or] expansion of competitors in any telecommunications retail market, by denying access to the essential service. [Emphasis added]²²²

355 These criteria are really a form of competitive effects test. The first criterion, which requires prevention of entry by a competitor, is similar to Telus' total prevention of competition standard. However, Cogeco refers to "entry by a potential competitor". Does this mean that a network access facility might be considered essential even if there is already a competing facilities-based provider in the market? I think not. Assuming that both of these facilities-based providers could supply network access services to a potential wholesale access-based competitor, neither could, acting individually, prevent entry by the competitor. For entry to be prevented, both must refuse. Thus neither would have the market power of the sort required by Cogeco's test. (Of course, refusal by both (whether the refusal is independent or coordinated) would prevent entry, assuming the potential competitor truly requires wholesale access.)

356 Shaw is also a cableco and a facilities-based competitor. Shaw proposes a definition that is intended to be based on the existence of market power in both the upstream and downstream markets. Shaw begins with three propositions:

a) If the owner of the facility faces effective competition downstream from firms that do not rely on it, then it has no ability to prevent or lessen competition downstream.

b) If there is no or limited competition downstream, then the owner of the facility can only prevent or lessen competition downstream if it has market power upstream.

c) The owner of the facility will have market power upstream if it faces no, or only limited, competition upstream, and it is not economically or operationally feasible for competitors to duplicate the facility.²²³

357 Shaw derives a three-part test from these three propositions:

It follows that a wholesale service is essential if:

(1) there are no or a limited number of competitors in a downstream market that do not rely on the wholesale service;

(2) there are no or a limited number of competing suppliers of the wholesale service; and

(3) it is not operationally or economically feasible for competitors to duplicate the wholesale service.²²⁴

358 Shaw provides little explanation of this test in its evidence. However, Shaw's responses to interrogatories are helpful. In particular, Shaw provides two important clarifications about its non-duplicability element. First, Shaw says that it is the functionality, not the facility itself, that counts: "duplication of a service or facility does not necessarily require that a competitor deploy the same technology".²²⁵ Second, Shaw effectively answers the question "essential for whom" with "competitors", not "a competitor". Shaw states that if a competitor can self-supply a facility with different technology, replacement of the wholesale facility is practical and feasible. In other words, if one competitor can do it, so can others.²²⁶

359 Shaw's formulation of the first two requirements is somewhat different from the way a competition lawyer would put them. Strictly speaking, for instance, one does not assess the state of competition just by looking at the number of competitors in the market. The first element, however, rightly discounts the presence of competitors that rely on the facility owner.

D. Analysis and conclusions

360 A preliminary question is: should the CRTC use the same test as the Competition Bureau or Tribunal would use in dealing with an abuse of dominance case involving denial of access to an essential facility? Or, should the CRTC develop its own test designed to deal with the exigencies of the CRTC's mandate?

361 In the Local Competition Decision, the CRTC adopted an explicitly competition policy-based test. But the CRTC appears to have recoiled from the consequences of that test, and adopted a case-by-case approach.

362 In this proceeding, while CLECs such as Primus, MTS Allstream, and Rogers, propose tests that to varying degrees appear to rely on competition law concepts, their tests are designed to ensure continued wholesale access for competitors that rely on it, without regard to whether such reliance or even those competitors are necessary for competition in downstream markets. In short, these tests violate the fundamental principle of competition law that it is competition, not competitors that is to be protected.

363 The evidence shows that progress has been made towards the CRTC's goal of fostering competition in telephony, albeit perhaps not in the way that the CRTC foresaw. But the existing regulatory structure has a tendency to coddle individual competitors. In order to fulfil its mandate to rely on market forces as much as possible, the CRTC should adopt a test that is consistent with the principle that competition, not competitors, is to be protected. Although competition authorities are sometimes criticized for being too interventionist, generally speaking, competition law is premised on reliance on market forces. Imperfect though it may be, competition law is, in theory, designed to protect those market forces from anti-competitive conduct, but otherwise to let the competitive process determine who wins and who loses.

364 Consequently, I favour relying on a test derived from competition law concepts, particularly, those applicable to section 79. Although the refusal to deal provision is similar to the essential facilities doctrine, I do not think it provides a workable test for determining when wholesale access should be mandated.

365 That being said, the CRTC is not applying section 79 and is not bound to apply any particularities of its structure or judicial interpretation that cause difficulty in the regulatory context. For instance, section 79's reliance on exclusionary, predatory or disciplinary purpose should not, in my view, be grafted onto the CRTC's essential facilities test. Similarly, the Canada Pipe Appeal's narrow approach to business justifications would not be inappropriate here.

366 I propose the following test:

a) The owner of the facility is vertically integrated;

b) The owner is dominant over the product market for the facility;

c) The functionality provided by the facility is essential for reasonably efficient competitors of the owner in a downstream market;

d) It is not practical or feasible for competitors to duplicate the functionality provided by the facility;

e) The owner either has refused to provide access to the facility, or the owner and a competitor have failed to reach agreement on access;

f) The facility can feasibly be provided or made available to competitors; and

g) Not mandating wholesale access to the facility would substantially lessen or prevent competition in a downstream market (or, put more positively, mandating wholesale access would substantially increase competition in the downstream market).

367 It should be apparent that this test has much in common with the US test as outlined by Professor Robinson, except that I have used Canadian competition law concepts. This test is also very similar to the test proposed by the Competition Bureau. This test appears to have more elements than either of those tests. However, both Professor Robinson and the Bureau packed several elements into one at various places in their tests. I have separated the elements out for the sake of clarity.

368 I will discuss each element briefly below. The following discussion assumes the analysis of section 79 conducted above.

369 As a practical matter, disputes over essential facilities in the telecom industry will involve vertically integrated facilities owners. That being said, as the Bureau recognizes in the TAB, the vertical integration need not be structural; the facilities owner could achieve the same result through an exclusive contract with a preferred downstream firm.²²⁷

370 This element involves the same analysis as the dominance element of the abuse of dominance provision in the Competition Act (s. 79(1)(a)). That is, the analysis begins with determining what product market the facility is in. Because this analysis considers whether there are good substitutes for the facility in question, if competitors are able to self-supply or purchase from another supplier equivalent functionality to that provided by the facility, that product will likely be in the same product market as the facility in question. If the alternative facility offers substantial excess capacity, or can be expanded, then, as the Bureau points out in the TAB, the owner's ability to exercise market power may be reduced or non-existent, and 1/n market shares may be appropriate.²²⁸

371 Both direct and indirect evidence of substitutability can be used. Some parties assert that prices under the current wholesale access régime are too low, such that CLECs are under-incented to build facilities or obtain them elsewhere. The traditional hypothetical monopolist, or SSNIP, test may not be much use when asking whether another facility is a good substitute for a facility currently subject to mandatory access. That is because the SSNIP test assumes that the price is at or near the competitive level. If it is not, adjustments must be made. However, we probably do not know with certainty what adjustments to the price are necessary (the ILECs say the price is too high; the CLECs say it's too low).

372 Barriers to entry must also be considered, as if there are no barriers to entry, it is not possible for the owner to exercise market power. It seems to be generally acknowledged, however, that the telecom industry is characterized by barriers to entry.

373 "Dominant", as discussed above, is a somewhat lower threshold than the monopoly threshold favoured by Professor Robinson. However, in cases decided to date under section 79, the dominant firms enjoyed extremely high market shares. If the owner's market share is below 50%, it will not be dominant.

374 I have left out the Bureau's requirement for dominance by the owner in the downstream market. One can imagine scenarios where the owner is not dominant downstream, but uses the facility to achieve dominance. However, as these scenarios are not very likely, I agree that it is a useful screen. What we care about here is competition in the downstream market. Consequently, once a finding is made that the owner is not dominant downstream, or the downstream market is competitive, the inquiry should generally stop there.

375 However, the assessment of the downstream market for purposes of this screen should not include competitors that rely on the owner of the facility for access. Because of mandatory wholesale access, we do not know if the facility owner would have voluntarily provided access on terms that the competitor would accept; nor we do know if the facility owner would continue to provide access if it was not compelled to. Counting wholesale access-based competitors could lead to identifying as non-essential facilities that are in fact essential. In other words, if there the are only access-dependent competitors downstream, once access is no longer mandated, the owner might cut off the competitors and achieve market power downstream.

376 In the Local Forbearance Variation Order, the government directed the CRTC to adopt, for purposes of local forbearance, the Bureau's test for finding that an ILEC is not dominant. In my view, this is an appropriate test for identifying situations where the ILEC is not dominant downstream. If the test is not met, this does not necessarily mean that the ILEC is dominant, however, although it probably will be.

377 I have phrased this requirement so that it is clear that the answer to "essential for whom" is "competitors", not "a competitor", and that we are talking about reasonably efficient competitors. This flows from the maxim in competition law that competition, not competitors, is to be protected.

378 I used the phrase "functionality provided by the facility" to emphasize that the analysis needs to look at the functionality, not just the particular facility or technology.

379 If it is not economical to duplicate the facility or obtain it from a third party, then the facility cannot feasibly be duplicated. Apart from the cost involved, operational, technical, or legal constraints may render a facility non-duplicable.

380 Once again, it is the functionality that is important, not the particular facility. If there are alternatives to the facility that are good substitutes and are thus in the same product market, then the facility can feasibly be duplicated.

381 I do not intend by using the word "functionality" that the product market analysis should be confined to considerations of functional interchangeability. Other criteria for determining the relevant product market should be considered. That being said, I think that functional interchangeability may be one of the most important factors in this analysis.

382 Interestingly, the regulatory approach applied to lower layers in the network architecture affect the duplicability of higher layers. Thus, mandating access to support structures and getting tough with municipalities that obstruct construction of facilities will make it easier for CLECs to build facilities.

383 Similarly, ILECs are to some extent in control of barriers to entry. For instance, even if access to support structures continues to be mandated, ILECs may be able to put roadblocks in the way of CLECs' using those facilities. Assuming it to be accurate, Primus' account of Bell Canada's obstruction of its use of conduit in Toronto would be an example.²²⁹ Paradoxically, if ILECs want to be free of the obligation to share local loops and similar facilities, it is in their interest to facilitate access to support structures.

384 When considering whether to mandate access ex ante, this element can be presumed.

385 However, to the extent that the CRTC adopts an ex post régime, a refusal is, of course, necessary.

386 The alternative, "or the owner and a competitor have failed to reach agreement on access" might strike some as odd. Should the requirement not be that the owner has refused to provide access on reasonable terms?

387 There are two reasons for stating the alternative requirement as I have. First, if the facility is not essential, then the owner does not have to grant access or offer access on reasonable terms. If the owner wants or needs the revenue associated with granting access to the facility, then the market will force the owner to offer access on reasonable terms. Consequently, second, there is no point getting hung up on whether the ILEC or the CLEC were being reasonable in their negotiations. Before dealing with the extremely difficult question of what terms are reasonable, we should determine whether the facility is essential. If it is, and access is to be mandated, reasonable terms will need to be determined, of course.

388 There should be a business justification defence for the ILEC. It should be broader than the business justification defence recognized by the Federal Court of Appeal in Canada Pipe.

389 That being said, I suspect that cases where the business justification defence would be relevant would be rare. If the objection is that granting access is not economic for the owner, the simple answer is that the price paid by the competitor should be high enough to make it economic. If granting access would force to the owner to build facilities, and it is economic to do so (in the sense that the competitor is prepared to pay the price), then it is unlikely that the facilities were non-duplicable (unless the reason they are non-duplicable is operational or technical, for instance, if land or buildings owned by the ILEC were needed to build the facility, and the ILEC was not prepared to grant access to the land or buildings).

390 This element comes from the substantial prevention or lessening of competition requirement in the abuse of dominance provision (s. 79(1)(c)). Usually a denial of access to an essential facility will involve a prevention of competition, because it prevents entry or expansion.

391 This is a lower threshold than the total prevention of competition standard adopted by Professor Robinson.

392 A number of parties have either expressly or implicitly raised the question of how many independent networks we need for there to be competition.

393 Some point to the finding of the TRP Report that three national wireless carriers does not seem to be enough to ensure competition:

The smaller number of mobile providers in Canada — and the fact that all three national wireless service providers are also owned by large telecommunications service providers that also provide wireline services — may mean that there is less competition in the Canadian wireless market than in the U.S. market, which consequently has resulted in higher prices, less innovation, lower uptake and lower rates of usage.²³⁰

394 Some parties to this proceeding have argued that an ILEC/cable duopoly will not provide sufficient competition. In her submission to the TRP, the Commissioner of Competition took a different view:

24. Assuming, for the purposes of this question, that the market does evolve into a form of duopoly between Incumbent Local Exchange Carriers (ILECs) and cable companies, the implications for the telecommunications market could be substantial improvements over the monopoly conditions that have, for so long, dominated this sector.

25. The Bureau considers that the potential positive outcomes of a cable/ILEC duopoly might include a decreased likelihood for unilateral effects (i.e., that a firm could profitably raise prices above competitive levels) due to the degree of rivalry between the cable companies and the ILECS. Aggressive price competition and innovation are other potential benefits that could be expected since most of the costs of service provision (i.e., network costs) by either the ILECs or the cable companies are fixed and sunk, the cost to provision VoIP would be incremental and the services offerings would be of similar quality.²³¹

395 The Bureau then states that two competing networks may provide sufficient competition where:

i. [There are] At least two unaffiliated networks, where the entrant is capable of providing access at a lower cost to each potential location where access to the PSTN is demanded. The area for which forbearance applies will be the geographic region for which the two networks have relatively similar geographic footprints, either actual or potential locations that can be added to a network at relatively low cost.

ii. From the perspective of consumers, both networks have similar features available and equivalent quality of access.

iii. Industry characteristics are such that the coordinated exercise of market power is not likely to be a significant issue.²³²

396 This test is similar to, though not exactly the same as, the Bureau's test for lack of dominance that was adopted by the government in the Local Forbearance Variation Order. The requirement in the above test that the entrant be able to provide access at a lower cost seems to have been replaced with a requirement that the entrant have similar or lower costs than the incumbent's variable costs in the Bureau's local forbearance submission.

397 If Canada's relatively low levels of wireless penetration reflect lack of competition in the wireless market, as the TRP Report suggests, then it should be noted that Canada has the highest broadband penetration of all of the G7 countries (Canada, United States, United Kingdom, France, Italy, Germany and Japan), at 51.4%.²³³ Canada's broadband penetration is higher by a wide margin: the other G7 countries range from 30.9% (Italy) to 43.9% (Japan). The percentage of broadband connections provided by DSL was the second lowest in Canada, at 46.1% (the US was lower at 42.2%, but its rate of broadband penetration is much lower than Canada's, at 38.0%). The CRTC attributes these differences to the choice of facilities-based providers in Canada²³⁴ As I understand it, residential broadband internet is largely a duopoly between cablecos and ILECs, with some resellers. These numbers suggest that a cableco/telco duopoly can result in strong facilities-based competition.

398 In any event, although Canada's wireless penetration rate continues to be among the lowest of OECD member countries, average revenue per minute is also among the lowest. The CRTC attributes this to the high-volume minute plans that Canadian carriers have adopted from US carriers. The CRTC notes that this is an indication of affordability of wireless services in Canada, and is indicative of the level of competition in Canada.²³⁵

VI What facilities are essential?

A. Characteristics of the telecommunications industry in 2007

399 It is common ground that the telecommunications industry is in a state of rapid transition driven mainly by technological advances.

400 As the TRP Report notes, perhaps the most important change currently underway is the shift to internet protocol ("IP") to carry voice, video and data. The TRP Report points out that this development permits a separation of network infrastructure from the applications and content layers. Consumers will be able to mix and match services from different providers.²³⁶ Thus, for instance, the emergence of access-independent VoIP has permitted consumers to buy telephone service (an application) from a variety of providers other than the provider of their broadband internet connection.

401 The TRP Report also describes how the shift to IP leads to a consolidation of industries into a new, converged telecommunications marketplace. The Panel predicts a four-way convergence of traditional telecom, entertainment and content, consumer electronics, and computing and software. That is, each of these industries participates in this new telecommunications marketplace. The Panel describes how today consumers download music onto mobile devices using a number of different service providers.²³⁷

402 While to some extent these developments are in their infancy, they are gathering steam.²³⁸ The CRTC reports that:

Non-legacy services such as Ethernet, Virtual Private Network (VPN) IP-VPN, Internet and wireless service grew at an annual rate of approximately 15% since 2002 whereas legacy service revenues declined by 5%39 annually to the point where, as displayed in Figure 3.4.1, by year end 2006 over half of the industry's revenues were generated from the newer, non-legacy services.²³⁹

403 The result will be a major shift in the structure of the telecommunications industry in Canada. According the TRP Report, a structure is emerging of a telecommunications market that has many different types of participants: infrastructure providers (such as cable, ILECs, wireless broadband), applications providers (such as TV and telephony providers²⁴⁰), and content providers (such as music, TV content, and the like).

404 Dale Hatfield (Rogers' expert) echoes the themes in the TRP Report. He says that all traditional networks, including wireline telephony, cable television, and wireless telephony, are moving in the same direction, which can be summed up in five points:

First, and perhaps most fundamentally, all three types of platforms are evolving from the analog format to end-to-end digital transmission and switching/routing. Indeed, this conversion has largely been completed with the exception of certain "last mile" facilities. Second, all three types of providers are, for the most part, seeking to extend broadband digital facilities deeper into their network and, ultimately, directly to customer locations. Third, all three types of providers are evolving toward networks that employ packet switching and statistical multiplexing with increased "intelligence" at edge of the network rather than the interior. Fourth, in combination, the previous three developments lead to networks that are capable of supporting all applications – voice, data, image, video and rich multimedia combinations thereof. Fifth, all three types of providers are using the Transmission Control Protocol/Internet Protocol ("TCP/IP") family or suite of standards and protocols as the means of logically organizing their respective platforms and as a way of routing the packets of information – voice, data, image and video as the case may be – between different platforms and over diverse types of transmission networks (e.g., cable modems or Digital Subscriber Lines – "DSL") and transmission media (e.g., fiber optic cable, coaxial cable, or twisted-pair copper cable). [Emphasis in original.]²⁴¹

405 Mr. Hatfield describes how the TCP/IP protocol suite is made up of layers:²⁴²

Application Layer
Transport Layer
IP Layer
Data Link Layer	Network Inter- FACE lAYER
Physical Layer
Support Structures

Figure 1 – Representation of the Internet Protocol Suite (TCP/IP)

406 This is essentially the same as the four-layer model of network architecture described by the TRP Report. The layers identified by the Panel are:

a) Content layer: this is the media – music, tv, data, voice, etc.

b) Applications layer: this is what the user interacts with to get the content.

c) Signal or transport layer: "To make the physical medium carry traffic, a variety of signal transmission/modulation schemes are used. This constitutes the signal or transport layer of the network."

d) Physical layer: this is the medium that caries the signal (wire, fibre, wireless).²⁴³

407 Mr. Hatfield states that there are greater barriers, both economic and operational to duplication of the lower layers in the protocol stack. Economies of scale are greater at the bottom than at the top. Thus, he observes, it is not only uneconomic, but socially undesirable, to duplicate telephone poles. But entry is much easier at the application layer and above.²⁴⁴

408 There seems to be widespread agreement that the "stepping stone" approach, whereby competitors are given wholesale access in hopes that they will develop their own facilities, has not led either to widespread development of facilities or to competitive downstream residential markets.

409 The TRP Report states bluntly:

There is no evidence in Canada that the CRTC's "stepping-stone" strategy has provided an effective transition to greater reliance by entrants on their own facilities. There is, on the other hand, reason to believe these policies have distorted the behaviour and incentives of new entrants in Canadian telecommunications markets.²⁴⁵

410 Rather, as a number of the parties have pointed out in their evidence, it is the entry of cablecos into local competition through VoIP products that has spurred competition. The data in the July 2007 CRTC Telecommunications Monitoring Report show how quick and dramatic this entry has been.

411 The following table updates the table at ¶33 of the Bureau's Evidence. The Bureau points out that cablecos entered the local residential market in 2005 and doubled the market share attributable to CLECs in that year. Cablecos have nearly re-doubled the CLEC market share, from 7.47% to 14.09%, in 2006.

412 The number of lines provided by competitors using owned-lines more than doubled in just one year, from 0.8 million lines in 2005 to 1.7 million lines in 2006.²⁴⁶ Now, competitors use their own facilities for 71% of the residential lines they provide, and for 41% of the business lines they provide. As the Bureau notes, as recently as 2004, competitors used unbundled local loops for 80% of their residential lines.²⁴⁷

413 The trends identified by the TRP Report and Mr. Hatfield suggest that in the not too distant future, households will connect to the communications network through broadband internet, whether hosted on a coaxial cable, telephone wire, fibre, or perhaps wireless. Applications such as telephony will be delivered over that connection. In that case, what's happening in the market for broadband internet may, in the long run, be more important than what's happening in the market for local residential lines. The CRTC has previously held that the market for residential internet services is competitive.

414 In 2006, cablecos supplied 54.2% of residential broadband internet connections, down slightly from 58.3% in 2002. Incumbent telcos accounted for 41.5%, up slightly from 39.7% in 2002. Other TSPs accounted for only 4.4%, up from 2% in 2002.²⁴⁸ While market shares are not everything, these numbers suggest that neither incumbent telcos nor cablecos have market power. They also suggest that as telephony transitions to IP, telcos may not, in fact, have the upper hand. Indeed, the greater bandwidth offered by coaxial cable may give cablecos an important competitive advantage over ILECs.

415 As discussed above, both the TRP Report and Mr. Hatfield's report discuss how communications networks are converging on the IP standard and, at the same time, products are disaggregating. While we are not there, yet, the way that households connect to communications networks will be through IP over a broadband connection. Telephony will become an application that is delivered over a broadband connection.

416 This analysis has profound implications for how we define product and geographic markets in communications networks.

417 First, local telephone service will become a different product. Currently, if you buy local telephone service from an ILEC, you buy a bundle of a connection to the telephone network and local telephone service itself. The ILEC will also sell you broadband internet as an adjunct to if you want it. The advent of VoIP has already disaggregated the network connection product from the local telephone service product for many customers. Now you can buy broadband internet from a cableco, and you can choose to bundle VoIP with it, or you can buy VoIP from an access-independent provider like Primus or Vonage.

418 The same process is occurring with television, albeit much more slowly. Still, it is becoming more common to download or stream television programs and movies over the internet.

419 Of course, this process of disaggregation is far from complete. The 85% of Canadian households that purchase local telephone service from an ILEC still do not differentiate between network connection and local telephone service. As well, cablecos still sell broadband internet as an adjunct to their core business, cable television. However, if the TRP Report is correct, this will change. Network connection through broadband internet will become the basic product through which households connect to communications products (applications) and receive content.

420 Second, applications products like "local" telephone service and television may cut loose from their physical and thus geographic moorings. This is already happening. Primus and Vonage offer access-independent telephone service. You can even choose a different area code from the one you live in. As well, it is possible (though I have not tried it) to replace local telephone service with Skype, a VoIP service offered by a Finnish company.²⁴⁹ It would a logical development for both telcos and cablecos to begin to offer VoIP service on an access-independent basis as well as to their broadband customers.

421 I want to stress that this future is not here yet. However, realignment in the definition of product and geographic markets will follow from the changes predicted by the TRP Report.

422 Is wireless in the same product market as landline telephony? When the Bureau examined this question at the time of Rogers' acquisition of Fido, it concluded that wireless and landline telephony are two different markets.

423 A number of the ILECs point to evidence that increasing numbers of Canadians only have a mobile phone to argue that they are now in the same product market. The Companies point out that wireless substitution is around 5% in Canada, and as high as 10% in some urban areas.²⁵⁰

424 The Tribunal's approach to product market definition in Canada Pipe may be instructive here. Canada Pipe makes and sells cast iron drain, waste and vent ("DWV") pipes and relating hardware. The evidence showed that plastic DWV pipes were increasingly being substituted for cast iron, and was in fact a better product for many applications. However, for certain applications, cast iron had to be used. The fact that "cast iron still pays a distinct role" and was treated as a separate product by distributors, contractors, and even Canada Pipe, led the Tribunal to define the market narrowly as cast iron DWV products.²⁵¹

425 Clearly an increasing number of people rely on wireless only. It may be, however, that some, or perhaps many, of those people needed or wanted wireless for the mobility it offers, despite its higher cost, and then found that it did not make sense to pay for a landline as well.

426 However, the evidence suggests that for many, if not most, wireless and landline telephony remain different products. First, most wireless customers also have a landline, given the relatively low percentage of people who rely on wireless only.

427 Second, Canada's relatively low wireless penetration rate suggests many landline consumers want the characteristics associated with a landline phone, or are not prepared to pay more to get the additional features that wireless telephony offers.

428 Third, while landline and wireless telephony both allow one to talk over the telephone and access the internet, they have somewhat different characteristics. Landline telephones connect to a fixed place (although access-independent VoIP is changing this). It is possible to put several telephones on the same line without incurring extra charges. Local calls are free and long distance is cheap (assuming one subscribes to a long distance plan). Dial up internet calls are essentially free as well, although broadband requires a subscription.

429 By contrast, wireless telephones connect to a person; they are highly personal devices intended to be carried around. They offer a more features than landlines (SMS, email, and music downloads, for instance), but you can't plug extra phones onto the same line. Local calls are metered (although many plans include unlimited evenings and weekends) and long distance is much more expensive. Data usage is metered and expensive.

430 Another consideration is that wireless telephony is in a state of rapid transition. Rogers has already debuted a 3.5 G product called Rogers Vision in the Golden Horseshoe. Vision is based on the HSDPA standard and offers enhanced services like video calling, and much faster data transmission, for example.

431 While a more detailed analysis is called for, I doubt that the wireless and wireline telephony are in the same product market.

432 I would add that internet access over cellphones is almost certainly not in the same product market as broadband internet over cable or telephone lines. It is much more expensive and slower (although, according to the ads, EVDO can achieve true broadband speeds, and HSDPA has the potential to offer speeds of around 3-7 mbps).²⁵²

B. Ex post or ex ante regulation?

433 Presently, wholesale access is mandated ex ante.

434 The Companies have proposed moving to an ex post, complaints-based system. That is, the CRTC would not mandate access to any facilities ex ante (except interconnection and 911 services, which the Companies treat as an instance of social and technical regulation). The Companies rely on Dr. Taylor's evidence to the effect that the economic costs and distortions associated with ex ante regulation are higher than those associated with ex post regulation.

435 However, there are also costs associated with ex post regulation. Ex post regulation could lead to hundreds, perhaps thousands of individual proceedings, if ILECs refuse to grant access to competitors to facilities the competitors think they need. The costs of such a multiplicity of proceedings would likely exceed the cost of a proceeding to determine the essentiality of a category of facilities.

436 Another difficulty associated with ex post regulation is the time it takes to resolve disputes could delay entry or expansion by competitors.

437 New Zealand's experience is instructive here. New Zealand privatized its monopoly ILEC, Telecom Corporation of New Zealand, in 1990 and opened the door to competition. It did not mandate access, relying instead on oversight by the competition authority and the courts. This resulted in litigation, notably between Clear Communications Ltd. and Telecom that went all the way up to the Privy Council. Clear sued Telecom after a failure of negotiations that led to a refusal in October 1991 by Telecom to supply wholesale facilities to enable Clear to supply telephone service to, ironically, the courts. The case was finally resolved by a decision of the Privy Council in 1994.²⁵³ Concerns that recourse to the courts and the competition authority was too slow were among the reasons cited for moving to a mandatory wholesale access régime in 2001.²⁵⁴

438 In its Supplementary Evidence, the Bureau discusses five regulatory models: (i) full competitive analysis ex ante and (ii) ex post; (iii) proxy analysis ex ante; (iv) hybrid of ex ante and ex post approaches; and (v) framework for negotiation.²⁵⁵

439 The Bureau suggests that

The best outcome for this proceeding would be the development of accurate proxy rules that identify when certain facilities (e.g., access differentiated by capacity, and transport) are likely to be essential. These proxy rules would base the determination of whether a facility is essential ex ante on easily observable and verifiable structural characteristics.²⁵⁶

440 I favour a hybrid of ex ante and ex post approaches. An ex ante approach is appropriate where it can be said with confidence that the criteria for mandating access are met, where there is broad agreement among all industry participants that access should be mandated, or where access is to be mandated for social or technical reasons.

441 Otherwise, access should not be mandated ex ante. The CRTC should, however, establish a rapid process for resolving disputes over access, and make interim access orders where appropriate. (It must be kept in mind that interim access orders will affect the parties' incentives, however. Where there is no interim order, the ILEC has an incentive to delay. Where there is one, the CLEC has an incentive to delay.)

442 The development of accurate proxy criteria would certainly be helpful. It may speed up dispute resolution. Perhaps more importantly, it would enhance predictability, which will promote settlement of disputes without recourse to dispute resolution.

C. Interconnection

443 All parties appear to agree that access should be mandated for facilities necessary for interconnection of networks, as well as emergency services (911). All carriers, not just CLECs, need interconnection, although the incentives are not equal. There is some disagreement over exactly which interconnection services are essential and which are not.

444 Although the point was not specifically addressed, my impression from the economic evidence in this proceeding in relation to the incentives and disincentives created by mandatory access is that regulatory overreach in respect of interconnection is less problematic than it is with wholesale access and perhaps also support structures.

445 Given that interconnection is of vital importance to all carriers, and very much in the public interest, I would err on the side of mandating more, rather than fewer, interconnection services. If a mandating access to a particular interconnection service promotes the public policy of ensuring a high level of interconnection, and does not reduce incentives for the development of facilities, then interconnection should probably be mandated.

D. Wholesale access

446 In my view, the essential facilities test will not be met for most, if not all, wholesale access facilities used for residential telephone service in urban areas. About 89% of Canadian households have access to a competing facilities-based provider, their local cableco. Like telcos, cablecos have nearly ubiquitous networks and a base of customers to market their VoIP product to. It is noteworthy that the presence of facilities-based competitors in cities, including wireless, is what led the CRTC to forbear from local rate regulation in most Canadian cities in July 2007.

447 In rural areas that lack cable coverage, wholesale access likely will remain essential. Some rural areas do have terrestrial wireless broadband services. If those services are able to support VoIP, are priced competitively, and do not depend on the ILECs facilities, then they may qualify as a facilities-based service provider.

448 Two-way satellite broadband internet over Telesat's Ka band is now available throughout Canada. However, this service only increases availability of broadband internet in Canada by one percentage point because of capacity limitations.²⁵⁷ According to the Xplornet website, this service is compatible with VoIP. I do not know what the quality of VoIP over this service is.²⁵⁸ Pricing of Xplornet's service is much more expensive than landline based broadband.²⁵⁹ A customer would not switch from an ILEC-provided local telephone service to satellite broadband plus VoIP in response to a SSNIP in the price for local telephone service, or, for that matter, the price for landline based broadband. Xplornet's service only makes sense for people who cannot get DSL or cable internet. For those people, adding VoIP may make sense.

449 The case is not so clear in business markets. There are sharp disagreements between the parties and the evidence conflicts. A few facts stand out.

(i) Enterprise customers

450 As Dr. Taylor points out, there are two different kinds of business customers: small and enterprise. Small business customers likely have similar characteristics to residential customers. They may need a simple line or two plus an internet connection.

451 However, enterprise business customers have different needs and buy communications services differently. They may have many telephone lines that they want linked through a PBX or VoIP system onsite, or through a co-located or managed service. They may want toll free lines. They may want special features to enable a call centre, or automated attendants. They may have a lower tolerance service interruption than a residential customer. They may have multiple offices and want services that facilitate calling between offices. They may want one provider who can deliver services coast to coast. This last fact could justify defining a market around national accounts.²⁶⁰

452 Dr. Taylor points out that such customers may buy communications services through RFPs, receive customized attention, and enter into multi-year deals with negotiated discounts. This is consistent with press releases that I have seen that suggest that major telcos like Bell, Telus, and MTS Allstream are competing to offer highly managed business telephone and internet services.

453 I would note that enterprise businesses do not need to be that large. My law firm of 13 lawyers and about the same number of staff has communications requirements that tend to place it in the enterprise category. For example, about five years ago I conducted a process similar to an RFP and replaced our old PBX and multiple business lines with a 3Com NBX VoIP system that connects to line capable of handling over 20 calls at once. Telus is our local and long distance provider, but the line was (and still is, to my knowledge) leased from Bell. Telus' service involved an installation cost and a three-year contract, but was cheaper than the services it replaced.

454 The asymmetric nature of my firm's DSL internet access has led me to realize that business internet needs are different from residential. Relatively low upload speeds are not, I suspect, a problem for most residential customers. Slow upload speeds make accessing a network over VPN quite tedious. It is also difficult to transfer very large files, as we increasingly must do.

455 While it is dangerous to generalize from one example, I suspect that small enterprises that rely heavily on communications services (such as law firms, accounting firms, consulting firms, public relations firms, etc) and/or rely on VPN access would have needs not dissimilar to those I have described.

(ii) Extent of facilities-based competition

456 The CRTC Telecom Monitoring Report shows that in 2006, ILECs accounted for about 84% of local business lines, ILECs operating out of territory, about 10%, and CLECs, about 6%. Competitors' overall share has increased from about 11% in 2002 to 16% in 2006. Interestingly, competitors enjoy a slightly higher market share in business markets than in residential.

457 The prevalence of owned facilities is much lower in business markets than in residential: 41% vs 71%. Conversely, reliance on local loops leased from ILECs is much higher in business markets than in residential: 37% vs 19%. Unfortunately, the monitoring report does not break these figures out between small business and enterprise customers.

458 As the above figures suggest, the competitors in business markets are different. Cablecos are not as present in business markets as in residential markets, with the possible exception of EastLink and Vidéotron. Most of the competition seems to be coming from ILECs operating out of territory, and the evidence in this proceeding suggests that most of the owned facilities belong to ILECs operating out of territory.

459 I note that The Companies, in their supplementary evidence, cite numerous public statements from MTS Allstream and Rogers in their annual reports and press releases to the effect that business markets are competitive and that they are winning business customers. This research is impressive, but evidence that MTS Allstream and Rogers are winning business customers is not particularly helpful without knowing whether they are winning them with their own, or leased facilities.

460 For instance, Rogers provides business wireline telephony in 72 ILEC exchange areas, including most major Canadian cities. However, all of its business wireline telephony is provided through wholesale access (leased loops or PRI). Rogers does not serve these customers using its cable network.²⁶¹ As a result of its acquisition of assets of GT Group Telecom, Rogers now has fibre into approximately 1,000 business locations in eastern Canada.²⁶²

461 I have not seen any evidence concerning the extent to which MTS Allstream or Telus provide services through their own facilities.

462 Distributel provides local telephony to business through its access-independent VoIP product and resold business lines. It does not provide any business lines through its own facilities.²⁶³

463 Hydro utilities have also entered the business market, typically providing high speed fibre connections, Ethernet, private line, and similar services. Examples include Toronto Hydro Telecom's fibre-based internet and VoIP services in Toronto; Blink Communications (Oakville Hydro) in Oakville;²⁶⁴ Telecom Ottawa and Trytel (Hydro Ottawa), which offers these services in Ottawa and Cornwall.

464 I am aware of two other internet service providers in Toronto that appear, from their websites to be facilities-based, or at least partly so: Cogent²⁶⁵ and Beanfield.²⁶⁶

465 As well, anyone paying attention to manhole covers while walking through Toronto's financial district would notice that Telus, Allstream, Rogers, and Beanfield all have facilities underground, in addition to Bell and Toronto Hydro. There are still Group Telecom manhole covers, but I do not know who owns that fibre now.

(iii) Conclusion on business markets

466 Business markets continue to exhibit reliance on leased lines and other wholesale facilities. Strictly speaking, this is not probative; such reliance may be a symptom of a system that incents reliance on wholesale facilities over development of facilities. The question is whether competitors can duplicate facilities economically.

467 There are indications that facilities-based competition is developing even for enterprise customers. However, there is insufficient information about the extent of facilities-based competition to draw firm conclusions. As well, different geographic markets may be in different stages of development.

468 While the parties have filed detailed cost information in confidence, I lack the expertise to estimate whether facilities can be built economically.

469 Consequently, I agree with the conclusion the Bureau came to in its Supplementary Evidence:

53. As noted in the Bureau's Evidence, the record of competition shows that while entry by the cable companies is likely sufficient to control the market power of ILECs with respect to residential customers, the current competitive situation in markets for business services appears very different.

[...]

56. More generally, the Bureau cannot, at this stage in the proceeding and with the information available to it, apply its definition to many specific wholesale facilities to determine, ex ante, whether access to them should be mandated. Accordingly, the Bureau can only recommend an ex post approach for determining whether to mandate access to those facilities. Under such an approach, which is described above, specific facilities would be deemed essential only where potential entrants could demonstrate upon request that the Bureau's test is met.

470 The ex post approach recommended by the Bureau is consistent with the presumption against regulation that the CRTC has been directed to apply. CLECs should bear the burden of showing mandatory access should be continued in particular markets. Evidence that there are no facilities-based competitors and that it is not economical to build facilities will be required. There will need to be a sufficiently long transition period to allow wholesale access dependent CLECs either to build facilities or apply to the CRTC for continued access (or both!).

(iv) Do ILECs have incentives to offer wholesale access?

471 The Companies argue that even without mandatory wholesale access, facilities owners have incentives to provide wholesale services to competitors. The Companies point to services they voluntarily offer on a wholesale basis, and the fact that Bell West uses such non-mandated wholesale services. The Companies also rely on the fact that no ILEC in Canada is able to offer services in all geographic markets.²⁶⁷

472 On the other hand, it appears from the Local Competition decision that wholesale access was not negotiated between ILECs and would-be entrants following the Review of Regulatory Framework decision; it had to be mandated.

473 I would have liked to have seen more economic evidence on this point. It seems intuitive to me that ILECs with large territories that include major cities like Telus and Bell would have an incentive to grant each other wholesale access. Each has important markets to bring to the table. It is less apparent that Bell or Telus would have incentives to grant access to SaskTel or MTS Allstream.

474 One could argue that if Telus and Bell have an incentive to grant each other wholesale access, that may be enough, because having granted each other wholesale access, refusing it to others may count as a discriminatory refusal to deal and be more susceptible to being characterized as an anti-competitive act under the Competition Act's abuse provisions.

475 It also seems intuitive to me that where facilities-based competition exists, all facilities owners may have incentives to grant wholesale access to non-facilities based competitors. A facilities owner might grant such access if it calculated that although the competitor would take market share from all participants in the downstream market, the new wholesale revenue would exceed the revenue lost downstream. I lack the expertise to predict whether this would be the case, however.

476 There is evidence from the wireless telephony market that such incentives may exist, however. Several resale wireless providers currently operate in Canada. Also, other firms in other industries routinely engage in behaviour analogous to wholesale access. For instance, car makers that have excess capacity may contract to build cars for other car makers that do not.²⁶⁸

E. Support structures

477 There is broad agreement that access to support structures should continue to be regulated, because of the public interest in avoiding duplication of telephone poles and trenches.

478 A major concern in this proceeding is incenting the building of new facilities, or at least not disincenting the building of new facilities. It is widely (though not universally) accepted that an overly generous mandated wholesale access regime will incent reliance on wholesale access and disincent building new facilities.

479 As well, some CLECs have emphasized the barriers to building new facilities that relate to access to support structures. While generous mandated access to support structures may disincent duplication of support structures (which may be social desirable in any event), it will remove barriers to building new facilities and perhaps reduce the cost of building new facilities.

480 I would also note that the Telecommunications Act grants powers to the CRTC and privileges to carriers to assist in construction of facilities. Put bluntly, the CRTC can use federal paramountcy to overrule provincial and municipal authorities. To the extent that municipalities discriminate in favour of incumbents or make life difficult for CLECs, as has been alleged in this proceeding, the CRTC should not hesitate to use its powers.

F. Next generation facilities

481 The CRTC should not attempt to speculate about whether next generation facilities will be essential or not. The Policy Direction applies a presumption in favour of market forces and against regulation.

482 I also agree with the Bureau that it is highly unlikely that new facilities would be essential. As the Bureau points out, if the new facility competes with existing products, then by definition it is not essential.²⁶⁹ Even if it does not, the fact that the owner of the facility built it strongly suggests that the facility can be duplicated by rivals. Even if the facility cannot be duplicated, forcing the firm that developed it could seriously undermine the incentives to innovate.²⁷⁰

VII How should essential facilities be priced?

A. Regulate or negotiate?

483 I prefer a negotiate-first approach because it results in something closer to market based pricing than regulated pricing.

484 That being said, this approach presents dangers. Facilities owners may have an incentive to delay by making impossibly high demands.

485 These dangers can largely be mitigated through strict timelines and a speedy arbitration system. As the Companies point out, final offer arbitration has worked in other industries, including pricing of access transport facilities, which may be in some ways analogous to communications facilities.

B. Pricing principles

486 My review of the evidence, particular the expert evidence, has convinced me that, while there may be approaches to pricing of essential facilities that are better in theory than cost plus, they are too difficult to implement in practice.

487 The question, however, is, what costs, and what "plus". I do not feel qualified to address this in any detail as I am not an economist. Whatever price is chosen, however, it should not result in a squeeze. That is, while I do not support retail minus pricing, the wholesale should be less than the retail price for services that use the wholesale service.

VIII When should future reviews be conducted?

488 In theory, future reviews should not be necessary; as circumstances change, parties should bring the appropriate applications.

489 Thus, where the CRTC adopts an ex post régime, a party may apply for mandatory access to a particular facility in a particular place. If the CRTC grants access, the owner of the facility should be entitled to apply to end mandatory access if the competitive landscape changes such that the facility is no longer essential.

490 The same applies to facilities where access is mandated ex ante: parties can apply to end mandatory access to a particular facility in a particular place.

491 As the history after the Local Competition decision shows, however, there is a temptation to grant access to facilities in order to help competitors. As a result, periodic reviews of the mandatory access régime should be conducted. I would recommend a five year interval.

IX What regulatory régime should apply to non-essential services?

492 Mandatory access to non-essential services should be phased out. The transition period should be long enough to allow competitors that rely on access to build facilities or exit in an orderly fashion.

ALL OF WHICH IS RESPECTFULLY SUBMITTED

September 6, 2007

________________________________
Michael Osborne

Notes:

¹ See www.thelitigator.ca.

² See Competition Bureau press release dated December 21, 1999, Competition Bureau Announces It Will Not Oppose Acquisition of Canadian Airlines, http://www.competitionbureau.gc.ca/internet/index.cfm?itemID=619&lg=e

³ See s. 2(1):

⁴ Telecommunications Act 1993, S.C. 1993, c. 38, s. 7.

⁵ SOR/2006-355

⁶ Interpretation Act, R.S.C. 1985, c. I-21, s. 11.

⁷ Policy Direction, s. 1(c)(ii)

⁸ SOR/2007-71

⁹ Telecom Decision CRTC 2006-15, Forbearance from the regulation of retail local exchange services, ¶242.

¹⁰ Local Forbearance Variation Order, s. 2

¹¹ Telecom Decision CRTC 97-8, Local Competition, ¶73-74:

¹² Evidence of the Commissioner of Competition, Telecom Public Notice CRTC 2005-2.

¹³ See p. 17.

¹⁴ See Canada (Director of Investigation and Research) v. Tele-Direct (Publications) Inc., [1997] C.C.T.D. No. 8 at ¶539: "section 79 is not intended to condemn a firm merely for having market power. Instead, it is directed at ensuring that dominant firms compete with other firms on merit and not through abusing their market power".

¹⁵ Canada (Commissioner of Competition) v. Canada Pipe Company Ltd., 2006 FCA 233 ¶25-26 ("Canada Pipe Appeal")

¹⁶ Canada (Director of Investigation and Research) v. NutraSweet Co. (1990) 32 C.P.R. (3d) 1 at 28-29; [1990] C.C.T.D. No. 17 ("NutraSweet")

¹⁷ NutraSweet, at 31-33; Canada (Director of Investigation and Research) v. Laidlaw Waste Systems Ltd. (1991) 40 C.P.R. (3d) 289, [1992] C.C.T.D. No. 1 (Comp. Trib.) ("Laidlaw"); Canada (Director of Investigation and Research) v. The D&B Companies of Canada Ltd. [1995] C.C.T.D. No. 20 (Comp. Trib.) ("Nielsen"). In Nielsen, McKeown J. wrestled with the problem that s. 79(1)(a) uses the phrase "class or species of business", while s. 79(1)(c) uses the term "market". The use of different terms normally gives rise to an interpretive presumption of different meaning, but the Tribunal in NutraSweet had pointed out that "the logic of the section is better if the product market is precisely defined in connection with question of control rather than being partially dealt with under para 79(1)(a) and then revisited in para 79(1)(c)". McKeown J. concluded the phrase "class or species of business" means "product market". Parliament did not use "market" because "market" has a geographic dimension, and Parliament used "throughout Canada or any part thereof" to express the geographic dimension of the market, and to limit it to Canada.

¹⁸ Laidlaw, 324; Canada (Commissioner of Competition) v. Superior Propane Inc. [2000] C.C.T.D. No. 15 (Comp. Trib.) ("Propane 1") at ¶48.

¹⁹ Laidlaw, 316.

²⁰ Propane 1, ¶47; Tele-Direct, ¶225.

²¹ Canada (Director of Investigation and Research) v. Southam Inc. (1992), 43 C.P.R. (3d) 161 (Comp. Trib.) ( "Southam"); Tele-Direct, ¶225.

²² Tele-Direct, ¶76; Nielsen, 241.

²³ Economists refer to this as "elasticity of demand". The "own-price elasticity of demand" looks at what happens to the demand for the product under consideration in response to price increases. If the demand remains the same, that is, is "inelastic", the product likely constitutes a product market. The "cross-price elasticity of demand" compares the relative demand of products A and B in response to changes in price. If consumption of A rises when the price of B rises, and the consumption of B rises when the price of A rises – that is, consumers switch from A to B and vice versa – then A and B are likely in the same market. Own-price elasticity of demand is considered to be a more reliable indicator of market delineation Propane 1, ¶59-63, Commissioner of Competition v. Canada Pipe, 2005 Comp. Trib. 3, ¶70. See also Canada (Director of Investigation and Research) v. Southam Inc., [1997] S.C.R. 748 at ¶16. However, evidence of elasticity of demand must be approached with care in abuse of dominance cases, where the allegation is that the respondent controls a market. This is because a firm that has a monopoly can be expected to raise prices to the point at which demand becomes elastic, that is, to the point where consumers are willing to switch to products that are not close substitutes. See the discussion on the "cellophane fallacy", below.

²⁴ Nielsen, 241; Tele-Direct, ¶76; Canada (Director of Investigation and Research) v. Southam Inc. [1995] 3 F.C. 557 (C.A.).

²⁵ The Tribunal highlighted the lack evidence of switching in Tele-Direct, ¶170; and the low cross-elasticities between aspartame and its supposed competitors in Nutra-Sweet.

²⁶ Tele-Direct, ¶111.

²⁷ Enforcement Guidelines on the Abuse of Dominance Provisions, Competition Bureau, 2001, §3.2.1(a) ("Abuse Guidelines").

²⁸ Tele-Direct, ¶98. The Supreme Court similarly compared automobiles and tanks: Southam, SCC at ¶72.

²⁹ In Nielsen, the Tribunal found that scanner-based data was better than audit and warehouse shipment data. In Tele-Direct, an examination of the physical characteristics of Yellow Pages helped show that they served a completely different purpose from other media from an advertiser's perspective. In Nutra-Sweet, the technical characteristics of aspartame were quite different from those of other sweeteners and helped define the market.

³⁰ Tele-Direct, ¶126-172.

³¹ This factor has not played an important role in any of the contested abuse cases, but the cost of switching from propane to other fuels was an important factor in determining the relevant market in Propane 1.

³² In NutraSweet, the Tribunal found that the best way to determine whether other sweeteners constrained the price of aspartame was to compare the price of aspartame in markets where there was competition between different suppliers of aspartame with those where there was no such competition. NutraSweet, at 19. Similarly, Canada Pipe lowered its price in response to entry by other cast iron pipe suppliers, but not in response to plastic pipe suppliers, which suggested that iron pipe was a separate product market. Canada Pipe, at ¶97-101.

³³ Canada (Commissioner of Competition) v. Canada Pipe, 2005 Comp. Trib. 3, at ¶112.

³⁴ NutraSweet, at 19-20.

³⁵ Canada (Commissioner of Competition) v. Canadian Waste Services Holdings Inc., 2001 Comp. Trib. 3, ¶69 ("CWS").

³⁶ Laidlaw, at 320.

³⁷ CWS, at ¶90-91.

³⁸ United States v. E.I. duPont de Nemours & Co., 351 U.S. 377 (1956).

³⁹ CWS at 92; Abuse Guidelines, 3.2.1.

⁴⁰ Laidlaw, at 318-320; Nielsen, at 253 ; Tele-Direct, ¶130.

⁴¹ TAB, §2.2, p. 6

⁴² TAB, §2.4, p. 7

⁴³ Competition Bureau Technical Backgrounder, Acquisition of Microcell Telecommunications Inc. by Rogers Wireless Communications Inc., http://www.competitionbureau.gc.ca/internet/index.cfm?itemID=257&lg=e

⁴⁴ The choice of this phrase raises some interpretive difficulties. For instance, a firm that dominates the world market for a product but has a small market share in Canada would be beyond the reach of s. 79, even if it leveraged its dominance over the world market to increase its market power in Canada: Nielsen, at 230-231. It is also open to be argued that s. 79(1)(a) precludes the Tribunal from delineating a market larger than Canada for purposes of s. 79, even if the geographic market is in fact larger than Canada. This argument does not appear to have been made or accepted in past cases, however.

⁴⁵ Abuse Guidelines, 3.2.1(b).

⁴⁶ TAB, §2.6

⁴⁷ See for instance NutraSweet, at 28. Virtually every Tribunal case recites this definition.

⁴⁸ Tele-Direct, at ¶226.

⁴⁹ Tele-Direct, at ¶285-286. As well, in Tele-Direct, the Tribunal recognized the difficulties inherent in converting accounting profit to economic profit: ¶273.

⁵⁰ CWS, at ¶74-83.

⁵¹ NutraSweet, at 19.

⁵² Canada Pipe, at ¶136.

⁵³ Laidlaw, at 327-328.

⁵⁴ Tele-Direct, at 289-297.

⁵⁵ CWS, at ¶78.

⁵⁶ Tele-Direct, at ¶225.

⁵⁷ Tele-Direct, at ¶18, 231; Laidlaw, at 325; Nielsen, at 254. However, in Canada Pipe, the Tribunal went on to observe that this prima facie conclusion must be supported by findings on including barriers to entry: at ¶138.

⁵⁸ Laidlaw, at 317.

⁵⁹ Abuse Guidelines, 3.2.1(d). In the case of joint dominance, the Bureau's threshold is 60%.

⁶⁰ For example, in Laidlaw, the Tribunal discussed using: weight of refuse dumped, number of containers, and number of collection trucks. In the airline industry, one could look at available seat miles, revenue passenger miles, flights, segments, seats, etc.

⁶¹ This occurred in Laidlaw.

⁶² TAB, §3.5

⁶³ Tele-Direct, ¶232. See also Canada Pipe: "where barriers to entry are non-existent, even a very large market share will not support a finding of market power", at ¶138.

⁶⁴ Laidlaw, at 330-331.

⁶⁵ NutraSweet, at 51.

⁶⁶ Tele-Direct, ¶243, Canada Pipe, at ¶156.

⁶⁷ Canada Pipe, ¶141. There is a debate about sunk costs as a barrier to entry. In Southam, the Tribunal said that sunk costs alone are not enough; there must also be economies of scale. Director of Investigation and Research v. Southam Inc. (1992), 43 C.P.R. (3d) 161 at 281-82. In Tele-Direct, the Tribunal added that sunk costs in combination with a likely response by the incumbent created a significant barrier to entry. Since incumbents can usually be expected to respond, this amounts to a finding that sunk costs are a barrier to entry when the entrant faces a dominant incumbent. Tele-Direct, at ¶246-253.

⁶⁸ TAB, §1.4

⁶⁹ In Canada Pipe, the Tribunal referred to cost of entry as an entry barrier, apart from sunk costs. Canada Pipe, at ¶142-143. This runs contrary to the Tribunal's statement in Southam that sunk costs alone are not a barrier to entry.

⁷⁰ NutraSweet, at 27; Southam, at 281-282.

⁷¹ NutraSweet, at 27.

⁷² For instance, Tele-Direct had access to listings, an established reputation, and a valuable trademark, all of which raised barriers to entry. Tele-Direct, at ¶254-267. Canada Pipe offered a full product line and had a strong national presence. Canada Pipe, at ¶144.

⁷³ In Laidlaw and Nielsen, exclusive contracting practices locked up the market, raising barriers to entry. Laidlaw, at 331; Nielsen, at 254-255, 259-266, 277.

⁷⁴ Canada Pipe, at ¶160.

⁷⁵ Canada (Director of Investigation and Research) v. Hillsdown Holdings (Canada) Ltd. (1992) 41 C.P.R. (3d) 289 (Comp. Trib.).

⁷⁶ Laidlaw, at 327.

⁷⁷ NutraSweet, at 34.

⁷⁸ Canada Pipe Appeal, ¶78-79.

⁷⁹ Canada Pipe Appeal, ¶70-73.

⁸⁰ Nielsen, at 257; Laidlaw, at 342-343; Tele-Direct, at ¶540.

⁸¹ See for instance: Tele-Direct, at ¶602-605; NutraSweet, at 40; Nielsen, at 260, 264.

⁸² In Tele-Direct, subjective evidence of anti-competitive intention was not enough to persuade the Tribunal that "overwhelming intensity of competitive response" was an anti-competitive act.

⁸³ Canada Pipe Appeal, ¶90.

⁸⁴ Canada Pipe Appeal, ¶90.

⁸⁵ Canada Pipe Appeal, ¶91

⁸⁶ To be fair, however, it must be noted that several of the enumerated acts in s. 78(1) use expressions like "for the purpose of disciplining or eliminating a competitor".

⁸⁷ Laidlaw, Nielsen, NutraSweet all involved exclusive contracts.

⁸⁸ NutraSweet, at 42; Laidlaw, at 307.

⁸⁹ NutraSweet; Nielsen.

⁹⁰ Laidlaw, at 343-344, 309-314. Laidlaw's use of this tactic in enforcing its exclusive contracts attracted the Tribunal's outrage. The Tribunal approved of R.H. Bork's statement that "As a technique for predation, sham litigation is theoretically one of the most promising". R.H. Bork, The Antitrust Paradox (New York: Basic Books, Inc., 1978) at 347, as cited by Laidlaw at 344.

⁹¹ Laidlaw, at 331-339.

⁹² NutraSweet, at 43-44. Predatory pricing has been shown in one abuse case: Canada (Commissioner of Competition) v. Air Canada. Although this case was brought under airline-specific provisions, the Tribunal's approach to the cost standard may inform future cases alleging predatory pricing as an anti-competitive act.

⁹³ Tele-Direct's refusal to deal with consultants who competed with Tele-Direct's own distribution channel escaped a finding that it was anti-competitive only because the costs Tele-Direct would incur in dealing with them justified its refusal: Tele-Direct, ¶717-733.

⁹⁴ NutraSweet, at 45-46.

⁹⁵ Canada (Director of Investigation and Research) v. Bank of Montreal (1996), 68 C.P.R. (3d) 527 (Comp. Trib.) ("Interac"). Members of Interac had excluded many potential participants in the Interac network through its by-laws and fees. However, as these were consent proceedings, the Tribunal presumed that the anti-competitive acts identified in the application were anti-competitive; it did not do its own analysis. The decision is thus of less precedential value in identifying anti-competitive acts than are the decisions in contested cases.

⁹⁶ Tele-Direct used its control over telephone directory space to increase its market power over telephone directory advertising services through tied selling and discriminatory practices. However, Tele-Direct's practice of tied selling was dealt with under the tied selling provisions. Canada (Director of Investigation and Research) v. Xerox Canada Inc. (1990), 33 C.P.R. (3d) 83 (Comp. Trib.). Xerox refused to sell parts to "independent service organizations" to eliminate competition in the market for servicing photocopiers. This case was brought under the refusal to deal provisions, but it might have been possible to frame it as an abuse of dominance case, given Xerox's market power in high and medium volume segments of the photocopier market (at 96).

⁹⁷ See §4.2.

⁹⁸ See §4.2.1.

⁹⁹ In July 2007, the Commission fined the Spanish ILEC, Telefonica SA, €151 million for a margin squeeze between its wholesale prices charged to competitors, and retail prices charged to competitors. EU Press Release IP/07/1011, 4 July 2007, Commission fines Telefónica over €151 million for over five years of unfair prices in the Spanish broadband market. Interestingly, the Commission found that cable operators were not able to constrain Telefonica's market power due to their limited coverage, and their market share had been in decline since 2001. See Memo/07/274, Commission decision against Telefónica - frequently asked questions. In 2004, the Commission obtained commitments from Deutsche Telekom to terminate a margin squeeze in relation to local loops used for broadband internet access. EU Press Release IP/04/281, 1 Mar 2004, Competition probe leads to decrease in tariffs for broadband access via line sharing in Germany. As well, the EU has challenged roaming rates charged by T-Mobile and its parent Deutsche Telekom, Vodaphone, O2.

¹⁰⁰ G. Werden, "The Law and Economics of the Essential Facility Doctrine", 32 St. Louis U. L.J. 433 at 449-450.

¹⁰¹ Paragraph 78(1)(i) refers to "selling articles at a price lower than the cost of acquisition". As services are not "articles", predatory pricing of services would not fall under this provision. Nevertheless, predatory pricing of a service by a dominant firm would likely be found to be an anti-competitive act.

¹⁰² William Baumol is generally credited with the avoidable costs measure. W. Baumol, "Predation and the Logic of the Average Variable Cost Test", 39 J. Law & Econ. 49 (1996).

¹⁰³ France Télécom SA c. Commission des Communautés européennes, T-340/03, 30 Jan 2007.

¹⁰⁴ France Télécom, at ¶130.

¹⁰⁵ France Télécom, at ¶183.

¹⁰⁶ France Télécom, at ¶227.

¹⁰⁷ Tele-Direct, ¶610

¹⁰⁸ NutraSweet, at 35.

¹⁰⁹ NutraSweet, at 35.

¹¹⁰ Canada Pipe Appeal at ¶36-39. See also NutraSweet, at 47; Nielsen, at 266.

¹¹¹ Canada Pipe Appeal at ¶51-57.

¹¹² Abuse Guidelines, §5.3.2.

¹¹³ Canada Pipe, ¶268-269. The Tribunal's reasoning on this point is not fully developed, however. The Federal Court of Appeal did not expressly overturn the Tribunal's decision on this point, nor did it consider this aspect of the test.

¹¹⁴ Except in the case of dominant airlines: s. 79(3.1) - (3.3) allow the Tribunal to impose "administrative monetary penalties" ("AMP") of up to $15 million on a dominant airline that it finds has engaged in a practice of anti-competitive acts. There have been proposals to make AMPs generally available in abuse cases.

¹¹⁵ Laidlaw, at 351.

¹¹⁶ Southam, SCC at ¶83-89.

¹¹⁷ Tele-Direct, at ¶761-763.

¹¹⁸ Now known as the Commissioner of Competition.

¹¹⁹ Quoted in Tele-Direct, at ¶587.

¹²⁰ Tele-Direct, ¶588-589.

¹²¹ Tele-Direct, ¶724, 726.

¹²² Tele-Direct, ¶733.

¹²³ Tele-Direct, ¶753.

¹²⁴ Paragraph 78(1)(k) was not part of s. 78 at the time.

¹²⁵ Tele-Direct, ¶66.

¹²⁶ Ziff's Principles of Property Law, (4^th ed, Carswell, 2006) at 2.

¹²⁷ Ziff's, at 5.

¹²⁸ Regulations Respecting Anti-Competitive Acts of Persons Operating a Domestic Service, SOR/2000-324

¹²⁹ Airline Guidelines, ¶60.

¹³⁰ Airline Guidenlines, ¶61

¹³¹ Airline Guidelines, ¶62.

¹³² Airline Guidelines, ¶63.

¹³³ Airline Guidelines, ¶65.

¹³⁴ Airline Guidelines, ¶66-67. If the dominant firm were expected to build facilities to supply to its competitors, it should be entitled to recover the whole cost plus a profit margin from the competitor. If the competitor can afford to pay this, then the question arises: why could the competitor not build the facilities itself? The Airline Guidelines do not discuss this point, however.

¹³⁵ TAB, §4.2.2

¹³⁶ For example, a discriminatory refusal to deal, where the owner of the facility provides access to some, but not to rivals. According to Einer Elhauge, the refusal to deal cases in the US Supreme Court can be explained by this theory. See E. Elhauge, "Defining Better Monopolization Standards", 56 Standford LR 253 (2003). Termination of an existing supply relationship has also been suggested. The EU DG Competition discussion paper on the application of Article 82 of the Treaty to exclusionary abuses proposes a lower threshold for cases involving termination of supply than refusal to start supplying. In the latter case, the discussion paper requires that the input be "indispensable". Compare ¶218 with §9.2.2.

¹³⁷ TAB, §4.2.2

¹³⁸ This was essentially the approach taken by the Tribunal in Tele-Direct, ¶663.

¹³⁹ Abuse Guidelines, Appendix III.

¹⁴⁰ TAB, §4.2.2

¹⁴¹ TAB, §4.2.

¹⁴² Robert H. Bork, The Antitust Paradox: A Policy at War with Itself (New York: Basic Books, 1978) at 372-74; Werden, "The Law and Economics of the Essential Facility Doctrine", at 470, 472.

¹⁴³ Economic Advisory Group on Competition Policy, An economic approach to Article 82, p. 43.

¹⁴⁴ The EAGCP Report gives as example: "Refusal to deal increases the market power of a dominant firm only if it was unable to fully exploit its monopoly power over the bottleneck good beforehand. For example, this is the case if the dominant firm has a problem committing to charging all the downstream firms the monopoly price." (at p. 45).

¹⁴⁵ Tele-Direct, at ¶539.

¹⁴⁶ For a list of private applications, see the List of Cases Filed by Type on the Tribunal's website (www.ct-tc.gc.ca).

¹⁴⁷ Canada (Director of Investigation and Research) v. Chrysler Canada Ltd. (1989), 27 C.P.R. 1 (Comp. Trib.).

¹⁴⁸ Canada (Director of Investigation and Research) v. Xerox Canada Inc. (1990), 33 C.P.R. (3d) 83 (Comp. Trib.).

¹⁴⁹ B-Filer Inc. v. The Bank of Nova Scotia, 2006 Comp. Trib. 42.

¹⁵⁰ Canada (Director of Investigation and Research) v. Chrysler Canada Ltd. (1989), 27 C.P.R. (3d) 1 (Comp. Trib.)

¹⁵¹ B-Filer, ¶79

¹⁵² Chrysler, at 18.

¹⁵³ Chrysler, at 23; B-Filer, at ¶79.

¹⁵⁴ B-Filer.

¹⁵⁵ See for instance: Sears Canada Inc. v. Parfums Christian Dior Inc. and Parfums Givenchy Canada Ltd., 2007 Comp. Trib. 6 (a fraction of 1%); Sono Pro Inc. v. Sonotechnique P.J.L., 2007 Comp. Trib. 18 (10%); Broadview Pharmacy v. Pfizer Canada Inc., 2004 Comp. Trib. 23 (11%).

¹⁵⁶ Xerox, at 116.

¹⁵⁷ Xerox, at 116-117. Xerox argued that the use of the plural "suppliers" in s. 75(1)(b) meant that a single supplier market was not caught by the section. This was really a variation on its argument that the market was the final market for photocopiers. The Tribunal rejected this contention, holding that the plural "suppliers" includes the singular.

¹⁵⁸ Xerox, at 116.

¹⁵⁹ B-Filer, ¶147.

¹⁶⁰ B-Filer, ¶186-187.

¹⁶¹ B-Filer, ¶208.

¹⁶² B-Filer, ¶211.

¹⁶³ B-Filer, ¶213.

¹⁶⁴ Chrysler, at 28.

¹⁶⁵ Chrysler, at 24-27; B-Filer, at ¶231ff.

¹⁶⁶ Telecom Decision CRTC 94-19.

¹⁶⁷ Telecom Decision CRTC 94-19, §II.D

¹⁶⁸ Telecom Decision CRTC 97-8

¹⁶⁹ Telecom Decision CRTC 97-9, ¶149.

¹⁷⁰ Telecom Decision CRTC 98-22.

¹⁷¹ Telecom Decision CRTC 98-9, ¶19-21.

¹⁷² Telecom Order CRTC 2001-184, ¶28-29, 34-35.

¹⁷³ Telecom Decision CRTC 2002-34.

¹⁷⁴ Telecom Decision CRTC 2002-34, ¶167-169.

¹⁷⁵ Telecom Decision CRTC 2002-34, ¶44-46.

¹⁷⁶ Telecom Decision CRTC 2002-34-¶72-73.

¹⁷⁷ Telecom Decision CRTC 2002-34, ¶139, 145, 155-160.

¹⁷⁸ Independent Members of the Canadian Association of Internet Providers – Digital Subscriber Line Internet services by Bell Canada and Bell Nexxia, Telecom Decision CRTC 2002-37, ¶26-27.

¹⁷⁹ Telecom Decision CRTC 2005-6

¹⁸⁰ Telecom Decision CRTC 2005-6, ¶65.

¹⁸¹ Telecom Decision CRTC 2005-6, ¶240.

¹⁸² Telecom Order CRTC 2007-20

¹⁸³ Telecom Order CRTC 2007-20, ¶77-78.

¹⁸⁴ Telecom Order CRTC 2007-20, ¶82; Telecom Decision CRTC 2005-6, ¶208.

¹⁸⁵ Competition Bureau Supplementary Evidence, ¶17.

¹⁸⁶ G Robinson, The Role of The Essential Facilities Doctrine in Competition and Regulatory Policy, ("Robinson Report"). ¶20.

¹⁸⁷ Robinson Report, ¶19.

¹⁸⁸ Competition Bureau Supplementary Evidence, ¶36.

¹⁸⁹ Competition Bureau Supplementary Evidence, ¶39.

¹⁹⁰ Telus Evidence, ¶41; Weisman Report, ¶35.

¹⁹¹ Robinson Report, ¶15.

¹⁹² Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 US 398 (2004) at 410-411.

¹⁹³ Weisman Report, ¶42; Robinson Report, ¶16.

¹⁹⁴ My understanding is that the same holds true of section 2 of the Sherman Act in the US.

¹⁹⁵ Robinson Report, ¶18.

¹⁹⁶ Weisman Report, ¶44.

¹⁹⁷ Robinson Report, ¶20.

¹⁹⁸Robinson Report, ¶20.

¹⁹⁹ See Robinson Report, ¶29, where Professor Robinson quotes with approval the Bureau's statement in the TAB that non-duplicability means, "that such an entrant would not find it feasible to enter or compete effectively if it had to self-supply the facility."

²⁰⁰ Robinson Report, ¶21.

²⁰¹ Robinson Report, ¶23.

²⁰² Robnson Report, ¶28

²⁰³ 15 USC 1, §2

²⁰⁴ See for example Canada (Commissioner of Competition) v. Canadian Waste Services Holdings Inc., 2001 Comp. Trib. 3, where the Tribunal uses the phrase "substantial prevention of competition".

²⁰⁵ Rogers Evidence, ¶74.

²⁰⁶ For example, in its Merger Enforcement Guidelines, the Bureau outlines certain thresholds below which it will not challenge a merger. See ¶4.11-4.12.

²⁰⁷ Rogers' Evidence, ¶161.

²⁰⁸ Rogers' Evidence, ¶163.

²⁰⁹ Rogers' Evidence, ¶213-215.

²¹⁰ MTS Allstream Evidence, ¶121-122.

²¹¹ MTS Allstream Evidence, ¶127.

²¹² MTS Allstream Evidence, ¶129.

²¹³ MTS Allstream Evidence, Appendix A, ¶13. The next paragraph shows how quickly things have changed: in 2004, before the entry of cablecos, reliance on leased loops was at 80%. In short, reliance on leased loops by CLECs fell by 50% in only one year.

²¹⁴ MTS Allstream Evidence, ¶96-104.

²¹⁵ Policy Direction, s. 1(a)

²¹⁶ Primus Evidence, ¶142.

²¹⁷ Cogeco Evidence, ¶15.

²¹⁸ Cogeco Evidence, ¶21, 50.

²¹⁹ Cogeco Evidence, ¶59.

²²⁰ Cogeco Evidence, ¶61. However, in paragraph 60, Cogeco says that a facility is essential where "it is not practical or feasible [...] for a competitor to substitute or self-supply a network facility to enter or compete effectively in a given retail telecommunications market...". Note – "a competitor".

²²¹ Cogeco Responses to CRTC Interrogatories, No. 105.

²²² Cogeco Evidence, ¶68. I think there are some words missing in the second criterion. I doubt Cogeco intended to say "preserve or enhance expansion". I have inserted in square brackets the words that I think express what Cogeco is trying to say.

²²³ Shaw Evidence, ¶21-22.

²²⁴ Shaw Evidence, ¶23.

²²⁵ Shaw Response to CRTC Interrogatories, p. 2.

²²⁶ Shaw Response to CRTC Interrogatories, p. 2.

²²⁷ TAB, §4.2.2, footnote 53.

²²⁸ TAB, §3.5, see in particular footnote 42.

²²⁹ Primus Evidence, ¶123ff.

²³⁰ TRP Report, p. 1-21.

²³¹ Competition Bureau, Comments of the Commissioner of Competition – Telecommunications Policy Review, August 15, 2005, ¶24-25.

²³² Competition Bureau, Comments of the Commissioner of Competition – Telecommunications Policy Review, August 15, 2005, ¶29.

²³³ CRTC, CRTC Telecommunications Monitoring Report, July 2007, §3.3, table 3.3.1. 93% of Canadian households have access to broadband internet, and 65% of these subscribe to it (p. 77).

²³⁴ CRTC, CRTC Telecommunications Monitoring Report, July 2007, §3.3, table 3.3.1.

²³⁵ CRTC, CRTC Telecommunications Monitoring Report, July 2007, p. 94.

²³⁶ Telecommunications Policy Review Panel, Final Report, March 2006, available at www.telecomreview.ca, at p. 1-24 - 29.

²³⁷ Telecommunications Policy Review Panel, Final Report, March 2006, available at www.telecomreview.ca, at p. 1-30 – 32.

²³⁸ For instance, on September 22, 2006, Shaw Communications reported that it is signing up a new Shaw Digital Phone customer every 96 seconds.

²³⁹ CRTC, CRTC Telecommunications Monitoring Report, July 2007.

²⁴⁰ Applications providers would include cable, satellite and IP TV; ILECs, CLECs, access-independent applications providers such as Vonage, Skype, and Primus, and wireless resellers such as Virgin Mobile.

²⁴¹ Hatfield Report, ¶3.

²⁴² Hatfield Report, ¶5.

²⁴³ TRP Report, p. 1-28. I have inverted the order in which these layers are presented in the report so that they correspond to Mr. Hatfield's presentation.

²⁴⁴ Hatfield Report, ¶20-21.

²⁴⁵ TRP Report, p. 3-35.

²⁴⁶ CRTC Telecommunications Monitoring Report, July 2007, p. 45.

²⁴⁷ Bureau Evidence, ¶35.

²⁴⁸ CRTC Telecommunications Monitoring Report, July 2007, p. 70.

²⁴⁹ Skype does not offer 911 service, however, and it is complicated to figure out exactly what you need to buy to do this.

²⁵⁰ The Companies Evidence, ¶23.

²⁵¹ Canada Pipe, ¶112.

²⁵² At these speeds, HSDPA could compete with landline broadband internet if it were priced competitively.

²⁵³ Telecom Corporation of New Zealand Ltd v Clear Communications Ltd, [1995] 1 NZLR 385.

²⁵⁴ New Zealand, Ministerial Inquiry into Telecommunications, Final Report, 2000, p. 24.

²⁵⁵ Bureau Supplementary Evidence, ¶50.

²⁵⁶ Bureau Supplementary Evidence, ¶52.

²⁵⁷ CRTC Telecommunications Monitoring Report, July 2007, p. 60.

²⁵⁸ I have doubts about it. Satellite internet necessarily involves high latency. Xplornet's website warns: "Latency refers to the time it takes for signals to travel to and from the satellite. (Any signal you send from your computer via satellite has to travel over 22,000 miles to space and back again on both the forward and return trip in order to deliver your information.) This typically creates a sub-second lag. For this reason, Xplornet's satellite service may not be ideal for any Internet activity requiring real time data. (Examples include: Online trading and gaming. Essentially any online activity requiring real-time data transfers)".

²⁵⁹ Xplornet's plans start at $54.99 for 512 Kbps. 1.0 Mbps costs $89.99.

²⁶⁰ The Tribunal has defined markets based on national accounts, for instance, in Propane at ¶81.

²⁶¹ Rogers(The Companies)12Apr07-1, -2.

²⁶² Rogers (The Companies)12Apr07-19.

²⁶³ Distributel(The Companies)12Apr07-1, -2 (abridged).

²⁶⁴ Blink is now advertising in Toronto as well. I do not know whether it offers services there.

²⁶⁵ www.cogentco.com

²⁶⁶ www.beanfiled.com

²⁶⁷ The Companies Evidence, ¶6, 75, 90; Appendix 12.

²⁶⁸ For example, Toyota announced in 2006 that it would be building Camrys at Subaru's assembly plant in Illinois.

²⁶⁹ Bureau Evidence, ¶73.

²⁷⁰ See Elhauge, "Defining Better Monopolization Standards", at 274.

Date Modified: 2007-10-01