Application by Telstra Corporation Limited ABN 33 051 775 556
[2010] ACompT 1
•10 MAY 2010
AUSTRALIAN COMPETITION TRIBUNAL
Application by Telstra Corporation Limited ABN 33 051 775 556
[2010] ACompT 1
Citation: Application by Telstra Corporation Limited ABN 33 051 775 556 [2010] ACompT 1 Parties: APPLICATION FOR REVIEW OF THE FINAL DECISION OF THE AUSTRALIAN COMPETITION AND CONSUMER COMMISSION UNDER SECTION 152BU OF THE TRADE PRACTICES ACT 1974 IN RELATION TO THE ORDINARY ACCESS UNDERTAKING SUBMITTED BY TELSTRA CORPORATION LIMITED FOR THE UNCONDITIONED LOCAL LOOP SERVICE FOR BAND 2 AREAS and TELSTRA CORPORATION LIMITED ABN 33 051 775 556 File number: 1 of 2009 Tribunal: JUSTICE JR MANSFIELD (DEPUTY PRESIDENT)
R STEINWALL
RF SHOGRENDate of determination: 10 May 2010 Catchwords: TRADE PRACTICES – telecommunications – access regime – access undertaking – rejection of access undertaking by ACCC – review of decision of ACCC – whether TSLRIC+ approach reasonable – whether modelling by TEA Model version 1.3, including assumptions upon which that modelling was based, reasonable – consideration of particular assumptions and inputs into modelling – degree of network optimisation/ “scorched node” approach – depreciation – value of weighted average cost of capital (WACC) – operating and maintenance costs – long term interests of end users – consideration of non-price terms of undertaking – desirability of certainty in terms of undertaking – decision of ACCC affirmed – Trade Practices Act 1974 (Cth), Part XIC, s 152AB, s 152AH and s 152CF Legislation: Trade Practices Act 1974 (Cth)
Telecommunications Act 1997 (Cth)
Telecommunications National Code 1994 (Cth)Cases cited: Telstra Corporation Limited [2006] ACompT 4
Telstra Corporation Ltd (No 3) [2007] ACompT 3
Re Optus Mobile Pty Limited & Optus Networks Pty Limited [2006] ACompT 8
Telstra Corporation Limited v Australian Competition Tribunal [2009] FCAFC 23
R v Hunt; Ex Parte Sean Investments Pty Ltd (1979) 180 CLR 322
R v Toohey; Ex parte Meneling Station Pty Ltd (1982) 158 CLR 327
Seven Network Limited [2004] ACompT 10
Seven Network Limited (No 4) [2004] ACompT 11
Re Telstra Corporation Ltd (No 1) (2006) 204 FLR 384
Application by Telstra Corporation [2009] ACompT 1
“Access Dispute between Primus Telecommunications Pty Limited and Telstra Corporation Limited – Unconditioned Local Loop Services, Final Decision of ACCC, December 2007”
Verizon Communications Inc v FCC (2002) 535 U.S. 467 (2002)
Re Gasnet Australia (Operations) Pty Limited [2003] ACompT 6Dates of hearing: 24, 25, 26, 27 and 28 August 2009 Place: Adelaide (via video link to Melbourne and Sydney and Canberra) Category: Catchwords Number of paragraphs: 572 Counsel for the Applicant: N Hutley QC, C Moore and J Hewitt Solicitor for the Applicant: Gilbert + Tobin Counsel for the Australian Competition & Consumer Commission: M Sloss SC and A McClelland Solicitor for the Australian Competition & Consumer Commission: DLA Phillips Fox Counsel for Chime Communications Pty Ltd, Agile Pty Ltd, Adam Internet Pty Ltd, PowerTel Limited, Request Broadband Pty Ltd: M Hoyne Solicitor for Chime Communications Pty Ltd, Agile Pty Ltd, Adam Internet Pty Ltd, PowerTel Limited, Request Broadband Pty Ltd: Herbert Geer & Rundle Counsel for Optus: S Free Solicitor for Optus: Minter Ellison
IN THE AUSTRALIAN COMPETITION TRIBUNAL
File No 1 of 2009
RE: APPLICATION FOR REVIEW OF THE FINAL DECISION OF THE AUSTRALIAN COMPETITION AND CONSUMER COMMISSION UNDER SECTION 152BU OF THE TRADE PRACTICES ACT 1974 IN RELATION TO THE ORDINARY ACCESS UNDERTAKING SUBMITTED BY TELSTRA CORPORATION LIMITED FOR THE UNCONDITIONED LOCAL LOOP SERVICE FOR BAND 2 AREAS
BY: TELSTRA CORPORATION LIMITED ABN 33 051 775 556
Applicant
TRIBUNAL:
JUSTICE JR MANSFIELD (DEPUTY PRESIDENT)
MR R STEINWALL
MR RF SHOGREN JJDATE OF ORDER:
10 MAY 2010
WHERE MADE:
ADELAIDE (VIA VIDEO LINK TO MELBOURNE AND SYDNEY AND CANBERRA)
THE TRIBUNAL ORDERS THAT:
1.The decision of the Australian Competition and Consumer Commission made on 22 April 2009 is affirmed.
Note:Settlement and entry of orders is dealt with in Order 36 of the Federal Court Rules.
The text of entered orders can be located using Federal Law Search on the Court’s website.
IN THE AUSTRALIAN COMPETITION TRIBUNAL
File No 1 of 2009
RE: APPLICATION FOR REVIEW OF THE FINAL DECISION OF THE AUSTRALIAN COMPETITION AND CONSUMER COMMISSION UNDER SECTION 152BU OF THE TRADE PRACTICES ACT 1974 IN RELATION TO THE ORDINARY ACCESS UNDERTAKING SUBMITTED BY TELSTRA CORPORATION LIMITED FOR THE UNCONDITIONED LOCAL LOOP SERVICE FOR BAND 2 AREAS
BY: TELSTRA CORPORATION LIMITED ABN 33 051 775 556
Applicant
TRIBUNAL:
JUSTICE JR MANSFIELD (DEPUTY PRESIDENT)
MR R STEINWALL
MR RF SHOGRENDATE:
10 MAY 2010
PLACE:
ADELAIDE (VIA VIDEO LINK TO MELBOURNE AND SYDNEY AND CANBERRA)
INTRODUCTION........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ... [1] THE ISSUES........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .... [28] UNCONDITIONED LOOP SERVICE........ ........ ........ ........ ........ ........ ........ ........ ........ . [42] BACKGROUND TO THE PRESENT APPLICATION........ ........ ........ ........ ........ .... [67] The ACCC Pricing Principles........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ..... [67] Previous Telstra undertakings and ACCC decisions........ ........ ........ ........ ........ ........ [77] TSLRIC+........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ...... [94] The meaning of TSLRIC+........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ... [94] Application of TSLRIC+........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ..... [98] i. The cost concept........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ . [98] ii. TSLRIC+ access pricing........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [103] iii. Scorched node........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [104] iv. Valuation of assets........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ... [113] v. Allocation of common fixed costs........ ........ ........ ........ ........ ........ ........ ........ .... [117] ALTERNATIVE TECHNOLOGIES FOR THE PROVISION OF RETAIL VOICE AND BROADBAND SERVICES........ ........ ........ ........ ........ ........ ........ ........ .... [118] Optical fibre........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [125] i. Background........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ [125] ii. HFC networks........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ... [128] Wireless........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [137] THE ACT........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [140] The Terms of the 2008 Undertaking........ ........ ........ ........ ........ ........ ........ ........ ... [159] TSLRIC + METHODOLOGY........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ....... [180] The views of the parties........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ [182] TSLRIC+ and the reasonableness criteria........ ........ ........ ........ ........ ........ ........ ........ . [188] Choice of technology........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .... [200] THE TEA MODEL........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [230] Costs of a hypothetical new entrant........ ........ ........ ........ ........ ........ ........ ........ ........ .... [230] Conclusions regarding the TEA Model........ ........ ........ ........ ........ ........ ........ ........ ...... [238] Conclusions regarding the proposed monthly charge of $30........ ........ ........ ........ .... [247] DEPRECIATION........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .... [250] WEIGHTED AVERAGE COST OF CAPITAL........ ........ ........ ........ ........ ........ ........ .. [339] The WACC parameters........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ....... [346] OPERATING AND MAINTENANCE COSTS........ ........ ........ ........ ........ ........ ........ ... [499] Overview of the top-down methodology........ ........ ........ ........ ........ ........ ........ ........ ..... [507] Calculating the O&M factors........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ...... [512] Calculating O&M costs........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ [526] The areas of dispute........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ..... [529] (a) The methodology used to estimate operating costs and the annualized capital cost for network support and indirect assets........ ........ ........ ........ ........ ........ ........ ...... [530] (b) The allocation of operating expenses for fibre-related assets, multiplexing systems and local switching........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ .. [540] (c) Overlap between asset categories........ ........ ........ ........ ........ ........ ........ ........ ........ ... [544] (d) Whether the methodology used to calculate operating costs is susceptible to compounding errors........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ..... [549] (e) The methodology of calculating overhead loading (including transparency and efficiency of the overhead loading figures advanced by Telstra)........ ........ ...... [551] The Tribunal’s views on O & M costs........ ........ ........ ........ ........ ........ ........ ........ ........ . [564] CONCLUSION........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ [568] REASONS FOR DECISION
INTRODUCTION
Telstra Corporation Limited (Telstra) has applied to the Australian Competition Tribunal under s 152CE(1) of the Trade Practices Act 1974 (Cth) (the TP Act) for review of a decision of the Australian Competition and Consumer Commission (the ACCC) under s 152BU(2) of the Act to reject an access undertaking offered by it. The ACCC Final Decision was notified to Telstra on 28 April 2009, but is dated 22 April 2009 (the Final Decision).
The ACCC decision was made in respect of an ordinary access undertaking lodged with the ACCC by Telstra on 3 March 2008 (the 2008 Undertaking). It specifies some of the terms and conditions under which Telstra would meet its standard access obligations (SAOs) in respect of the unconditioned local loop service (ULLS) for Band 2 areas (for practical purposes, the metropolitan areas of the major cities within Australia and excluding the Band 1 areas, which are the central business districts). The Band 2 areas are those exchange areas set out in detail in the annexure to the 2008 undertaking.
The 2008 Undertaking proposes a monthly charge of $30 for each Telstra ULLS connected at an exchange building in a Band 2 exchange service area. If accepted, the 2008 Undertaking will remain until 31 December 2010.
The ACCC of course participated in the hearing.
Optus Networks Pty Limited (Optus) was given leave to participate in the proceeding. Optus took an active role in the proceeding before the ACCC and on this application.
In addition, a number of other entities were given leave to participate in the proceeding. They are Chime Communications Pty Ltd; Agile Pty Ltd; Adam Internet Pty Ltd; PowerTel Limited and Request Broadband Pty Ltd. Each of those entities receives the Telstra service in issue from Telstra. Each has notified an access dispute to the ACCC pursuant to s 152CM(1) of the TP Act which relates to Telstra’s terms and conditions of access to that service. Depending on the outcome of this application, the ACCC may proceed to make a determination in those ULLS access disputes different from the terms of Telstra’s proposed undertaking. Those parties, other than Optus, were jointly represented. In these reasons for decision they are collectively called “the Interveners”.
The Tribunal’s powers on the review are set out in s 152CF(1) of the Act. Relevantly, s 152CF(1) provides:
On a review of a decision of the Commission under subsection 152BU(2), … the Tribunal may make a decision:
(a) in any case – affirming the Commission’s decision; or
…
(c)in the case of a review of a decision of the Commission under subsection 152BU(2), … to reject an undertaking – both:
(i) setting aside the Commission’s decision; and
(ii)in substitution for the decision so set aside, to accept the undertaking;
…
The 2008 Undertaking is said by Telstra to be based on the costs to recover the forward-looking economically efficient incremental costs of providing the ULLS over the long run, plus a contribution towards common costs shared by the ULLS. This approach is known as the total service long run incremental cost “plus”, or TSLRIC+. Telstra submitted that its proposed monthly charge is supported by the Telstra Efficient Access Model (TEA Model), and that the TEA Model properly and appropriately models TSLRIC+.
The ACCC first declared the ULLS pursuant to s 152AL(3) in Part XIC of the TP Act in August 1999. Once a service is declared, carriers and carriage service providers supplying the declared service to themselves or others are subject to the SAOs as set out in s 152AR of the Act. The ACCC re-declared the ULLS in July 2006 for a period of three years, with effect from 1 August 2006. Those Declarations were duly published.
The service description of the ULLS is contained in Appendix A of the first Declaration and is defined in the same terms in the ACCC Final Determination of July 2006 on its Declaration inquiry for the ULLS (and other services), as well as in Appendix 2 of that document. It reads as follows:
The unconditioned local loop service is the use of unconditioned communications wire between the boundary of a telecommunications network at an end-user’s premises and a point on a telecommunications network that is a potential point of interconnection located at or associated with a customer access module and located on the end-user side of the customer access module.
A “communications wire” is defined as:
A copper-based wire forming part of a public switched telephone network.
A “customer access module” is also defined in Appendix A as:
A device that provides ring tone, ring current and battery feed to customers’ equipment. Examples are Remote Subscriber Stages, Remote Subscriber Units, Integrated Remote Multiplexers, Non-Integrated Remote Integrated Multiplexers and the customer line module of a Local Access Switch.
More simply, in its reasons for its Final Decision, the ACCC described the ULLS in the following terms:
The ULLS is most commonly a twisted copper wire pair that provides a communications path between the telephone exchange and the consumer or business end-user premises. As a result, the ULLS is an essential input used by other telecommunications access seekers in combination of their own equipment to provide competitive telephony and high-speed broadband services to consumers and businesses.
It was said in the course of submissions that the ULLS also includes a small, but irrelevantly small, element of aluminium wire. It is not necessary to refer to that further.
As noted, by the 2008 Undertaking, Telstra proposed a monthly charge of $30 to be paid by access seekers for access to the ULLS in the Band 2 metropolitan areas. At the time of that undertaking, that proposed charge was substantially above the then currently regulated access charge of $16.75. The proposed charge was, however, significantly lower than a higher figure estimated by Telstra as the appropriate one, using its proposed cost model. In fact, the TSLRIC+ cost estimated by version 1.3 of the TEA Model for Band 2 is $46.54 (subsequently adjusted slightly by reason of an acknowledgment made by Telstra in the course of the hearing). Its proposed monthly charge is therefore only about two-thirds of the figure which its modelling produced. Other non-price terms were also proposed. It is necessary to refer to them in a little detail later in these reasons.
The ACCC rejected the 2008 Undertaking because, according to the Executive Summary of the Final Decision, it:
· is unlikely to promote the long term interests of end-users, as it will not promote competition and will not encourage the economically efficient use of, and investment in, infrastructure;
· will result in Telstra recovering more revenue than is necessary to promote Telstra’s legitimate business interests;
· will harm the interests of access seekers and persons who have rights to use the service;
· contains price terms which will exceed the direct costs of providing access;
· does not have a material effect on the operational and technical requirements necessary for the safe and reliable operation of telecommunications services; and
· is not likely to facilitate the economically efficient operation of the ULLS.
The Final Decision by the ACCC followed the release of a Discussion Paper on the 2008 Undertaking on 4 June 2008 (the Discussion Paper), and subsequently on 13 November 2008, the publication of a draft decision entitled “Assessment of Telstra’s Unconditioned Local Loop Service Band No 2 Monthly Charge undertaking” (the Draft Decision). The Draft Decision also proposed to reject the 2008 Undertaking.
Under s 152BV of the TP Act, the ACCC is directed not to accept such an undertaking unless, relevantly, it is satisfied that the undertaking is consistent with the SAOs that are applicable to the carrier or provider; and unless it is satisfied that the undertaking, if it deals with price or the method of ascertaining price, is consistent with any Ministerial pricing determination; and unless the ACCC is satisfied that the terms and conditions specified in the undertaking are reasonable.
On this application, the Tribunal stands in the shoes of the ACCC. Its decision is taken to be a decision of the ACCC: s 152CF(2). It, too, must not accept the 2008 undertaking unless it is “satisfied that the terms and conditions specified in the undertaking are reasonable”: s 152BV(2)(d).
The summary by the ACCC of its reasons set out above at [14] reflects its assessment of the reasonableness of the 2008 Undertaking.
Section 152AH deals with the topic of reasonableness. It provides:
(1)For the purposes of this Part, in determining whether particular terms and conditions are reasonable, regard must be had to the following matters:
(a)whether the terms and conditions promote the long-term interests of end-users of carriage services or of services supplied by means of carriage services;
(b)the legitimate business interest of the carrier or carriage service provider concerned, and the carrier’s or provider’s investment in facilities used to supply the declared service concerned;
(c)the interests of persons who have rights to use the declared service concerned;
(d)the direct costs of providing access to the declared service concerned;
(e)the operational and technical requirements necessary for the safe and reliable operation of a carriage service, a telecommunications network or a facility;
(f)the economically efficient operation of a carriage service, a telecommunications network or a facility.
(2)Subsection (1) does not, by implication, limit the matters to which regard may be had.
Certain of those matters are further explained in other provisions of the TP Act.
Part XIC of the TP Act contains the Telecommunications Access Regime. It commences with s 152AA containing its simplified outline.
Then s 152AB describes the object of the Part. Section 152AB(1) says that its object is to promote the long-term interests of end-users of carriage services or of services provided by means of carriage services (the LTIE).
Section 152AB(2) of the TP Act provides that, in determining whether a particular thing promotes the LTIE of either carriage services or services supplied by means of carriage services (together defined as listed services), regard must be had to the extent to which the thing is likely to result in the achievement of three specified objectives. They are:
(c) the objective of promoting competition in markets for listed services;
(d)the objective of achieving any-to-any connectivity in relation to carriage services that involve communication between end-users;
(e)the objective of encouraging the economically efficient use of, and the economically efficient investment in:
(i) the infrastructure by which listed services are supplied.
…
Section 152AB(3) provides that s 152AB(2) is intended to limit the matters to which regard may be had.
However, s 152AB(4) and (5) provides that, in determining the extent to which a particular thing is likely to result in the achievement of the objective of promoting competition in markets for listed services, regard must be had to the extent to which the thing will remove obstacles to end-users of listed services gaining access to listed services. That does not limit the matters to which regard may be had on that topic.
Section 152AB(6), (7), (7A) and (7B) of the Act concerns the objective in s 152AB(2)(e) of encouraging the economically efficient use of, and the economically efficient investment in, the infrastructure by which listed services are supplied. Those subsections direct the ACCC to have regard, inter alia, to the following matters:
(a)whether it is, or is likely to become, technically feasible for the services to be supplied and charged for, having regard to:
(i) the technology that is in use, available or likely to become available; and
(ii)whether the costs that would be involved in supplying, and charging for, the services are reasonable or likely to become reasonable; and
(iii)the effects, or likely effects, that supplying, and charging for, the services would have on the operation or performance of telecommunications networks;
(b)the legitimate commercial interests of the supplier or suppliers of the services, including the ability of the supplier or suppliers to exploit economies of scale and scope; and
(c)the incentives for investment of the infrastructure by which the services are supplied and any other infrastructure by which the services are, or are likely to become, capable of being supplied, having regard (without limitation) to the risks involved in making the investment.
That does not, by implication, limit the matters to which regard may be had.
The tasks of the Tribunal, having regard to those provisions, was discussed in Telstra Corporation Limited [2006] ACompT 4; Telstra Corporation Ltd (No 3) [2007] ACompT 3 (Telstra No 3); Re Optus Mobile Pty Limited & Optus Networks Pty Limited [2006] ACompT 8; and Telstra v Australian Competition Tribunal [2009] FCAFC 23 (Telstra v ACompT). It is not necessary to refer to those decisions in detail at this point. It is sufficient to note that, if the Tribunal is satisfied that the LTIE as explained in the TP Act, commencing with s 152AB(1), lies with acceptance of the 2008 Undertaking, the Tribunal has no residual discretion to refuse to accept the 2008 Undertaking.
The Tribunal is not bound by the rules of evidence: s 103(1)(c) of the TP Act. It is, however, confined by s 152CF(4) to considering its decision on the material that was before the ACCC or referred to it. There was some debate between the parties as to the extent of that material. The Tribunal has separately discussed that matter later in these reasons. In any event, nothing really turns on the resolution of that dispute.
Telstra submitted that the Tribunal’s task on that material is to compare the “future with” the 2008 Undertaking (if it is accepted) and the “future without” the 2008 Undertaking (if it is rejected), and to consider which state of affairs is in the LTIE.
THE ISSUES
In the course of the hearing, the parties very helpfully prepared a list of 31 issues to be addressed. They proved to be of considerable assistance in focusing attention to the significance of particular submissions and of the material before the Tribunal. To some degree, they overlapped or intersected. The Tribunal has not, therefore, dealt discretely with each of those issues. It is convenient to note first the principal headings under which the individual issues appeared. They included:
·The Act
·Telstra’s Undertaking
·Role of TSLRIC+ and Historical costs
·The TEA Model
·Asset Lives
·Depreciation
·WACC
·O&M
·International Benchmarking
·Retail Pricing
·Other Cost Models
·Other Issues
·Support for $30 Charge
The Tribunal has found it necessary to discuss only some of them in detail. Before doing so, it is helpful to record more about the nature of the proceeding.
The main question in issue between the parties was the reasonableness of the proposed access charge of $30 per month. To demonstrate its reasonableness, having regard to the terms of s 152AH, some form of modelling was – it was commonly agreed – an appropriate step. That was the means by which Telstra’s earlier proposed access charges in its earlier undertakings were sought to be justified.
It was common ground that the integrity of any form of modelling would depend not only upon the model itself, but also (and obviously) upon the data input into the model. Telstra, for its part, was clearly conscious of the need to justify not simply the form of modelling but the data inputs. It was also conscious, as its senior counsel acknowledged, that it was vulnerable to the ACCC (or the Tribunal) deciding that a particular data input, or perhaps several of them, was or was not reasonable. The result might well affect the modelled output, the access charge. It did not want to have the 2008 Undertaking rejected for some relatively minor difference about the reasonableness of particular data inputs, so affecting the assessment overall of the reasonableness of the outcome produced by its modelling. Its monthly charge of $30 therefore simply allows for some “tolerances” or difference of views on those matters. There is no reason why Telstra should not have based its 2008 Undertaking on such a “broad brush” reduction from the modelled outcome. However, the consequence is that it may be necessary – at least to a degree – to address a series of cascading questions. Some are fundamental. Some are a matter of detail.
The first fundamental question is whether the use of TSLRIC+ is an appropriate means of assessing the reasonableness of the proposed monthly charge. As part of this issue, it will be appropriate to address the extent to which the historical costs of Telstra in establishing the ULLS provide either a basis for, or a cross-check for, a means of determining the reasonableness of the proposed monthly charge.
The next fundamental question is whether modelling the TSLRIC+ by the TEA Model version 1.3 is still an appropriate means of providing an outcome which may inform the reasonableness of the proposed monthly charge. The modelling by the TEA Model version 1.3 was done on the basis of an all copper network. It is clear enough that any future provision of services such as are provided by means of the ULLS, ie using the ULLS as an input to services provided to end-users, would not be done on that basis. There are alternatives as new technology emerges. The Tribunal refers to that matter later in its reasons.
The next stage of consideration, if it were required, involves particular assumptions or inputs into the TEA Model version 1.3 modelling. Different inputs will produce different outcomes. The parties identified a number of inputs into the TEA Model version 1.3 modelling, or assumptions made to select the inputs, which were contentious and which had a sufficient sensitivity to produce a significantly different output. Given the basis upon which Telstra has selected its proposed monthly charge for the 2008 Undertaking, Optus and/or the Intervenors have understandably adopted the approach of challenging some of the sensitive inputs or assumptions underlying them, as well as attacking the TEA model approach more fundamentally. Using the description of those matters adopted by the parties, they relate to:
(a) the degree of network optimisation (the “scorched node approach”);
(b) the extent of trench sharing;
(c) the surface barrier assumption for breakouts and reinstatement costs;
(d) the use of non-tapered architecture;
(e) cabling down one side of the street;
(f) inclusion of lead-in costs; and
(g) the equipment prices used in the TEA Model.
In anticipation of the possibility that some of those inputs or assumptions might be regarded by the Tribunal as not reasonable, Telstra provided in its submissions in reply a “sensitivity test of the TEA Model results” showing some of the changes in the modelled outcome which each of them individually may produce, adopting the change proposed by the respective respondent parties.
That table included sensitivity analysis relating to two of the more significant assumptions or inputs into the TEA Model version 1.3.
One relates to the asset life of the copper cable used in the ULLS, and the annuity used in calculating depreciation on the ULLS. The inputs used by Telstra attracted significant adverse comment. There is a potential for quite a substantial adjustment to the modelled monthly charge, depending upon whether the Tribunal is of the view that Telstra’s input is reasonable in all the circumstances.
The other group of inputs which have the potential for quite a substantial adjustment to the modelled monthly charge, depending on the Tribunal’s view as to the reasonableness of Telstra’s inputs, are those concerning the weighted average cost of capital (WACC).
Finally, in the constellation of inputs or assumptions into the TEA Model version 1.3 are a series of matters relating to the operating and maintenance of the ULLS. As the parties identified them, they were:
(a)the methodology used to estimate operating costs and the annualised capital cost for network support;
(b)the allocation of operating expenses for fibre related assets, multiplexing systems and local switching;
(c) overlap between asset categories;
(d)the transparency of the methodology used by Telstra to calculate operating costs;
(e)whether the methodology used to calculate operating costs is susceptible to compounding errors; and
(f)the methodology of calculating an overhead loading (including transparency and efficiency of the overhead loading figure advanced by Telstra).
For reasons appearing below, the Tribunal has not needed to consider in detail the quality of each of the inputs into the TEA Model version 1.3 set out above. It has, however, made certain observations and findings about some of the more significant ones. As is apparent, certain of those matters had greater potential significance than others. Because Telstra adopted a “broad brush” approach to deciding the $30 monthly charge proposed in the 2008 Undertaking, and apparently one which gave it substantial room for adjustment of a number of inputs, the Tribunal has addressed certain of those inputs or assumptions which may have the greater impact upon the modelled outcome if the Tribunal does not consider that they are reasonable. If, contrary to the Tribunal’s view, the modelling of TSLRIC+ by the TEA Model version 1.3 were an appropriate basis to determine the reasonableness of the 2008 Undertaking, and if the Tribunal concluded that certain of those more significant inputs or assumptions are not reasonable and that reasonable inputs or assumptions would produce a modelled outcome or access charge in the order of $30 only, in any event, it would not be necessary to give such detailed consideration to the range of other contentious inputs or assumptions.. If there were little or no remaining flexibility in the modelled outcome to allow for further adjustment of the access charge, that is if the modelled outcome thereby becomes quite sensitive to any further adjustment of an input or assumption which might further reduce it, the Tribunal would not need to quantify precisely the consequence of such an adjustment to reach the view that the access charge proposed is not reasonable. In that event, as was common ground, the Tribunal should not be satisfied that the 2008 Undertaking was itself reasonable.
Telstra did not contend that, if the Tribunal did not regard the access charge of $30 per month as reasonable, nevertheless the 2008 Undertaking could itself be regarded as reasonable.
UNCONDITIONED LOOP SERVICE
When an access seeker such as Optus seeks access to a ULLS in respect of a particular customer, Telstra will take the copper wire which leads from that customer’s premises back to the local exchange, disconnect it from Telstra’s equipment (assuming that the customer is a Telstra customer) and connect it to the equipment of the access seeker which is housed in the exchange building. More precisely, although the wire remains terminated on the Main Distribution Frame (MDF) in the exchange, instead of the termination point being connected by a jumper to Telstra’s equipment, a jumper instead connects the wire to the access seeker’s equipment. The wire is now connected to the network of the access seeker and is not connected to Telstra’s network. The access seeker obtains use of the wire to provide certain services it may wish to provide, which typically will comprise telephony and ADSL internet. The wire is “unconditioned” in the sense that no electronic devices are attached to the line between the customer’s premise and the MDF. Further, Telstra does not control the operation of the line in any way – it carries no signal or current from Telstra.
The copper wire used in the Customer Access Network (CAN) is typically comprised of a number of individual copper wires divided into “pairs”. The CAN is described in a little more detail below. Each wire has a plastic insulation covering and the two wires are twisted together to form the pair. One copper wire pair is required to provide the ULLS service.
The ULLS is a service provided over copper. Fibre or wireless technology cannot support the ULLS as defined in the current Declaration. It provides that the ULLS is “the use of unconditioned communications wire”, where “communications wire is a copper based wire forming part of a public switched telephone network”. The same or similar services that are supplied to end users via the ULLS (being telephony and broadband internet services) can be supplied over optical fibre and wireless technology, as discussed later in these reasons.
The ACCC considers that the telephony and broadband internet services provided over wireless generally are not substitutes for the telephony and broadband internet services provided via the ULLS.
A copper wire network may incorporate “pair gain” systems, which use equipment attached to the copper pairs in the CAN, which allows the copper pairs to service more than one end-user. However, copper pairs that are attached to “pair gain” systems are not suitable for the supply of ULLS as they are “conditioned”. That is, such wires are not “unconditioned communications wire” for the purposes of the ULLS Declaration. For an access seeker to use the ULLS to provide a service to an end-user, there must be an uninterrupted metallic (ie copper or aluminium) pair of wires between the customer’s premises and the access seeker’s equipment located at the Customer Access Module (CAM). Where there is a “pair gain” system at some point along the metallic path, this requirement will not be met.
The ULLS gives an access seeker the use of the copper pair without any dial tone or carriage service. This allows the access seeker to use its own equipment in an exchange to provide a range of services, including traditional voice services and digital subscriber line (DSL), to end-users connected to that exchange. Access seekers can therefore supply local and long-distance telephony services and other standard (circuit-switched) voice services, without relying on the resale of Telstra’s wholesale and local call services.
The ULLS has no prescribed bandwidth as the access seeker receives the twisted copper pair without conditioning or specific carriage technology. The access seeker may deploy its own infrastructure in Telstra’s exchange to supply a range of downstream services, including the supply of high bandwidth data communications and voice services.
A DSLAM is a Digital Subscriber Line Access Multiplexer. It is a device required to connect a subscriber to a DSL service, and to connect an end user to voice services via the ULLS.
The copper wire CAN forms part of the Public Switched Telephone Network (PSTN). The PSTN consists of telephone circuit switches connected by transmission systems on one side of a MDF and the CAN comprising customer access lines on the other side. The telephone switches are generally specific to a set of services. The transmission systems are typically shared by all of the switches.
The PSTN is Telstra’s largest telecommunications network. As at 30 June 2008, there were approximately 9.86 million fixed line services connected to Telstra’s PSTN. The physical elements of Telstra’s PSTN comprise the CAN, local telephone exchanges and the Inter Exchange Network (IEN).
The CAN is that part of the PSTN that connects the terminal device at each end user’s premises back to the access point with the IEN, typically at the local telephone exchange. The “network termination point” or terminal device at the end-user’s premises is typically the first telephone socket on the customer’s premises. The terminal device is the customer's equipment (for example, telephone handset or fax machine), which typically connects to the first socket. The first socket is the Network Boundary of the carrier’s CAN. The CAN provides connectivity between the customer’s location and the exchange. The point of demarcation is typically the first wall socket. In a commercial building or large multi-dwelling unit (MDU), that is the building distributor or MDF. In a small number of cases a Network Terminal Device is used.
The part of the CAN that is serviced by a particular exchange is referred to as an Exchange Service Area (ESA). An ESA comprises the whole of the CAN which is connected to that particular, and only that particular, local telephone exchange.
There are 584 Band 2 ESAs. Band 2 ESAs are defined as those with more than 108.4 Services in Operation in a square kilometre area which is not a Band 1 area (Band 1 areas are the central business districts of NSW, Queensland, South Australia, Victoria and Western Australia). As noted above, the Band 2 areas are the metropolitan areas outside of the central business districts of NSW, Victoria, South Australia Queensland and Western Australia. They cover 67% of services in operation (SIOs) (approximately 6.9 million SIOs from a total of 10.2 million lines), 70% of the population, but only 0.2% of the land mass.
As at August 2008, there were [X] SIOs in Band 2.
As at March 2009, the CAN comprised 170,291 kilometres of trenching, 375,482 kilometres of copper cable sheath, 168,933 manholes and, Telstra says, some 3,860,630 pits.
The CAN comprises 2 component networks referred to as the main network (also called the feeder network) and the distribution network.
The main network comprises the cables and associated equipment which are located between the local telephone exchange and what is known as a point of cross connection. The point of cross connection is typically a pillar. The use of the term “point of cross connection” refers to the fact that such a point is one at which a connection between the main and distribution networks can be readily made.
The distribution network is comprised of the cables and associated equipment which are located between the pillar and each end-user’s premises. The copper wire pairs are split off along the streets to serve individual premises.
A pillar contains “terminal units” which allow for the connection of copper wire pairs in the distribution network to copper wire pairs in the main network. The collections of end-user premises that are connected to one particular pillar comprise what is referred to as a Distribution Area (DA). Each DA is served by one, and only one, pillar.
The terminal units contain a set of connection terminals for each main network pair that is connected to the pillar. The pairs from the distribution network can be connected to any one of those main network pairs to obtain a continuous circuit back to the local telephone exchange. The pillar therefore provides an important point of flexibility in the network as it allows any main network pair to be connected to any pair in the distribution network back to the local telephone exchange. As a result, it is possible to connect new end-users to the distribution network in a particular DA without having to make any, or any significant, changes to the main network cables.
It should be noted that there is now some debate about the capacity of the CAN to meet additional demand satisfactorily.
A segment of cable known as a “lead-in” is used to make the connection from the distribution cables which are located in the street, to the network boundary point at the end-user’s premises. That is, in a standard residential situation, the lead-in is the copper pair that is specific to that individual dwelling that connects back to the distribution cable.
Lead-ins are installed either underground or aerially. In today’s legislative and regulatory environment, lead-ins tend to be placed underground, although aerial cabling remains a possibility. However, as the ACCC stated in the Final Decision, underground cabling would now most likely be required because of local council requirements. Prior to 1997, aerial cabling could be used because the Telecommunications Act 1991 (Cth) and the Telecommunications National Code 1994 (Cth) did not distinguish between underground and aerial cabling.
The Government has indicated that it may relax regulatory burdens in respect of aerial cabling that relates to the fibre rollout of the proposed National Broadband Network (NBN). Specifically, the Government has indicated that it would allow optical fibre to be rolled out overhead on existing poles to expedite the process and to reduce the costs of deploying the network.
A copper wire cable of the kind used in telecommunications networks such as the CAN is typically comprised of a number of individual copper wires grouped into “pairs”. Each wire has a plastic insulation covering and the two wires are twisted together to form the pair. The pair count of a cable indicates the number of pairs contained within the cable. One copper wire pair is required to provide one unconditioned local loop service.
BACKGROUND TO THE PRESENT APPLICATION
The ACCC Pricing Principles
Section 152AQA of the TP Act requires the ACCC to determine principles relating to the price of access to a declared service. The determination may also contain price-related terms relating to access to the declared service. Such a determination must be made at the same time as, or as soon as practicable after, the ACCC declares or varies a declared service. The ACCC is also required to publish a draft of the determination, invite submissions on the draft and consider any submissions received, before it makes a final pricing determination.
In its July 1997 access pricing principles paper, the ACCC stated that pricing based on total service long-run incremental cost (TSLRIC) to recover the efficient costs of a “forward-looking” network will satisfy the broad statutory criteria, including the reasonableness criteria under s 152AH of Part XIC of the Act.
Following the first declaration of the ULLS by the ACCC in July 1999, the ACCC finalised pricing principles for the ULLS in March 2002. In its 2002 Pricing Principles, the ACCC stated that TSLRIC should be applied in the costing of provision of the ULLS. The ACCC also stated that, in practice, TSLRIC is usually defined to include a contribution to indirect costs. The ACCC now considers that the past rationale for implementing TSLRIC or TSLRIC+ may be less relevant at the present time. TSLRIC+ is explained in the next section of these reasons.
In July 2006, as part of the ACCC’s Final Determination in respect of the Declaration Inquiry for the ULLS, the ACCC published its Draft Pricing Principles Determination for the ULLS. In November 2007, the ACCC made its final ULLS Pricing Principles Determination as required by s 152AQA of the Act. The ACCC did not then specify indicative prices in the 2007 ULLS Pricing Principles Determination.
The ACCC at that time stated that, historically, it has been of the view that TSLRIC+ pricing “is consistent with the price that would prevail if an access provider faced effective competition, and that it usually best promotes the long-term interests of end-users” and best accords with the relevant legislative criteria.
In June 2008, the ACCC made the 2008 ULLS Pricing Principles and Indicative Prices Determination. That determination included indicative prices on ULLS monthly charges to apply until 31 July 2009, and stated that ULLS pricing should be based on TSLRIC+ methodology. The ACCC said that, until it has settled upon its own fixed network cost model, indicative prices based on TSLIRC+ pricing principles should be based on the PSTN Ingress and Egress Model II (PIE II) network cost model. The ACCC used the PIE II network cost model in making its 2008 ULLS Pricing Principles and Indicative Prices Determination.
However, in making the 2008 ULLS Pricing Principles and Indicative Prices Determination the ACCC also stated that:
The ACCC continues to hold concerns about the transparency of the PIE II model but
considers that, given the benefits of issuing indicative prices, it would not be appropriate to wait until an alternative cost model is available and tested.The ACCC believes that, with reservations and appropriately considered inputs, the PIE II model can be used to set indicative prices for the ULLS.
The PIE II model was used in preference to the TEA Model, which was at that time untested.
The indicative prices for ULLS monthly charges on a per service, per month basis for Band 2 for the period to 31 July 2009 are as shown:
Year
2005-06
2006-07
2007-08
2008-09
Monthly Charge
$12.30
$13.70
$14.30
$16.00
The indicative prices for ULLS monthly charges for Band 2 prior to 2005/06 were higher than the period 2005/06 to 2008/09. The indicative prices for the ULLS monthly charges for Band 2 were as shown:
Year 2000/01 2001/02 2002/03 2003/04 2004/05 Monthly Charge $35 $35 $35 $22 $22
Telstra said that it relied upon the ACCC’s 2007 ULLS Pricing Principles in constituting the TEA Model, version 1.3, used in calculating the economic cost of building the CAN for the supply of the ULLS. It claimed that it had taken into account “real world natural, man-made and legal obstacles to deploying the CAN”. It submitted that its modelling, reviewed and supported by experts as suitable to accurately estimate the TSLRIC+ price of supplying the ULLS, provided the presently most realistic network cost model for that purpose.
Previous Telstra undertakings and ACCC decisions
On 9 January 2003, Telstra lodged an undertaking with the ACCC in relation to the supply of the ULLS (2003 Undertaking). The material lodged by Telstra in support of the 2003 Undertaking included Telstra’s economic costing model, the PIE II model. The 2003 Undertaking covered the 2003-04, 2004-05 and 2005-06 financial years. Telstra proposed a monthly price of $13, $22, $40 and $100 for the ULLS in Bands 1, 2, 3 and 4 respectively.
In October 2003, the ACCC published its Final Determination for model price terms and conditions of the PSTN, ULLS and DSL services. In that Final Determination, the ACCC stated that the indicative access price for ULLS per SIO per month for Band 2 areas for 2003-04, 2004-05 and 2005-06 was $22. Telstra subsequently withdrew the 2003 Undertaking and submitting a replacement undertaking on 14 November 2003 (2003 Replacement Undertaking). In October 2004, the ACCC published a draft decision to reject the 2003 Replacement Undertaking. Telstra subsequently withdrew the 2003 Replacement Undertaking and submitted a revised ULLS monthly charge undertaking on 13 December 2004 (2004 Revised Undertaking). On 21 December 2005, the ACCC issued a final decision to reject the 2004 Revised Undertaking.
On 23 December 2005, Telstra lodged a ULLS monthly charge undertaking proposing a single (average) price of $30 per month (2005 Undertaking). The single price was averaged across Band 1, Band 2, Band 3 and Band 4. In support of the 2005 Undertaking, Telstra relied on the PIE II Model to estimate its network costs.
The PIE II Model is a total long-run incremental cost model. It is said to determine, on the basis of various inputs, the network elements which would be necessary to construct a CAN, the costs of those elements, the annual amounts necessary to recover those capital costs, the operational and maintenance and indirect costs applicable to the CAN and the proportion of the CAN costs that relate to the ULLS.
For the purposes of the 2005 Undertaking, the PIE II Model was used to determine a uniform ULLS monthly access charge of $30 across all four Bands (Band 1, Band 2, Band 3 and Band 4), based on the average of its estimated efficient costs of supplying the ULLS in those four areas for each of the years of the 2005 Undertakings.
Telstra claimed that “[t]he PIE II model is the best available model for estimating the TSLRIC of ULLS”.
In August 2006, the ACCC rejected Telstra's 2005 Undertaking. On 17 May 2007, the ACCC's decision was affirmed by the Australian Competition Tribunal: Telstra (No 3).
Under s 152CM of the Act, parties may notify the ACCC to arbitrate an access dispute if the parties are not able to agree on the terms and conditions of access to the declared service. Seven access seekers notified the ACCC of a total of eight access disputes with Telstra regarding the ULLS monthly charges. There was overlap in the terms of access that were disputed. After seeking the parties' views, the ACCC held a joint arbitration to consider the terms of access that were commonly disputed between the access disputes. On 28 June 2007, the ACCC provided to the parties a draft final determination and accompanying consultation paper. The draft final determination reflected the ACCC’s preliminary views pending consideration of the parties’ submissions. In December 2007, March 2008 and April 2008, the ACCC made final determinations in the arbitration of the eight disputes between Telstra and access seekers regarding the supply of the ULLS. These final determinations specified the monthly charges under which Telstra supplied the ULLS to access seekers, expiring on 30 June 2008. As at August 2008, the ACCC was arbitrating over 12 ULLS access disputes, all of which involved ULLS monthly charges. There are currently a number of access disputes before the ACCC.
Since the PIE II model was originally designed to reflect a network as then in place and to calculate network costs for the period 2001-02 to 2004-05, in order to calculate costs for 2005-06 and beyond a procedure was required to take costs from years up to and including 2004-05 and extrapolate them to 2005-06 and beyond. Telstra provided the ACCC with an updated PIE II model and amended the underlying databases with updated information for 2006 to 2008. In its Statements of Reasons in respect of the various arbitration determinations, the ACCC noted that it had used Telstra’s updated PIE II model, populated with the ACCC’s preferred inputs to estimate network costs. In calculating prices for 2007-08, the ACCC used the risk-free rate at the start of the 2007-08 financial year to re-calculate the appropriate Weighted Average Cost of Capital (WACC). The PIE II model was then used by the ACCC to recalculate prices relevant to that year (in contrast to previous years where the historical risk-free rate was used in the ULLS final determinations). In determining indicative prices for 2008-09, the risk-free rate as of 26 May 2008 was used to determine the appropriate WACC. Network costs for each year of the PIE II model were calculated and then a trend line applied to obtain cost estimates for 2008-09. In June 2008, the ACCC stated that, until it had consulted and settled upon its own fixed network cost model, it considered that indicative prices based on TSLRIC+ pricing principles should be based on the PIE II network cost model. There were significant problems with the PIE II model. It is not presently necessary to recite them.
The arbitrated monthly access charges for Band 2 ULLS were $12.30 for 2005/06, $13.70 for 2006/07, $14.30 for 2007/08 and $16.00 for 2008/09. The ACCC prices that applied to industry from 2004/05 to 2006/07 ranged from $22 to $17.
The ACCC also made its ULLS Pricing Principles Determination of June 2008 using the PIE II model. The indicative price was $12.30 for 2005/06, $13.70 for 2006/07, $14.30 for 2007/08 and $16.00 for 2008/09.
In December 2007 Telstra lodged an ordinary access undertaking relating to the terms and conditions of supply of the ULLS for Band 2 Areas with the ACCC (December 2007 Undertaking). This undertaking was withdrawn on 3 March 2008.
On 3 March 2008, Telstra replaced the December 2007 Undertaking with a new ordinary access undertaking relating to the terms and conditions of supply of the ULLS for Band 2 areas with the ACCC. That is the 2008 Undertaking. The 2008 Undertaking, as noted above, has led to the present application.
Prior to the 2008 Undertaking, the ACCC used the n/e/r/a model and the PIE II model to assess Telstra’s ULLS undertakings. The Telstra Efficient Access model (TEA Model) was developed by Telstra to support its 2008 Undertaking. It is referred to later in these reasons. The n/e/r/a model was commissioned by the ACCC in 1999 for assessing PSTN prices. It was further modified in 2002 for use in assessing ULLS prices. The ACCC has not used the n/e/r/a model in its regulatory price-setting or price-assessing roles since 2002. The ACCC did not rely on the n/e/r/a model in rejecting the ordinary access undertaking in respect of the ULLS given to the ACCC by Telstra on the 2005 Undertaking.
On the basis that both the ACCC and Telstra considered the n/e/r/a model was “significantly out of date”, the Tribunal (in considering the ACCC's final decision in relation to the 2005 Undertaking) was also not satisfied that the n/e/r/a model provided a reasonable estimate of the efficient network costs associated with the providing the ULLS for purposes of considering the 2005 Undertaking: Telstra (No.3) at [377].
The TEA Model was used to replace Telstra's PIE II model for the purpose of providing estimates of the cost of the ULLS. It was first provided to the ACCC by Telstra in support of its December 2007 Undertaking.
As with any cost model, the TEA Model necessarily involves the making of a number of assumptions, which impact the reasonableness of the cost model.
TSLRIC+
The meaning of TSLRIC+
TSLRIC is an acronym for “total service long run incremental cost”. The “+” refers to the allocation of common fixed costs, or contribution to indirect costs. It can be understood by breaking the concept into its components.
(a)“Total service” refers to the cost of production of an entire service, not to the cost of a particular unit. However, the cost is usually expressed on a per-unit basis by dividing by the number of units supplied;
(b)“Long run” means that the concept refers to a period where all factors of production can be varied, as opposed to the short run, where the amount of at least one factor of production is fixed;
(c)“Incremental cost” means that the concept refers to the additional costs of supplying the service over and above the situation where the service was not supplied, assuming the scale of all other production activities remains unchanged; and
(d) “+” refers to the allocation of common fixed costs, or contribution to indirect costs.
The terms “common costs” and “indirect costs” are interchangeable (they both refer to costs not directly attributable the production of any one service).
TSLRIC+ is also an attributable cost concept as it refers only to those costs that can be attributed to the production of the service. However, in the case of ULLS, it is produced using production elements shared with customer access lines and the Integrated Services Digital Network (ISDN), and these costs are rolled-in and shared over all lines.
The existence of common fixed costs (costs common to two or more services and therefore costs which do not form part of the incremental costs of any of the individual services to which they are common) means that, if all services were priced on the basis of TSLRIC+, total revenues would fall short of total cost. To prevent such a shortfall, common fixed costs are allocated and recovered via a mark up on TSLRIC+.
Application of TSLRIC+
i. The cost concept
TSLRIC+ is a broad theoretical concept, which can be implemented in a number of different ways. TSLRIC+ can be based on forward-looking economic costs, although it does not have to be. There is nothing to prevent TSLRIC+ from being estimated on the basis of costs incurred in the past (historical cost) to provide the service. However, in order to create efficient pricing signals it is common ground that it is preferable to use estimates of cost that will be incurred in the future. Forward-looking economic costs are the costs of providing the service using the best available and commercially proven technology and efficient production practices. Such costs are derived using current asset prices.
Like all models, TSLRIC+ is dependent on the quality of the cost inputs and assumptions made in relation to them for its utility. They must reflect “forward-looking best available technology and practices consistent with a real-world network architecture and actual conditions”. The input values must be consistent with each other and with the purpose of the modelling. They must reasonably allocate joint and common costs. They must take account of costs incremental to providing the ULLS to as many potential customers as possible. Hence, its utility depends on the quality of decisions about how to measure and allocate costs. The input can obviously affect the output as an access price very significantly, as can the choice of pricing methodology. Clearly, achieving satisfactory inputs is both complex and critical. The parties accepted that it is appropriate to have regard to the existing network design in determining TSLRIC+ estimates, and they also agreed that should be done having regard to the nature of the declared service. There was significant debate on a range of those issues between the parties. One of those issues was whether it was appropriate to model the cost of providing the ULLS as an exclusively copper wired CAN when considering the LTIE, and so to determine whether the 2008 Undertaking was reasonable, as referred to in s 152AH(1)(a) and s 152AB(2) of the TP Act. Another concerned the capacity of the TEA Model, as it was used, to meaningfully determine or assist in determining those matters.
Hence, whilst Telstra maintained that a standard TSLRIC+ principle is that the cost of a network of an efficient operator should reflect today’s values, so that an operator should only be able to recover costs necessary for maintaining future real-asset values in a competitive market, and therefore, the asset valuation should be derived from current cost accounting methodologies, neither Optus nor the other Interveners agreed, and the ACCC’s attitude was also qualified.
The ACCC’s position is that, for a forward-looking TSLRIC+ estimate, the better estimate is on equipment costs at today’s costs, i.e. replacement costs with the modern equivalent assets. Those assets should have the same service potential, and so represent the best (least-cost) option under current or best-in-use technology. Where replacement cost is difficult to quantify, historical cost (appropriately inflated) or reproduction cost (to replace existing plant in substantially the same form at current prices) might be used to provide an estimate of replacement cost, provided the current assets embody the most efficient technology of providing the service. Optus and the other Interveners agreed generally with the ACCC.
TSLRIC+ may also be estimated by reviewing the historical and current costs of operators, or by applying an optimised cost model using forward-looking (current) costs. Optus said the perspective should be narrower in relation to Telstra’s avoidable costs: its operating and maintenance costs during the period of the proposed undertaking, and the cost of any required renewal or replacement of infrastructure or equipment. It said there was a case for using a regulated asset value on the basis of historical cost accounting valuation, ie the written down value.
ii. TSLRIC+ access pricing
The ACCC states that the implementation of TSLRIC+ depends on how costs are measured and allocated, and the parameter values and underlying network assumptions used to produce cost estimates. Optus submitted that, in principle, the application of a traditional implementation of a TSLRIC+ framework would value all assets at the cost of a Modern Equivalent Asset (MEA). A MEA is the lowest cost asset built with the latest available, proven technology which can provide the equivalent service potential as the service which is being costed. The ACCC has acknowledged that “[i]f the rolling out of fibre closer to the customer makes the prospects of efficient duplication more remote, then some of the key rationales for a TSLRIC+ approach to pricing will be less relevant”. Allocative efficiency is promoted by setting marginal consumption signals equal to short run marginal cost. Bypass of an infrastructure asset occurs when a rival invests in competing infrastructure that provides substitute services. Those matters are discussed below.
iii. Scorched node
The first step in TSLRIC+ modelling is to determine the configuration and technology of the network. There are three main approaches to determining network configuration:
(a)Existing network design, which maintains the locations of existing network nodes and uses the types and volumes of equipment currently in place at and between nodes, regardless of whether the existing design and technology is efficient;
(b)Scorched earth, which is based on the most efficient (i.e. least cost) network architecture, sizing, technology and operating practices that are currently available. If this approach is taken, network nodes can be relocated in order to build an optimal network and minimize the costs of access lines, switching and interoffice transport; and
(c)Scorched node, which maintains the network nodes in their current positions but uses efficient technology and volumes of equipment in and between the current node locations. There is no universally accepted notion of which “nodes” ought to be used in the scorched node approach.
“Nodes” are key features of the network. In the TEA Model, the following components of the existing network are retained: the telephone exchange location, DA boundaries, pillar locations and customer locations. Certain contentions were put that, by reason of that starting point in the modelling, the TEA Model could not usefully provide an output access charge which could or should satisfy the requirements of s 152BV(2)(d) – the reasonableness test – as informed by s 152AH.
There was also significant disagreement as to whether the TEA Model can, and should, use an optimised subset of the existing Distribution and Main cable nodes (using the existing rights of way).
What constitutes a “node” for the purposes of implementing a scorched node approach to TSLRIC+ estimation itself depends on the assumptions made. The ACCC considers that a “scorched node” approach is based on the existing network, but optimised between and within the nodes. The nodes that are to be fixed and the degree of “scorching” in a scorched node approach is a choice that is up to the modeller. In general, the more nodes that are fixed, the less optimisation that occurs. The scorched node assumptions typically made in TSLRIC modelling are that the exchange and customer locations are fixed but the other aspects of the network (e.g. cables, trenches and ducts etc) are subject to optimisation. Marsden Jacob Associates, Review of the TEA Model – A Report prepared for Competitive Carriers Coalition, 12 August 2008, stated that “[t]he scorched node assumption typically used in TSLRIC modelling fixes the exchange location, but it allows other aspects of the network (cables, trench, duct etc.) to be variable and subject to optimisation”.
Hence, there are different optimisation methods. The ACCC and Optus expressed concerns about aspects of the optimisation in the TEA Model. Optus said that, as the scorched node approach maintains some elements of the existing network and legacy decisions which may not be the most efficient in the current period, it will give a high estimate of TSLRIC+ as compared to a scorched earth model. Under a scorched earth approach all network elements (e.g. pillars, exchanges, routes) are subject to most efficient/best practice analysis and optimisation. This approach generally gives the lowest possible estimate of TSLRIC+ as it removes all inefficiencies associated with the historical development of the network through time. At the other end of the spectrum, a network can be based on the current costs of the existing firm. This will give the highest possible estimate of TSLRIC+ because it does not allow for any optimisation beyond what is already contained in the actual existing network. Those parties said that the TEA Model is not based on a conventional scorched node approach, but is a hybrid between standard bottom-up and top-down model approaches. For what it is worth, we note that the TEA Model’s approach on this issue is apparently contrary to that adopted in a number of jurisdictions in unbundled local loop models: see eg Network Strategies, Report for Optus: Review of Telstra’s TEA Model Version 1.1 – ULLS Undertaking (5 September 2008).
NERA Economic Consulting (NERA) stated that standard practice in TSLRIC+ models is to employ a scorched node approach. This involves taking the existing number and location of network nodes but assuming best in use technology and efficient volumes of equipment within and between these nodes.
Ovum Consulting (Ovum) stated in August 2008 that a scorched-node approach is one that bases the costs of the network on the existing network topology while a scorched-earth approach is based on a network with an ideal network topology that would meet the demands of a fully efficient operator. Ovum further stated that the network modelled by the TEA Model is based on a scorched-node approach, as the main nodal locations are fixed. Ovum stated that the modelled network was a fair starting point but that the model should be modified to eliminate Telstra’s inefficiencies. Ovum concluded in August 2008 that the TEA Model (then version 1.0) did not reflect a network of an efficient operator and so failed to meet standard TSLRIC+ principles.
Ovum later stated in February 2009 that the TEA Model, version 1.2, was working as originally described by Telstra and that the dimensioning of cables, ducts, pits, manholes, cable joints, cable gauges and pillars were all appropriate for a “scorched node” model of a copper access network for an efficient operator. These calculations include efficiency gains over the existing network. Ovum stated that although a “modified scorched node” approach may be beneficial in terms of efficiency, this would be a major undertaking and that Ovum did not advocate doing it. Ovum stated that the “scorched node” approach, in which the Distribution Areas are fixed, is satisfactory for determining the costs of an efficient operator.
It is not clear that the Ovum statement of February 2009 relates to engineering issues rather than to economic principles. Moreover, Ovum states that the TEA Model is a “scorched node” model where the pillar locations are fixed. Ovum uses the language “modified scorched node” to describe the situation where only some of the nodes are used as a starting point for the network design. Ovum indicates that a “modified scorched node” (as described by it) “may yield further efficiencies but would be a substantial undertaking and is probably not justified for the purpose of the access undertaking”. It also noted that “[i]n Europe and across the world many regulators have adopted [what Ovum describe as] a modified scorched-node approach”.
iv. Valuation of assets
It should be noted that TSLRIC+ modelled by the TEA Model, at best, is an estimate of what it would cost a hypothetical firm to replace Telstra’s current network in today’s conditions in a very short period of time using a single vintage of the best technology currently in use. That would be impossible to achieve in the real world.
Implementing the TSLRIC+ concept requires the making of decisions about how to measure and allocate costs that potentially can have as large an effect on the access price as the choice of pricing methodology.
For example, a “scorched earth” approach to estimating TSLRIC+ assumes that nothing is fixed (and the cost estimate is based on the costs of the network that would be built if no network existed). The ACCC appears to accept that it is appropriate to take into account the existing network design in determining TSLRIC+ estimates, but says that the extent to which it is appropriate to do so depends on the nature of the declared service. Optus did not accept that as a legitimate starting point.
Standard practice is to value assets using the cost of replacing them with the MEA. The MEA is the lowest cost asset, providing at least equivalent functionality and output to the asset being valued.
v. Allocation of common fixed costs
TSLRIC+ requires an allocation of common fixed costs to services using some kind of mark up. The Interveners, however, say that the appropriateness of such inclusion is a matter for debate amongst economists. In its Pricing Principles, the ACCC stated that it is appropriate to allocate common fixed costs when discussing the application of TSLRIC+. Most TSLRIC+ models use an equal proportionate mark up to allocate common fixed costs to different services.
ALTERNATIVE TECHNOLOGIES FOR THE PROVISION OF RETAIL VOICE AND BROADBAND SERVICES
Broadband and telephone services can be delivered in the following ways:
(a) through metallic wires;
(b) through optic fibres;
(c) wirelessly, either by satellite or a mobile phone style of network; and
(d)through fixed wireless means, such as WiMAX technology (which uses an air interface (as an alternative to copper or fibre) to connect a broadband service.
Hybrid fibre-coaxial cable (HFC) networks are also capable of providing broadband and telephone services. An HFC network is a network consisting of both fibre optic cabling and coaxial cabling. Fibre optic cable is used at the trunk level and may be used up to a service area or up to the curb. Coaxial cable is then run from either of these points into the end-user premises. Both analogue and digital services may be run over cable. The capability of HFC networks to provide both voice and broadband services depends on the manner in which they are configured.
There has been some reduction in fixed line usage over recent years. Fixed lines in Australia peaked at 11.66 million in 2004. They fell by some 740,000 between 2004 and 2007. It is not clear whether these numbers include HFC lines or are confined to copper fixed lines. If those statistics include HFC customers, then the decline in Telstra’s fixed line customers may be larger than the 6% which they indicate. Over the period 2003-04 to 2006-07, demand for existing access lines on Telstra’s network (retail and wholesale basic access and ULLS) fell by a significant number each year.
Telephony customers on Optus’ HFC network were 502,000 at 30 June 2004 and 513,000 at 30 June 2007. This represents an increase of 2.2%. However, telephony customers on Optus’ HFC network have grown 3.7% from 30 June 2007 to 30 June 2008.
At 31 December 2007, there were 6,933,814 SIOs in Band 2. At 31 December 2008, there were 6,816,695 SIOs in Band 2. The compound annual growth rate, as assessed by Telstra, was -1.4%.
The TEA Model does not take into account the forecast decline in future demand.
ULLS uptake for all bands, however, has seen positive growth since June 2005. Between 30 September 2007 and 31 December 2008, a very significant percentage of the growth in ULLS SIOs was attributable to Band 2. At 31 December 2007, there were 10,264,732 SIOs across all bands. At 31 December 2008, there were 10,115,537 SIOs across all bands. The compound annual growth rate is assessed by Telstra at 1.2 percent.
Optical fibre
i. Background
Optical fibre uses light waves for the transmission of all forms of telecommunications traffic, permitting the carriage of traffic at very high speeds with little interference. At the local access network level, it is used for transmission between network nodes and switches, and also for access to the premises of high-volume business subscribers and multi-storey office buildings.
If a telecommunications carrier were to build a CAN today, it is more likely that optical fibre cable technologies would be used. Optical fibre cable is not copper based and therefore not a “communications wire” for the purposes of the Declaration for ULLS. Optical fibre can be rolled out to the kerb or to the home. In Canberra, TransACT offers a fibre to the kerb network. Optical fibre connections to the home or building were not widely available in 2008. As discussed in the next section of these reasons, that is a significant matter.
It is accepted that, to ensure that all developers install networks for the future using fibre optic technology, the Government proposes to mandate the use of fibre optic infrastructure to the home and workplace in greenfield estates across Australia that are approved after 1 July 2010. Indeed, as the Interveners urged, the evidence indicates that it is very likely, if not inevitable, therefore that optical fibre would be used rather than copper to build a CAN today. There is evidence that fibre cables are much smaller, so they could be laid via micro trenching compared to the trenching required for larger copper cables, but to the extent that is relevant there may nevertheless be access difficulties for that trenching. In addition, by way of contrast, there is evidence that copper cables are particularly susceptible to corrosion; in order to keep moisture out they need to be grease filled or filled with pressurised air. There is no corresponding requirement for fibre. Finally, if the TEA Model were to include fibre wherever it is more cost effective than copper wire cable, the effect would be that the longest local loops would be provisioned with fibre. This is because those loops represent the copper loops with the highest capital cost because they require the most copper wire cable. Telstra argues, therefore, that if the TEA Model assumed that the longer loops were provisioned by fibre, this would significantly understate the cost of provisioning those longer loops using copper (so that they were capable of being used for ULLS) which would result in the average cost of loops calculated by the TEA Model for the purpose of estimating the cost of ULLS being significantly below the true cost of providing the ULLS to access seekers.
ii. HFC networks
In metropolitan areas of Sydney, Melbourne and Brisbane, Telstra and SingTel Optus have deployed their own HFC networks.
A typical HFC network consists of an optical fibre network that feeds nodes and a coaxial cable network from the nodes to customer premises. Each node has up to 4 coaxial outputs.
In the Final Decision, the ACCC stated that in Band 2 ESAs, the proportion of ESAs with the presence of HFC competitors was a significant percentage by December 2008. Optus commenced its rollout of an aerial HFC network across Australia in the mid 1990s. Optus’ data on its HFC network from some years ago indicated that Telstra and the Optus HFC networks at that time passed 2.2 million homes. In March 2008, Optus stated that the Telstra HFC networks pass around 2.5 million homes.
The Optus HFC network is in residential areas in metropolitan Sydney, Melbourne and Brisbane. The majority of Telstra’s ESAs which overlap with the SingTel Optus HFC network are in ULLS Band 2 areas. Telstra estimates that the HFC network covers 227 ESAs in Band 2, being 38% of Telstra’s Band 2 ESAs. In 75% of these ESAs, Optus’ HFC network services 50% or more of the dwellings in each ESA.
It is important to note, however, that the mere fact that a HFC network may be present in an ESA, does not indicate the extent to which the network may be available in that ESA. That is, the HFC is said to be “present” in an ESA if just one premises in that ESA is serviced by the HFC, and there is no evidence of any wholesale services being supplied using either Telstra’s nor Optus’ HFC.
To provide services to premises not currently connected to the Optus HFC network, Optus has to make a physical connection to these premises which includes lead-in cabling, and more significant infrastructure installation in the case of MDUs. The ACCC has previously noted there are significant barriers to infill expansion of the Optus HFC network, namely, some houses are genuinely unserviceable by either Optus or Telstra using their respective HFC networks, and there are significant barriers to connections of MDUs to the HFC. Optus stated that it supplies residential customers via its HFC network, rather than Telstra’s CAN, where the premises are determined to be “serviceable” by the HFC network.
Of the 2.2 million homes passed by the Optus HFC network, Optus treats approximately 800,000 homes as “unserviceable”. The Optus HFC network passes approximately 514,000 dwellings located in MDUs that are not currently serviced. The balance of approximately 286,000 are Single Dwelling Units (SDUs). Optus does not supply services via HFC to MDUs. Optus classified approximately 36% of homes passed by the Optus HFC network as unserviceable. Telstra classified approximately 6.58% of homes passed by the Telstra HFC network as unserviceable.
In addition, Telstra uses its HFC network to service MDUs only for cable broadband and Pay TV services. It uses the CAN to deliver telephony services to MDU customers.
Optus stated that it faces a range of costs to connect homes passed by its HFC network. Some homes are less costly to connect and serve using HFC; others are more costly due to, for example, difficult terrain, or being a long distance from the HFC cable. In order to minimise production costs, Optus is more likely to serve the more costly homes using Telstra wholesale services. SingTel Optus considers that this approach is technically efficient because the cost of serving a particular customer is minimised.
Wireless
Telstra says that the O&M expenses included in the calculation of the O&M factors were ongoing O&M expenses only, as they were the expenses relevant to the ongoing ULLS monthly charge. All once-off costs such as installation costs were removed from the analysis. The ACCC disagrees – It says that a number of expenses not relevant to the cost of the ULLS remain in the calculation of O&M factors, including switching equipment and optical fibre.
Telstra maintains that the equipment in the model which is classified as “switching equipment” comprises the terminating blocks and associated cable racking on the main distribution frame. It says that these blocks are not part of the switch and are required to terminate CAN services. The “optical fibre” relates to fibre in the main cable which is included in the TEA model so that the costs of main conduit and trenching are appropriately shared between fibre and copper main cable networks.
Calculating O&M costs
After calculation of the relevant “factors” the TEA Model calculates O&M costs broadly as follows:
(a)costs associated with the direct assets are calculated by multiplying the O&M factor by the TEA model’s estimate of the investment value of each category of direct asset. For example, the factor for ducts and pipes is multiplied by the TEA model’s cost estimate for ducts and pipes.
(b)the costs associated with Network Support Assets are calculated by multiplying the capital cost factor for each relevant category of asset (eg, Network Land) by the Network Asset Capital Investment value allocated to each direct asset (for example, ducts and pipes) to produce an operating cost for that category of asset.
In addition to the costs associated with direct, network support and indirect assets, Telstra also calculates factors for “Indirect Expenses”. These are the additional operating costs from Telstra’s overall business that the TEA model attributes to the ULLS. In the annual cost sheet, the Indirect Expenses are calculated by multiplying the factor for each relevant category of asset (eg, Product and Customer Costs) by the O&M cost for each direct asset (eg, Ducts and Pipes).
Each of the capital costs and operating costs are added to produce a monthly cost per line. This is shown on the right hand side of the annual cost sheet against the line item “Monthly Cost”.
The areas of dispute
As indicated, six separate areas of dispute were identified. For convenience, in our consideration, we have addressed some of those areas together, rather than sequentially.
(a) The methodology used to estimate operating costs and the annualized capital cost for network support and indirect assets
As indicated, in order to estimate operating costs, the annual cost sheet multiplies the cost factors by the TEA Model’s estimated investment value for each category of direct assets (ducts and pipes, copper cable etc). The ACCC argues that if the values of those assets increase (for example through an accounting re-valuation), then the TEA Model assumes that operating costs associated with the ULLS will also increase. This approach it says is unreasonable because there is no basis for assuming that all operating costs associated with a service would be in any way linked to the valuation of the assets used in the supply of the service.
However Telstra maintains that when factors are used to calculate O&M, it is the factor which provides the basis for calculating an O&M expense related to the cost of the assets supplying the service.
In addition, the ACCC, Optus and the other Intervenors argue that the use of historical cost data to estimate operating costs is inherently unreasonable. In a forward-looking implementation of TSLRIC+, the optimised network is modelled as new. Operating expenses, particularly the maintenance component, would be lower in a new network than in Telstra’s legacy network. Although Telstra has made a so-called 'forward-looking adjustment' to ducts and pipes and copper cables for the purpose of calculating the cost factors. They assert that Telstra has made no adjustment to the historical costs it has used to calculate the factors to account for the fact that those costs were incurred in relation to an aged network. In those circumstances, they argue that the cost factors derived by Telstra from its historical data are likely to be inflated.
Telstra maintain that this argument lacks foundation. The O&M expenses calculated using the factors are, consistent with TSLRIC+ methodology, intended to apply for the entire lives of the assets used to construct the network. An analysis (to be provided separately) conducted using both the historical costs and indexed historical costs (from the Current Cost Accounts) shows that Telstra’s capital has been depreciated by approximately 50%, indicating that they are halfway through their lives. Accordingly, Telstra’s expense costs have been applied to a network of average age which Telstra says is appropriate for the TSLRIC+ analysis.
Optus broadly argues that the TEA model uses an inappropriate methodology to calculate O&M factors and therefore overestimates the O&M costs of an efficient network-operator.
The factors for O&M costs are derived from Telstra’s RAF. Optus argues that as the value for network capital costs in the RAF is based upon the wholly depreciated value of assets, while the O&M costs in the RAF are not depreciated, the TEA model would estimate a value for O&M costs that exceeds actual cost. Telstra rejects this assertion. It says that the mark-up factor for O&M costs uses a value for network capital costs (plant and equipment) based on the total value “prior to depreciation”. Telstra also considers that Optus ignores the forward-looking adjustment that was applied to ducts and pipes and copper cable and in so doing accounted for 96% of the O&M expenses.
Optus and the other Interveners maintain that (c) by using Telstra’s RAF there is an implicit assumption that the unit investment cost per line of ULLS Bands 1, 3 and 4 are the same as Band 2. Rural O&M costs, such as those experienced in Band 4 and parts of Band 3, are likely to be higher than urban O&M costs experienced in Bands 1 and 2. Optus says that in its experience, the average service assurance activity cost with respect to attending a rural fault is [X] higher than the cost of attending an urban fault.
Telstra rejects this argument. It says that the application of a constant factor to investments that vary by band necessarily means that the cost per line per band will differ. As the investments per line in Bands 3 and 4 are higher than the investments per line in Band 2, the calculated O&M costs are higher in those bands than they are in Band 2.In any event, Telstra argues that developing band specific factors is not consistent with the normal process used in TSLRIC+ models. Telstra does not contend that the O&M costs associated with Band 2 assets are the same as those associated with assets in other bands. As the investment cost per line in Band 2 are lower than they are in Bands 3 or 4, the application of company-wide factors results in lower O&M costs in Band 2 than in Bands 3 and 4.
In much the same way as the ACCC has argued, Optus and the other Intervenors also consider that the O&M factors have been calculated based on Telstra’s historical costs associated with its legacy network, which does not recognise that a new entrant’s network with modern equipment would be cheaper to maintain than Telstra’s legacy network.
Optus states that while Telstra states that O&M costs are, on average, [X] below actual O&M costs allocated to the ULLS in the RAF, after removing costs that should not be allocated to the ULLS, such as multiplexing equipment, the TEA model actually produces O&M costs which are 6.5% higher than those in the actual O&M costs allocated to the ULLS in the RAF.
(b) The allocation of operating expenses for fibre-related assets, multiplexing systems and local switching
The ACCC says that the annual cost sheet used in the TEA Model lists items of plant and equipment as being direct assets which, in fact, have no role to play in the provision of the ULLS, specifically optical fibre cables, multiplexing systems and local switching). It argues that neither the capital costs nor the operating costs associated with those assets should have been included in the annual cost sheet. Similarly, the annual cost sheet includes operating costs for Network Power Systems despite the fact that the provision of the ULLS does not require power.
Optus and the other Interveners raise similar concerns.
In its response Telstra says that for the main network, the TEA Model takes account of the cost of some fibre and related equipment so as to allow the sharing of trench and duct costs by fibre and copper lines to be taken into account in the main network. In any event, it says the contribution of the fibre and related equipment to O&M was minor - at least 96% of the O&M expenses related to ducts and pipes and copper cables.
Telstra also says that the operating costs for Network Power Systems is on of the network support asset categories that has been “reversed out” of the direct asset categories to calculate O&M and indirect factors. As the construction of the RAF requires Telstra to allocate this, along with the other support assets associated with buildings and power, to communications plant and equipment categories, Telstra maintains that it is appropriate to also account for this asset category in the calculation of the cost of providing the ULLS.
(c) Overlap between asset categories
The ACCC asserts that the methodology used by Telstra to calculate operating costs lacks transparency. In particular, some of the categories of assets and expenses used in the annual cost sheet appear in more than one place. For example, “Land”, “Buildings” and “Building Improvements” appear as both “Network Assets” and “Indirect Assets”. Further, “Network Management Systems” are included as a “Network Asset Cost”, “Software” is included as an “Indirect Asset Cost” and “Information Technology” is included as an “Indirect Expense”. The ACCC says that each of those costs appear to relate to the same thing and that the Tribunal cannot be satisfied whether there is any overlap between those costs.
Telstra disputes this. It says that on the “Annual Cost Summary” worksheet, the asset categories “Network Land”, “Network Buildings” and “Network Building Improvements” appear as “Network Assets”, whereas the asset categories of “Land”, “Buildings” and “Building Improvements” appear as “Indirect Assets”.
As explained in Telstra’s Cost Factor Study, the Network Support Assets (also called “Network Assets”) are those support assets associated with buildings and power which the construct of the RAF required Telstra to allocate to communications plant and equipment categories. In order to calculate O&M and indirect factors, Telstra says it had to reverse this allocation to allow the separate identification of costs associated with support assets. Accordingly, the asset categories that appear as “Network Assets” are the support asset categories that Telstra “reversed out” of its direct assets. In contrast, the asset categories that appear as “Indirect Assets” are taken from the Fixed Asset Statements of the RAF. These asset categories are distinct and there is no overlap.
The ACCC also argues that Telstra has made a number of “adjustments” to its historical costs before using them to calculate the O&M factors. Some of those adjustments were made it says in order to – as the ACCC put it – “shoe horn” the categories of costs in the RAF into the categories of assets and expenses used in the annual cost sheet. Further adjustments were made to add depreciation and remove other costs. The ACCC is not satisfied that Telstra has provided a clear explanation of how the quantum of these adjustments was determined.
Telstra considers that the so-called “shoe horning” process and the adjustments made as a consequence were necessary to ensure that the O&M and indirect cost factors calculated were accurate. It says that the quantum of adjustments was explained in the detailed worksheets accompanying Telstra’s Cost Factor Study. In any event, Telstra says that the adjustments that the ACCC criticises had the effect of substantially reducing the indirect expenses and the indirect assets and as a result the indirect expense factor and the indirect asset factor.
(d) Whether the methodology used to calculate operating costs is susceptible to compounding errors
The ACCC and the other Interveners assert that the methodology used by Telstra to calculate operating costs is inherently susceptible to compounding errors. This is because Telstra’s methodology involves calculating costs by “building” on preceding calculations. For example, the calculation of direct O&M costs involves applying a factor to the value of direct assets. If the value of direct assets is inflated, then direct O&M costs will be inflated by a similar magnitude. Further, because indirect expenses are calculated by multiplying the indirect expenses factor by direct O&M expenses, an error at any of the preceding stages of the calculation (ie, the valuation of direct assets and the calculation of direct O&M factors) will also result in indirect expenses becoming inflated.
Telstra says that the ACCC has not identified any actual compounding errors. Rather it says that this submission is merely a restatement of the ACCC’s criticism of the top-down approach and should be rejected.
(e) The methodology of calculating overhead loading (including transparency and efficiency of the overhead loading figures advanced by Telstra)
The ACCC argues that Telstra has applied an overhead loading of X% to equipment costs to account for network planning, logistical support, warehousing and field supervision associated with the construction of the CAN. That is, the values of the direct assets used in the TEA Model it says are inflated by that percentage to include these overhead expenses. The ACCC maintains that the Tribunal should not be satisfied that the approach used by Telstra to estimate and account for overhead expenses is reasonable.
Telstra maintains that it is appropriate and reasonable for overhead costs to be capitalised by applying an overhead loading factor to direct asset amounts since these overhead costs are associated with the construction of the CAN. Telstra says that the statements of its experts indicate that indirect overhead is directly attributable to the building of the CAN, which is a capital asset. It also says that this is the manner in which they have been treated in the audited financial accounts of Telstra.
In its support Telstra submits that in his statement, [X] (Telstra’s senior financial analyst) has stated that a detailed review of the calculation of the internal overhead process confirmed that the activity type codes used in the derivation of the indirect overhead complied with the Australian Accounting Standards and Telstra’s Corporate Accounting policies.
Telstra argues that it applied a conservative estimate of the indirect overhead. The overhead loading factor was calculated in August 2008 to be X and was subsequently updated slightly upwards in December 2008 to Y. Despite this, a factor lower than the X was used in the TEA Model.
The ACCC also maintains that the methodology used by Telstra to generate the overhead loading is not transparent and that the statement of [X] (Telstra’s operations manager) is opaque and lacking in detail. [X] refers to the use of “Allocation Matrices” to allocate overhead costs between various initiatives. The ACCC says that this allocation methodology has not been explained, nor were the Allocation Matrices furnished to the ACCC.
Work on the CAN occurs either by Telstra’s internal workforce or its external contractors. The ACCC refers to the statement of [X] that the overheads associated with external contractors are lower than the overheads associated with Telstra’s internal workforce. However the ACCC says that he does not specify to what degree. As such the ACCC suggests that it is unclear why it would be reasonable to assume that the overhead applicable to Telstra’s internal workforce should be used in the TEA Model.
In its response, Telstra maintains that the statement by [X] explains a complex process as clearly as possible. Telstra also says that the fact that it did not provide the Allocation Matrices to the ACCC is of little consequence, as [X] states that the allocations (of overhead to each project or initiative within each construction program, which constitute the capital build program) contained in the Allocation Matrices were very closely aligned to the total expected expenditure for that financial year.
Telstra points out that contrary to the ACCC’s submission, the overhead applicable to Telstra’s internal workforce is not in fact used in the TEA Model. Telstra points to [X] statement that the majority of the direct capital work on the CAN is carried out by Telstra’s external contractors. Also, as the overhead loading is intended to be applied to the cost of capital network (undertaken by external contractors) the “support overhead” (representing the costs associated with internal support activity, such as management and supervision costs) is excluded from the calculation of the overhead loading factor. The exception is for overheads which relate to the whole Integrated Network Access Program which are allocated across all initiatives regardless of whether the work is undertaken by an internal workforce or external contractors).
The ACCC says that Telstra’s approach assumes that it is appropriate to capitalise overhead expenses by inflating the value of the direct assets used to provide the ULLS by a fixed percentage. The ACCC argues that this is not an appropriate way to treat internal labour costs that are not directly attributable to those direct assets. To the extent that it is reasonable to allow for overhead expenses, the ACCC says that they ought to be accounted for as expenses, and not capitalised.
The ACCC considers that the manner in which Telstra has accounted for overhead loading raises the potential for errors in the calculation of the overhead loading to be magnified in the calculation of other costs of providing the ULLS.
Telstra’s response is that the ACCC has provided no evidence of any errors. Furthermore, the processes involved in the calculation of overhead loading, as described in the statements by experienced Telstra employees, has been done in accordance with Australian Accounting Standards and Telstra’s Corporate Accounting policies.
The ACCC refers to the statement of [X] that the costs comprising Telstra’s internal overheads are recorded in Telstra’s accounts as expense items. If that is correct, then the ACCC argues that there will be double counting when the TEA Model estimates operating costs, because Telstra draws expenses from its RAF accounts to calculate the cost factors. If overhead expenses are included in the RAF accounts, then those expenses will inflate the numerators in each of the factor formulas, leading to higher factors. However, if the TEA Model’s estimate of the value of direct assets has already been marked-up by a fixed percentage to account for those expenses, then the ACCC says that Telstra’s methodology for estimating operating costs will account for those expenses twice - once by inflating the value of the direct assets, and again by inflating the cost factors which are multiplied by the value of those direct assets.
Telstra argues that the ACCC has not accurately characterised [X] statement - Telstra says that [X] statement indicates that when the costs comprising Internal Overhead are incurred they are initially recorded in Telstra’s accounts as expense items rather than capital items. The statement explains the monthly capitalisation routine in which these costs are accumulated and allocated across the various capital projects in the Integrated Asset program. As internal overhead is recorded in Telstra’s accounts as capital, not as an expense, Telstra says that the double counting asserted by the ACCC does not occur.
The Tribunal’s views on O & M costs
Since ducts and copper cables account for some 96% of O&M expenses, the Tribunal has confined itself to considering that element. The Tribunal has three major concerns with Telstra’s estimation of these expenses.
The first is that the factors used, as described above, are calculated for all bands and thus represent averages across all bands, while it appears that O&M per unit of investment may be lower in Band 2 than on average. It is not to the point that, as Telstra says, O&M expense as estimated is lower for Band 2 than for Bands 3 and 4 because investment is higher for Bands 3 and 4. It may still not be relatively low enough.
The second concern is with the assumption that, once factors are calculated, their application implies the assumption that O&M varies with investment, including capitalised overhead expenses. While a top-down approach may or may not be required because of data limitations, such an assumption seems too far from what might be expected to obtain in practice to be a reliable basis for estimation.
Thirdly, the Tribunal is not persuaded that the treatment of capital costs in calculating factors (as the denominators), on the one hand, and as the multiplicand with the factor to produce the O&M estimate, on the other hand, is consistent. The mixed use of historical and forward-looking costs is confusing, with the potential to produce higher factors (because of lower denominators) and higher O&M estimates (because the factors multiply higher investment cost estimates). It may be a consequence of this approach that the estimated O&M expenses in total are well above Telstra’s actual O&M expenses. For reasons explained above in relation to the TEA model implementation of the TSLRIC+ approach, the Tribunal cannot be satisfied that such an outcome meets the legislative reasonableness criteria.
CONCLUSION
The Tribunal, for the reasons given, is not satisfied that the 2008 Undertaking is reasonable.
Primarily, it takes the view that the scorched node modelling used in the TEA Model version 1.3 makes assumptions about the location of the infrastructure of the ULLS and CAN that are not appropriate.
In addition, from its consideration of the topics of depreciation, the WACC, and the operating and maintenance costs, it does not consider that the assumptions and inputs into the TEA Model are appropriate in the respects identified. Whilst the Tribunal has not precisely applied the sensitivity analysis to those elements of the make-up of Telstra’s modelled access charge, the Tribunal takes the view that the use of assumptions or inputs which would be appropriate would take the modelled access charge (assuming the scorched node approach is appropriate) to a figure so close to the access charge proposed in the 2008 Undertaking that the Tribunal would not be satisfied that the 2008 Undertaking is reasonable. Telstra applied a very “broad brush” to the modelled outcome for the access charge. If those three elements cause the Tribunal such concern about the validity of the modelling – in terms of the reasonableness of the output – that it is not satisfied that the proposed $30 monthly access charge is reasonable, it becomes unnecessary in any event to separately consider each of the less significant elements of the TEA Model assumptions and inputs, with their potential cascading effect. Telstra sought to stand on the TEA Model with its assumptions and inputs, and then applied a “broad brush” reduction to the output. The Tribunal has formed a view about those three elements which leaves it unpersuaded that the “broad brush” figure then arrived at is a reasonable one, without separately considering the other elements.
As also appears, the Tribunal is also concerned that the TSLRIC+ approach in the circumstances may no longer be an appropriate one for the reasons given. That matter of principle was raised by Optus, and with perhaps lesser emphasis by the other Intervenors. It was not necessary for the purposes of its decision for the Tribunal to finally resolve that question.
The Final Decision of the ACCC should be affirmed.
I certify that the preceding five hundred and seventy-two (572) numbered paragraphs are a true copy of the Reasons for Decision (public version) herein of the Honourable Justice Mansfield, R Steinwall & RF Shogren. Associate:
Dated: 10 May 2010
0
4
3