AUSTRALIAN COMPETITION TRIBUNAL Application by APA GasNet Australia (Operations) Pty Limited (No 2) [2013] ACompT 8 Citation: Application by APA GasNet Australia (Operations) Pty Limited (No 2) [2013] ACompT 8 Parties: APA GASNET AUSTRALIA (OPERATIONS) PTY LIMITED File number: ACT 2 of 2013 Tribunal: MANSFIELD J (PRESIDENT) PROFESSOR D ROUND (MEMBER) MR R STEINWALL (MEMBER) Date of judgment: 18 September 2013 Catchwords: COMPETITION LAW – National Gas Law (NGL) – review of access decision of Australian Energy Regulator (AER) – transition of access arrangement under National Third Party Access Code for Natural Gas Pipeline Systems to NGL – new access arrangement commenced on 1 July 2013 – whether previous access arrangement left an Interval of delay so rule 92 of the National Gas Rules (NGR) applied and AER properly took tariffs received during interval of delay into account in fixing reference tariffs – previous access arrangement provided revisions commencement date is "the later of 1 January 2013 and the date on which approval of revisions to this Access Arrangement take effect" Held: previous access arrangement continued to operate to 30 June 2013, so no interval of delay existed for rule 92 to apply COMPETITION LAW – National Gas Law (NGL) – review of access decision of Australian Energy Regulator (AER) – determination of opening capital base at commencement of earlier access period under rule 77(2)(a) NGR – adjustment for difference between estimated and actual capital expenditure incurred in that opening capital base – whether adjustment included adjustment for revenue earned on difference as well as the difference in capital expenditure Held: adjustment should not include adjustment for revenue earned on difference in capital expenditure COMPETITION LAW – National Gas Law (NGL) – review of access decision of Australian Energy Regulator (AER) – depreciation allowance – consideration rule 89 NGR – limited discretion of AER – depreciation proposal of access provider did not include inflation adjustment so its revenue profile would be greater in earlier years and tariff profile over period flatter – AER satisfied that proposal did not meet requirement of rule 89(1)(a) as it did not promote efficient growth in market for reference services – common anticipation of significant lessening of reference tariffs from previous access arrangement period – consideration of capacity constraints Held: AER properly exercised limited discretion and refused to approve proposed depreciation schedule in circumstances COMPETITION LAW – National Gas Law (NGL) – review of access decision of Australian Energy Regulator (AER) – rate of return on equity – methodology to determine rate of return on equity discussed – s 87 NGR – consideration of dividend growth model assessment Held: methodology and inputs into modelling by AER not shown to be erroneous COMPETITION LAW – National Gas Law (NGL) – review of access decision of Australian Energy Regulator (AER) – consideration of interaction of ss 65, 258 and 261 of NGL and rules 59-64 NGR Legislation: National Gas (Victoria) Act 2008 (Vic) National Gas (South Australia) Act 2008 (SA) The National Gas (Victoria) Act 2008 (Vic) Gas Pipelines Access (South Australia) Act 1997 (SA) Gas Pipelines Access (Victoria) Act 1998 (Vic) National Gas (Victoria) Act 2008 (Vic) Gas Pipelines Access (South Australia) Law Judiciary and Navigation Acts (1921) 29 CLR 257 Cases cited: Application by APA GasNet Australia (Operations) Pty Ltd [2013] ACompT 4 Application by ElectraNet Pty Limited (No 3) [2008] ACompT 3 Application by Energy Australia [2009] ACompT8 Application by DBNGP (WA) Transmission Pty Ltd (No 3) [2012] ACompT 14 Application by Envestra Ltd (No 2) [2012] ACompT 3 Re Epic Energy South Australia Pty Ltd [2002] ACompT 4 Application by Ergon Energy Corporation Limited (Street Lighting Services) (No 6) [2010] ACompT 14 Re East Australian Pipeline Limited [2005] ACompT 1 Application by ActewAGL Distribution [2010] ACompT 4 Application by Jemena Gas Networks (NSW) Limited (No 3) [2011] ACompT 6 R v Young (1999) 46 NSWLR 681 Marshall v Watson (1972) 124 CLR 640 Parramatta City Council v Brickworks Ltd (1972) 128 CLR 1 Jones v Wrotham Park Settled Estates [1980] AC 74 Philip Morris Inc v Adam P Brown Male Fashions Pty Ltd (1981) 148 CLR 457 Fencott v Muller (1983) 152 CLR 570 Date of hearing: 13, 14, 15 & 16 August 2013 Place: Adelaide (heard in Melbourne) Category: Catchwords Number of paragraphs: 311 Counsel for the Applicant: CA Moore SC and B Kremer Solicitor for the Applicant: Gilbert + Tobin Lawyers Counsel for the Australian Energy Regulator: M O'Bryan SC, C Horan and R Scheelings Solicitor for the Australian Energy Regulator: Australian Government Solicitor IN THE AUSTRALIAN COMPETITION TRIBUNAL ACT 2 of 2013 re: APPLICATION UNDER SECTION 245 OF THE NATIONAL GAS LAW FOR A REVIEW OF A FULL ACCESS ARRANGEMENT DECISION MADE BY THE AUSTRALIAN ENERGY REGULATOR IN RELATION TO APA GASNET AUSTRALIA (OPERATIONS) PTY LIMITED PURSUANT TO RULE 64 OF THE NATIONAL GAS RULES by: APA GASNET AUSTRALIA (OPERATIONS) PTY LIMITED Applicant tribunal: MANSFIELD j (pRESIDENT) PROFESSOR D ROUND (MEMBER) MR R STEINWALL (MEMBER) DATE OF DIRECTIONS: 18 SEPTEMBER 2013 WHERE MADE: ADELAIDE (VIA VIDEO LINK TO MELBOURNE) THE TRIBUNAL DIRECTS THAT: 1. The parties confer with a view to agreeing where possible on the orders which the Tribunal should make to give effect to the reasons for decision published this day. 2. The parties by 2:00 pm on 27 September 2013 either provide to the Tribunal the form of orders to give effect to the reasons, or separately file and serve the form of orders which they respectively propose, together with such written submissions as they may be advised in support of those orders. IN THE AUSTRALIAN COMPETITION TRIBUNAL ACT 2 of 2013 RE: APPLICATION UNDER SECTION 245 OF THE NATIONAL GAS LAW FOR A REVIEW OF A FULL ACCESS ARRANGEMENT DECISION MADE BY THE AUSTRALIAN ENERGY REGULATOR IN RELATION TO APA GASNET AUSTRALIA (OPERATIONS) PTY LIMITED PURSUANT TO RULE 64 OF THE NATIONAL GAS RULES by: APA GASNET AUSTRALIA (OPERATIONS) PTY LIMITED Applicant TRIBUNAL: MANSFIELD j (pRESIDENT) PROFESSOR D ROUND (MEMBER) MR R STEINWALL (MEMBER) DATE: 18 SEPTEMBER 2013 PLACE: ADELAIDE (VIA VIDEO LINK TO MELBOURNE) REASONS FOR JUDGMENT 1 Section 7 of the National Gas (Victoria) Act 2008 (Vic) provides that the National Gas Law (NGL), as set out in the Schedule to the National Gas (South Australia) Act 2008 (SA), applies as a law of Victoria. The National Gas (Victoria) Act 2008 (Vic) commenced operation on 1 July 2008. Section 26 of the NGL in turn gives the National Gas Rules (NGR) the force of law in Victoria. 2 By application dated 20 May 2013, APA GasNet Australia (Operations) Pty Limited (APA GasNet) applied under s 245(1) of the NGL for review of the Access Arrangement Decision of the Australian Energy Regulator (AER) to give effect to its proposed revisions to an access arrangement for the Victorian Transmission System (VTS) pursuant to rule 64 of the NGR. Leave to make the application was given by the Tribunal on 5 July 2013: Application by APA GasNet Australia (Operations) Pty Ltd [2013] ACompT 4. 3 APA GasNet is the owner of the VTS. APA GasNet is wholly owned by APT Pipelines Limited, part of the publicly listed APA Group. The APA Group owns and operates a number of gas transmission pipelines in Australia. 4 The VTS consists of 45 licensed pipelines and associated facilities supplying to the Melbourne metropolitan area, country Victoria and supply to New South Wales and South Australia. The VTS also transports gas across the system and into New South Wales at Culcairn. 5 The VTS is a covered pipeline pursuant to item 6 of Schedule 3 of the NGL. 6 The VTS is a "declared transmission system" pursuant to section 39 of the National Gas (Victoria) Act 2008 (Vic). The effect of the VTS being a declared transmission system includes: (a) that the Australian Energy Market Operator (AEMO) is given relevant functions in respect of the VTS under section 91BA of the NGL; and (b) pursuant to section 91BE of the NGL, APA GasNet is required to have an agreement in place (a service envelope agreement) with AEMO for the control, operation, safety, security and reliability of the VTS. 7 Section 27 of the NGL prescribes the functions and powers of the AER under the NGL and NGR. 8 Section 132 of the NGL provides that a covered pipeline service provider must submit to the AER, for approval by the AER under the NGR, a full access arrangement or revisions to an applicable access arrangement that is a full access arrangement, in respect of the pipeline services the service provider provides or intends to provide in the circumstances specified by the NGR and within the period of time specified by the NGR. 9 There is a process by which the AER makes a decision in relation to such a submission by APA GasNet [see [13] ff]. Ultimately, on 29 April 2013, the AER published its decision entitled Access Arrangement Decision: APA GasNet Australia (Operations) Pty Ltd 2013-17 (Reviewable Decision). The revisions made by the AER were incorporated into the access arrangement (including terms and conditions and the access arrangement information) and were published by the AER with the decision. 10 The Reviewable Decision is the subject of the present application. BACKGROUND TO REVIEWABLE DECISION 11 On 25 June 2008, the relevant regulator at the time, the Australian Competition and Consumer Commission (ACCC), published a decision pursuant to section 2.41(b) of the National Third Party Access Code for Natural Gas Pipeline Systems (Gas Code) which approved the revisions to the access arrangement to apply to the VTS to take effect from 9 July 2008 until revisions to that access arrangement were next approved (2008-13 Access Arrangement). In contentions APA GasNet called it the 2008-2013 Access Arrangement because, it contended, that access arrangement did not expire until 30 June 2013. The Gas Code was set out in Schedule 2 of the Gas Pipelines Access (South Australia) Act 1997 (SA). It was applied as law in Victoria pursuant to the Gas Pipelines Access (Victoria) Act 1998 (Vic) (later repealed by the National Gas (Victoria) Act 2008 (Vic)). 12 Pursuant to the transitional provisions in the NGL (clause 26 of Schedule 3), the 2008-12 Access Arrangement is deemed to be a full access arrangement approved by the AER under a full access arrangement decision. Pursuant to those transitional provisions, the 2008-12 Access Arrangement is an "applicable access arrangement" for the purposes of rule 52 of the NGR as it is a full access arrangement that has taken effect after being approved or made by the AER under the NGR as from the commencement day of the NGL, being 1 July 2008. It will be necessary to refer to the transitional provisions in more detail later in these reasons for decision. 13 Pursuant to the Gas Code, the 2008-12 Access Arrangement was required to include a "Revisions Submission Date" (clause 3.17(a)). The Revisions Submission Date was stated to be 31 March 2012. 14 Under rule 52 of the NGR, APA GasNet was required to submit, and on 2 April 2012 (as 31 March 2012 was a Saturday) did submit, an access arrangement revision proposal to the AER (APA GasNet Access Arrangement Proposal). 15 Pursuant to rule 59 of the NGR, the AER was required to make, and did make, an access arrangement draft decision in relation to the APA GasNet Access Arrangement Proposal entitled Access Arrangement Draft Decision: APA GasNet Australia (Operations) Pty Ltd 2013-17 dated September 2012 and published on 11 September 2012 (Draft Decision). 16 Rule 60 of the NGR sets out that APA GasNet may submit additions or other amendments to the access arrangement proposal in the period fixed by the AER, subject to the following conditions and requirements: (a) APA GasNet was only permitted to make the amendments necessary to address matters raised in the Draft Decision unless the AER otherwise approved further amendments; and (b) if APA GasNet submitted amendments, APA GasNet was also required to provide the AER (together with the amendments) a revised proposal incorporating the amendments. 17 On 9 November 2012, APA GasNet submitted additions and other amendments to the AER (APA GasNet Access Arrangement Proposal Revisions). 18 Pursuant to rule 62 of the NGR, the AER was required to make an access arrangement final decision in relation to the APA GasNet Access Arrangement Proposal Revisions. 19 On 15 March 2013, the AER published its access arrangement final decision under rule 62 of the NGR entitled Access Arrangement Final Decision: APA GasNet Australia (Operations) Pty Ltd 2013-17 (Final Decision). The AER's Final Decision was to refuse to approve the APA GasNet Access Arrangement Proposal Revisions. 20 Pursuant to rule 64(1) of the NGR, if, in an access arrangement final decision, the AER refuses to approve an access arrangement proposal, the AER must itself propose revisions to the access arrangement for the relevant pipeline. 21 In the AER's Final Decision, the AER proposed revisions to the access arrangement to apply to the covered pipeline. These revisions were set out in the AER's Final Decision (AER proposed revisions). 22 Pursuant to rule 64(4) of the NGR, the AER must, within two months after the Final Decision, make a decision giving effect to its proposed revisions. 23 The Reviewable Decision of 29 April 2013 was the final step in the process. 24 In making the Reviewable Decision, the AER was required by s 28(1) of the NGL to perform or exercise that function or power in a manner that will or is likely to contribute to the achievement of the national gas objective in s 23 of the NGL. 25 Section 23 of the NGL sets out the national gas objective (NGO) in the following terms: The objective of this Law is to promote efficient investment in, and efficient operation and use of, natural gas services for the long term interests of consumers of natural gas with respect to price, quality, safety, reliability and security of supply of natural gas. 26 The AER was also required by s 28(2)(a) of the NGL to take into account the revenue and pricing principles in s 24 of the NGL when making those parts of the Reviewable Decision that related to a reference tariff. The revenue and pricing principles are defined in section 24 of the NGL. 27 An application for review by the Tribunal is limited in that: (a) the applicant must demonstrate that the AER made an error of one of the 4 kinds described in section 246(1) of the NGL; (b) the applicant must not raise any "matter" (by way of evidence or submissions) that was not "raised in submissions in relation to the reviewable regulatory decision before that decision was made"; and (c) the Tribunal must not consider any material other than that specified in section 261(1) and (2) of the NGL, which material essentially comprises the material that was before the AER. 28 The grounds of review to which APA GasNet is therefore limited by s 246(1) of the NGL are that: (a) the AER made an error of fact in its findings of facts, and that error of fact was material to the making of the decision; (b) the AER made more than 1 error of fact in its findings of facts, and those errors of fact, in combination, were material to the making of the decision; (c) the exercise of the AER's discretion was incorrect, having regard to all the circumstances; or (d) the AER's decision was unreasonable, having regard to all the circumstances. 29 The Tribunal's task is to determine whether the AER erred in the ways specified in section 246(1). It is not to substitute a decision which the Tribunal may prefer to make on the material before the AER: Application by ElectraNet Pty Limited (No 3) [2008] ACompT 3 at [64] and [69] (ElectraNet (No 3)); Application by Energy Australia [2009] ACompT 8 (Energy Australia) at [70]; Application by DBNGP (WA) Transmission Pty Ltd (No 3) [2012] ACompT 14 (DBNGP (No 3) at [25]. THE ISSUES ON REVIEW 30 APA GasNet applies for review of the Reviewable Decision in respect of the following matters: (1) Interval of delay: the AER determined that an interval of delay had arisen with respect to the making of the Reviewable Decision and therefore rule 92(3) of the NGR applied to allow the AER to take the operation of rule 92(3) into account in fixing reference tariffs for the new access arrangement period (Interval of Delay Decision). (2) Opening capital base as at 1 January 2013: the AER determined that the opening capital base as at 1 January 2013 should exclude not only the amount of the difference between the estimated capital expenditure and actual capital expenditure for 2007, but also the associated return on capital in respect of that difference (Opening Capital Base Decision). (3) Depreciation and indexation of the capital base: the AER determined that the depreciation schedule should be based on indexed values for the capital base (Depreciation and Indexation Decision). (4) Rate of return on equity: the AER, in calculating the rate of return on equity, did not correctly apply the Capital Asset Pricing Model (Rate of Return on Equity Decision). 31 Each of those issues was addressed by APA GasNet and the AER in their helpful written and oral submissions. They are separately considered below. However, in respect of certain of those issues, two particular matters were raised by the AER which should be noted at this point. 32 The first concerns the application of s 258(2) of the NGL. It prohibits any party other than the AER from raising before the Tribunal "any matter that was not raised in submissions in relation to the reviewable regulatory decision". 33 In the context of s 258(2), the word "matter" is a synonym for "issue": Application by Envestra Ltd (No 2) [2012] ACompT 3 (Envestra (No 2)) at [115]; it is a "controversy or thing in dispute": DBNGP (No 3) at [299]. The effect of section 258(2) is therefore to restrict the issues that the Tribunal may consider and that an applicant can raise in a review. If an issue cannot be identified as arising out of matters raised in submissions considered by the AER, it is not permitted to be raised in the review: Re Epic Energy South Australia Pty Ltd [2002] ACompT 4 (Epic Energy) at [24]-[25] regarding a similar provision of the Gas Pipelines Access (South Australia) Law (GPAL); Application by Ergon Energy Corporation Limited (Street Lighting Services) (No 6) [2010] ACompT 14 (Ergon (No 6)) at [36]-[37] regarding the GPAL; Re East Australian Pipeline Limited [2005] ACompT 1 (EAPL) at [2]-[4], [9]-[10] regarding the GPAL. 34 Whether a matter was "raised in submissions in relation to the reviewable regulatory decision" may depend on the level of abstraction at which a relevant "matter" is defined. In DBGNP (No 3) at [305], the Tribunal observed: The Tribunal adopts the purposive construction of s 258(2) referred to in the decisions discussed. The process leading to an access arrangement decision by a regulator involves a series of steps designed to secure the identification of contentious issues and then resolution by the regulator in the light of the material and submissions of the covered pipeline provider and those entities whose interests may be affected by the decision. In practical terms, it is a matter for assessment by the Tribunal in the particular circumstances whether the issue which is in contention is one where, as a result of that process, an informed decision has been made by the regulator and which is the subject of the review. 35 In certain respects, the AER contends that APA GasNet seeks to argue certain aspects of one or other of the four issues referred to in contravention of s 258(2), including in circumstances where APA GasNet is said to have made the particular submissions only after the Final Decision and where the AER in the circumstances properly declined to consider it: rule 64(3) of the NGR. 36 The second aspect concerns s 261 of the NGL. It regulates the matters that are required, and permitted, to be considered by the Tribunal in reviewing a reviewable regulatory decision. 37 The relevant provisions of section 261 are as follows: 261—Matters to be considered by Tribunal in making determination (1) Subject to this section, the Tribunal, in reviewing a reviewable regulatory decision, must not consider any matter other than review related matter. (2) The Tribunal, in reviewing a reviewable regulatory decision, must (a) in all cases, have regard to any document— (i) prepared, and used, by the original decision maker for the purpose of making the reviewable regulatory decision; and (ii) that the decision maker has made publicly available; and (b) … In addition, if in a review, the Tribunal is of the view that a ground of review has been established, the Tribunal may allow new information or material to be submitted if the new information or material— (a) would assist it on any aspect of the determination to be made; and (b) was not unreasonably withheld from— (i) in all cases, the original decision maker when the decision maker was making the reviewable regulatory decision; and ... (6) In the case of a review of a reviewable regulatory decision of the AER that is a decision to make a full access arrangement decision in place of an access arrangement that the AER did not approve, the Tribunal may consider the reasons of the AER for its decision not to approve the access arrangement. (7) In this section review related matter means— (a) the application for review and submissions in support of the application; and (b) the reviewable regulatory decision and the written record of it and any written reasons for it; and (c) in the case of a reviewable regulatory decision that is an applicable access arrangement decision—any document, proposal or information required or allowed under the Rules to be submitted as part of the process for the making of the decision; and (d) any written submissions made to the original decision maker before the reviewable regulatory decision was made …; and (e) any reports and materials relied on by the original decision maker in making the reviewable regulatory decision …; and (f) any draft of the reviewable regulatory decision …; and (g) any submissions on— (i) the draft of the reviewable regulatory decision or the reviewable regulatory decision itself considered by the original decision maker; or (ii) …; and (h) the transcript (if any) of any hearing conducted by the original decision maker for the purpose of making the reviewable regulatory decision. 38 In short, the Tribunal's review is confined to the material that was before the AER in making its decision: Application by ActewAGL Distribution [2010] ACompT 4 (ActewAGL) at [28]. 39 In this review, the reviewable regulatory decision for the purposes of s 261 is the Reviewable Decision but it is accepted that its reasons may be informed by the Final Decision. As recognised by the Tribunal in DBNGP (No 3), submissions provided to the AER prior to the Reviewable Decision may be review related matter, even if the submissions were not considered by the AER (because they were received after the completion of the consultation undertaken in connection with and prior to the making of the Final Decision): DBNGP (No 3) at [31]. Under s 261, the Tribunal may have regard to such materials, but is not required to do so. 40 Again, in certain respects the AER contends that APA GasNet is seeking to rely on material on which it is not entitled to rely upon, and that to that extent the Tribunal should not entertain its submissions. 41 In the Introduction to its written submissions, the AER included a section under the heading "Issues Raised and Submissions made After the Final Decision". The AER objected to APA GasNet relying on "various late and unsolicited submissions, particularly insofar as the submissions raise new issues". It provided by an annexure to its submissions a list of the documents to which it objected. They related to the Depreciation and Indexation Decision, the Opening Capital Base Decision, and the Rate of Return on Equity decision. 42 The AER's process of decision-making is prescribed by the NGR. After receiving an access arrangement revision proposal, the AER must make a final decision within 13 months: rule 13. In the interim period, it must invite external submissions: rule 58; make a draft decision (including fixing a revision period): rule 59. At the time of the draft decision, the AER must fix a time to make a revised access arrangement proposal, and then for further written submissions not less than 20 days after the revision period: rule 59(c)(ii) and (iii). It must then consider the responses and it may have a hearing: rule 61. After the final decision is made: rule 62, if the AER does not approve the revised access arrangement proposal, it must make a decision within a further two months giving effect to its own access arrangement proposal: rule 64(3) says that the AER may (but is not obliged to) consult on its proposal. 43 The AER's concern is that APA GasNet made further submissions after the Final Decision of 15 March 2013 (and after the date for the receipt of submissions fixed under rule 59(c)(iii) as 7 January 2013). It seems prepared to accept that APA GasNet could make submissions up to 7 January 2013, that is after the close of the revision period fixed under rule 59(c)(ii) as 9 November 2012. After the Final Decision, written submissions received by the AER may constitute review related matter: s 261(7)(d) of the NGL, but it is clear the AER is not required to consult during the two month period between the Final Decision and the Reviewable Decision: rule 64(3), and it is not required to consider any submissions received after the period fixed under rule 59(c)(iii): s 65, NGL. 44 In that way, the relevant provisions seek to ensure that the access provider is given a fair opportunity to present material to the AER in support of an access arrangement proposal or revised proposal, third parties are given a fair opportunity to consider and comment upon those proposals, and on the draft decision of the AER, and the AER is given an appropriate time to consult as necessary and to form its views. It is not necessary in this matter to go beyond those general comments; in particular, it is not necessary to consider the contention that the AER may be precluded from considering late submissions from the access provider because that may give rise to procedural fairness obligations to third parties. That is a matter to be addressed if, and when, it arises directly. The Tribunal has referred above to its comments in DBNGP (No 3). 45 The submissions of the parties, including the schedule attached to the written submissions of the AER, proceeded on the assumption that "submissions" may include fresh evidentiary material as well as what the access provider says in relation to them. 46 The Tribunal has not placed any real weight on any of the material listed in that schedule in its reasoning and conclusions on the Opening Capital Base Decision or the Depreciation and Indexation Decision, simply because the parties' contentions did not require it to do so. 47 In relation to the Rate of Return on Equity Decision, the two documents objected to are the letter from APA GasNet of 28 March 2013 and the enclosed report of Competition Economists Group (CEG) of March 2013. 48 The material before the Tribunal shows that APA GasNet submitted with its Access Arrangement Proposal detailed expert reports relating to the topic of the Rate of Return on Equity from CEG, SFG Consulting (SFG), NERA Economic Consulting (NERA) and Capital Research Pty Ltd (CR). Prior to the Draft Report, the AER had obtained expert reports from Associate Professor Lally and from McKenzie and Partington of the Securities Industry Research Centre of Asia-Pacific (SIRCA) Ltd (McKenzie and Partington). 49 Then APA GasNet with its Access Arrangement Proposal Revisions submitted further reports from CEG, NER and SFG together with expert reports from PwC (Economics) (PwC), Professor Gregory, Professor Wright and Ernst & Young. 50 It is not surprising that, leading up to the Final Report, the AER obtained further expert reports from Lally, and McKenzie and Partington, as well as an expert report from Cambridge Economics Policy Associates Ltd (CEPA). 51 Those documents were considered and variously referred to in the Final Decision. It cannot have been intended that the AER should not be entitled to seek expert comments on the further expert material submitted by APA GasNet. Nor is there any indication that the AER's inquiries should be confined to the particular experts it had previously engaged. 52 As we remark later in these reasons, to some degree the process of expert response to expert response could continue more or less indefinitely, perhaps prolonged by the ongoing dialogue being in writing. Nevertheless, as noted above, the time prescriptions applicable to the AER mean that it could not continue indefinitely. Given the extensive exchange of experts' views, it is understandable that, despite the regulatory strictures referred to above, there may be more that an access provider wants to say outside the process prescribed, before the access arrangement decision is made, so that it may be seen as review related material. The two documents referred to in [47] above fall into that category. 53 The AER declined to take that material into account. It was entitled to adopt that position, because the prescribed process does not provide an access provider the right, after the Final Decision, to adduce further material and have it considered by the AER. In this matter, it is not necessary to determine whether there are no circumstances in which, despite that situation, otherwise belated material before the AER is review related material capable of demonstrating error on the part of the AER. That is because, as appears in the section of these reasons addressing the Rate of Return on Equity Decision, the Tribunal does not consider that that further material demonstrates reviewable error by the AER, either taken with all the material on the issue or taken alone. INTERVAL OF DELAY DECISION Background 54 The issue concerns the operation of rule 92 of the NGR. Rule 92 provides that a full access arrangement must include a reference tariff variation mechanism that is designed to equalise (in terms of net present values (NPVs)) forecast revenue from reference services over the access arrangement period. 55 Rule 92 relevantly provides as follows: However, if there is an interval (the interval of delay) between a revision commencement date stated in a full access arrangement and the date on which revisions to the access arrangement actually commence: (a) reference tariffs, as in force at the end of the previous access arrangement period, continue without variation for the interval of delay; but (b) the operation of this subrule may be taken into account in fixing reference tariffs for the new access arrangement period. 56 As noted, the 2008-2012 Access Arrangement was approved by the ACCC under the Gas Code. Clause 2 of the 2008-2012 Access Arrangement (headed "Term") provided for a commencement date (1 January 2008) and a "Revisions Commencement Date" specified in the following terms: The Revisions Commencement Date is the later of 1 January 2013 and the date on which approval of revisions to this Access Arrangement take effect. 57 Section 3.17 of the Gas Code provided that an access arrangement must include a "Revisions Commencement Date", being a date upon which the next revisions to the access arrangement are intended to commence. 58 Consequently, APA GasNet contends that, consistent with section 3.17, the Revisions Commencement Date specified in the 2008-2012 Access Arrangement was not 1 January 2013 but the later of 1 January 2013 and the date on which approval of revisions to the access arrangement take effect. The 2013-2017 access terms under the Reviewable Decision took effect on 1 July 2013. APA GasNet says that rule 92 of the NGR was not engaged because no revenue equalisation adjustment could properly have been made by the AER because there was no interval of delay in the sense that term is used in rule 92. 59 The AER took a different view in the Reviewable Decision. It said there was a temporal gap between access arrangements for the pipeline, so the reference tariffs in force at the end of the previous access arrangement continued during the gap but those tariffs had to be taken into account in fixing tariffs for the next access arrangement. 60 It took the position that APA GasNet's 2008-2012 Access Arrangement specified reference tariffs only until 31 December 2012. As the 2013-17 access arrangement (the subject of this review) did not take effect until 1 July 2013, there was a gap (the interval of delay) of 6 months. As a matter of uncontested fact, the tariffs prevailing at 31 December 2012 were higher than the tariffs determined under the 2013-17 access arrangement by the Reviewable Decision. As APA GasNet continued to apply, and receive the benefit of, the 2012 tariffs in the period 1 January – 30 June 2013, applying rule 92(3), the AER adjusted the reference tariffs commencing on 1 July 2013 to account for the higher tariffs that were charged during the interval of delay. 61 As noted, APA GasNet disputes that any such adjustment should have been made. APA GasNet also raises a further matter that the AER says was not raised by the Application: that the adjustments made by the AER pursuant to rule 92(3) contained calculation errors in any event. 62 In relation to the alleged calculation errors, the AER has considered the matters raised by APA GasNet (even though they were not included within its Application). The AER agrees that the adjustments within the 2013-17 access arrangement contain calculation errors. If the Tribunal reaches the view that rule 92(3) was applicable and empowered the AER to adjust the initial 2013 reference tariffs, the AER proposes to correct the calculation errors pursuant to rule 68 of the NGR. Consideration 63 Division 8 of Part 9 of the NGR governs the specification of reference tariffs (tariffs applicable to reference services) over the course of an access arrangement period. 64 Rule 92 is set out above. 65 Rule 92(3) is the primary rule concerning the conversion of total revenue into tariffs: it stipulates that the reference tariff variation mechanism must be designed to equalise (in terms of NPVs): (a) forecast revenue from reference services over the access arrangement period (ie the revenue forecast to be earned by the service provider having regard to the applicable reference tariffs and service volume forecasts); and (b) the portion of total revenue allocated to reference services for the access arrangement period (ie the regulated or allowable revenue determined under rule 76). 66 The issue of fact upon which the AER is said, erroneously, to have based the Reviewable Decision in this respect is that there was a period of time (1 January 2013 to 30 June 2013) when the relevant revision commencement date stated in the 2008-2012 Access Arrangement had elapsed and before the new reference tariffs under the Reviewable Decision came into force. Or, put another way, the issue of fact is whether the "end of the previous access arrangement period" was 31 December 2012, so that the tariffs recovered by APA GasNet from 1 January 2013 to 30 June 2013 were recovered by operation of Rule 92(3). 67 To address that question, it is first desirable to note in a little more detail the transitional provisions applicable. 68 The Gas Code was repealed with effect from 1 July 2008 and replaced with the NGL and the NGR. The notion of a "revisions commencement date" was replaced with a "revision commencement date" [singular]. The relevant transitional provisions in the NGL and NGR provided for the continuing in operation of the 2008-2012 Access Arrangement as follows: (a) section 26 in Schedule 3 to the NGL provides that a current access arrangement approved in a relevant final decision under the Gas Code "is deemed to be a full access arrangement approved by the AER under a full access arrangement decision"; (b) clause 3(9) of Schedule 1 to the NGR relevantly provides that: …a date designated in a transitional access arrangement as a revisions commencement date will be taken to be a revision commencement date for the purposes of the rules. (c) clause 1(1) of Schedule 1 to the NGR defines "transitional access arrangement" in a manner that includes an access arrangement deemed to be a full access arrangement pursuant to section 26 of Schedule 3 to the NGL. 69 Rule 3 of the NGR contains the following definition in relation to a "revision commencement date": Revision commencement date for an applicable access arrangement means the date fixed in the access arrangement as the date on which revisions resulting from a review of an access arrangement are intended to take effect. 70 That composite expression is not used elsewhere in the NGL or the NGR. However, the term "applicable access arrangement" is defined in s 2 of the NGL as follows: Applicable access arrangement means a limited access arrangement or full access arrangement that has taken effect after being approved or made by the AER under the Rules and includes an applicable access arrangement as varied – (a) under the Rules; or (b) by an access determination as provided by this Law or the Rules … 71 Under rules 48(1)(i) and 49(1) of the NGR, a full access arrangement (other than a voluntary access arrangement) must contain a review submission date (ie a date on or before which an access arrangement revision proposal is to be submitted) and a revision commencement date. Rule 50(1) then provides that, as a general rule a review submission will fall four years after the access arrangement took effect or the last revision commencement date; and that a revision commencement date will fall five years after the access arrangement took effect or the last revision commencement date. Rule 50 also provides that the AER must accept a proposal to fix a review submission date and a revision commencement date in accordance with the general rule, but may also approve dates that do not conform with the general rule if satisfied that they are consistent with the national gas objective and the revenue and pricing principles. 72 It is clear enough that in general terms the 2008-2012 Access Arrangement as approved by the ACCC under the Gas Code in its decision dated 30 April 2008 had a commencement date of 1 January 2008 and an anticipated end date of 31 December 2012. It referred to the previous Access Arrangement periods as commencing on 1 January 2003 and ending on 31 December 2007. 73 It is also clear enough that, when APA GasNet made the APA GasNet proposal to the AER on 2 April 2012, it anticipated that the revised access arrangement would apply for five years from 1 January 2013. 74 When the AER released its Draft Decision in September 2012, the Draft Decision specified the timetable for consultation in respect of the decision and a revised proposal from APA GasNet. It was contemplated that, following the consultation, the AER would release its final decision in March 2013. The AER at that time observed that there would be an interval of delay; that the AER had taken into account the operation of rule 92(3) in fixing reference tariffs for the 2013-17 access arrangement period; and that the 2013 reference tariffs would take effect from 1 July 2013. 75 In the APA GasNet Access Arrangement Proposal Revisions of November 2012, APA GasNet disagreed with the AER's decision to apply rule 92(3) for the reasons given in the present Application for Review. APA GasNet argued that there was no interval of delay within the meaning of the rule because clause 2.3 of the 2008-12 Access Arrangement stated that the revision commencement date was the later of 1 January 2013 and the date on which approval of revisions to that access arrangement take effect. So the matter was squarely then raised. 76 In the Final Decision, the AER maintained its decision to apply rule 92(3) in respect of the period from 1 January to 30 June 2013. It observed: APA GasNet's 2008-2012 access arrangement provides no tariffs beyond 31 December 2012. However, the revisions to that access arrangement will not take effect until 1 July 2013. APA GasNet proposes that, in the interim, it will maintain its 2012 tariffs. The 2012 tariffs are higher than the 2013 tariffs. As a result, unless there is an adjustment to APA GasNet's 2013 tariffs, it will receive a revenue shortfall / windfall (sic). 77 The application of rule 92(3) in this matter resulted in the AER reducing the initial 2013 reference tariffs. This was because the reference tariffs determined under the 2013-17 access arrangement were lower than the tariffs prevailing at the end of the 2008-12 Access Arrangement. During the period 1 January to 30 June 2013, APA GasNet continued to charge the tariffs that were in force 31 December 2012 under the previous access arrangement period. Applying rule 92(3)(b), the AER took account of the higher tariffs charged in the period 1 January to 30 June 2013 in determining the tariffs to apply for the remainder of 2013. 78 It is the Tribunal's view that, at least in respect of the 2008-2012 Access Arrangement, there was no interval of delay upon which rule 92(3) could operate. 79 It is correct to say that generally under Division 8 of Part 9 of the NGR, the focus is on the conversion of total revenue under rule 76 into reference tariffs. Rule 92 within that Division is headed "Revenue Equalisation" and rule 92(2) stipulates the primary rule to give effect to that objective. It may well be that the AER as a matter of practice will require specification of a precise "revision commencement date" to best achieve that objective. It may also be arguable that, apart from the special circumstances applicable in this matter (by reason of the transition of the applicable access arrangements from the Gas Code to the NGL regime), the requirement of rule 92 is the imposition of a precise revision commencement date. It is not necessary to decide that issue. 80 The distinguishing features in this matter are first the wording of clause 2 of the 2008-2012 Access Arrangement quoted above, and second how the transitional provisions operate in relation to that Access Arrangement. 81 The AER in its submissions accepted that the literal meaning of clause 2 indicates that it was intended to avoid any interval of delay. But, it contends, such a construction is inconsistent with the meaning of the term "revision commencement date" in the Gas Code and in the NGR, and is inconsistent with other terms of the 2008-2012 Access Arrangement as approved by the ACCC. 82 For present purposes, the Tribunal is prepared to assume that the concept of a "revision commencement date" [a term not separately defined in the NGR] is a date fixed in the access arrangement as the date on which revisions resulting from a review of an access arrangement are intended to take effect. That is, under the NGR, it is assumed that the revision commencement date is a fixed future date. That would be consistent with the general purpose of rule 92(2) and would give appropriate scope for rule 92(3) to operate. It would facilitate the regulatory task of the AER if the previous access arrangement period does not have an uncertain duration. However, it is not by the policy or terms of the NGL that the proper scope and operation of the 2008-2012 Access Arrangement are to be assessed. That is obviously so because the NGL and the NGR were not, at the time of the 2008-2012 Access Arrangement, the applicable legislative or regulatory provisions. 83 Those provisions are found in the Gas Code as then in force. Under clause 3.17 of the Gas Code, an access arrangement was required to include a date upon which the next revisions to the access arrangement are intended to commence, defined as the "Revisions Commencement Date". There is a not insignificant difference between the definition of "revision commencement date for an applicable access arrangement" in rule 3 of the NGR and the definition in clause 3.17 of the Gas Code, because there is no use of the words a "date fixed" but the more general expression noted above. Moreover, as far as was identified in the submissions, there is no equivalent to rule 92(3) of the NGR in the Gas Code. 84 The fact that the submissions of the AER relied significantly upon the terms and effect of rule 92(3) to support the proposition that the "date fixed" must be a specified date, rather than a date capable of being fixed and determined with certainty by future events, is important. It signifies a regulatory recognition of the prospect that the revision commencement date under the NGR may be earlier than the date on which the revisions to an access arrangement come into operation. The potential hiatus is addressed by rule 92(3). The absence of such a provision under the Gas Code, where the risk of such a hiatus was no doubt equally capable of anticipation, provides both a sound practical reason why clause 2 of the 2008-2012 Access Arrangement should have been expressed as it was, and as well a sensible reason why clause 3.17 of the Gas Code should permit and accommodate the expression of a Revisions Commencement Date in such terms. 85 That would explain why the then regulator, the ACCC, was prepared to determine the 2008-2012 Access Arrangement in the terms in which it did, including clause 2. Such a means of expression, that is, expressing a commencement date by reference to the later of two occasions (for example as a specified date or the occurrence of a later event, such as the issuing of a particular approval) is commonplace in commercial transactions. 86 There is, in our view, no other provisions of the 2008-2012 Access Arrangement which indicate any intention on the part of APA GasNet or the ACCC that clause 2, expressing the Revisions Commencement Date as the later of two things (and in the present circumstances 1 July 2013), should for some reason be read down. Indeed, given the foreseeable risk that the setting of tariffs for the then anticipated next revision of the access arrangement might not occur by 1 January 2013, it was an appropriate and sensible step to address the potential hiatus. 87 Nor is there anything in the transitional provisions referred to above which would dictate that, by reason of the transition from the Gas Code regime to the NGL and NGR regime, the plain meaning of the words in clause 2 of the 2008-2012 Access Arrangement were or are to be read down as meaning 1 January 2013. The transitional provisions deem that document to be a full access arrangement under the NGL, but otherwise leave it to operate according to its terms. There is no reason to think that the transitional provisions were intended to, or did, alter its substantive terms. They were to establish its status as a starting point for the application of the new regime to it for the purposes of the next revision of the access terms. 88 The consequence, in the view of the Tribunal, is that there is no interval of delay upon which rule 92(3) of the NGR operated in respect of the period 1 January 2013 to 30 June 2013. 89 It follows that the Tribunal is satisfied on this issue that the Reviewable Decision involved an error of fact, namely that there was an interval of delay between 1 January 2013 and 30 June 2013 in respect of which it was required to make a decision in accordance with rule 92(3) of the NGR. It is also satisfied there was no proper basis for the adjustment to the tariff calculation to account for any over-recovery or under-recovery of tariffs during that period. 90 The AER accepted that, if the Tribunal took that view, it would be appropriate for the matter to be reconsidered by it. It is not necessary to address whether that error can or should be categorised as reviewable under any other provision of s 246 of the NGL. 91 The parties accept that, in the circumstances, it is appropriate for the Tribunal to publish its reasons for decision so that they may address the orders which should be made by the Tribunal. They may be able to identify the necessary amendments to the Reviewable Decision to give effect to the Tribunal's conclusion. 92 As noted above, that may also provide the parties with the opportunity to address the calculation errors which APA GasNet asserts and which, to some extent, the AER accepts. OPENING CAPITAL BASE DECISION 93 Under the NGR, the total revenue for each regulatory year of an access arrangement period is derived by applying a building block approach. One of the building block items is a return on the projected capital base for the year (rule 76(a)). 94 Under rule 78, the projected capital base comprises: (a) the opening capital base; plus: (b) forecast conforming capital expenditure for the period; less: (c) forecast depreciation for the period; and (d) the forecast value of pipeline assets to be disposed of in the course of the period. 95 The calculation of the opening capital base is addressed in rule 77. For present purposes rule 77(2) relevantly applies: If an access arrangement period follows immediately on the conclusion of a preceding access arrangement period, the opening capital base for the later access arrangement period is to be: (a) the opening capital base as at the commencement of the earlier access arrangement period (adjusted for any difference between estimated and actual capital expenditure included in that opening capital base); plus: (b) conforming capital expenditure made, or to be made, during the earlier access arrangement period; plus: (c) any amounts to be added to the capital base under rule 82, 84 or 86; less: (d) depreciation over the earlier access arrangement period (to be calculated in accordance with any relevant provisions of the access arrangement governing the calculation of depreciation for the purpose of establishing the opening capital base); and (e) redundant assets identified during the course of the earlier access arrangement period; and (f) the value of pipeline assets disposed of during the earlier access arrangement period. 96 Under rule 77(2)(a), the starting point in calculating the capital base for a current access arrangement period is the opening capital base at the commencement of the earlier access arrangement period, adjusted for any difference between estimated and actual capital expenditure included in that opening capital base. 97 The dispute between the parties ultimately relates to the nature of the adjustments that are permitted under rule 77(2) – specifically the adjustments permitted by the words in the parenthetical phrase in paragraph (a) of that rule. 98 APA GasNet's estimated capital expenditure for 2007 was $93.9 million (nominal). Its actual capital expenditure for 2007 was $72.4 million (nominal). However when the ACCC made its final decision in relation to the 2008-2012 Access Arrangement, the figure for actual capital expenditure for 2007 was not known. Therefore, the opening capital base specified for the 2008-2012 Access Arrangement included the $93.9m (nominal) in estimated capital expenditure for 2007. 99 After adjusting for inflation, the difference between estimated and actual capital expenditure for 2007 was approximately $20.0 million (nominal). As APA GasNet's total revenue under the 2008-2012 Access Arrangement included a return on the capital base (and thus a return on the estimated amount of 2007 capital expenditure), APA GasNet obtained a return over the five years of this access arrangement period on this $20 million of approximately $13.2 million ($2012). 100 Under Rule 77(2)(a), in specifying the opening capital base for the 2013-2017 Access Arrangement, APA GasNet took the opening capital base for the 2008-2012 Access Arrangement and adjusted for the difference between estimated and actual capital expenditure by removing the figure of $20 million. The AER agrees with this adjustment. 101 However, the AER says that there should be a further adjustment of $13.2 million to remove the return earned on the $20 million. The AER explained its reasons for making that adjustment in its Draft Decision as follows: While APA GasNet reduced its opening capital base by about $20 million ($nominal) to account for the difference between estimated and actual capex in 2007, it did not remove the benefit of the return on capital associated with this difference from the capital base. The AER considers that this benefit should be removed in these circumstances because allowing the difference would create an incentive for APA GasNet to overestimate its capex for the final year of the access arrangement period. Table 2.6 shows that the return on this difference that was earned by APA GasNet over the 2008–2012 access arrangement period is $13.2 million ($nominal). The AER therefore amends APA GasNet's proposed RFM to remove the return on the difference between estimated and actual net capex for 2007 from the capital base. This adjustment removes the benefit APA GasNet received by applying the rate of return to the estimated capex instead of the lower actual capex which APA GasNet incurred. The AER's decision to remove the rate of return from APA GasNet's capital base is consistent with the approach adopted by the AER in previous gas access arrangement decisions. This is also consistent with the Australian Competition Tribunal's decision upholding the AER's removal of the return on difference between estimated and actual capex from Jemena Gas Networks' capital base. (Draft Decision, Part 2, Attachments, pp 18 – 19) 102 The AER received e-mails from APA GasNet raising an alternative argument regarding the calculation of the proposed adjustment to the opening capital base. For reasons the Tribunal discusses later, the AER did not consider those arguments. Rather, in its Final Decision the AER maintained its view that the capital base should be adjusted to remove the return of $13.2 million. The AER said: The adjustment prevents APA GasNet from gaining/losing from any difference between estimated and actual capex for the final year of an access arrangement period. This means APA GasNet has no incentive to overestimate capex for that final year, or to defer efficient expenditure. Conversely, the adjustment does not impose additional penalties on APA GasNet if its actual final year expenditure exceeds its estimate. … The AER considers its proposed revision will result in an appropriate balance to encourage efficient investment in APA GasNet's network. It will do so by removing the incentive to overestimate or defer efficient expenditure during the final year when an access arrangement review is occurring. As a result the AER considers that this will promote the long term interests of consumers of natural gas with particular respect to price. (Final Decision, Part 2: Attachments, pp 27 – 28) 103 The Draft Decision and Final Decision show that the AER reached its conclusion on broadly two grounds. First, that its approach removes any incentive for a service provider to overestimate capex in the final year of an access arrangement period. This promotes the long-term interests of consumers of natural gas with respect to price, which is part of the NGO. 104 Second, that its conclusion is consistent with a similar approach taken by a differently constituted Tribunal in Application by Jemena Gas Networks (NSW) Limited (No 3) [2011] ACompT 6 (Jemena). 105 APA GasNet argues that the AER made a number of errors. In summary, that rule 77(2) makes express and exhaustive provision for the matters to be included in the opening capital base. It does not provide for the further adjustment made by the AER. Also, the AER did not have regard to the fact that the relevant capital expenditure was (except for a minor amount) actually incurred between 2008 and 2012. APA GasNet also maintains that if the AER was (contrary to the terms of rule 77(2)) entitled to make the adjustment, then the $13.2 million is excessive. Does the NGR provide for the further adjustment? 106 In Its Final Decision, the AER observed that there was a difference between the NGR (which contains no express provision dealing with the relevant adjustment) and the National Electricity Rules (NER), which contains an explicit provision permitting an adjustment in the following terms: (3) The previous value of the regulatory asset base must be adjusted for the difference between: (i) the estimated capital expenditure for any part of a previous regulatory control period where that estimated capital expenditure has been included in that value; and (a) (ii) the actual capital expenditure for that part of the previous regulatory control period. This adjustment must also remove any benefit or penalty associated with any difference between the estimated and actual capital expenditure. (Schedule 6.2.1(e)(3) to the NER) 107 The AER in its Final Decision acknowledged that rule 77(2)(a) of the NGR does not explicitly specify this adjustment in the same manner as the NER. Nevertheless it considered there is nothing prohibiting the adjustment being made. It also said that there is no sound economic reason why capital base approaches and incentives regarding final year capex estimates should be applied differently to gas and electricity service providers. 108 At least part of the AER's rationale for permitting the adjustment in the case of gas is that it promotes the NGO which looks at the efficient investment and operation of natural gas services for the long-term interests of consumers of natural gas. The rule is therefore directed at addressing the economic cost that may be imposed on service providers or on consumers by reason of a difference between estimated and actual capex, including in the opening capital base. 109 APA GasNet argues that the adjustment removes the incentive for a service provider to achieve efficiencies in capital expenditure. For example, consider a service provider that estimates a project will cost $100m in the final year of the access arrangement period, but then manages to deliver the project for $80m in the new access arrangement period. Under the current form of the NGR the benefits of this $20m saving will be shared between consumers and the service provider. By contrast, the approach adopted by the AER will not provide any benefit to the service provider in such a circumstance, and there is no incentive for the service provider not to spend the full $100m, which would be to the detriment of consumers. 110 As the AER explained in the Final Decision, it considers that the adjustment under rule 77(2)(a) has no effect on the incentive to achieve efficiencies in capital expenditure. For instance, during the 2002-2007 access arrangement period, APA GasNet earned revenue based upon a capital base that included a forecast of capital expenditure to be incurred during that period. The forecast of capital expenditure was made prior to the commencement of that access arrangement period for each year of that period, including 2007. If APA GasNet were able to reduce the forecast capital expenditure or defer it, the revenues it earned (based on the forecast) exceeded the costs actually incurred. That framework creates an incentive to achieve efficiencies in capital expenditure, including in 2007. 111 Rather rule 77(2)(a) is said to be directed to a narrower issue – the fact that the opening capital base of an access arrangement period may require estimation of the capex in the final year of the previous period. This would typically occur because the modelling and other work in finalising an access arrangement would need to be complete before final year capex actual figures are known. 112 The AER also justified its decision on the basis that the adjustment removed the incentive for a service provider to overestimate the final year capex for inclusion in the opening capital base of the next access arrangement period. The incentive is said to exist because the service provider earns a return on the amount of the opening capital base during that next access arrangement period. 113 In its Final Decision, the AER suggests that the incentive for a service provider to over-estimate capex results from an information asymmetry, because a service provider has more available information to estimate final year capex than the AER. 114 APA GasNet suggests that the potential for a service provider to over-inflate its estimates is limited by the fact that these estimates are carefully reviewed by the AER and the AER may also require it to be supported by a statutory declaration, as was the case here. Even though APA GasNet signed a statutory declaration to verify its estimates, the AER says they are uncertain by nature. The uncertainty may cause APA GasNet to choose the higher end from a possible range of values. 115 These arguments were advanced, not as an answer of themselves, but rather in support of a particular construction of rule 77(2)(a). In the case of the AER those arguments point to an interpretation of the rule that supports the adjustment it made. For APA GasNet, those arguments favour an interpretation that does not permit the adjustment. That indicates that there are policy considerations supporting either argument. 116 Ultimately whether the adjustment made by the AER was permitted by rule 77(2)(a) is a matter of construction. The Tribunal accepts that ordinarily, assessing the impact of a particular interpretation may (together with other matters) shed some light on the provision. However in that task, the Tribunal here gains little value from the suggestion that the rule was designed to encourage (or discourage) efficient capex or provides an incentive (or disincentive) to overinflate capex estimates. This is because the design of a regime with or without those features is ultimately a matter for policy makers. The Tribunal can conceive a regime that may or may not include some or all of those features. Were the Tribunal to favour an interpretation of the rule merely because it fostered (or did not foster) those matters, it would be undertaking a role which is properly a matter for policy makers. We use the expression "policy makers" in recognition of the fact that the design of the gas regime involves the input of many jurisdictions, under the co-operative arrangements that underpin that regime. 117 Leaving aside its "policy" arguments, APA GasNet maintains that as a matter of construction the language of rule 77(2)(a) indicates that the list of matters there is exhaustive and does not permit the adjustment made by the AER. The rule sets out the process of addition and subtraction that must be followed. The only relevant adjustment concerns the $20 million, not the further $13.2 million. 118 As indicated, the AER in its Final Decision found support for its adjustment in the Tribunal's decision in Jemena. APA GasNet rejects the Tribunal's approach in Jemena on a number of grounds. 119 First, that the circumstances in which a Court or Tribunal may add missing words to legislation are extremely limited. APA GasNet referred the Tribunal to Cooper Brookes (Wollongong) Pty Ltd v Federal Commission of Taxation (1981) 147 CLR 297 in support of the general principle that if, when reading the provision "as part of the whole instrument, its meaning is clear and unambiguous, generally speaking 'nothing remains but to give effect to the unqualified words'": Gibbs CJ at 304. 120 It is suggested that there is no ambiguity in the present case. In order to depart from the plain meaning of the words of the rules in NGR, it must be clear beyond question both that the literal meaning does not give effect to the intention of the legislature and that some departure from the literal meaning will fulfil that intent (Stephen J at 310). Alternatively it must be shown that the plain meaning of the words results in an operation that "is capricious and irrational", such that it can be concluded that the drafter "could not have intended such an operation and that an alternative interpretation must be preferred" (Mason and Wilson JJ at 321). 121 That is said not to be the case here because rule 77 clearly sets out a formula for calculating the opening capital base and there is no basis upon which its operation could be said to be capricious or irrational. Nor is there any basis to conclude that a different interpretation, in which additional powers were vested in the AER, must have been intended. 122 Counsel for APA GasNet submitted that filling in the gaps, when there is no capriciousness or irrationality in the plain meaning of the words ( in order to rectify perceived shortcomings in an otherwise clear statutory scheme), runs the risks alluded to by Spigelman CJ in R v Young (1999) 46 NSWLR 681 at 686 (Abadee and Barr JJ agreeing): The proposition that a court can introduce words into an Act of Parliament offends a fundamental principle of our constitutional law. It is no part of the function of any judge to amend legislation. The task of the courts is to determine what Parliament meant by the words it used, not to determine what Parliament intended to say … . In order to construe the words actually used by Parliament, it is sometimes necessary to give them an effect as if they contained additional words. This is not, however, to introduce words into the Act. This involves the construction of the words actually used. Judicial statements which appear to have been prepared to countenance something more than this, should be so understood. 123 APA GasNet referred to Marshall v Watson (1972) 124 CLR 640 where Stephen J at 649 (with Menzies J concurring) observed that: Granted there may seem to be lacking in the legislation powers which it might be thought the Legislature would have done well to include, it is no power of the judicial function to fill gaps disclosed in legislation; as Lord Simonds said in Magor and St Mellons R.D.C. v. Newport Corporation, 'If a gap is disclosed, the remedy lies in an amending Act' and not in a 'usurpation of the legislative function under the thin disguise of interpretation'.(at 646 and 649, noting also at 644). 124 The same view was adopted in Parramatta City Council v Brickworks Ltd (1972) 128 CLR 1. Gibbs J, with whom all other members of the High Court relevantly agreed, said at 12: "if the words of the sub-section are unambiguous we must give effect to the intention which they reveal and it is for the legislature and not for the courts to fill any gap that may unintentionally have been left in the statute." 125 In Jemena the Tribunal referred to the passage from Jones v Wrotham Park Settled Estates [1980] AC 74 in which Lord Diplock said (at 105-106): First, it was possible to determine from a consideration of the provisions of the Act read as a whole precisely what the mischief was that it was the purpose of the Act to remedy; secondly, it was apparent that the draftsman and Parliament had by inadvertence overlooked, and so omitted to deal with, an eventuality that required to be dealt with if the purpose of the Act was to be achieved; and thirdly, it was possible to state with certainty what were the additional words that would have been inserted by the draftsman and approved by Parliament had their attention been drawn to the omission before the Bill passed into law. Unless this third condition is fulfilled any attempt by a court of justice to repair the omission in the Act cannot be justified as an exercise of its jurisdiction to determine what is the meaning of a written law which Parliament has passed. Such an attempt crosses the boundary between construction and legislation. It becomes a usurpation of a function which under the constitution of this country is vested in the legislature to the exclusion of the courts. 126 APA GasNet maintains that this statement referred to in Jemena does not directly refer to the high standard required before it can be concluded that a matter had been inadvertently overlooked to justify reading the statute other than according to its plain terms. Nor is it said to be compatible with Parramatta City Council v Brickworks Ltd above. APA GasNet suggests that it is not apparent that any of the conditions referred to by Lord Diplock are satisfied in the present case. 127 APA GasNet also rejects the view expressed in Jemena that a similar approach was intended for both electricity and gas regulation. Indeed it suggests that the opposite may be true as there are many examples of provisions which differ between the two regimes. 128 In Jemena, the Tribunal said that for the AER to adjust the capital base (in that case) to remove the rate of return, it must show that there is a gap in the rules which the Tribunal is permitted to fill. 129 Counsel for the AER said the Tribunal's language in Jemena of 'filling the gap' should be understood as merely an exercise in construction. As Spigelman CJ said in R v Young at 688: The construction reached in this way will often be more clearly expressed by way of the addition of words to the words actually used in the legislation. The references in the authorities to the court `supplying omitted words' should be understood as a means of expressing the court's conclusion with clarity, rather than as a description of the actual reasoning process which the court has conducted. In all cases, what the court has done is to construe the words actually used in their total context. When the authorities are so understood, the additional words proposed in the present case are plainly impermissible. 130 Under the AER's approach additional words in the NER are not required in order for the adjustment to be made. This is because, on the AER's construction, the preposition "for" in Rule 77 (2)(a) is broader than the preposition "by" and does not lead to a mechanical operation requiring that the adjustment be limited merely to the subtraction of the difference between actual and estimated capex. That is, the AER says that those words are sufficient in themselves to permit the AER to deduct the return on unspent capital. 131 In the AER's view, although acknowledging the concerns of Spigelman CJ in R v Young, it is nevertheless sometimes necessary to give a provision effect, as if it contained additional words. This is not introducing words into the Act – rather it involves the construction of the words actually used. 132 It is clear that there are many differences in the drafting of the NGR and the NER. That those differences exit is not disputed. In light of that it is perhaps hardly surprising that there are also differences in the provisions between the two regimes in the words used (or omitted) when dealing with the treatment of the return on forecast and estimated capex. As the Tribunal has explained at [116] and [117], it is not helpful here to reach a position of construction by merely surmising the policy intent of the drafters. That would lead the Tribunal into the very area about which Spigelman CJ cautioned in R v Young at 686. 133 The Tribunal does not accept that the word "for" in the parenthesised clause "adjusted for any difference between estimated and actual capital expenditure included in that opening capital base" has the extensive scope of operation contended for by the AER. 134 The adjustment is between: (1) the estimated capital expenditure included in the previous opening capital base; and (2) the actual capital expenditure included in the previous opening capital base. 135 That is readily understood. The previous opening capital base must be fixed by the regulator in a process leading up to the end of that regulatory period. To the extent that, in that "lead up" period, there is ongoing capital expenditure, necessarily the expenditure to the end of that period must be estimated as the final amount cannot be quantified until a time after an access arrangement decision is finalised. It is accepted that the estimate may not prove to be accurate. The actual expenditure in that "lead up" period will only be certain after the expiration of that period. The Tribunal was informed that the AER has processes by which it aims to be satisfied about the reliability of the estimate. 136 It is obviously sensible for any difference between the estimated and actual capital expenditure to be reflected in an adjustment to the opening capital base for the next regulatory period. 137 The question is whether the word "for" was intended also to require the adjustment to reflect the extent to which the access provider has benefited from the return earned during the previous regulatory period. The AER took the view that it did. 138 The following contextual and textual matters point to the opposite conclusion. (1) The concepts being addressed in Div 4 relate to capital rather than revenue. Even where the AER makes a determination under rule 80 (in relation to proposed capital expenditure by a service provider), the prevailing tariffs do not immediately alter to reflect expenditure that may have been determined by the AER. Rather the determination operates as a ruling that the expenditure will meet the new capital expenditure criteria. (2) The determination of access tariffs for a regulatory period does not respond directly to the time when capital expenditure (whether estimated: rule 77(2)(9) or forecast: rule 78(b)) is incurred, so there is no special reason to think that the use of the word "for" is intended to operate to have that effect; (3) The estimated capital expenditure not actually incurred by the commencement of the next access period may well be spent soon after its commencement (there may be good reasons for the expenditure being delayed), and there is no obvious reason to think that in that circumstance it was intended that the tariff setting and so the return on capital should be significantly reduced retrospectively at the commencement of the next regulatory period. (4) The structure of rule 77(2)(a) itself is focused on determining the opening capital base for the previous regulatory period, and the difference specified is between two capital figures. (5) (i) the parenthesised clause could have been expressed, as a matter of grammar in several ways: as it stands, using "for", or using "by", or by using an active form of the verb "adjust" such as "by adjusting the difference"; (ii) if the intention were to have required the pullback of the extra return on capital as the AER proposes, that intention could have been readily expressed (by some formula such as that in the NEL). (6) the use of the word "for" is a grammatically neutral term, in the sense that it is capable of requiring either the simpler adjustment contended for by APA GasNet or the more complex adjustment contended for by the AER: there is a range of alternate dictionary options to illustrate that (Macquarie Dictionary, 5ed, 649; SOED, 4ed 1993, Vol 1 996-997), so there is no textual reason to think that the word used of itself has the particular meaning contended for by the AER. 139 It is part of the efficiency incentive that the projected capital base for an access period should include forecast conforming capital expenditure, so the more expansive meaning of the simple word "for" does not have contextual support. 140 In Jemena, the relevant discussion is at [53]-[55], after reciting briefly the respective contentions of the access provider and the AER. Resolution of the dispute was approached in a different way from that presented on this application on behalf of the AER. It was described at [53] as whether the AER had shown that there was a gap in the rules which the Tribunal was permitted to fill. The Tribunal proceeded on the basis that that gap should be filled because it was intended that there was intended "to be consistency and commonality between electricity and gas regulation" at [54], and because the AER's approach "minimises the incentive to overestimate or underestimate the amount of capital expenditure". The Tribunal departs from that conclusion, but with great caution. For reasons which are apparent, it does not think that the rules under the NEL and under the NGL are so similar that their content should routinely be treated as identical whatever different wording is used. Moreover, as noted, senior counsel for the AER in this matter did not put the AER's contention as requiring the filling of a legislative gap. The Tribunal also does not think that in this particular respect at least it should "fill a gap"; it could be only in the clearest of circumstances that that should be done. The NGR show how the legislative arm of government expresses how, in this respect, the NGO is to be measured and satisfied. So, as was put to the Tribunal, the resolution of the issue depends on what work is given to the word "for" in rule 77(2)(a). 141 For the reasons given, the Tribunal does not consider that it carries the meaning contended for by the AER. As was said in Jemena in [56], if that were intended, it would be desirable for the rules to be amended to expressly provide such an adjustment. 142 In the Tribunal's view, the AER was not entitled to make the adjustment of $13.2 million to the opening capital base for the previous access period under rule 77(2)(a). In doing so, its error is properly categorised as an error of fact (the extent of the difference authorised by the adjustment required by rule 77(2)(a)) or as an unreasonable decision because it reflects an adjustment not authorised by that rule or as a wrong exercise of discretion for the same reason. 143 The parties are agreed that, if the Tribunal takes that view, they should have an opportunity to formulate the appropriate orders to give effect to it. Was there a miscalculation of the adjustment? 144 APA GasNet argues that even were the AER entitled to make the adjustment, the amount of $13.2 million was overstated. In view of the conclusion reached by the Tribunal that the AER was not entitled to make the adjustment, the issue does not strictly arise. However as this was the subject of submissions, it may be helpful if the Tribunal provided its views. That may also be a prudent course to the extent it may have any bearing on the orders to be formulated by the parties or in giving practical effect to them, once formulated. 145 APA GasNet's argument is that the AER's adjustment was overstated because the majority of projects which accounted for the difference between estimated and actual expenditure in 2007 were completed or substantially completed during the currency of the 2008-2013 Access Arrangement. By removing from the opening capital base the entirety of the return on the difference the AER has removed much more than the "benefit" associated with this difference. 146 For example, APA GasNet had estimated an amount of $54.3 million ($2012) in 2007 for a project involving augmentation of the Brooklyn Lara Loop. In 2007, the expenditure on this project was $46.4 million ($2012). Relative to the estimated amount, APA GasNet had "underspent" by $7.9 million. It had forecast an amount of $16.8 million ($2012) on the Brooklyn Lara Loop project in 2008. However, in 2008, it actually spent $18.6 million (a difference of $1.8 million relative to the forecast amount), with $2.4 million spent in 2009, $2.5 million in 2011 and a further $0.6 million in 2012 ($2012). Therefore, the majority of the 2007 "underspend" amount was actually incurred in 2008 and 2009. The figure of $13.2 million therefore significantly overstates the benefit obtained by APA GasNet from the postponement of the expenditure, and indeed penalises it for postponing the relevant capital expenditure. 147 APA GasNet says that a proper calculation of the "benefit" obtained would be to calculate the return earned on a project included in the estimated 2007 expenditure allowance in the interval between 2007 and when the project (or its equivalent) is undertaken. For example, if an item of the estimated 2007 expenditure which was not expended in 2007 was expended in 2008, then there would be 1 year of returns which did not reflect actual expenditure, not 5½ years of such returns, which is assumed under the approach adopted by the AER. 148 The AER's response is twofold. First, it says that the arguments raised by APA GasNet were not raised in submissions and APA GasNet is precluded by s 258 (2) of the NGL from raising them. Second that the adjustments proposed by APA GasNet cannot practically be applied under the regulatory framework. 149 The AER stated its intention to make the adjustment under rule 77(2)(a) in its Draft Decision of September 2012. Under rules 59(5)(c)(ii)and (iii), the Draft Decision specified the date of 9 November 2012 as the date by which APA GasNet was to submit a revised access arrangement proposal, and the date of 7 January 2013 as the date by which interested parties were to make submissions in response to the Draft Decision and the revised proposal. 150 The AER says that in its access arrangement revised proposal APA GasNet did not advance any arguments concerning the calculation of the amount of the adjustment. It was not until 21 February 2013 that it raised the argument concerning the calculation of the adjustment. It was then that the AER advised APA GasNet that it was not practicable for the AER to consider the new information. It relied on s65 of the NGL in electing not to consider that material. Section 65 permits (but does not require) the AER to consider a submission it receives after the period specified in a notice inviting submissions. 151 Section 258(2) provides: A party (other than the original decision maker) to a review under this Division may not raise any matter that was not raised in submissions in relation to the reviewable regulatory decision before that decision was made. 152 The AER suggests that the words in s 258(2) "submissions in relation to the reviewable regulatory decision" should be construed as submissions received within the timeframes in Division 8 of Part 8 of the NGR. That Division prescribes the procedure for dealing with a full access arrangement proposal. If the provision were not construed in this way it is argued that service providers would be entitled to make submissions at any time prior to the publication of the AER's access arrangement proposal. That would undermine the regulatory framework. 153 The construction of s 258(2) was considered by a differently constituted Tribunal in DBNGP (No 3). The construction point arose in the context there of the approach to determining bond yields. As the Tribunal said at [294], there are two possible extreme positions of construction, neither of which are ideal: At one extreme, one might take the view that s 258(2) is intended to confine the review to the particular precise and confined argument put by an applicant to the regulator on a particular topic. At the opposite extreme, one might take the view that s 258 enables an applicant to raise entirely fresh and unconsidered grounds to review a decision on a topic provided the applicant put any argument on the topic, even if it did not pursue that argument. It becomes a question of a sensible constructional choice: neither of those extremes is likely to have been intended by the relevant legislature. 154 The expression "matter" in s 258(2) has been considered in a number of other contexts including under ss 75-77 of the Constitution where the jurisdiction of the High Court and other Courts exercising federal jurisdiction is expressed by reference to the word "matter". In that context it has been explained to mean the subject matter for determination in a proceeding, where there is an immediate right, duty or liability to be established or determined: DBNGP (No 3) at [298] citing Re Judiciary and Navigation Acts (1921) 29 CLR 257 at 265; Philip Morris Inc v Adam P Brown Male Fashions Pty Ltd (1981) 148 CLR 457 at 507-509 and Fencott v Muller (1983) 152 CLR 570. 155 In DBNGP (No 3) the Tribunal acknowledged at [299] that it would not be appropriate simply to adopt that expansive meaning. However that reasoning provides some guidance: But it may be taken to indicate that "matter" in s 258(1) and (2) extends to the controversy or thing in dispute, so long as it is further confined (in the case of the regulator) to a controversy or thing in dispute that relates to a ground of review or relates to a matter raised in support of a ground of review, and (in the case of an applicant) to a controversy or thing in dispute that was raised in submissions by the applicant. 156 The Tribunal also referred to the approach adopted in Epic Energy, which concerned the Gas Pipelines Access (South Australia) Act 1997 (SA). Section 39(2)(b) is in terms relevantly the same as s 258(2). In interpreting that provision the Tribunal said at [24]: Section 39(2)(b) limits the matters to which recourse may be had to those that may be identified in the submissions which, in fact, were made prior to decision. The matters include the subject matters raised, the issues raised and the materials relied upon in support of the position or proposal put forward in the submission as being relevant to the decision being made. Thus, if any matter, whether by way of argument or evidentiary material, cannot be identified as broadly arising out of a matter fairly raised in the submissions to the relevant Regulator before the decision under review was made, it will not be permitted to be raised in the review. This is not to say that a reformulation of an argument or contention previously put to the relevant Regulator on material which was before it before the decision was made would be excluded. 157 It is not necessary to refer in more detail to the other decisions surveyed by the Tribunal that bear on the construction of s 258(2). It is sufficient to note that having undertaken that review the Tribunal in DBNGP (No 3) at [305] favoured the purposive approach noted at [34]. 158 The Tribunal sees merit in that approach. However in the view the Tribunal takes here, it is ultimately not necessary to reach a concluded view on whether that approach would have permitted APA GasNet to raise the issue of miscalculation errors. This is because, even if the construction of s 258(2) obliged the AER to consider the miscalculation issues, the Tribunal is satisfied that the regulatory regime did not require the AER to consider them in the manner submitted by APA GasNet. 159 The AER argues that the adjustment proposed by APA GasNet is incapable of practical implementation within the regulatory framework. APA GasNet's proposal would require the AER to identify projects that were the subject of the 2007 capex estimate and to determine the amounts of the estimate that were not expended in 2007. If those amounts were subsequently expended, the AER would also need to determine the amount and the timing of that expenditure. 160 That exercise may well be the consequence of APA GasNet's submission. However whether or not it is inconvenient or impractical as the AER suggests, is not an answer to whether it ought properly to be undertaken, if it is required under the regulatory regime. 161 Rather, in the Tribunal's view the reason that APA GasNet's arguments cannot succeed is that the regulatory regime does not require that they be considered as APA GasNet has proposed. To explain this view it is useful to briefly outline the operation of the regulatory framework with respect to the setting of tariffs. 162 Under the framework, tariffs for reference services are set to enable the service provider to earn sufficient revenue to cover the efficient cost of providing those services. The building block approach is used to determine the efficient level of costs to provide the reference services and therefore the amount of revenue required. 163 One of the components of the building block approach is the return on the projected capital base for the year (rule 76(a)). The Tribunal has already referred to rule 78 which specifies the components of the projected capital base. They include the opening capital base plus forecast conforming capital expenditure for the period. Under the regime those forecasts must be arrived at on a reasonable basis and represent the best forecast possible in the circumstances. The forecast capital expenditure is carefully assessed for its accuracy. This is because once that forecast capital expenditure is approved, the service provider earns a return on it for the access arrangement period. 164 That is, forecast capital expenditure is 'locked' for the duration of the approved access arrangement period. It is in the setting of the opening capital base for the subsequent period, that the regulator will, as required by rule 77(2), assess the amount of the conforming capital expenditure actually made during the earlier access arrangement period. However there is no requirement until this time for the regulator to undertake a detailed analysis matching the forecasts in the prior period with the actual figures. 165 It is possible to surmise why the regime is so designed. The process for the approval of an access arrangement takes a considerable amount of time. A large number of submissions and expert reports are received on many aspects of the building block approach, including on forecasts. Under such a regime there must be some finality to the process until the issues are again 'opened up' in a subsequent access arrangement period. Depreciation and Indexation Decision 166 Pursuant to rule 76 of the NGR, one of the building blocks used to determine total allowable revenue for each regulatory year of the access arrangement period is the depreciation allowance. 167 The depreciation allowance may be calculated using either indexed (real) or unindexed (nominal) values for the asset base. Whether or not asset values are indexed will affect the quantum of the depreciation allowance. 168 Where unindexed (nominal) asset values are used and a nominal rate of return is applied to the asset base to calculate the return on capital, the depreciation allowance for each asset class is simply calculated as the nominal asset value divided by the remaining asset life (the APA GasNet approach). 169 However, where indexed (real) asset values are used and a nominal rate of return is applied to the asset base to calculate the return on capital, the depreciation allowance for each asset class is calculated by reference to: (a) the real asset value (i.e. after upwards inflation adjustment) divided by the remaining asset life; less (b) the amount of inflation of the asset value for that year (the AER approach), 170 The adjustment in (b) is necessary because the application of a nominal rate of return to an inflation adjusted asset base will lead to double compensation for inflation unless the amount of inflation is deducted from the depreciation allowance. When compared to the APA GasNet approach, the AER approach leads to a lower depreciation allowance in early years, a higher depreciation allowance in later years, and a higher figure for return on capital throughout. 171 A third approach is to use real asset values and a real rate of return. This is the approach adopted by regulators in the UK. Under this approach, there is no need to make the additional inflation adjustment to the depreciation amount. This UK approach, when compared to both the APA GasNet approach and the AER approach, leads to a higher figure for depreciation (because it is the indexed asset value divided by the asset life, but without a downwards adjustment for inflation), and a lower figure for the return on capital. 172 Over the life of the capital base, all three approaches will deliver the same revenue outcome in NPV terms. The only difference lies in the profile of revenue recovery over the life of the capital base. 173 Depreciation is dealt with in rule 88 and following of the NGR. Rule 88 requires that the depreciation schedule sets out the basis on which the pipeline assets constituting the capital base are to be depreciated for the purpose of determining a reference tariff. 174 Rule 89 deals with the content of the depreciation schedule and relevantly provides that: (1) The depreciation schedule should be designed: (a) so that reference tariffs will vary, over time, in a way that promotes efficient growth in the market for reference services; and (b) so that each asset or group of assets is depreciated over the economic life of that asset or group of assets; and (c) so as to allow, as far as reasonably practicable, for adjustment reflecting changes in the expected economic life of a particular asset, or a particular group of assets; and (d) so that (subject to the rules about capital redundancy), an asset is depreciated only once (ie that the amount by which the asset is depreciated over its economic life does not exceed the value of the asset at the time of its inclusion in the capital base (adjusted, if the accounting method approved by the AER permits, for inflation)); and (e) so as to allow for the service provider's reasonable needs for cash flow to meet financing, non-capital and other costs. (2) Compliance with subrule (1)(a) may involve deferral of a substantial proportion of the depreciation, particularly where: (a) the present market for pipeline services is relatively immature; and (b) the reference tariffs have been calculated on the assumption of significant market growth; and (c) the pipeline has been designed and constructed so as to accommodate future growth in demand. (3) The AER's discretion under this rule is limited. 175 It is clear from rule 89 that various methods of depreciation could potentially be used in the determination of reference tariffs. There is no method specified as a default or standard approach in rule 89. Rather, all that is required is that any depreciation approach that is proposed satisfies the criteria in sub-rule (1). 176 An important feature of rule 89 is that it is a limited discretion rule: see rule 89(3). As provided in rule 40(2), the AER may not withhold its approval to an element of an access arrangement proposal governed by a limited discretion rule if it is satisfied that the element complies with applicable requirements of the NGL and NGR and is consistent with any applicable criteria prescribed by the NGL and NGR. This is so even if the AER concludes that an alternative proposal might better satisfy the applicable criteria or might produce a better outcome. Indeed, a depreciation schedule is the very example provided in the note to rule 40(2) explaining the operation of a limited discretion rule. 177 APA GasNet challenges the depreciation schedule determined by the AER. Specifically, APA GasNet seeks review of the AER's determination that, in calculating the depreciation, the projected capital case is to be indexed for inflation. The AER did not approve APA GasNet's proposal to calculate the depreciation allowance on a nominal value of the capital base (ie without inflation indexation). The effect of APA GasNet's proposal, in comparison to the depreciation methodology that had applied in the previous access arrangement period, was to bring forward revenues. In other words, APA GasNet would receive more revenue early in an asset's life, and less revenue later in the asset's life. That revenue profile would be reflected in the reference tariffs. 178 The AER formed the view that APA GasNet's proposal did not satisfy criterion (a) in rule 89(1) of the NGR which required the depreciation schedule to be designed so that reference tariffs will vary, over time, in a way that promotes efficient growth in the market for reference services. The AER also did not accept a submission advanced by APA GasNet that its depreciation proposal was required by criterion (e) to allow for its reasonable needs for cash flow. 179 APA GasNet contends that the AER's decision involved two types of error: (a) First, APA GasNet contends that the AER misconstrued its task in reviewing the proposed depreciation schedule. It contends that the AER's task was limited to assessing the proposed profile of depreciation and the proposed asset lives. APA GasNet also contends that the AER failed to apply the limited discretion required by rule 89(3) and erroneously engaged in an assessment of which approach to depreciation better satisfied the relevant criteria. (b) Secondly, APA GasNet contends that the AER's decision was based on insufficient evidence, a failure to adequately take into account evidence provided by APA GasNet or errors of fact. 180 In the APA GasNet Access Arrangement Proposal of March 2012, APA GasNet proposed a depreciation schedule in which the capital base was not adjusted for inflation beyond 31 December 2012. This was a change from the depreciation methodology that had applied in the previous access arrangement period. In its proposal, APA GasNet did not explain the reasons for its proposed approach. However, it noted that while the annual returns on and of capital in a particular year would differ between the two methodologies, the NPV of the returns, over the life of the assets would be the same under either approach, or on changing from one approach to the other. 181 Given certain assumptions, it is accepted that the depreciation methodology proposed by APA GasNet, and the depreciation methodology determined by the AER, would produce the same NPV of returns over the life of the assets. For present purposes, the material difference between the two methodologies lies in the profile of the returns (and therefore tariffs) over time. 182 Under the methodology determined by the AER (and which applied in the previous access arrangement period), the overall return (the aggregate of depreciation and return on capital) is relatively flat. The return on capital is, in relative terms, higher because the capital base is indexed for inflation; the depreciation allowance is, in relative terms, lower because it is reduced by the inflation of the capital base. 183 APA GasNet's proposed methodology results in a higher initial overall return which decreases substantially over time. This is brought about because the capital base reduces in value more sharply over time by the absence of indexation for inflation, and the lower capital base reduces the return on capital over time. 184 On 15 June 2012, the AER sent an information request to APA GasNet, requesting it to provide any reasons why it had proposed a change in depreciation methodology from the previous access arrangement period. On 20 June 2012, APA GasNet replied stating that its depreciation proposal complied with rules 73 and 88 and "therefore no further explanation is necessary". At a meeting between APA GasNet and AER on 12 July 2012, the AER sought further information from APA GasNet about the reasons for its proposed depreciation methodology and for the changed approach. In a subsequent email dated 31 July 2012, APA GasNet explained that, as a listed business, it has to manage its cash flow and as it expected that there will be reduction in the rate of return under the next access arrangement then under consideration, it was looking for legitimate means to maintain as much as possible of the GasNet cash flow. 185 In the Draft Decision, the AER did not accept APA GasNet's proposed methodology because it considered it did not meet the requirements of rule 89(1)(a) concerning the promotion of efficient growth of the market for reference services. The AER made the following observations: (a) compared to the current depreciation methodology, APA GasNet's proposed methodology brings forward cash flows for APA GasNet by requiring customers to pay a greater proportion of an asset's costs earlier in its life; (b) APA GasNet had not offered any explanation for its change of approach other than a desire to bring forward cash flows to off-set an anticipated reduction in the rate of return that would be allowed by the AER; and (c) APA GasNet's proposed methodology is likely to inhibit efficient growth of the market by reason of inefficient asset utilisation (a revenue profile having higher prices early in an asset's life is likely to unnecessarily discourage demand), unnecessary high prices in the short to medium term (which would be likely to discourage gas usage and downstream investment) and inefficient management of assets (faster depreciation may create incentives for APA GasNet to replace assets sooner than may otherwise be the case). 186 In its Access Arrangement Revised Proposal dated November 2012, APA GasNet did not accept the revisions to the depreciation methodology requested by the AER and maintained its original proposal. APA GasNet gave three reasons for maintaining its proposal: (a) there was no requirement for indexation of the capital base under the NGR; (b) APA GasNet believed that its approach provided for variation in reference tariffs over time in a way that promoted efficient growth in the market for reference services (and its approach was preferable to the AER's approach in that regard); and (c) the AER's approach did not satisfy rule 89(1)(e) and the adoption of APA GasNet's approach was necessary to meet APA GasNet's reasonable needs for cash flow to meet financing, non-capital and other costs. 187 In the Final Decision, the AER maintained its view that the depreciation methodology proposed by APA GasNet was inconsistent with the criteria in rule 89(1)(a). It formed the view that APA GasNet's proposed methodology leads to tariffs varying, over time, in a way that promotes inefficient growth in the market for reference services. In respect of the efficiency criterion, the AER reached three primary conclusions: (a) First, efficient growth in the market for reference services ordinarily requires variations in tariffs to reflect variations in costs in the short to medium term. In the present case, costs had fallen by reason of a reduction in the cost of capital and in the capital base (as a consequence of actual capital expenditure being less than forecast in the prior period). It was not efficient to use the depreciation methodology to fill in revenues that had been reduced to reflect changes in efficient costs so as to maintain existing tariffs in the short to medium term. (b) Second, the capacity concerns raised by APA GasNet were insufficient to justify, from an efficiency perspective, the proposed change of depreciation approach. The AER said in the Final Decision: "A change of depreciation approach is an inefficient response to the emerging areas of constraint. It would impact all tariffs and all customers regardless of where the constraints are emerging". (c) Third, the AER decided that its proposed depreciation methodology was consistent with rule 89(1)(e) as it allowed for APA GasNet's reasonable needs for cash flow to meet financing, non-capital and other costs. 188 The summary of the errors of the AER asserted by APA GasNet was that the AER made: (a) errors of fact in relation to the assessment of spare capacity on the VTS, the relevance of this to the choice of depreciation method, and the ability of APA GasNet to adjust tariffs; (b) an error of discretion, in that the decision was made on the basis of a misapprehension of the extent of the AER's discretion – specifically, the AER proceeded on the basis that it was entitled to compare the relative merits of APA GasNet's proposed depreciation method with its own preferred approach, an inquiry that is not permitted under a limited discretion rule; and (c) an unreasonable decision, in light of the absence of any basis for rejecting APA GasNet's proposed depreciation method. 189 It was agreed between the parties that, if error within the scope of s 246(1) was made out, the most efficient course would be for the parties to identify the consequential amendments to the 2013-2017 Access Arrangement to give effect to the Tribunal's decision which would then be incorporated in the Tribunal's determination varying the AER's decision. 190 Put at a more general level, APA GasNet says the AER's approach was to identify reasons at a theoretical level that supported what the AER described as the "standard approach" to depreciation, and that its approach was preferable to APA GasNet's approach. By reason of that approach, the AER did not properly engage with whether, in the particular circumstances of the present case, the APA GasNet approach failed to satisfy the criteria. 191 In the Final Decision, the AER stated that it considered that the "standard depreciation approach" (i.e. the approach adopted by the AER) "will generally lead to tariffs varying, over time, in a way that promotes efficient growth in the market for reference services. In most circumstances, this would imply that sunk costs are recovered as evenly as possible over an asset's life and that revenues (and tariffs) are relatively flat". The AER observed proximately that one situation in which a flat tariff profile is unlikely to lead to efficient growth in the market is "where capacity is reached and no augmentation occurs". The AER said that "in assessing whether tariffs should rise to reflect emerging capacity constraints, it is important to have an idea of the likely impact those constraints would have on costs". 192 As it was a common anticipation by the time of the Final Decision that tariffs were not expected to rise in the 2013-2017 period, APA GasNet says it was not relevant for the AER to consider whether and how tariffs should rise to reflect emerging capacity constraints. The only question was how rapidly tariffs should fall and in what pattern. 193 It is correct to say that the AER identified that capacity constraints are relevant to an assessment of a desirable tariff path. It went on to conclude that capacity concerns were insufficient to justify the proposed change of depreciation approach. It accepted that APA GasNet may have some emerging capacity constraints, but thought that constraints were localised and could be managed in various ways, including through tariff structures. It considered that a change in depreciation approach was an inefficient response to the emerging areas of constraint as it would impact all tariffs and all customers regardless of where the constraints were emerging. 194 APA GasNet argued that such an analysis by the AER illustrated that it had adopted an erroneous approach, because it reflects an assessment of selecting an outcome preferred by the AER rather than deciding whether the APA GasNet approach did not satisfy criterion (a). Because this decision was a limited discretion decision, the AER had first to be satisfied that APA GasNet's proposal was not consistent with criterion (a) before it could substitute an alternative depreciation prescription: see rule 40. It is not enough for the AER simply to decide that its proposed depreciation prescription might better satisfy the requirements of the NGL and in particular criterion (a) in rule 89(1). 195 APA GasNet says then that the AER did not conclude that there was something about the tariff path produced by the APA GasNet Proposal that did not promote efficient investment, but assumed that the "standard" approach to depreciation was the correct approach, so that any other approach would need some strong reason to justify the change, and then found that the APA GasNet Proposal could not be justified by capacity constraints because these were capable of being addressed by other means. This inverts the correct approach. 196 In terms of the words used, the AER did not misunderstand the nature of the decision it had to make. The AER specifically concluded in the Final Decision that the APA GasNet Proposal did not meet the requirements of the NGR which require tariffs to vary, over time, in a way that promotes the efficient growth in the market for reference services. It found that the APA GasNet Proposal would lead to tariffs varying over time in a way that promotes inefficient growth in the market for reference services, and explained why it had reached that conclusion. It said (at 99); Considering the modelling scenarios presented in APA GasNet's proposal, the AER's analyses show that APA GasNet's proposed approach would lead to substantially higher tariffs for customers over the next three access arrangement periods. This is despite the expectation of falling demand in the short run and relatively subdued demand over the medium term. The standard approach, which is consistent with APA GasNet's current approach, leads to tariffs tracking forecast cost changes over time. However, APA GasNet's proposed approach results in higher starting tariffs that decrease regardless of the direction of costs in the scenarios modelled. This does not send an appropriate signal for asset utilisation. These scenarios are discussed further below. Further, the AER considers that the change of approach is not an efficient response to emerging capacity constraints. The AER engaged Frontier Economics (Frontier) to review APA GasNet's analysis. Frontier found the claim of constraints to be overstated and that demand over the medium term is likely to be subdued. Where constraints are emerging, they are localised. A change of depreciation approach is not an efficient or effective response to such issues. APA GasNet should not be using the depreciation approach to increase its cash flows to offset falls in other building block costs (such as the return on capital). Its reasonable cash flow needs do not require the change of depreciation approach. Such short term objectives can also lead to cash flow problems in the future. 197 It is convenient to note that the final paragraph in the quoted passage above relates to the consideration of criterion (e) in rule 89(1). 198 APA GasNet's contention that the AER simply substituted its preferred view for that put forward by APA GasNet cannot succeed, at least on the face of things. It is, however, necessary to consider the particular contentions on matters of fact made by APA GasNet. 199 Having regard to the change by APA GasNet in the depreciation formula in the APA GasNet Access Arrangement Proposal from the depreciation formula used in the 2008-2012 Access Arrangement, it was not surprising that the AER sought the reasons for the change. It was also entitled to assess their significance as the AER was charged with the responsibility of deciding whether it was satisfied that the APA GasNet proposal did not satisfy criterion (a) in rule 79(1). 200 One of the principal challenges to the way in which the AER reached that level of satisfaction was an attack on the limited extent to which it placed weight upon capacity constraints. 201 It is said to have failed to consider the impact of lower tariffs as stimulators of demand, and in turn as putting pressure on capacity and then in turn requiring the availability of funds for further significant capital investment. 202 The evidence presented by APA GasNet was capable of showing that the current capacity of the VTS is only just sufficient to meet demand on a 1-in-20-year peak day. For 2013, peak day system utilisation is forecast to be 99% (based on the 1-in-20-year peak day demand forecast for 2013). It is evident that capital expenditure is required if it is necessary to expand the capacity of the VTS in order to meet future demand. It is projected that by 2017, demand will exceed system capacity on the 1-in-20-year peak day (i.e. there will be insufficient capacity to meet peak day demand). The AER had approved capital expenditure projects for the 2013-2017 access arrangement period which will have the effect of expanding the capacity of the system, so that it maintains sufficient capacity to meet 1-in-20-year peak day demand. 203 The most heavily utilised or "capacity constrained" zone of the VTS is the Northern Zone as that zone is fully utilised in the sense that there is no spare capacity) around one day every month. In addition, the South-West Pipeline is also fully utilised on winter peak days (around one to two days every year). Also, in 2013, utilisation of the Longford Zone on peak days was projected to be 96%, meaning that it has 4% spare capacity on these days. 204 One matter addressed in APA GasNet's material was that some customer demand was particularly price sensitive, and that falling tariffs would be likely to lead to increased demand. In that circumstance, where tariffs will fall significantly in the next period, under the AER's approach to depreciation (APA GasNet says) they will fall more suddenly. APA GasNet therefore argued that the AER's consideration of capacity constraints in the Final Decision did not take into account the likely response of price-sensitive customers to significant reductions in transmission tariffs. If demand is sensitive to changes in transmission tariffs, then a significant fall in such tariffs will see a significant increase in demand. If, by contrast, demand (either generally or from particular types of customers or in particular regions) is not sensitive to changes in tariffs, then the adoption of a particular depreciation path has no likely impact on relevant investment decisions to be made by the non-price sensitive customers, and APA GasNet could not be said to have failed to satisfy the relevant criterion. 205 The argument then proceeds as follows: In a situation of capacity constraints, a significant fall in tariffs is likely to lead to inefficiencies in the market for reference services. Demand will likely increase, giving rise to greater capacity constraints. If APA GasNet makes investment in the network to alleviate the capacity constraints caused by the increased demand, then this investment will be included in the capital base for later periods and tariffs will accordingly rise, thus throttling off demand and giving rise to the risk of under-utilised assets – assets wastefully constructed. On the other hand, if APA GasNet does not invest in expanded capacity, then capacity constraints will prevent APA GasNet from meeting customer demand at the lower prices, in circumstances where prices will be higher later under the AER approach than under the APA GasNet approach. 206 APA GasNet asserts that these matters were not considered by the AER in its Final Decision, so its approach to the impact of capacity constraints was therefore incorrect. 207 APA GasNet also asserts that the AER analysis of existing capacity constraints was flawed. The AER's expert, Frontier Economics, concluded that there existed "significant levels of spare capacity" in the VTS. 208 The basis for Frontier Economics' position on capacity constraints was identified as being: (a) AEMO has recently revised downwards its forecasts of 1-in-20-year peak day demand, from which Frontier Economics inferred that there must now be excess capacity in the VTS; and (b) Frontier Economics appears to disagree with AEMO's view of likely future capacity constraints. Despite AEMO's conclusion that constraints are likely to arise in parts of the VTS over the next five to 10 years, Frontier Economics said that in its view, the demand from gas-powered generation will not be as significant as projected by AEMO; and (c) Frontier Economics also queried whether the 1-in-20-year peak day demand standard applied by APA GasNet to determine capacity constraints was appropriate. 209 The question to which the evidence concerning capacity constraints was relevant is whether and to what extent capacity constraints on the VTS have increased, or will increase during the access arrangement period, so that the long-run marginal costs of reference services would make it efficient to increase tariffs. 210 Frontier Economics in discussing the economic relevance of capacity constraints to tariffs said: The LRMC of reference services provided by the VTS will vary according to the level of spare capacity on the system and the cost of augmentation required to meet an increment of demand growth. As these variables may vary across the VTS, the LRMC of reference services will also tend to vary to some extent on a locational basis. As noted above, APA GasNet's proposed approach to depreciation would likely increase regulatory depreciation in the near term and reduce it in later years relative to under the current real approach. Other things being equal, this would translate into higher reference tariffs in the near term and lower reference tariffs in the longer term than would otherwise be the case. In our view, whether such changes are appropriate and consistent with the requirements of the NGR fundamentally depend on: • The current average LRMC of reference services across the VTS: • Whether the average LRMC is likely to rise, fall or remain fairly steady over the access arrangement period and beyond. Both of these variables will depend in large part on the nature of existing and approaching capacity constraints and the costs of addressing those constraints relative to historical levels of capital expenditure on the VTS. 211 In the Tribunal's view, the AER was entitled to have regard to that evidence. In addition, it was entitled to identify that APA GasNet could contend that there is no material spare capacity on the VTS only in the sense that it has defined capacity. As the VTS has been designed to meet 1 in 20 year peak day demand, it follows that, when assessed against 1 in 20 year peak day demand, it will not have spare capacity. 212 It may also be accepted that that definition is a rational one for the purposes of AEMO as it is a quasi-regulatory requirement imposed by AEMO. The means by which that is effective was explained by APA GasNet in the course of its submissions, including by the Service Envelope Agreement between it and AEMO. However, the Tribunal agrees with the AER that adopting that measure of capacity means that it does not routinely follow that the VTS is experiencing capacity constraints so that the long-run marginal costs are currently high or increasing. The capacity measure advanced by APA GasNet has no obvious or necessary bearing on that question. The material adduced by APA GasNet did not show that the AER was in error in deciding that it was not satisfied that higher reference tariffs were required to address higher marginal costs associated with capacity constraints. 213 There was material available to the AER, in particular the views of Frontier Economics, to support the propositions that the average long run marginal costs of access to the VTS are relatively low at the present time, although it is likely to gradually rise from the early 2020s; that to the extent there are constraints in the Northern Zone resulting in higher long-run marginal costs, APA GasNet has scope to rebalance tariffs by charging higher reference tariffs to Northern Zone customers, and that if and when constraints on the VTS need to be addressed through augmentation, the costs and timing of such investment are not likely to have a large impact on the overall long-run marginal costs of the network. Hence, the AER's conclusion on all the material that it was not satisfied that marginal costs will rise substantially over the short to medium term, and that any significant increase in long-run marginal costs is likely to occur well into the future, were available to it. 214 As to the Northern Zone, however, Frontier Economics also observed that APA GasNet had scope to rebalance tariffs by charging higher reference tariffs to Northern Zone customers within the bounds of its overall revenue cap. In its Final Decision, the AER agreed with the opinion of Frontier Economics. The AER observed: A change of depreciation approach is a very blunt instrument to deal with localised constraints and will be an inefficient (and potentially ineffective) response to such concerns … refinements to the tariffs could have been proposed if capacity constraints were becoming a significant issue in particular locations. These increases in costs could have been reflected in the relevant customer tariffs, rather than raising overall revenue (and therefore all tariffs) through the change of depreciation approach". The Tribunal observes that not only is the change of depreciation approach not targeted at the area of constraint, it is a disproportionate response to the potential problem (in terms of the amount of revenue affected). 215 APA GasNet argued that the AER erred in that passage in relying on the ability to rebalance tariffs. There is very little scope to rebalance tariffs because rule 95 of the NGR requires that reference tariffs set at the commencement of each period reflect the embedded cost of providing transportation services to users or classes of users. Rule 95, however, allows the tariff for a particular reference service to be based on the costs directly attributable to that reference service and other costs on a basis approved by the AER, so the structure of tariffs can be altered to address capacity issues (in terms of the mix of capacity and volume charges). 216 It has not been necessary for the Tribunal to address in detail the analysis of the views of Frontier Economics, and the criticisms of those views, put forward by APA GasNet. Section 246 does not entitle the Tribunal to review all the material before the AER simply for the purpose of substituting its view of the factual findings it might have made for those of the AER: see DBNGP (No 3) at [11]. 217 There is substantial agreement about what is required in terms of tariff paths to promote efficient growth in the market for reference services. 218 The economic experts for both the AER and APA GasNet (PwC for APA GasNet and Frontier Economics for the AER) generally agreed that, subject to tariffs reflecting long-run marginal cost, recovery of any remaining costs should be so as to minimise distortion of demand. PwC states that efficient pricing entails ensuring that the marginal cost of consumption is signalled to consumers. To the extent that there are non-marginal (fixed) costs of supply, then these should be spread across consumers in a way that minimises distortion of consumption decisions. PwC states that "in other words, efficient pricing… essentially entails devising tariffs that minimise the demand distortion that results from having to recover non-marginal costs through increases in price above marginal cost". Frontier Economics similarly states that: "subject to reference tariffs reflecting the average LRMC of system usage, any remaining regulated revenues should be recovered in a way that minimises the impact on the demand for reference services". 219 APA GasNet says that, therefore, a tariff path which involved a sudden increase (or decrease) in tariffs would be more likely to promote efficient growth in the market if there were a sudden increase (or decrease) in long-run marginal cost. It therefore contends that because the AER's approach will deliver a more immediate and dramatic reduction in tariffs, it can only have done so on the basis that there has been a very dramatic reduction in long-run marginal cost. Otherwise, its view is erroneous, and its conclusion that APA GasNet's proposed depreciation schedule did not satisfy criterion (a) of Rule 89 lacked any reasoned basis. 220 With regard to the stability of the tariff path resulting from APA GasNet's proposed methodology, APA GasNet's November 2012 submission was based upon a tariff path presented in nominal dollar terms. In its Final Decision, the AER concluded that the correct economic approach was to consider the tariff path in real dollar terms. This view was supported by the AER's consultant, Frontier Economics and by APA GasNet's consultant, PwC. The resulting tariff path was set out in the Final Decision. In respect of the tariff path scenarios modelled by APA GasNet, the AER concluded: Regardless of the scenario chosen as the best representation of the future costs, real tariffs decrease under APA GasNet's approach over the entire 20 years modelled, whereas real costs are either flat or increasing in those scenarios. In contrast, the standard approach (after an initial reduction in 2013 due to other cost reductions) leads to flat or increasing real tariffs consistent with all the scenarios presented and the cost trends assumed for each of those scenarios. 221 Then, in respect of the tariff path that results from the AER's approved depreciation methodology, the AER says that it is not correct that there is no evidence that long-run marginal cost has fallen to a large extent. The reduction in tariffs in 2013 was caused by reductions in APA GasNet's costs (as assessed within the building block methodology). The evidence shows that the cost of capital had declined and the value of APA GasNet's capital base had declined from the previous access arrangement period because its actual capital expenditure was substantially less than had been forecast for the previous period. In reference to the reductions in APA GasNet's costs, the AER concluded that APA GasNet's depreciation methodology did not promote efficient growth in the market for reference services because it had the effect of insulating customers from the cost reductions (by bringing forward cash flows in the short to medium term). 222 In the light of those considerations, the Tribunal is of the view that no reviewable error on the part of the AER has been made out. The AER directed itself to the correct question and reached the view that it was not satisfied that APA GasNet's proposed methodology promoted efficient growth in the market for reference services. The findings of fact made by the AER are not shown to have been erroneous. The conclusion based upon those findings is not shown to have been unreasonable in all the circumstances. There has not been an exercise of a discretion, that is the limited discretion allowed for under rule 79(3), which was misunderstood or which was incorrect in all the circumstances. 223 Nor has it been shown that the AER erred by failing properly to take into account the material adduced by APA GasNet to support its claim that the AER depreciation methodology would not satisfy criterion (e) of rule 89. APA GasNet relied upon a report by Australian Ratings that expressed opinions on the likely credit rating that would be attributed to APA GasNet based on Standard & Poor's criteria and methodology. In its Final Decision, the AER gave consideration to the Australian Ratings' report. The AER said that its analysis is sensitive to contentious assumptions which in several instances the AER did not accept. The sensitivity of the conclusions to these assumptions was untested, so the AER gave little weight to it to assess revenue decisions. Otherwise, there was little or no material to suggest that criterion (e) would not be met by the AER's depreciation methodology. It is clear that a service provider should be provided with a reasonable opportunity to recover at least the efficient costs incurred in providing reference services and complying with regulatory obligations. APA GasNet produced no cogent evidence in support of its contention that the tariff profile that would result from the AER's depreciation methodology would fail to satisfy its reasonable cash flow needs. 224 The depreciation methodology approved by the AER results in a stable tariff path. The reduced tariffs in 2013 are a result of lower costs (including lower cost of capital and lower capital base); the lower tariffs are not the result of the depreciation methodology. 225 The AER's conclusions with regard to reliance upon the Australian Ratings report do not disclose error. 226 For these reasons, the attack upon the Reviewable Decision in respect of the Depreciation and Indexation Decision does not succeed. RATE OF RETURN ON EQUITY DECISION 227 Rule 87(1) of the NGR requires that the rate of return on capital is to be commensurate with prevailing conditions in the market for funds and the risks involved in providing reference services. Rule 74(2) provides that a forecast or estimate must be arrived at on a reasonable basis and must represent the best forecast or estimate possible in the circumstances. 228 Rule 87(2)(b) provides that in determining a rate of return on capital a well-accepted approach that determines the cost of equity and the cost of debt, such as the Weighted Average Cost of Capital (WACC) is to be used, and that these should be estimated by the use of a well-accepted financial model such as the Capital Asset Pricing Model (CAPM). 229 The AER determined, with no demurral by APA GasNet, that the CAPM should be used to estimate the cost of equity for the purposes of determining the WACC referred to in rule 87(2). 230 Under the CAPM, the rate of return on equity, Ke, is expressed in terms of the algebraic identity: Ke = Rf + ß * (Rm - Rf) where: Rf is the nominal risk-free rate; ß is the equity beta; and Rm is the expected return on the market portfolio of equities 231 The right hand term Rm - Rf in this identity is known as the Market Risk Premium (MRP), and is measured as the excess of the expected future return over the risk-free rate (typically the rate on 10-year Commonwealth Government Securities (CGS) that investors require to invest in a well-diversified portfolio of risky assets. The expected return on the market is the same as the expected return on the average firm. Clearly, the MRP is an input into the calculation of the cost of equity for the purposes of the WACC referred to in rule 87(2)(b). 232 Rf, the nominal risk-free rate, is readily observable, as the market for 10-year CGS is liquid. In contrast, as the MRP is an expected or forward-looking parameter, it is by definition not observable in advance of reality. Therefore its future value must be estimated in some way. The Tribunal noted in Envestra (No 2) at [146] that, as is the case with any variable whose values must be estimated for a future time period, there is unlikely ever to be one unique "correct" value of the MRP. Reasonable minds can be expected to differ on the data and technique(s) to be used in its calculation. 233 Accordingly, regulatory price determinations typically involve considerable disagreement as to how the MRP can best be calculated at any given point of time, to apply to a future access arrangement period. 234 That has certainly been the case in this matter. The disagreement involved many issues, and particularly whether the MRP should be estimated as a number in its own right, as contended for by the AER and the basis of its determinations of MRP for some time, or whether it should be measured by way of separate estimation of its two component parts, as was submitted by APA GasNet. This is a question for which there exists no definitive answer, either in theory or as a matter of empirical estimation. 235 APA GasNet's submission to the AER led to a major disagreement as to the basis on which Rm, one of the two separate components of the MRP, could be measured. There was no major disagreement in this matter as to the basis for measuring the relevant Rf value to include in the CAPM formula. 236 APA GasNet submitted that the AER's figure for the cost of equity was both incorrect and illogical, and its method that reduced this return in line with the lower risk-free rate was "inappropriate". It contended that the AER had paid too little regard to the Dividend Growth Model (DGM) as a valid means of estimating the MRP component of the cost of equity. 237 APA GasNet submitted that the MRP should be 8.72 percent, based on a series of expert reports produced for it. The AER rejected these arguments and exercised its discretion to adopt an MRP of 6 percent, on the basis that no persuasive evidence had been provided to convince it to depart from either its previous estimation methods or the rate of 6 percent which it had used consistently as the value of the MRP for some time. APA GasNet's proposal and the AER's determination 238 In its Revised Access Arrangement Proposal, APA GasNet noted that it was common ground between it and the AER that, in accordance with the requirement in the Rules that the rate of return was to be commensurate with prevailing conditions in the market for funds, all elements of the CAPM formula that are time-variant should ideally reflect prevailing market conditions; that an appropriate value for the equity beta was 0.8; and that the nominal risk-free rate should be determined as the prevailing yield on 10-year CGS, averaged over an agreed 20-day period, leading to a rate of 3.22 percent. 239 In its original Access Arrangement Proposal dated March 2012, APA GasNet proposed an MRP of 8.5 percent. It supported this proposal with expert consulting reports from NERA, CEG and CR. CEG in particular supported its proposed estimation of the MRP by the use of a DGM methodology, which APA GasNet pressed in this hearing. In its Revised Proposal following the AER's Draft Decision, dated November 2012, APA GasNet proposed a revised MRP of 8.72 percent, estimated using the DGM methodology provided by its consultant CEG. 240 In its Draft Decision, the AER gave detailed consideration to the arguments and evidence presented by APA GasNet and its consultants. It decided that a preferable estimate of the MRP was 6 percent, relying upon a wide range of evidence. 241 This evidence came from six main sources: (a) estimates of historical excess returns, which ranged from 4.9 percent to 6.1 percent if calculated using an arithmetic mean, and from 3.0 percent to 4.7 percent based on the geometric mean; (b) surveys of market practitioners, which consistently supported 6 percent as the most commonly adopted value for the MRP; (c) forward-looking methods like those based on the DGM and implied volatility estimates, both of which yielded highly variable results and were subject to significant limitations. The DGM estimates provided on behalf of APA GasNet suggested that the MRP could lie in the range of 5.9 percent to 8.4 percent, while it was reported that implied volatility-based MRP estimates indicated that the MRP was currently below its historical average level; (d) the advice of the AER's consultants, which supported an estimate of 6 percent; (e) recent Tribunal decisions, which concluded that a 6 percent MRP was reasonable; and (f) recent decisions of other Australian regulators, which also supported a 6 percent estimate. 242 In its Final Decision, the AER gave further detailed consideration to the arguments and evidence presented by APA GasNet and its consultants. After assessing all the evidence, it maintained its view that the preferable estimate of the forward-looking MRP for the next access arrangement period was 6 percent. 243 As a reasonableness check, the AER observed that regulated assets have generally been sold at a premium to the value of the regulated asset base, indicating that its method of determining the MRP, and thus the cost of equity, was acceptable to the market and was therefore reasonable. Grounds of review 244 APA GasNet argues that the AER made a number of errors of fact, exercised its discretion incorrectly, or reached a conclusion that was unreasonable. Broadly those errors relate to the application of the CAPM model, and the inputs to that model. 245 First, APA GasNet contends that the AER estimated the MRP directly as a whole identity, rather than estimating the expected market return and subtracting from that the risk-free rate of return. Second, the AER either did not have regard to, or failed to give sufficient weight to, the evidence from APA GasNet's experts regarding the use of the DGM to estimate the expected market return, as an input into estimating the MRP, a methodology which APA GasNet appears to assume is the only one that faithfully estimates the MRP under the theoretical CAPM formula. 246 APA GasNet took issue generally with the method by which the AER estimated the MRP, submitting that it erred in estimating it as a single number rather than by its two component parts, and also that in this process it implicitly used a different risk-free rate to that used as a standalone value in the first term in the CAPM identity, a figure which had been agreed to by both parties. It submitted that it was critical to use a consistent value for the risk-free rate in both places where it appears in the formula. 247 The AER, APA GasNet contended, had used the agreed-upon risk-free rate in the first part of the CAPM identity, but then, given its method of deriving the MRP as single rather than a two-component figure, it had combined the prevailing risk-free rate (which happened to be very low) with an MRP that was derived from historical measures (when the risk-free rate was higher) such that the two risk-free rate figures in the formula were inconsistent. 248 In APA GasNet's opinion, the use by the AER of historical average excess returns to estimate the MRP (which is disputed by the AER) represented an incorrect use of the CAPM model. 249 There was no disagreement between APA GasNet and the AER that an acceptable way to determine the MRP was to consider the level of excess market returns over a risk-free rate. Rather, the differences between them concerned the "right" methodology to employ, and the relevant time period from which the forward-looking estimates were derived and constructed. 250 It was contended by APA GasNet that the AER erred in its assessment of the DGM method that it had advocated, on account of that fact that it had failed to deal properly with the evidence presented in support of using the DGM, especially the evidence of Dr Hird of CEG, and that by ignoring this method, it had failed to derive the proper expected market return. 251 APA GasNet further submitted that the AER was in error in the way it estimated the MRP, because the method it used resulted in a value of the MRP that was not commensurate with the prevailing conditions in the market for equity funds. This, APA GasNet submitted, was unreasonable and relied illogically on regulatory precedent. 252 Counsel for APA GasNet asserted in the hearing that the AER had not estimated the MRP using the CAPM. When asked what model the AER had used, he characterised it as "the hybrid historic-current model with two inconsistent risk free rates". Essentially, APA GasNet's submission was that the AER had failed to understand the task of estimating the MRP, which, to be true to the CAPM model, required the estimation of the difference between the expected market return on equity, a forward-looking value, and the current risk-free rate of return. 253 Under rule 40 (3) of the NGL, the AER may withhold approval of a rate of return proposed by a service provider if, in the regulator's opinion, a preferable alternative exists. 254 It is not sufficient for APA GasNet to persuade the Tribunal that its preferred figure of 8.72 percent for the MRP should be accepted. It is required under s 246 (1) of the NGL to demonstrate that the AER made one or more errors of fact, or that it incorrectly exercised its discretion, having regard to all the circumstances, or that the AER's decision was unreasonable, having regard to all the circumstances. Unless one or more of these grounds can be made out, the Tribunal, in accepting a figure other than the 6 percent value determined by the AER, would be merely reaching a different conclusion as to the preferable result. The mere fact that the Tribunal might believe that a different rate is preferable does not entitle it to substitute its preferred MRP for that of the AER, unless a ground of review has been made out. See the Tribunal's comments in DBNGP (No 3) at [326]. APA GasNet's contentions regarding the "correct" measurement of the MRP 255 APA GasNet relied upon three main issues in challenging the AER's estimation of the MRP. First, the fact that it did not calculate the two component terms independently, which resulted in its estimated MRP being a theoretically invalid hybrid estimate that was not true to the logic of the CAPM. Second, that it rejected the use of the DGM to estimate the expected market return, which, given the agreed value between the parties as to the relevant risk-free rate to use, would have permitted the determination of the MRP that was consistent with the CAPM. Third, that it erred in rejecting evidence that there was a negative relationship between the MRP and the risk-free rate. 256 APA GasNet variously challenged the AER's MRP estimate of 6 percent as an "historical", "fixed", "long-term average" or "constant" measure of the MRP. However the AER had explained at length in the Final Decision why its estimation process produced a value for the MRP that was none of these. Rather, it was derived on the basis of an extensive range of evidence. In contrast, APA GasNet contended that only one method, the DGM methodology, should be used to calculate the expected market return on equity, from which the risk-free rate is subtracted to yield the MRP. 257 APA GasNet's contentions relied largely on its experts' arguments that favoured DGM estimates, and rejected any of the alternative estimating methods used in combination by the AER. The AER rejected the use of the DGM on the basis of numerous other experts' reports. 258 As stated by the Tribunal in Envestra (No 2) at [147-148] in relation to estimating the MRP: In circumstances such as this it is incumbent upon Envestra, if it is to show a ground of review, to demonstrate why the choice of methodology made by the AER was not reasonable in all the circumstances. While it is true that Envestra pointed to a large body of expert evidence that supported its contention that 6.5 percent would have been a more appropriate choice for the MRP, there was also a large body of evidence that supported the AER's chosen rate." "It is not sufficient for Envestra to persuade the Tribunal that 6.5 percent should be preferred. It must demonstrate the unreasonableness of the decision made by the AER. Unless this can be done, the Tribunal would be merely reaching a different conclusion as to the preferable result. The mere fact that the Tribunal may prefer a different rate does not entitle it to substitute its preferred MRP for that of the AER unless a ground of review has been made out. In all the circumstances of this matter, it was reasonably open to the AER to choose an MRP of 6 percent. 259 Therefore it is for APA GasNet, if it is to demonstrate a ground of review, to explain how the AER erred in its choice of methodology. While it is true that APA GasNet produced a several expert reports that supported its contention that an MRP of 6 percent was too low, there was also a large body of evidence that supported the AER's assessment that the MRP should be set at 6 percent. The calculation of the MRP as a whole, or as the difference between its two elements 260 A major criticism advanced by APA GasNet of the AER's approach was that it had combined the agreed-upon prevailing risk-free rate (which currently is very low) and a MRP that was derived from historical measures (when the risk-free rate was higher) such that the two risk-free figures implicit in the calculation of the MRP based on the CAPM identity were not consistent. 261 APA GasNet submitted that Professor Gregory had observed that this approach involved a very common error which had been discussed in recent UK regulatory appeals. In his opinion in such a case the CAPM being estimated could be expressed as the identity: Ke = Rf, current + ß * (Rm, historic - Rf, historic) and accordingly there is no theoretical validity in the use of such a model to estimate the MRP because of the two different risk free returns employed in the calculation. 262 The potential error of using two inconsistent risk-free rates was also identified by other APA GasNet experts including Professor Wright and NERA, and by the AER's expert CEPA. APA GasNet submitted that CEPA supported its argument, in its observation that the simplification of the term (Rm - Rf) to the MRP was appropriate only if the MRP is independent of the risk-free rate such that it is a constant over time and the expected market return moves in lock step with the risk-free rate. As CEPA put it in its March 2013 report: However, when the MRP is not independent of the risk-free rate (i.e. the value of the MRP depends on the value of the risk-free rate), the equation will not be fully identified unless the risk-free rate is the same as that used to identify the MRP (which in turn must be consistent with the expected market returns) . . . To be consistent with the SL CAPM, . . . the explicit estimate of the risk-free rate and that implicit in the estimate of the MRP must be the same. For example, if there is a relationship and the MRP is calculated when the risk-free rate was higher than its current level and this is added to the current risk-free rate, the calculation is not valid as the impact of the lower current risk-free rate has not been factored into the estimate of the MRP. (emphasis added by the Tribunal) 263 A key disagreement underlying this debate between the parties involves whether there is in fact a relationship between the MRP and the risk free rate in practice. We discuss this below. 264 The Tribunal agrees with all the submissions and reports made about the importance of using internally consistent values of the risk-free rate in estimating the cost of equity under the CAPM. Under different circumstances which do not apply in this matter, it noted in Re GasNet Australia (Operations) Pty Ltd [2003] ACompT 6 at [46]-[47], that While it is no doubt true that the CAPM permits some flexibility in the choice of the inputs required by the model, it nevertheless requires that one remain true to the mathematical logic underlying the CAPM formula. In the present case, that requires a consistent use of the value of rf in both parts of the CAPM equation where it occurs... 265 However, the Tribunal there was referring to the use of two different explicit values of the risk-free rate based on different bonds with differing maturities. The circumstances in the current matter are quite different. The AER did not calculate and use two different Rf values. It did not derive its estimate of the MRP by calculating both Rm and Rf separately. Rather, it chose as a matter of empirical practicality and reliability to produce an estimate of the MRP itself, and did not derive its estimate as the sum of its two separate components. The use of the DGM in determining the MRP 266 The MRP, and one of its components in the CAPM formula, the expected market return, are both forward-looking concepts that are subject to all the usual problems associated with producing an estimate of any parameter's future value. 267 In the Final Decision, the AER spent considerable time (Part 2, pp.79-84, and Part 3 Section B.6) assessing the reliability of the various methods of estimating these two variables. It noted the following issues in particular: It is well-known that DGM estimates are highly sensitive to the assumptions made. As noted by its experts Professor McKenzie and Associate Professor Partington, DGM estimates are sensitive to the assumed future growth rate, on which consensus is rarely found. The expected market growth rate in dividends per share, a key input into DGM analysis, is typically proxied by analysts' short-term forecasts of market-wide earnings per share growth, or by long-term expectations of GDP growth, or both. Associate Professor Lally, another of its experts, had advised that such proxies could be expected to produce an upward bias in MRP estimates. He suggested in his report of July 2012 that "using the DGM to estimate the MRP is worthy of consideration but as a complement rather than as a substitute for the AER's current approach". Despite the submission of APA GasNet to the contrary, the Tribunal cannot see in this statement any support for the DGM method to be used in its own right to estimate the MRP. The AER's 2009 review of WACC parameters under the NEL had considered DGM estimates. In that review, academics and industry representatives had considered DGM estimates should be used only as a cross-check on the reasonableness of other methods to estimate the MRP, rather than as the primary method. The AER noted in that review that regulators had been wary to lower the MRP when DGM estimates produced a value for the MRP that was below 6 percent. Different consultants had produced widely different DGM-based MRP estimates over a short period. In February and March 2012, APA GasNet's three consultants used DGM estimates to produce MRP estimates ranging from a low of 6.18 percent to a high of 9.56 percent. While DGM analysis is currently producing high MRP estimates, it produced MRP estimates just above 2 percent in 1994. For some time there has existed considerable scepticism on the evidence for a relationship between certain variables and the MRP. Some studies have indicated that the historical average provides the best forecast of excess returns. The AER disagreed with the submissions of APA GasNet that Lally was critical of their use. Indeed in his July 2012 report he stated that the historical averaging of excess returns is an estimation methodology that is suitable (in conjunction with other methodologies) for estimating the MRP for the next 10 years as well as for estimating the long-term average MRP. The use of historical averaging results may introduce a downward bias at the present time, but the effect is likely to be small relative to the standard deviation in the estimate and to possible upward bias in the methodology arising from significant unanticipated inflation in the 20th century. (emphasis added by the Tribunal) 268 The Tribunal observes that in advice provided to the AER by McKenzie and Partington in their December 2011 report, contrary to the submission of APA GasNet, they cautioned strongly about relying on MRP estimates derived from the DGM. They advised that there were "several unresolved difficulties with this approach, which in our opinion render it unsuitable as a primary method for estimating the MRP", and cited the work of others that supported their view. 269 Importantly, they concluded that an MRP estimated through the forward-looking DGM will be "largely a by-product of the chosen growth rates". Dividend growth clearly is a function of expected market growth, a figure upon which experts can be expected to differ, and a figure whose values could change from one day to the next, depending on changing market or economy-wide circumstances. 270 In a subsequent supplementary report produced in February 2012, McKenzie and Partington advised that we would only use implied cost of capital estimates, such as those derived from the dividend growth model as a reasonableness check and even then we would be circumspect in our use of such models. 271 As the Tribunal sees it, two distinct polar positions have been argued by APA GasNet and the AER. APA GasNet urges the adoption of forward-looking estimates of the MRP, based on estimates of expected market returns obtained from DGM estimation techniques, which are based heavily on future growth predictions for the market as the basis for its estimates of dividend growth. 272 However, the Tribunal is aware that surveys of the growth expectations of those close to the market produce estimates the basis of which cannot easily be determined, and which could be based on past experience or speculation about future events, or some combination of the two. 273 APA GasNet also asserts that the use of historical averages (whatever form of average may be chosen) of the MRP, or of the market return on equity, could be based on market conditions that do not reflect anticipated future market conditions for the firm or industry under examination. 274 This leads to the necessity of considering exactly how the AER estimated the MRP. 275 In its Final Decision at Part 3 p. 27, the AER argued that the criticisms of its method by several of APA GasNet's experts misunderstood how it had determined the MRP. It explained that it did not calculate a long term historical average MRP (incorporating a long-term average risk-free rate) and that conceptually it estimated a 10 year forward-looking cost of equity by determining an estimate of the 10 year forward-looking risk-free rate and a 10 year forward-looking MRP. This meant that its estimate of the MRP was internally consistent. 276 It was APA GasNet's contention that the proper application of the CAPM formula required the expected market return to be estimated separately. It agreed that it is well known that this parameter cannot be directly observed or measured, as it is a forward-looking estimate. It urged in its submission that the MRP is only properly expressed . . . as the difference of two measured values (provided they are measured on the same basis . . .; [the] MRP is not a figure or a concept or a property in its own right. It represents the difference at a particular point in time between two specific and measurable concepts. 277 APA GasNet submitted that the evidence used by the AER to measure the MRP was dominated by historical measures - historical excess returns, historical surveys (based on past practices in valuation and investment), and historical decisions of regulators. In contrast, it said, the forward-looking measures considered by the AER such as the DGM, were the subject of significant criticisms by the AER and were given little weight in its estimation process. 278 Against this, the AER submitted that it did consider forward-looking methods to estimate the expected market return but believed, based on advice from several experts, that such estimates lacked credibility. One option was to consider the use of historically measured excess returns, which measure the MRP over time using historically consistent and contemporaneous measures of the expected market return and the risk-free rate, as it can be argued that investors will base their forward-looking expectations on what has been observed in the past. The AER noted in its Final Decision (Part 2 pp. 75-76) that there was academic support for the use of such studies to estimate the MRP. 279 The fact that one of the sources on which the AER based its estimate of the MRP was an historical average cannot, without more, be taken to imply that the MRP figure that it derived was an historic measure. Indeed, the AER's independent experts had considered this issue and supported the AER's estimation method. This advice was reported in the Final Decision in Part 2, pp. 88-89 and Part 3, pp. 27-29. 280 Accordingly, the AER adopted what APA GasNet called "a primarily historical measure" of 6 percent, which APA GasNet suggested was its "traditional value" that did not reflect prevailing market conditions, and which involved the assumption that the current fall in the risk-free rate was matched by a fall in the expected market return. 281 The Tribunal notes that this value of 6 percent lies at the upper end of the range of the MRP values produced in the historical excess return studies, a fact revealed by the AER in is Final Decision (Part 2, Table 5.2, p. 76). 282 An MRP of 6 percent was endorsed by Lally in his March 2013 report after surveying a wide range of empirical studies. He concluded that these studies "reinforce the conclusion that the appropriate MRP estimate for Australia at the present time is 6 percent". 283 A key point in APA GasNet's submissions was that as the risk-free rate had been falling (a fact that was not in dispute), reflecting the increased demand for risk-free assets like CGS, there had been, in the words of a letter written by the Reserve Bank of Australia to the ACCC on 18 July 2012, a widening in the spreads between CGS yields and those on other Australia dollar-denominated debt securities, confirming: the market's assessment of the risk-free nature of CGS and reflects a general increase in risk premia on other assets . . . That said, market risk premia are unlikely to be stable through time. While it is a reasonably simple matter to infer changes in debt risk premia from market prices, it is less straightforward to do so for equity premia. In making use of a risk-free rate to estimate a cost of capital, it is important to be mindful of how the resulting relativity between the cost of debt and that of equity can change over time and whether that is reasonable. (emphasis added by Tribunal) 284 The Tribunal sees this statement as being rather more supportive of the AER's position than it is of APA GasNet's argument. Yet CEG, a consultant to APA GasNet, had reported that if the AER's approach was correct, yields on all other assets should fall in line with CGS yields, whereas if its approach was incorrect, the spread (risk premium) between CGS and other assets should have risen. CEG's research showed that the spread had increased significantly between CGS and four other classes of debt and equity: 10 year state government debt; AAA corporate bonds at 4 years (the longest maturity for the Bloomberg AAA curve); higher risk bonds; and the dividend yield on ASX-listed securities. 285 Another consultant to APA GasNet, Professor Gray of SFG, believed that the correct cost of equity was significantly higher than the one produced by the AER's method. APA GasNet submitted that on his evidence the AER's estimate of the required return on equity was the lowest on record, suggesting it regarded equity investors as being more prepared to make equity investments requiring lower returns than ever before. In contrast, Gray argued that dividend yields were currently high in historical terms, suggesting that the market was applying a high discount rate to dividends and earnings and that the required return on equity was high compared to the last 20 years. 286 APA GasNet also submitted that whereas debt risk premia were currently three or four times higher than pre-GFC levels by the AER's estimates, equity risk premia had not increased at all, indicating that a market that requires a three- to four-fold increase in risk premia when investing in debt securities in a benchmark firm requires no additional risk premium when investing in riskier equity securities in the same firm. This, it implied, could not be correct. The relationship between the risk-free rate and the expected market return 287 APA GasNet advanced a third argument against an MRP value of 6 percent, namely that its derivation involved an assumption that the decline in the risk-free rate was matched by a decline in the expected market return and that the AER did not have any evidence that this assumption was correct. A major part of these contentions related to the relevance of movements in the risk-free rate. 288 The AER submitted that its assessment did not depend on such an assumption. Rather, it had examined all the evidence available to it to estimate separately the risk-free rate and the MRP. 289 APA GasNet also contended, based on the advice of CEG and Gray, that the AER had not properly taken account of the spread (risk premium) between CGS and other assets, particularly other debt assets. It argued that the MRP cannot move independently of the debt risk premium (DRP), and that as research showed that the spread between CGS and corporate bond yields had increased, then the MRP should also be higher. 290 The AER had contrasting advice from McKenzie and Partington (that the required return on equity does not necessarily change if there is a change in debt default spreads. The DRP captures changes in debt default risk, and is a market-driven phenomenon that is observable. In contrast, the CAPM contains no default risk adjustment for the cost of equity, as in determining the cost of equity, expected cash flows are adjusted to reflect changes in the level of default risk. 291 The AER had in fact considered this evidence and APA GasNet's submissions concerning the postulated relationship between the unobservable MRP and the observable DRP. It came to the conclusion, supported by expert opinion (Final Decision, Part 3, section B.5.3, p. 48), that there is no consensus in academic literature on the direction or magnitude of the relationship between observed credit spreads and the MRP. The lack of academic consensus on the direction of any relationship casts doubt on the reliability of drawing any conclusions on the MRP from observable debt premiums. Moreover, the inability to reliably quantify the magnitude of any relationship limits its usefulness in a regulatory framework. 292 It was strongly contended by APA GasNet that the AER had erred in not accepting that there was a negative relationship between the MRP and the risk-free rate over time. 293 The AER did in fact pay considerable attention to this issue. It commissioned reports from several experts to consider both the theoretical support for the argument and the empirical evidence on the relationship. Their advice was that there was insufficient evidence to support any firm conclusion about the nature of the relationship between the MRP and the risk-free rate. A graph produced by CEG in its November 2012 report, showing MRP, CGS and cost of equity data for the period 1993 to 2012, was found by the AER not to be persuasive that a negative relationship existed between the risk-free rate and the MRP. 294 The AER therefore concluded that there was no justification for accepting that a negative relationship existed and so that in the face of low CGS yields, the MRP should be higher. 295 Another, related, major point of contention was whether a proven negative relationship existed between the two component terms of the MRP, the market return on equity and the risk-free rate. The risk-free rate is currently at a low level not seen for many years, albeit arguably not especially low in terms of the long-term average risk-free rate. 296 APA GasNet submitted that this negative relationship meant that in the presence of such a low risk-free rate, the inferred estimate of the market return should be higher than it has been in the past, to reflect this negative relationship. This argument carried with it undertones of an implicit assumption that the MRP over time is relatively stable. 297 Considerable time was spent by the academic experts for both sides reviewing the literature on this relationship. The empirical evidence appears to be far from presenting a concluded position. The AER reviewed this evidence at length, and was not confident that such a negative relationship existed, neither in terms of statistical significance nor in terms of its strength. Accordingly, it rejected the submission that a low risk-free rate should be balanced by a higher expected market return, and thus justifying a higher MRP than 6 percent. 298 The AER did not err in the terms specified in s 246(1) of the NGL in the terms specified in s 246(1) of the NGL merely because it disagreed with APA GasNet on the direction and strength of this relationship. Given the state of the academic evidence on the relationship, the AER's conclusion simply reflected a difference of opinion concerning an empirical issue that, in the opinion of the Tribunal, is far from settled. 299 Accordingly, the Tribunal finds no error in the AER's rejection of an alleged significant negative relationship between the risk-free rate and the expected market rate of return on equity. 300 In passing, the Tribunal notes that should such a relationship in fact exist, then it would follow that in the future when the risk-free rate increases, regulated entities, to be consistent, would need to argue for a correspondingly lower market return to be used in estimating the MRP 301 APA GasNet further sought to advance an argument that the AER had erred in reasoning and logic by relying on one particular report by Lally, in which he criticised reports provided by two of APA GasNet's experts, CEG and SFG. Given that we have found that the AER committed no error in rejecting the claimed negative relationship, there is no need to consider this argument at any length, other than to note that what really occurred was yet another example of disagreement between the experts, which is, as the Tribunal has noted many times, a far from uncommon event in matters such as this. Conclusion with respect to the MRP 302 The MRP is a forward-looking concept. By definition, it is not directly observable and so must be estimated or predicted. There are many techniques that are available to do this. Each method, by itself or in combination with other methods, is necessarily imperfect in estimating what is a logical, but nevertheless strictly theoretical, parameter. 303 Thus there can be no such thing as an empirically "correct" estimate of the MRP. It must be recognised that the best estimate of this parameter will be one based on careful judgment and evaluation of the competing estimation methods, and reflecting the expected market circumstances over the forthcoming period to which the access arrangement applies. The best estimate could well be based on different methods from time to time. 304 The Tribunal considers it worth repeating here what it said recently in Application by WA Gas Networks Pty Ltd (No 3) at [106] … the MRP is a forward-looking estimate. There is no single accepted econometric, mathematical or financial technique that can, uniquely, be deployed to ascertain an estimate of the MRP that can apply to any future period. Further, there are substantial debates among the experts as to how particular methodologies should be employed and the assumptions that are necessary to drive them effectively. These choices of methodologies and assumptions can significantly alter the resulting estimate. 305 Accordingly, in estimating a value for the MRP, the AER will need to exercise its discretion based on its own experience and previous decisions, the advice of its experts, historical data, its expectations of future market economic and financial conditions, and, of course, taking fully into its consideration the submissions and expert advice put to it by the regulated entity and any other parties granted standing in the matter. 306 As the Tribunal recently stated in DBNGP (No 3) at [326]: It is clear enough that a finding of fact by the ERA is not shown to be erroneous simply because there is material which could support a different finding of fact. Nor is a finding of fact by a regulator shown to be erroneous simply because the Tribunal might, if it considered the material afresh, prefer to make a different finding of fact. The reference to "error" clearly requires more. In some instances, the error will be apparent: there will be no material to support the finding, or the only material will support a different finding. However, where the finding is a complex one, that is one which involves the assessment of expert opinion material or conflicting material or (as here) conflicting expert opinion material, an error of fact is not shown simply because one expert opinion or set of opinions has been preferred over another expert opinion or set of opinions. Something more will be required, before the Tribunal is satisfied that there is an error of fact (and, necessarily, a material one) on the part of the regulator. 307 The Tribunal has noted on many previous occasions that its task is determine whether the AER erred in the ways specified in s 246(1) of the NGL. It is not to substitute a decision which it might prefer to make on the material before the AER: ElectraNet (No 3) at [64], [69] and [72]; DBNGP (No 3) at [25]. 308 It is the opinion of the Tribunal that APA GasNet did not establish, as required under s 246(1), any of these grounds with respect to the AER's estimation of the MRP. APA GasNet's complaint in reality concerns the result of the AER's investigations, and not the process. In all the circumstances of this matter, it was reasonably open to the AER to choose an MRP of 6 percent. CONCLUSION 309 For the above reasons, the Tribunal has concluded that APA GasNet has succeeded on its application on the Interval of Delay Decision and the Opening Capital Base Decision. 310 As noted, the parties were agreed that, if time permitted, they should be given the opportunity to consult with a view to proposing orders which would best facilitate giving effect to the decision of the Tribunal. 311 The Tribunal will allow them that opportunity. The Tribunal's orders must be made by 4 October 2013. In case further submissions are necessary, the Tribunal will allow the parties until 2:00 pm on 27 September 2013 either to submit the form of orders which will give effect to the Tribunal's reasons, or to submit in writing their respective arguments for the orders which they separately propose. I certify that the preceding three hundred and eleven (311) numbered paragraphs are a true copy of the Reasons for Decision herein of the Honourable Justice Mansfield (President), Professor Round and Mr Steinwall (Members). Associate: Dated: 18 September 2013