Application by APT Allgas Energy Limited (No 2) [2012] ACompT 5

statutory scheme 7 Section 7 of the National Gas (Queensland) Act 2008 (Qld) (Act) applies the NGL, set out in the Schedule to the the National Gas (South Australia) Act 2008 (SA), as a law of Queensland. 8 Section 26 of the NGL gives the National Gas Rules (NGR) the force of law in Queensland. 9 Allgas is a "service provider" within the meaning of section 8 of the NGL, in that it owns, controls or operates a scheme pipeline. The NGL defines "scheme pipeline" to include a "covered pipeline". Allgas's gas distribution network is a "covered pipeline" within the meaning of the NGL. 10 The AER is responsible for the economic regulation of pipeline services provided by service providers, including Allgas, by means of or in connection with a scheme pipeline. In particular, under Part 9 of the NGR, the AER is responsible for determining the total revenue for Allgas for each regulatory year of an access arrangement period for the provision by Allgas of reference services in Queensland. 11 Under Rule 52 of the NGR, Allgas was required to submit and, on 30 September 2010, did submit an access arrangement revision proposal for the access arrangement period from 1 July 2011 to 30 June 2016 to the AER for consideration in accordance with the NGR (Allgas's Access Arrangement Proposal). 12 Under Rule 59 of the NGR, the AER was required to make, and did make, an access arrangement draft decision in relation to Allgas's Access Arrangement Proposal entitled Draft Decision APT Allgas Access arrangement proposal for the Qld gas network 1 July 2011 - 30 June 2016, February 2011 (Draft Decision). 13 Rule 60 of the NGR entitled Allgas to submit additions or other amendments to Allgas's Access Arrangement Proposal to address the matters raised in the Draft Decision. The amendments which Allgas was permitted to make were limited to those amendments necessary to address matters raised by the AER in the Draft Decision, unless the AER approved further amendments: NGR r 60(2). 14 On 23 March 2011, Allgas submitted a revised access arrangement proposal to the AER (Allgas's Revised Access Arrangement Proposal). 15 Pursuant to Rule 62 of the NGR, the AER was required to make an access arrangement final decision in relation to Allgas's Revised Access Arrangement Proposal. Rule 62 of the NGR provides that an access arrangement final decision is a decision to approve, or refuse to approve an access arrangement proposal. 16 On 17 June 2011, the AER published the Final Decision. In the Final Decision, the AER refused to approve Allgas's Revised Access Arrangement Proposal and indicated it would publish its own revised access arrangement and access arrangement information. 17 Pursuant to Rule 64(4) of the NGR the AER decided to approve the access arrangement (including the access arrangement information) drafted by it for Allgas's Queensland gas distribution network. That decision is set out in the Access Arrangement Decision. 18 Subject to section 245 of the NGL, an affected or interested person or body, with the leave of the Tribunal, may apply to the Tribunal for a review of a reviewable regulatory decision. 19 An "affected or interested person or body" (as defined in section 244 of the NGL) includes a service provider to whom a reviewable regulatory decision applies. Allgas is the service provider to whom the reviewable regulatory decision, namely, the Access Arrangement Decision, applies and is therefore an "affected or interested person or body" within the meaning of section 244. 20 The NGL provides that a reviewable regulatory decision includes an applicable access arrangement decision, subject to the exception that a full access arrangement decision that does not approve a full access arrangement is not a reviewable regulatory decision: NGL s 244. 21 An applicable access arrangement decision includes a full access arrangement decision. A full access arrangement decision includes a decision of the AER under the NGR that: 1. approves or does not approve a full access arrangement proposal or revisions to an applicable access arrangement submitted to the AER under section 132 of the NGL or the NGR; or 2. makes a full access arrangement in place of a full access arrangement proposal which the AER does not approve in that decision. 22 The Access Arrangement Decision is a reviewable regulatory decision within the meaning of section 244 of the NGL because the Access Arrangement Decision is a full access arrangement decision in which the AER made a full access arrangement in place of a full access arrangement proposal which the AER did not approve in that decision.

GROUNDS FOR REVIEW 23 Subsection 246(1) of the NGL provides that an application for review under section 245(1) may only be made on the following grounds: 1. the original decision maker made an error of fact in the decision maker's findings of facts, and that error of fact was material to the making of the decision; 2. the original decision maker made more than one error of fact in the decision maker's findings of facts, and those errors of fact, in combination, were material to the making of the decision; 3. the exercise of the original decision maker's discretion was incorrect, having regard to all the circumstances; 4. the original decision maker's decision was unreasonable, having regard to all the circumstances. 24 Senior counsel for the AER made extensive submissions about the circumstances in which one or more of the grounds of review may be made out. The onus of making out a ground of review rests upon Allgas. 25 As the Tribunal has previously recognised, the Tribunal's task is not simply to substitute for a decision of the AER a decision which the Tribunal may prefer to make on the material before the AER, that is, on the review related material as determined in accordance with s 261: see Application by ElectraNet Pty Limited (No 3) [2008] ACompT 3 (ElectraNet (No 3)) at [64] and [69]; Application by Energy Australia [2009] ACompT 8 (Energy Australia) at [70]. So much follows from the terms of s 246(2) as explained in Australian Competition and Consumer Commission v Australian Competition Tribunal (2006) 152 FCR 33 (ACCC v ACT) per French, Goldberg and Finkelstein JJ at [176], a decision concerning the grounds of review under s 39(2)(a) of the Gas Pipelines Access Law established under the Gas Pipeline Access (South Australia) Act 1997 (SA). Those grounds of review are not materially different from those expressed in s 246(1) of the NGL. 26 The Tribunal in this matter remains mindful of its role as explained in ACCC v ACT. It is also mindful that Allgas is not permitted to raise any matter not raised before the AER: s 258 of the NGL. 27 The lengthy and helpful oral and written submissions of the AER on the scope of the available grounds of review, at least initially, were made on a general level. The Tribunal's observations that follow in relation to those submissions made below must also be seen in that context. They are made to acknowledge these submissions and to add some comments of the Tribunal. However, they must not be taken as the Tribunal's concluded views on matters concerning the scope or application of s 246(1) remote from the particular circumstances of this application. Each application for review must be considered on its own merits. Ultimately, it is for the Tribunal to identify any particular matter in respect of which Allgas has asserted reviewable error, and the Tribunal must then be persuaded that the error is one which is reviewable in terms of s 246(1). That is the Tribunal's task in this matter, as in like matters where the grounds of review are so confined. 28 However, as indicated, before embarking on that task, the Tribunal makes the following comments on the AER's more general submissions.

Error of fact 29 It is clear that findings of fact may concern the existence of a present or historical fact, being an event or circumstance. There is also some support for the proposition that an opinion about the existence of a future fact or circumstance, as well as opinions formed by the AER based upon approaches to the assessment of facts or methodologies which it has chosen to apply may also constitute findings of fact: ACCC v ACT at [171]; ElectraNet No 3 at [67]. 30 The inclusion of opinions about the existence of future facts within the meaning of "fact" is the subject of some controversy between the parties. In Application by ActewAGL Distribution [2010] ACompT 4 (ActewAGL) the Tribunal said that the inclusion of opinions in the meaning of "fact" was a radical meaning to be given to the word and that it was generally accepted "that an opinion is an inference which is drawn from facts": ActewAGL at [33]. 31 The nature of the AER's task under the NGL and NGR necessarily involves an assessment as to likely future occurrences and states of affairs, formed on the basis of expert opinion and evidence of current and historic facts. The phrase "finding of fact" should not be given a meaning that would render its applicability to the AER's functions minor and largely superficial. That is, "the term 'findings of fact' should be interpreted broadly enough to be meaningful in relation to the function of the" AER under review: ACCC v ACT at [171]. 32 It is clear, however, that the term "findings of fact" does not include the making of choices between permitted methodologies. Nor will a finding of fact be in error because it was based on the use of one methodology rather than another. Further, the weight to be given to competing regulatory considerations is not a finding of fact: ACCC v ACT at [171]. 33 Even if an expression of opinion could be a finding of fact under sections 246(1)(a) and (b), this ground of review is established only if the expression of opinion was erroneous, as revealed only by the review-related material. 34 In order to make out this ground for review, the applicant must not only establish an error of fact. It must also establish that the error (or the errors in combination) was (or were) 'material'. This requirement was not present in ACCC v ACT. In this context, an error of fact is 'material' if the relevant decision of the AER depends on or is based on the error: Ibrahim v Minister for Immigration and Citizenship [2009] FCA 1328 at [8].

Incorrect exercise of discretion 35 In ActewAGL at [34], the Tribunal considered the meaning of a "discretionary decision". It stated: It is most commonly applied to decision making which involves essentially a weighing up of relevant facts. First the decision maker finds the facts. Then the decision maker undertakes a weighing up process which involves taking into account considerations that are found to be relevant, assessing the weight to be given to those considerations so assessed and determining what, as a result of that process, is the right result. 36 This ground for review is not available merely because the Tribunal would exercise the discretion in a different way. In ElectraNet (No 3), the Tribunal stated at [72]: the concept of incorrectness extends beyond "Wednesbury unreasonableness", but on the other hand does not extend simply to where the Tribunal would have exercised the discretion in a different way. 37 In East Australian Pipeline Pty Ltd v Australian Competition and Consumer Commission (2007) 233 CLR 229 at [79] Gummow and Hayne JJ, (Gleeson CJ, Heydon and Crennan JJ agreeing at [13]) compared this ground of review with that of unreasonableness, and their Honours said that the two separate grounds could be understood by the explanation in House v The King (1936) 55 CLR 499 at 505. 38 The Full Court of the Federal Court in ACCC v ACT at [174] stated that the discretion of the regulator may be incorrectly exercised in the following ways: 1. an exercise of discretion based upon a misconstruction or misapplication of the relevant principles or methodologies or factors required to be considered by the statutory scheme; 2. an exercise of discretion affected by a failure to have regard to a mandatory relevant factor as prescribed by the statutory scheme; and 3. an exercise of discretion affected by the regulator taking into account a factor extraneous to those relevant by reason of the statutory scheme. 39 In ElectraNet (No 3) at [66], the Tribunal applied this statement albeit in the context of the National Electricity Law and Rules. 40 The Full Federal Court in ACCC v ACT noted at [175] that each of these matters is a traditional ground of judicial review. Where such a ground is made out in the context of an administrative review (as opposed to judicial review), it is appropriate to describe the exercise of the discretion based upon such an error as 'incorrect'. 41 The AER must also have regard to the matters required by the NGL and NGR to be included in an access arrangement: see, for example, NGR r 48. The AER must have regard to the service provider's "access arrangement proposal", being the terms and conditions about access to pipeline services proposed by the service provider. 42 The AER must have regard to its own reasons for refusing to approve the service provider's proposal. Rule 64(2) must be read in conjunction with obligations on the AER imposed by the NGL, such as the obligation to exercise its economic regulatory functions in a manner that will or is likely to contribute to the achievement of the national gas objective (NGO) and the obligation to take into account the revenue and pricing principles when performing certain economic regulatory functions: NGL s 28. 43 The AER submitted to the Tribunal that the various obligations imposed upon service providers when making access arrangement proposals and/or when generating and providing access arrangement information do not fall upon the AER as binding constraints upon its power. Applying this reasoning, the requirements expressed in Rules 74, 87 and 91 of the NGR do not bind the AER when it is proposing or making revisions to an access arrangement under rule 64(2). 44 Rule 64(2) of the NGR sets out the matters that the AER must have regard to in making its proposal for an access arrangement if it has refused an access arrangement proposal. These are: 1. the matters that the NGL and NGR require an access arrangement to include; 2. the service provider's access arrangement proposal; and 3. the AER's reasons for refusing to approve that proposal. 45 The AER need only have regard to the requirements of rules 74, 84 and 91 of the NGR, and others like them, to the extent that such consideration is necessary to have regard to the service provider's access arrangement proposal and the AER's reasons for refusing to approve it. 46 If the AER exercised its discretion on correct principles and if the particular exercise of the discretion was open to it within the framework of the legislation, the Tribunal is not empowered to set aside that decision simply because it thinks another decision would have been preferable: ACCC v ACT at [175]. 47 A decision which is not determined by reference to the applicable criteria in the NGL or the NGR is likely to have involved an incorrect exercise of discretion: Energy Australia at [68]. An incorrect exercise of discretion may occur where the exercise of the discretion is based upon a misconstruction or misapplication of relevant principles, methodologies or factors required to be considered by the NGL or NGR: ElectraNet (No 3) at [66]; see also ACCC v ACT at [174]. 48 Further, a decision affected by failure to have regard to a mandatory relevant factor prescribed by the NGL or the NGR, or affected by the regulator taking into account factors which are extraneous to those relevant under the NGL or NGR, may well involve an incorrect exercise of discretion: ElectraNet (No 3) at [66]; see also ACCC v ACT at [174]. 49 If the reasons for a decision contain a logical error or an unexplained discretionary choice made in reaching a conclusion, then the decision is likely to have involved an incorrect exercise of discretion, as well as also being unreasonable: Energy Australia at [67]. If factual error is made out and the exercise of discretion is based upon it, the exercise of discretion is incorrect ElectraNet (No 3) at [66]; see also ACCC v ACT at [175].

Unreasonableness 50 The ground for review in subsection 246(1)(d) of the NGL requires that the decision under review is itself unreasonable. Guidance as to the ambit of this ground was provided by the Tribunal in ElectraNet (No 3), where it said at [74]: The unreasonableness must be of the AER's decision itself, not of a step in its factual findings or its reasoning. It is important to recognise that it is the AER's decision which must be unreasonable having regard to all the circumstances before that ground is enlivened. 51 The concept of unreasonableness goes beyond Wednesbury unreasonableness but is still limited. There must be logical error or irrationality in the decision. To make out this ground, the AER's decision must not be justified by reference to its stated reasons. In ACCC v ACT, the Full Court stated at [178]: The concept of 'unreasonableness' imports want of reason. That is to say the particular discretion exercised by the [regulator] is not justified by reference to its stated reasons. There may be an error in logic or some discontinuity or non sequitur in the reasoning. It may be that the decision has an element of arbitrariness about it because there is an absence of reason to explain the discretionary choices made by the [regulator] in arriving at its conclusion. 52 The Tribunal repeated and applied this statement in ElectraNet (No 3) at [65] and in Energy Australia at [66]-[67]. 53 In ActewAGL, the Tribunal said that a decision will be unreasonable if it is arbitrary or capricious. It said at [35]: It is, we think, neither possible nor necessary to give an exhaustive definition of what is an unreasonable decision. At one extreme a decision that is arbitrary or capricious will plainly be unreasonable. At the other extreme, it will not be sufficient merely to reach a different decision to the first instance decision maker; in many areas reasonable persons can perfectly reasonably come to opposite conclusions. 54 A decision which is not determined by reference to the applicable criteria in the NGL or the NGR is likely to be unreasonable in all the circumstances: Energy Australia at [68]. A failure to take into account a matter which is required to be considered or consideration of a matter which is irrelevant may also give rise to a decision which is unreasonable: ActewAGL at [35]. 55 The unreasonableness ground in subsection 246(1)(d) and the incorrect exercise of discretion ground in subsection 246(1)(c) overlap to a certain extent: Energy Australia at [67] and [68]. For example, if the reasons for a decision contain logical error or an unexplained discretionary choice made in reaching a conclusion, then the decision may well be unreasonable: Energy Australia at [67]; see also ACCC v ACT at [178].

If a Ground of Review is Established 56 If the Tribunal is satisfied that a ground of review is established, the first matter that the Tribunal may consider is whether to allow new information or material to be submitted: NGL s 261(3). The Tribunal may allow new information to be submitted if the new information would assist the Tribunal and was not unreasonably withheld from the AER when it was making its decision. 57 Once the Tribunal has had an opportunity to consider any new material and the parties' submissions with respect to that material, the Tribunal must make a determination. Section 259(2) of the NGL provides the Tribunal with four options - it may: 1. affirm the AER's decision; 2. set aside the AER's decision; 3. vary the AER's decision; or 4. remit the matter to the AER to reconsider the matter in accordance with any direction or recommendation the Tribunal considers appropriate.

Debt Risk Premium 58 The first issue in this matter is the appropriate rate of return on debt to be used in determining Allgas's allowable revenue. Rule 72(1)(g) of the NGR provides that one of the items of information that must be provided to the AER by the service provider is the proposed rate of return, the assumptions on which the rate of return is calculated and a demonstration of how it is calculated. 59 The rate of return on capital is to be commensurate with the prevailing conditions in the market for funds and the risks involved in providing reference services: NGR r 87(1). In determining the rate of return on capital it is to be assumed that the service provider meets benchmark levels of efficiency and uses a financial structure that meets benchmark standards as to gearing and other financial parameters for a going concern and reflects in other respects commercial best practice: NGR r 87(2)(a). Further, a well-accepted approach that incorporates the cost of equity and debt, such as the Weighted Average Cost of Capital (WACC) and a well-accepted financial model, such as the Capital Asset Pricing Model, are to be used: NGR r 87(2)(b). 60 There was no disagreement between the parties as to the appropriate formula or model to be used in calculating the return on capital. Both Allgas and the AER accepted that the appropriate model for determining the return on capital was a nominal vanilla WACC, a weighted average of the pre-tax cost of debt and the post-tax cost of equity, calculated as follows: where: is the value of debt as a proportion of the value of equity and debt; is the nominal risk-free rate; DRP is the debt risk premium; is the equity beta; and MRP is the market risk premium. 61 The DRP is the margin above the nominal risk free rate that a debt holder would require for it to invest in a benchmark efficient service provider. This is determined by subtracting the yield on ten year Commonwealth Government bonds (the nominal risk free rate) from the yield payable on a reference bond. It is accepted by both parties that the relevant reference bond is an Australian issued BBB+ bond with a ten year maturity, issued by an Australian company. 62 Financial services provider Bloomberg publishes what is known as a fair value curve. A fair value curve plots estimates of bond yields against terms to maturity for a given credit rating. Relevantly, Bloomberg publishes a BBB fair value curve (taking into account BBB+, BBB and BBB- rated corporate bonds) for periods up to seven years (the Bloomberg curve). The Bloomberg curve is derived using proprietary methodology from a sample of Australian corporate bonds. As at 16 March 2011, the sample comprised 18 bonds with maturities of less than six years. 63 As the benchmark bond in this context is a ten year bond, the Bloomberg curve must be extrapolated from seven years to 10. This was done by incorporating the change in spread between Bloomberg's AAA rated estimates from seven to ten years into the Bloomberg BBB fair value curve, averaged over the 20 trading days ending 22 June 2010. The result of this methodology, uncontroversial between the parties, is referred to by the Tribunal as the extrapolated Bloomberg value (EBV). 64 Allgas proposed placing sole reliance on the EBV. That is, Allgas submitted that the point estimate of the Bloomberg curve extrapolated to ten years should be the sole determinant of the DRP. 65 The AER did not agree, however, and stated that only a 50% weighing should be given to the EBV, with the other 50% weight being placed on the yield observed on a BBB rated bond issued by APA Group with a ten year term maturing in July 2020 (the APA bond). 66 In simple terms, Allgas submits that the DRP should be determined solely by reference to the EBV of 18 bonds with maturities up to six years, whereas the AER submits that it should be determined by reference to an average of a single ten year bond selected by it and the EBV. The AER's averaging resulted in it adopting a value for the DRP of 3.64%. Allgas's proposed sole reliance on the EBV resulted in a value for the DRP of 4.37%.

Rejection of the sole use of the extrapolated Bloomberg value (EBV) 67 In its original access arrangement proposal, Allgas proposed that the DRP should be calculated by taking an average of the value arrived at from the extrapolated Bloomberg curve and a value arrived at from a similar curve published by the Commonwealth Bank of Australia (the CBA Spectrum Curve). The CBA Spectrum Curve was last published on or about 8 September 2010. Allgas relied on an expert report by Synergies Economic Consulting Pty Ltd to support this proposal. 68 In the Draft Decision, the AER rejected this approach and instead based the estimate of the DRP on an average of the EBV and the APA bond. At the time of this decision, the AER was able to observe the yields for three long-dated bonds (of up to 11 years maturity), including the APA bond. The AER believed that this sample provided a basis for rejecting the use of the EBV alone and supported the use of an average of the EBV and the single ABA bond. 69 In response to the Draft Decision, Allgas submitted its Revised Access Arrangement Proposal, in which it proposed an estimate of the DRP based solely on the EBV. In support of this proposal Allgas submitted an expert report prepared by Australia Ratings. 70 After publishing the Draft Decision, the AER was able to observe yields on four additional long-dated bonds, bringing the number of bonds additional to those contained on the extrapolated Bloomberg curve to seven bonds. The AER believed that these yields supported placing greater weight on the APA bond at the expense of the EBV and wrote to Allgas on 23 May 2011 proposing to estimate the DRP on the basis of placing a 70% weight on the APA bond and a 30% weight on the Bloomberg curve (the May 23 letter). 71 Allgas responded to the May 23 letter by reiterating its submissions in support of placing sole reliance on the EBV and urging the AER not to place any greater weight on the APA bond. 72 In responding to the AER's May 23 letter, Allgas relied on an expert report prepared by Competition Economists Group, referred to by its own shorthand way as CEG. Its report, authored by Dr T Hird, criticised the AER's approach to determining the reliability of the Bloomberg curve. In particular, the report suggested that the AER's analysis was flawed in that it excluded several long-dated bonds that were relevant to establishing the reliability of the EBV. The yields on these bonds, in CEG's analysis, supported sole reliance on the EBV. The Final Decision rejected sole reliance on the EBV and instead estimated the DRP on the basis of a simple average of the EBV and the APA bond. 73 It was accepted by both Allgas and the AER, and is clearly correct, that it is appropriate for the AER to investigate the reliability and accuracy of any index that is advanced as the preferred method for determining the DRP. Indeed, such an investigation should be undertaken with respect to any method put forward as the basis for calculating a critical value in determining the WACC or any other component of regulated revenue. 74 The issue in this matter, then, is not whether the AER erred in not accepting sole reliance on the EBV, but whether it strayed into reviewable error in the process of the review it undertook or the conclusions it came to.

The Bloomberg curve's historic performance 75 In the Final Decision, the AER placed significant emphasis on the behaviour of the Bloomberg curve since the onset of the global financial crisis (GFC). The AER claimed that this behaviour was "somewhat counterintuitive" because the current yields implied by the Bloomberg curve were higher at the time of the Final Decision than during much of the GFC, suggesting an increased perception of risk. The AER viewed this as counterintuitive on the basis of "substantial evidence" that indicated that debt market conditions had improved significantly since the GFC. 76 Allgas responded to this point, in submissions before the Tribunal, by pointing out that this analysis assumes that risk was appropriately priced and estimated prior to and during the GFC. Further, as Dr Hird pointed out in CEG's response to the May 23 letter, the world's debt markets were significantly restricted following September 2008, leading to a paucity of information on corporate bond yields in the year or so following. Once trading information again became available towards the end of 2009 it was not surprising to see bond yields increase. Dr Hird argued that the Bloomberg fair value estimates have been relatively stable from that time until the date of his report. 77 It is also submitted for Allgas that the evidence before the AER illustrated that investors' views about the appropriate level of compensation for exposure to risk have changed and the financial regulatory environment has also changed. In addition, Allgas submits that it is notorious that one of the causes of the GFC was a failure to correctly evaluate risk. 78 The AER submitted that its conclusion that the Bloomberg curve's performance was counterintuitive was not material to its decision to reject sole reliance being placed on this curve, as its decision was based primarily on the basis of a comparison of the EBV with long dated bonds. In addition, the AER maintained that there was significant evidence that market risk was at its peak during the GFC and has subsided since. 79 It is clear from the Final Decision that the AER did not rely on its conclusion about counterintuitive behaviour in deciding to reject the sole use of the EBV. For this reason, it does not constitute a material error of fact and Allgas has not made out a ground of review on that respect. It does appear, however, as noted in [79] below, that the AER paid insufficient regard to the reasons behind the Bloomberg curve's performance during the GFC. Little regard was had to the question of whether this performance was likely to provide an accurate assessment of the Bloomberg curve's current reliability: see Application by Jemena Gas Works (NSW) Ltd (No 5) [2011] ACompT 10 (Jemena No 5) at [66]. 80 If the Bloomberg curve were to display poor performance during market conditions that can be regarded as 'normal' or likely to be repeated, then there will be strong evidence for questioning its usefulness. The reasons for poor performance are crucial to an understanding of whether such performance should be taken as indicative of the Bloomberg curve's general accuracy. The AER did not, in either the Final Decision or the Draft Decision undertake the analysis necessary to form a proper understanding of this question. It intuitively considered that, for the purposes of its Access Arrangement Decision, the Bloomberg curve was not a directly reliable source of information. It is necessary to consider carefully where the AER went from that step. 81 At this point, it is sufficient for the Tribunal to express the view that the performance of the Bloomberg curve during and after the GFC alone would not necessarily have warranted its rejection. The unusual circumstances and market conditions, in particular the restriction of the debt market, that prevailed during the GFC are unlikely to persist for extended periods and might not therefore be viewed as indicative of the likely market conditions that would prevail during the majority of the ten year reference period. At most, the so called "counterintuitive" performance would warrant further investigation of the reliability of the Bloomberg curve.

Comparison of the Bloomberg curve with long-dated bonds 82 The AER placed great emphasis on the fact that, in determining the implied fair value yield at ten years using Bloomberg data, an extrapolated curve was being used. That is, the information was not published by Bloomberg as part of the BBB fair value curve. The method of extrapolation was, however, not in contention between the parties. 83 The AER's submissions suggest that, because the EBV is an extrapolation at the 10-year point, greater scrutiny should be applied to its reliability. This is no doubt correct. Allgas agreed with the proposition that it is appropriate for the AER to investigate the reliability of any measure it is using to determine the DRP. 84 The AER's analysis of the reliability of the Bloomberg curve in the Final Decision rested on a comparison of an estimated yield derived from the Bloomberg curve with a sample of long-dated bonds. In particular, the AER relied on the following figure which depicts an extrapolation from 31 May 2018 to 31 May 2021: Figure 1: Australian corporate bonds with maturities greater than five years and credit ratings from BBB to A- (Final Decision: Figure A.5) 85 What becomes immediately obvious on viewing the AER's preferred depiction of the data is that a significant proportion of the plotted bonds lie beneath the Bloomberg curve and, in particular, below its extrapolated portion. At first blush this may suggest that the Bloomberg curve, at least the extrapolated portion, overstates true yields. 86 The problem with this view, however, is that it ignores the generally held proposition that yield curves slope upwards, a view accepted by both parties. For the Bloomberg curve to accurately reflect the yields displayed by the sample of long-dated bonds preferred by the AER it would be required to slope downward from the point at which extrapolation commences. The required change in direction becomes even more apparent when Allgas's preferred figure is inspected, but ignoring the additional bond sample: Figure 2: Bloomberg BBB Fair Value Curve extrapolated to 15 years and reported yields (CEG's response to the May 23 letter). 87 The additional bond sample comprises the three bonds labelled "SUNCORP", the bond labelled "DBCT" at the far right of the figure, the two bonds labelled "VERO", the two bonds labelled "BKQLD" and the bond labelled "AMP". These bonds were not included in the AER's preferred figure, being Figure 1 above. 88 Rather than casting doubt on the reliability of the extrapolation, the AER's analysis, if accepted, would cast doubt on the overall reliability of the Bloomberg curve at all, as a reflection of Australian corporate bond yields. Perhaps a more plausible explanation for the apparent incongruity between the Bloomberg curve and the reported yields on the AER's sample of long-dated bonds is that the sample of bonds used by the AER is not complete or representative. The AER's view that the Bloomberg curve overstated relevant bond yields may have been one that was open to it had the only available evidence been the sample it selected. This was not, however, the case. 89 As observed in [71] above, in response to the May 23 letter, Allgas relied on an expert report from CEG (this report is the source of Figure 2 above), which identified additional bonds that, in CEG's analysis, should be included in the comparison sample. The AER claimed that it did not have sufficient time to fully analyse CEG's response to the May 23 letter before publication of the Final Decision. 90 If these additional bonds are included in the sample, the extrapolated Bloomberg curve over the entire 15-year period appears to be much more reliable in terms of being representative of the whole sample of bonds. In particular, several of the new bonds lie substantially above the Bloomberg curve, at periods to maturity greater than 10 years. 91 The new sample of bonds arose, in large part, because Dr Hird, the CEG Report's author, changed the methodology he used to determine the maturity date of callable bonds. Previously he had used the first call date as the appropriate date of maturity. For the later report, Dr Hird instead used the final call, or final maturity, date in circumstances in which the call option is unlikely to be exercised. The AER in submissions through counsel on the review accepted the correctness of this approach, but had several criticisms of the inclusion of these bonds in the sample, at least in their current form. 92 In ActewAGL at [39], the Tribunal said that, when using a sample of bonds to check the reliability of a fair value curve, the sample should be as large as possible. In ActewAGL, the Tribunal was considering the basis on which the AER attempted to decide which of three fair value curves was most representative. In that case, the Tribunal stated that a sample of five bonds was insufficient, particularly when these bonds only covered half of the relevant term to maturity: ActewAGL at [39]. In the present matter the AER used a sample of seven bonds to reach its conclusion in the Final Decision. These bonds were, however, bunched tightly across time. Little consideration was given to the relevance of shorter or longer dated bonds in testing the overall validity of the Bloomberg curve or the representativeness of the APA bond. While it may be accepted that bonds with a maturity of approximately 10 years may be particularly relevant, they should not be considered to the exclusion of all others in deriving a representative estimate of the DRP. 93 The inclusion or otherwise of the additional bonds put forward by the CEG report may be of crucial importance in determining whether the Bloomberg curve provides a reliable estimate of 10 year BBB+ corporate bond yields. It was not an appropriate response by the AER to publish the Final Decision without considering, and if appropriate, taking into account the suggestions made by CEG. 94 For these reasons, a detailed analysis as to whether inclusion of the bonds proposed by CEG should have been undertaken. If, as the AER submits, there are anomalies with the yields used, suitable adjustments might be required to be made. Likewise, any adjustments to remove the value of the call options might also be necessary. 95 In the Final Decision the AER said: In the limited timeframe available to assess CEG's proposal, the AER has been unable to adequately verify the reasonableness of CEG's changed methodology. Regardless, the AER considers that the additional bonds noted by CEG are immaterial for this final decision. 96 The AER did not therefore investigate or methodically analyse the validity of CEG's proposal. The AER's analysis only extended to noting that bonds issued by financial institutions often have higher yields than those issued by infrastructure service providers and that several of the bonds in CEG's proposed sample were subordinated debt. 97 As Allgas correctly submitted, the nature of the debt, that is subordinated or unsubordinated, and the industry of the issuer should be taken into account in the determination of the bond's credit rating. Similarly, the industry of the issuer is not relevant within the current structure of the AER's process. If the AER is to continue to use BBB+ rated corporate debt as its benchmark for determining the DRP, it is not reasonable for it to pick and choose which of the BBB+ bonds it deems to be appropriate without considering the significance of the other potentially relevant bonds. The analysis of all potentially relevant bonds as assessed by Bloomberg produces the EBV. If the AER were to decide that the EBV was an unreliable indicator for the purposes of deciding that DRP, it would be desirable in the longer term to develop an alternative coherent and consistent methodology, in consultation with the relevant regulated entities and other interested parties. Although the DRP must be determined at a particular point in time, the use of a consistent and acceptable methodology would ensure regulatory consistency, and in relation to particular matters would also facilitate efficient decision making and in turn perhaps reduce the number of reviews of the DRP decisions by the AER brought to the Tribunal. While such a task would be a complex and lengthy one, it is one the Tribunal commends to the AER. 98 However, given the time constraints upon its decision making, that was not an option available to the AER in this matter. It is necessary to determine whether the AER committed reviewable error in the manner of selection of the DRP on the material available to it, and upon which it was obliged to make its decision. 99 Focusing on the industry of the issuer and being selective in choosing which BBB+ rated corporate debt to use for determining the DRP is a novel technique by the AER. This would require a different approach from the AER and different submissions from service providers. 100 Rule 62(1) of the NGR requires the AER to consider all submissions made in response to the Draft Decision before making the Final Decision, as well as any other matters it thinks relevant. While, strictly, the May 23 letter did not form part of the Draft Decision, it was a significant course of action for the AER to propose a substantial variation to its approach from the Draft Decision without giving Allgas much of an opportunity to respond and, having received the response in a timely manner, in the circumstances explained by the AER, fail to give Allgas' response thorough consideration. 101 The Tribunal appreciates the constraints of the timetable to be observed by the AER and the risk that a regulated entity may 'game' the timetable. But here, the timing issue was one of the AER's own making. Indeed, this was an issue that could have been raised and addressed by the AER so that it did have sufficient time to fully analyse CEG's response to the May 23 letter. 102 The AER appears to have reached a hybrid position. It properly decided to review the reliability of the EBV. However, in doing so, it selectively relied on the APA bond. For the reasons given, the Tribunal considers that that was not appropriate. There had been identified to the AER a range of other bonds, some of which lay below the EBV and some above the EBV. Had the AER considered them, its caution about the limited use of the EBV may have been resolved. The hybrid position emerges from the fact that the AER nevertheless decided to rely on the EBV as one of the two significant inputs into its weighting process. It must therefore have regarded the EBV as relevant and meaningful. 103 The AER, having rejected placing sole reliance on the EBV, decided to use the APA bond as a component of an averaging process. After first considering placing a 70% weight on the APA bond it decided to determine the DRP based on an average of the APA bond and the EBV, giving a 50% weight to each. 104 The choice of the APA bond and the weighting applied to it are attended by the same error as the decision to reject the sole use of the EBV, namely the failure to have sufficient regard to the expert report of CEG. As noted, this report was prepared in response to the AER's proposal to place a 70% weight on the APA bond. No consideration appears to have been given to the question of whether the APA bond has been used by the market, broadly defined, as a reference for determining acceptable yields for bonds with similar credit ratings and terms to maturity. Indeed, no consideration has been given to the question of whether it is accepted practice in the market to estimate reference yields on the basis of a single bond. The lack of consideration of this issue runs counter to the emphasis placed on market use of, and reliance on, evaluative techniques in ActewAGL at [78] and reiterated in Jemena (No 5) at [64]. 105 The manner in which the weightings in the averaging process were assigned was also in error. As was said in ActewAGL and Jemena (No 5), the process of averaging needs to be given significant consideration and the allocation of weights should be done on a scientific basis. The Tribunal noted, in discussing the processing of averaging two fair value curves in Jemena (No 5) that: An average is a blunt instrument unless careful thought is given to the individual components and whether each should be given the same consideration, or weight, in the calculation of the average. A simple unweighted average gives each component the same weight. This will not always be appropriate, especially where (as here) the two fair value curves differ considerably over the relevant periods to maturity. 106 This point is apposite in the current matter. There is a substantial difference between the DRP implied by the EBV and that implied by the APA bond. To take a simple average of these, without consideration of the methodology behind the Bloomberg curve, or the relevance of selecting just the APA bond from the whole sample of bonds (including longer and shorter dated bonds) is to use a very heavy "blunt instrument" when a more nuanced, rigorous and sophisticated method is required. 107 The sole basis for the assignment of weights in this matter was the AER's opinion that the APA bond and the EBV (based as it is on the yields of 18 bonds) were each equally reliable as indicators of the benchmark DRP. No explicit consideration appears to have been given to the fact that the APA bond was being used as a quasi-proxy for the other bonds the AER deemed as appropriate comparators, nor is there any consideration given to the nature of the Bloomberg curve, which is derived from a sample of 18 bonds. 108 The placing of equal weight on the APA bond and the EBV estimate effectively assigns a 50% weight to one bond, the APA bond, and a 2.8% weight to each of the 18 bonds comprising the Bloomberg curve sample. The AER has not provided any analysis, or demonstrated any consideration, of whether this division of weights is correct. 109 The choice of weights in the averaging process is of vital importance. It should not be undertaken in an arbitrary or haphazard manner. Any decision on weighting must have a sound and reasoned logic. 110 In the view of the Tribunal, the decision to reject the adoption of the EBV on the basis of the APA bond and the weighting chosen amounts to reviewable error on the part of the AER. The reviewable grounds invoked by Allgas are primarily under s 246(1)(c) and (d), although it also asserts material errors of fact so as to enliven s 246(1)(a) and (b). In the view of the Tribunal, the AER in the circumstances made a decision which was unreasonable by adopting a rate of return for the DRP based upon the simple averaging of the EBV and the APA bond. There may be reasons why some or all of the bonds referred to in Allgas' response to the May 23 letter are inappropriate to be considered. However, the AER regarded those bonds as "immaterial" without sufficient grounds for doing so. Given the common understanding of both Allgas and the AER as to the nature of the bonds considered in the formulation of the Bloomberg curve (that is, before its extrapolation), there was no apparent reason to exclude those bonds because they were issued by financial institutions. That is not how the AER had proceeded in the past, when averaging the EBV and the extrapolated CBA Spectrum value. That step on the part of the AER, in the Tribunal's view, rendered its reviewable decision unreasonable because the consequences of its error were obviously of significant magnitude. The significance of the different approach is addressed in the reasons for decision of the Tribunal giving leave to Allgas to apply to review the AER reviewable decision: Application by APT Allgas Energy Pty Ltd [2011] ACompT 11. It is also an error capable of being expressed within the terms of s 246(1)(c), that is as an incorrect exercise of the AER's discretion in selecting the rate of return for the DRP as it did without considering the increased bond sample proposed by Allgas, particularly as the AER by the 23 May letter invited comment on a variation of the weighting proposed in the Draft Decision. 111 Thus, Allgas has successfully made out a ground of review: NGL s246(1)(c) and (d). Accordingly, the Tribunal must decide whether to vary the AER's decision, affirm the AER's decision or remit the matter back to the AER for further consideration: NGL s 259(2). Before doing so, there are a couple of other matters to be mentioned.

Cross checking 112 The AER undertook "cross-checks" of its estimates with actual cost of debt data. It was submitted for Allgas that this was inappropriate. 113 The regulatory regime, in determining allowable revenue, is structured on the basis of attempting to simulate a benchmark efficient service provider. Recourse to the actual cost of debt, in seeking to defend the reasonableness of the decision, is inappropriate in this context. The reasonableness or otherwise of a component of allowable revenue must be determined on the basis of the factors set out in the NGR.

Conclusion with respect to DRP 114 The Tribunal, of course, accepts that in the first instance it is for the AER to determine whether to rely upon the Bloomberg curve, or to accept the extrapolation of that curve in the manner done in the past. It is not obliged to do so, although given the past regulatory decisions it may be expected to do so unless there were sound reasons to depart from that practice. For the future, that is a matter for the AER. 115 In the longer term, as the Tribunal has said, it is open to the AER to adopt a different methodology. Consideration of the proper composition of the comparison sample of bonds, the methodology for deciding on the appropriate sample of bonds and the relevance of these bonds to its task should be undertaken by the AER in consultation with interested parties across the spectrum of entities in the industries it regulates, consumers of their services and other interested parties. 116 In this matter, for the reasons given, the AER was obliged to do the best it could on the information available. Having determined that the AER fell into reviewable error, the Tribunal may set aside or vary the Access Arrangement Decision, or it may remit the matter to the AER to make the decision again: s 259(2). If the Tribunal sets aside or varies the decision, it may perform all of the functions and exercise all of the powers of the AER: s 259(3). If the Tribunal considers remitting the matter, it must have regard to the nature and relative complexities of the Access Arrangement Decision, and the matter the subject of the review: s 259(4). 117 The Tribunal has decided to vary the Access Arrangement Decision by substituting for the DRP value determined by the AER a DRP value of 4.37% based upon the EBV. 118 The Tribunal has taken into account that the AER, in the course of considering any proposed access arrangement or its revision, is obliged to address a multitude of issues and ultimately, it is only those few selected by the regulated entity which may come before the Tribunal; the majority of the issues resolved by the AER are resolved to the satisfaction of the regulated entity. That is the case in this matter. The Tribunal is reluctant to remit the matter to the AER to make the decision again, even constrained by any directions or recommendations of the Tribunal. In this matter, as the Tribunal has noted, there will be no real opportunity for the AER to develop a coherent alternative methodology to determine the DRP in the time available, so the AER would be forced to make the best decision it could on the material available if the matter were remitted to it. The more substantial task of developing an alternative methodology would be time consuming and complex, and necessarily be one which to a degree at least would not be specific to the parties but affect other regulated entities. 119 The Tribunal has also taken into account that the AER regarded the EBV as having sufficient reliability to give it substantial weight. In making the decision about whether to discount it in some way, the material available to the AER gave it no clear economic path to follow. 120 Allgas provided to the AER strong evidence in support of the EBV, in particular by its response to the May 23 letter. The view of Dr Hird of CEG was that that material did not demonstrate any basis for the substitution of an alternative estimate for the EBV. As noted, the AER itself accepted the relevance of the EBV. Whilst the Tribunal accepts that the AER properly considered the reliability of the EBV, it has reached the view on the available material that there is no reason shown from the available material why the use of the EBV should not be adopted in this particular matter. There is no viable alternative methodology at present, other than making a decision on all the material. The observations of the Tribunal in ActewAGL at [74]-[78] suggest also that, on the existing material, it is appropriate to vary the decision in the manner indicated.

conclusion 121 The AER's decision to reject sole reliance on the EBV and to determine the DRP based on an average of the APA bond and the EBV amounts to reviewable error. For the reasons given above, the Tribunal proposes to vary the Access Arrangement Decision by substituting for the DRP rate there specified the rate of 4.37%. I certify that the preceding one hundred and twenty-one (121) numbered paragraphs are a true copy of the Reasons for Decision herein of the Honourable Justice Mansfield (President), Mr R Davey and Professor D Round.