AUSTRALIAN COMPETITION TRIBUNAL

Application by SA Power Networks [2016] ACompT 11

IN THE AUSTRALIAN COMPETITION TRIBUNAL

THE TRIBUNAL DETERMINES THAT:

1.    The Final Decision: SA Power Networks Distribution Determination 2015-16 to 2019-20, including attachments, is affirmed.

REASONS FOR DETERMINATION

THE TRIBUNAL:

INTRODUCTION

1    This is an application by SA Power Networks (‘SAPN’), pursuant to s 71B(1) of the Schedule of the National Electricity (South Australia) Act 1996 (the ‘NESA Act’), for review of a reviewable regulatory decision of the Australian Energy Regulator (‘AER’), filed with the Australian Competition Tribunal (the ‘Tribunal’) on 19 November 2015.

2    The decision under review is the final decision and distribution determination published by the AER on 29 October 2015 pursuant to cl 11.60.4(c) of the NER, entitled ‘Final Decision: SA Power Networks Distribution Determination 2015-16 to 2019-20’ (‘Final Decision’).

3    On 4 May 2016, SAPN was granted leave to apply for review of the Final Decision, with respect to the designated grounds of review referred to in an amended application for review filed with the Tribunal on 3 May 2016 (‘Review Application’). The Tribunal’s reasons for granting leave are discussed below.

Regulatory regime

4    SAPN is the operator of an electricity distribution network located in South Australia (‘South Australian distribution network’). SAPN is registered as a distribution network service provider (‘DNSP’) under cl 2.5.1 of the National Electricity Rules (‘NER’), and provides distribution network services by means of the South Australian distribution network.

5    Section 6 of the NESA Act applies the National Electricity Law (‘NEL’), set out in the Schedule of the Act, as a law of South Australia. Section 9 of the NEL gives the NER the force of law in South Australia.

6    Section 7 of the NEL sets out the national electricity objective (‘NEO’), which is to promote efficient investment in, and efficient operation and use of, electricity services for the long-term interests of consumers with respect to:

(1)    price, quality, safety, reliability and security of supply of electricity; and

(2)    the reliability, safety and security of the national electricity system.

7    Section 16(1)(a) of the NEL provides that in performing or exercising an AER economic regulatory function or power, which includes a function or power performed or exercised by the AER relating to the making of a distribution determination, the AER must perform or exercise that function or power in a manner that will, or is likely to, contribute to the achievement of the NEO.

8    Section 16(1)(d) of the NEL provides that if the AER is making a reviewable regulatory decision and there are two or more possible reviewable regulatory decisions that will, or are likely to, contribute to the achievement of the NEO, the AER must:

(1)    make the decision that the AER is satisfied will, or is likely to, contribute to the achievement of the NEO to the greatest degree (the ‘preferable reviewable regulatory decision’); and

(2)    specify reasons as to the basis on which the AER is satisfied that the decision is the preferable reviewable regulatory decision.

9    In addition, s 16(2)(a) of the NEL requires the AER to take into account the revenue and pricing principles (‘RPP’) when exercising a discretion in making those parts of a distribution determination relating to direct control network services.

10    Section 7A of the NEL sets out the RPP, as follows:

(2)    A regulated network service provider should be provided with a reasonable opportunity to recover at least the efficient costs the operator incurs in–

(a)    providing direct control network services; and

(b)    complying with a regulatory obligation or requirement or making a regulatory payment.

(3)    A regulated network service provider should be provided with effective incentives in order to promote economic efficiency with respect to direct control network services the operator provides. The economic efficiency that should be promoted includes–

(a)    efficient investment in a distribution system or transmission system with which the operator provides direct control network services; and

(b)    the efficient provision of electricity network services; and

(c)    the efficient use of the distribution system or transmission system with which the operator provides direct control network services.

(4)    Regard should be had to the regulatory asset base with respect to a distribution system or transmission system adopted–

(a)    in any previous–

(i)    as the case requires, distribution determination or transmission determination; or

(ii)    determination or decision under the National Electricity Code or jurisdictional electricity legislation regulating the revenue earned, or prices charged, by a person providing services by means of that distribution system or transmission system; or

(b)    in the Rules.

(5)    A price or charge for the provision of a direct control network service should allow for a return commensurate with the regulatory and commercial risks involved in providing the direct control network service to which that price or charge relates.

(6)    Regard should be had to the economic costs and risks of the potential for under and over investment by a regulated network service provider in, as the case requires, a distribution system or transmission system with which the operator provides direct control network services.

(7)    Regard should be had to the economic costs and risks of the potential for under and over utilisation of a distribution system or transmission system with which a regulated network service provider provides direct control network services.

11    Clause 6.12.1 of the NER provides that a distribution determination is predicated on a number of constituent decisions to be made by the AER. These include:

(1)    a decision on the DNSP’s building block proposal in which the AER either approves or refuses to approve the annual revenue requirement for the DNSP, as set out in the building block proposal, for each regulatory year of the regulatory control period;

(2)    a decision in which the AER either:

(a)    acting in accordance with cl 6.5.7(c), accepts the total of the forecast capital expenditure (‘capex’) for the regulatory control period that is included in the current building block proposal; or

(b)    acting in accordance with cl 6.5.7(d), does not accept the total of the forecast capex for the regulatory control period that is included in the current building block proposal, in which case the AER must set out its reasons for that decision and an estimate of the total of the DNSP’s required capex for the regulatory control period that the AER is satisfied reasonably reflects the capex criteria, taking into account the capex factors;

(3)    a decision on the allowed rate of return for each regulatory year of the regulatory control period in accordance with cl 6.5.2;

(4)    a decision on the value of imputation credits as referred to in cl 6.5.3; and

(5)    any other amounts, values or inputs on which the building block determination is based.

12    Clause 6.12.2 of the NER provides that the reasons given by the AER for a final distribution determination under r 6.11 must set out the basis and rationale of the determination, including:

(1)    details of the qualitative and quantitative methods applied in any calculations and formulae made or used by the AER;

(2)    the values adopted by the AER for each of the input variables in any calculations and formulae, including:

(a)    whether those values have been taken or derived from the provider’s current building block proposal; and

(b)    if not, the rationale for the adoption of those values;

(3)    details of any assumptions made by the AER in undertaking any material qualitative and quantitative analyses; and

(4)    reasons for the making of any decisions, the giving or withholding of any approvals, and the exercise of any discretion, as referred to in Ch 6 of the NER, for the purposes of the determination.

13    The exercise of the AER’s discretion in making distribution determinations is governed by cl 6.12.3 of the NER.

AER distribution determination process

Impact of the 2012 changes to the NER

14    Section 15 of the NEL prescribes the functions and powers of the AER, which include the making of distribution determinations under the NER.

15    Certain provisions of the NER relating to distribution determinations were amended by the Australian Energy Market Commission (‘AEMC’), and were published, and took effect from on 29 November 2012 (‘2012 rule changes’).

16    A number of revenue determination processes due to be completed in 2014 and 2015 were delayed through transitional provisions inserted into the NER as part of the 2012 rule changes (‘transitional provisions’). This delay afforded the AER the time to develop ‘Better Regulation’ guidelines in response to the 2012 rule changes (the ‘Better Regulation Guidelines’). The Better Regulation Guidelines were published in late 2013.

17    Key aspects of the transitional provisions applying to SAPN’s distribution determination process are as follows:

(1)    the date for submission of SAPN’s regulatory proposal was delayed by nine months, from 31 January 2014 to 31 October 2014;

(2)    the AER was not required to publish a draft distribution determination in relation to SAPN’s regulatory proposal, as it would have been required to do under cl 6.10.1 of the NER, in the absence of the transitional arrangements. Rather, following appropriate consultation on SAPN’s regulatory proposal, the AER was able to proceed immediately to making a distribution determination under cl 6.11.1 of the NER;

(3)    at the same time as the AER published its distribution determination under cl 6.11.1 of the NER (which it was required to by 30 April 2015), it was required to also publish an invitation for written submissions on the revocation and substitution of that distribution determination;

(4)    under cl 11.60.4(b) of the NER:

(a)    any person was allowed to make a written submission to the AER in relation to the revocation and substitution of the distribution determination within the time period specified by the AER, which must not be earlier than 45 business days after the making of that distribution determination; and

(b)    without otherwise limiting the manner in which SAPN’s may make such submissions, SAPN was allowed to make a submission in the form of revisions to the regulatory proposal that it submitted to the AER in relation to the distribution determination that was to be revoked.

(5)    by no later than 31 October 2015, the AER was required to revoke its distribution determination for the 2015-2020 regulatory control period and make a new distribution determination in substitution for the revoked determination which takes effect as at the date it is made; and

(6)    in making the new distribution determination, the AER was required to have regard to:

(a)    the matters it would be required to have regard to if it were making a final distribution determination under “current Chapter 6” subsequent to it making a draft distribution determination that is the same as the revoked determination including (except where (c) below applies) the regulatory proposal that was submitted to the AER in relation to the revoked determination;

(b)    written submissions received in relation to the revocation and substitution of the distribution determination;

(c)    any revisions to the regulatory proposal that was submitted to the AER in relation to the revoked determination and that are given to the AER under cl 11.60.4(b); and

(d)    any analysis undertaken by or for the AER that is published prior to the making of the distribution determination or as part of the distribution determination.

SAPN’s regulatory review process

18    Pursuant to cl 6.8.2 of the NER (as modified by cl 11.60.3 of the transitional provisions), SAPN was required to submit, and on 31 October 2014 did submit, a regulatory proposal to the AER for consideration in accordance with the NER (‘Regulatory Proposal’).

19    On 30 April 2015, pursuant to cl 6.11.1 of the NER, the AER published a decision and distribution determination entitled ‘Preliminary Decision: SA Power Networks Distribution Determination 2015-16 to 2019-20’ (‘Preliminary Decision’).

20    On 3 July 2015, pursuant to cl 11.60.4(b) of the NER, SAPN submitted a revised regulatory proposal to the AER for consideration in accordance with the NER (‘Revised Proposal’).

21    On 29 October 2015, the AER published its Final Decision, pursuant to cl 11.60.4(c) of the NER.

22    On 19 November 2015, SAPN filed an application for leave and review under s 71B of the NEL.

LEAVE TO APPLY FOR REVIEW

23    Section 71B of the NEL provides:

71B – Applications for review

(1)     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.

(2)     An application must—

(a)     be made in the form and manner determined by the Tribunal; and

(b)     specify the grounds for review being relied on.

24    The Tribunal accepts that, for the purposes of s 71B(1) of the NEL:

(1)    SAPN is an “affected or interested person or body” (as that term is defined in s 71A of the NEL); and

(2)    the Final Decision, the decision that is under review in the present application, is a reviewable regulatory decision, as it is a network revenue or pricing determination that sets a regulatory period.

25    In addition, it is accepted that there is no issue that SAPN’s application for review was made within the 15 business-day time limit prescribed by s 71D of the NEL.

26    The application for leave to apply for review was unopposed by the AER.

27    As previously mentioned, on 4 May 2016, SAPN was given leave to apply for review of the Final Decision on the grounds set out in its Review Application.

28    It is worth noting that the South Australian Council of Social Service (‘SACOSS’) also applied for leave to review the Final Decision, however the Tribunal dismissed SACOSS’ application for review on 2 May 2016: see Application by South Australian Council of Social Service Incorporated [2016] ACompT 8.

Requirements for the leave application

29    Section 71C of the NEL sets out the following two relevant requirements for applications made under s 71B of the NEL:

71C – Grounds for review

(1)    An application under section 71B(1) may be made only on 1 or more of the following grounds:

(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 that 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;

(d)    the AER’s decision was unreasonable, having regard to all the circumstances.

(1a)     An application under section 71B(1) must also specify the manner in which a determination made by the Tribunal varying the reviewable regulatory decision, or setting aside the reviewable regulatory decision and a fresh decision being made by the AER following remission of the matter to the AER by the Tribunal, on the basis of 1 or more grounds raised in the application, either separately or collectively, would, or would be likely to, result in a materially preferable NEO decision.

30    It is worth repeating that, pursuant to s 7 of the NEL, the NEO is to promote efficient investment in, and efficient operation and use of, electricity services for the long-term interests of consumers with respect to:

(1)    price, quality, safety, reliability and security of supply of electricity; and

(2)    the reliability, safety and security of the national electricity system.

31    The Tribunal accepts that SAPN’s Review Application was made only on the grounds of review specified in s 71C of the NEL.

32    The Tribunal has reached that conclusion having regard to the observations made in recent Tribunal decisions relating to applications for limited merits review by DNSPs in NSW and ACT. For example, the Tribunal accepts that:

… the line between the several available grounds of review is not necessarily always clear cut. Sometimes, it will be a clear line, and sometimes it will not. Moreover, there is no prescription in s 71C that, in particular facts and circumstances, there can be only one ground of review made out …

There is also no clear line between factual error, opinion, and discretionary judgment; one may feed into the other.

… error or errors – if accepted – may be a combination of error or errors of fact, wrongful exercise of discretion, and/or the outcome of an unreasonable decision. Because the characterisation of error or errors, if made out, will more clearly emerge in the course of considering the review related material and the submissions dealing with it, the Tribunal does not consider it appropriate or necessary to embark upon the careful textual analysis and criticism of the [DNSP’s] application at this point to describe the combination or permutation of alternative expressions of reviewable error in that application.

See Applications by Public Interest Advocacy Centre Ltd, Ausgrid, Endeavour Energy and Essential Energy [2015] ACompT 2 at [55], [57]; Application by ActewAGL Distribution [2015] ACompT 3 at [21].

33    Further, the Tribunal has had regard to the matters outlined in SAPN’s submissions on its application for leave to review (‘Leave Submissions’) and paragraphs [22]-[26] and [200]-[210] of SAPN’s Review Application. In considering these matters, the Tribunal accepted, in granting leave to apply for review, that SAPN had sufficiently addressed the manner in which a determination by the Tribunal in accordance with s 71C(1a) of the NEL may result in a materially preferable designated NEO decision.

Criteria for the grant of leave

34    Pursuant to s 71E of the NEL, the Tribunal must not grant leave to apply for review unless it appears to the Tribunal that:

(a) that there is a serious issue to be heard and determined as to whether a ground for review set out in section 71C(1) exists; and

(b) that the applicant has established a prima facie case that a determination made by the Tribunal varying the reviewable regulatory decision, or setting aside the reviewable regulatory decision and a fresh decision being made by the AER following remission of the matter to the AER by the Tribunal, on the basis of 1 or more grounds raised in the application, either separately or collectively, would, or would be likely to, result in a materially preferable NEO decision.

35    However, even where the Tribunal determines that the criteria in s 71E are satisfied, leave to apply under s 71B must not be granted “unless the amount that is specified in or derived from the decision exceeds the lesser of $5,000,000 or 2% of the average annual regulated revenue of the regulated network service provider”: s 71F(2) of the NEL.

36    The Tribunal may also refuse to grant leave on the grounds set out in s 71H(2) of the NEL, namely that:

…

(a)    without reasonable excuse—

(i)    failed to comply with a request (including a request for relevant information), or a direction, of the AER made under this Law or the Rules for the purpose of making the decision; or

(ii)    conducted itself in a manner that resulted in the making of the decision of the AER being delayed; or

(b)    misled, or attempted to mislead, the AER on a matter relevant to the AER’s decision.

37    There is no suggestion, however, by the AER or SAPN that any of the conditions contained in that provision are present to enliven the Tribunal’s discretion to refuse the granting of leave.

38    Each of the relevant criteria for leave in s 71E and s 71F(2) of the NEL are considered below.

Financial threshold (s 71F(2) of the NEL)

39    The affidavit of Luke Woodward affirmed on 19 November 2015, and made in support of SAPN’s Review Application sets out the 2% of the average annual regulated revenue of SAPN over the 2015-2020 regulatory control period specified in or derived from the Final Decision as approximately $15.9 million. As such the relevant threshold pursuant to s 71F(2) is $5 million.

40    As previously held by the Tribunal in Application by Energex Ltd (No 4) [2011] ACompT 4 (‘Energex (No 4)’):

(1)    the “amount that is specified in or derived from the decision” should not be read literally as meaning the total revenue to be derived by the service provider, rather, the amount at issue in light of the grounds upon which the AER’s decision is challenged: Energex (No 4) at [50]; and

(2)    when determining whether the financial threshold in s 71F(2) of the NEL is satisfied, all of the errors are to be taken into account; the threshold does not need to be satisfied for each ground: Energex (No 4) at [52]; see also Application by Jemena Gas Networks (NSW) Ltd (No 2) [2011] ACompT 5 at [3]; Application by ATCO Gas Australia Pty Ltd [2015] ACompT 7 at [21] (‘ATCO 2015’).

41    SAPN’s Leave Submissions set out the relevant amounts specified in or derived from the decision in respect of each of the topics that are the subject of grounds for review in SAPN’s Review Application. On the basis of these amounts, the Tribunal considered that the financial threshold was comfortably satisfied.

Serious issue to be heard and determined (s 71E(a) of the NEL)

42    SAPN adopted the submissions of Ausgrid, Endeavour Energy, and Essential Energy in Applications by Public Interest Advocacy Centre Ltd, Ausgrid, Endeavour Energy and Essential Energy [2015] ACompT 2, in relation to the legal principles relevant to the meaning of “serious issue to be heard and determined”. These submissions were accepted in ATCO 2015 at [23], where the following principles were noted as being of particular relevance:

(a)    The phrase ‘serious issue to be heard and determined’ has been correlated with the phrase ‘serious question to be tried’ in the context of the grant of interlocutory injunctions: Re Application by ElectraNet Pty Limited [2008] ACompT 1 at [39]-[42];

(b)    The relevant question is indeed whether an applicant has established that there is a serious issue to be heard and determined given the nature of the rights asserted by the applicant and ‘the practical consequences likely to flow’ from the grant of leave. In particular, the Tribunal has previously expressed the view that the threshold merely requires the applicant to ‘show that there is a sufficient prospect of success to justify in the circumstances it being given the opportunity’ to have the decision reviewed: Application by Envestra Ltd [2011] ACompT 12 at [21].

43    SAPN made extensive submissions on whether there was a serious issue to be tried in respect of:

(1)    gamma

(2)    allowed rate of return – return on debt

(3)    forecast inflation;

(4)    forecast bushfire safety capex;

(5)    forecast labour cost escalation;

(6)    operating expenditure (‘opex’) for increased asset inspections in bushfire risk areas; and

(7)    opex for “no access” poles inspections.

SAPN ultimately decided not to apply for review of the final two topics above relating to opex.

44    For each of these topics, SAPN set out the background to the issue, summarised the issue as arising in SAPN’s Regulatory Proposal and in the AER’s Final Decision, and identified the specific aspects of the relevant grounds of review to demonstrate that there was a serious issue to be heard and determined.

45    As will be apparent from the reasons in respect of each ground of review below, there were serious issues to be heard and determined on SAPN’s Review Application.

Prima facie case on materially preferable NEO decision (s 71E(b) of the NEL)

46    For each of the issues in dispute, SAPN made submissions on how rectification of an error in the AER’s distribution determination – if found to be an error – could result in a materially preferable NEO decision. SAPN then submitted in its Leave Submissions at [241]:

The immediate consequence of these errors is that the overall revenue allowance determined by the AER in its Final Decision is materially below that which would be required by an efficient entity to recover at least efficient costs and provide a commercial market return to investors across the 2015-20 regulatory control period. As noted in paragraphs 198 and 199 of the SA Power Networks Application, these are individually material in terms of the overall regulated revenue SA Power Networks is permitted to earn, and in aggregate are very significant – that is, approximately, a $285.6 million ($nominal) revenue reduction.

47    Further, SAPN also submitted that if the Final Decision is left uncorrected:

(1)    SAPN’s incentives and signals to undertake investment in the network and operations would be distorted due to a significant revenue shortfall;

(2)    the shortfall in the allowances for the operating and capex may lead to an inability to recover the efficient costs of meeting regulatory (or otherwise appropriate) safety obligations and requirements of the distribution system;

(3)    perceptions of regulatory risk will significantly increase; and

(4)    it will set a precedent (in the non-technical sense of that term) for under-compensation of SAPN (and potentially other businesses) now and in the future.

48    The Tribunal considered SAPN’s submissions in this respect. The Tribunal was satisfied that SAPN has made a prima facie case in accordance with s 71E(b) of the NEL.

49    On the basis of the above considerations, and as will be apparent further from the reasons in respect of each of the topics below, the Tribunal considered it was appropriate to grant leave to SAPN to apply for review.

COMMUNITY CONSULTATION PROCESS

50    The consultation process referred to in s 71R(1)(b) of the NEL is an additional procedural step which the Tribunal must take and, ideally, be accommodated within the target time prescribed by s 71Q of the NEL.

51    The Tribunal, having given leave to SAPN to apply for review in this matter on 4 May 2016, sought information from the AER as to groups or persons who might have an interest in the Tribunal’s review under s 71R(1)(b) of the NEL.

52    The Tribunal then conducted an extensive communication process with each of those groups or persons, to invite them to indicate:

(1)    whether they wished to consult with the Tribunal in relation to the Final Decision;

(2)    the nature of their proposed participation; and

(3)    how the consultation might best be carried out.

53    Having determined a protocol for the consultation, the Tribunal issued a Consultation Agenda under which it provided for those who wished to speak to the Tribunal on that occasion either personally or on behalf of an organisation, to do so.

54    The Tribunal conducted the consultation on 1 June 2016 at the Federal Court of Australia in Adelaide. As there was time remaining at the end of the consultation, opportunity was also provided for those who had not registered to participate, to make submissions to the Tribunal.

55    The transcript of that consultation process has been included by the Tribunal on its website. The following is a list of the entities and their representatives who made submissions during the consultation process, or in writing as a complement or supplement to oral submissions, or by providing written submissions after the consultation:

(1)    National Irrigators’ Council – Tom Chesson (Chief Executive Officer)

(2)    Central Irrigation Trust – Gavin McMahon (Chief Executive Officer)

(3)    UnitingCare Australia and Uniting Communities – Mark Henley (Manager Advocacy and Communications)

(4)    Business SA – Andrew McKenna (Senior Policy Adviser)

(5)    Riverland Energy Association – Brenton Paige (Association Member)

(6)    SACOSS – Jo De Silva (Senior Policy Officer)

(7)    The South Australian Financial Counsellors Association – Wendy Shirley (Executive Officer)

(8)    Riverland Wine – Chris Byrne (Executive Officer)

(9)    Major Energy Users, Inc, acting for Energy Users Coalition of SA – David Headberry (Public Officer)

(10)    South Australian Chapter of The Electric Energy Society of Australia – Martyn Pearce (Chairman)

(11)    Jubilee Almonds and Century Orchards – Brendan Sidhu (Chief Executive Officer)

(12)    The Better Drinks Co Pty Ltd – Kym Baldock (General Manager)

(13)    Renmark Irrigation Trust – Barry Schier (General Manager)

(14)    Energy Consumers Australia – Rosemary Sinclair (Chief Executive Officer)

(15)    Consumer Utilities Advocacy Centre

(16)    Bundaberg Regional Irrigators Group

56    All submissions at the consultation were made on behalf of organisations. For ease of reference, the contributions will be identified according to the organisation, rather than the organisation’s representative.

57    SAPN, the AER and the South Australian Minister for Mineral Resources and Energy (the ‘Minister’) as intervener did not participate in the consultation process. That was appropriate, of course, because they each participated in the hearing before the Tribunal.

58    In the course of the consultation, a number of issues of concern to consumers and consumer interests were identified and the participants’ submissions fell into several broad themes. It is, in the view of the Tribunal, helpful to summarise the submissions made in relation to the major themes, as identified by the Tribunal. It is important to note that neither the entirety of the submissions, nor every topic raised will be captured, and the broad themes as identified by the Tribunal will only be in summary form.

59    It is useful to note the starting point from which the majority of the submissions were made. It was generally submitted that, when considered according to the elements of the NEO – price, quality, safety, reliability and security of supply of electricity – the only element with which consumers were dissatisfied was price. As the Bundaberg Regional Irrigators Council (‘BRIG’) submitted:

In terms of the elements of the NEO, BRIG believes that price is the most important and has the greatest impact on consumers and their long term interests. Price has been neglected in favour of investment in recent determinations.

60    Similarly, in survey of businesses conducted by Business SA, 87% of responses ranked reduction in electricity prices as most important, with less than 1% ranking it as either the lowest, or second lowest priority.

61    While each of the topics for review were referred to at the consultation, those that received most attention were gamma and taxation, forecast bushfire safety capex, and forecast labour cost escalation.

Impact on local industry, businesses and the local community

62    The majority of the participants drew connections between the cost of electricity and its impact on the relevant industry and the broader community. One of the key messages of participants was the importance of agriculture in South Australia, and therefore, the region’s dependency on irrigation and electrical energy to operate irrigation for agricultural purposes. Participants impressed upon the Tribunal the centrality of the agricultural industry to the region’s prosperity, and therefore the importance of its protection into the future. As was submitted by The Better Drinks Co Pty Ltd, the whole Riverland region relies on what is pumped out of the river and therefore it affects all industries in the area. Participants also submitted that with other local industries, such as manufacturing, shrinking, more pressure was placed on agricultural industries to keep the local economy healthy.

63    The participants identified specific ways in which the cost of electricity impacted not only on their own businesses but also more broadly on the local industry and, as a result, the local community. For example, Central Irrigation Trust submitted that to absorb the higher costs of electricity, the business were required to either run down assets or reduce opex or capex. National Irrigators’ Council submitted that, for many businesses, the only place that increasing electricity prices could be “worn” was in reducing the number of employees of the business. This had the ripple effect on households, already struggling with “some of the world’s highest energy prices”, which necessarily had to reduce their spending at local shops and restaurants, in turn harming the other local industries. Similarly, BRIG submitted that:

The intentional optimization and deliberate gaming of network profitability has resulted in a rate of price escalation that is unethical and is harming business and domestic consumers across the NEM with standards of living suffering appreciably and less money available for reinvestment in our farms.

64    While a number of participants outlined the measures they were taking to manage increasing electricity prices, improve efficiency and remain competitive, several participants noted that electricity prices were something beyond the consumers’ control. South Australian Financial Counsellors Association (‘SAFCA’) submitted that the steep increase in electricity prices during the previous regulatory period had not been matched by the community’s capacity to pay.

65    A number of participants described the “Swiss cheese” effect of increased electricity prices, being that higher electricity prices were forcing some consumers to seek alternative sources of energy “off the grid”. In the long run however, this would mean fewer consumers to pay for the network. It was in this context that participants submitted that from a policy perspective, increasing electricity prices which drove consumers away from the grid seemed an “absurd” strategy for SAPN itself.

Competition in international and domestic markets

66    Several participants addressed the Tribunal on how increasing electricity prices, and energy costs in general, had, and would continue to undermine the competitiveness of South Australian businesses if something was not done to address them. The high export rate of produce from irrigated agriculture in South Australia meant that the almond, wine and juice producers were competing with producers from Chile, Argentina, South Africa, North America and across Europe. Therefore, to continue to be a significant wealth generator for South Australia, agriculture producers would have to remain competitive both domestically and internationally.

67    For businesses such as The Better Drinks Co Pty Ltd, which operated in a highly competitive market both domestically and internationally, it was submitted that there was a very limited ability to pass on increased costs to its customers, and so the increase in costs had to be absorbed by business itself. Central Irrigation Trust submitted that Australia’s competitive advantage was being quickly squandered through rising energy costs. Business SA submitted that the pending structural change on South Australia’s economy with the exit of the automotive manufacturing industry in 2017 provided further impetus for energy costs to not contribute to an already uncompetitive cost base, compared with interstate and international competitors.

Long term interests of the consumer

68    The notion of the “long term interests of consumers” (‘LTIC’), as the factor underpinning the NEO, was a focal point of many of the submissions. In considering the factors relevant to the LTIC, UnitingCare Australia and Uniting Communities (‘UnitingCare’) sought to juxtapose the “supply-side” perspective of the LTIC contended for in SAPN’s Review Application with the “demand-side” perspective of network users. Therefore, it was submitted that the elements of the LTIC, as specified in the NEO, should be interpreted in the following way:

    Price – end consumer pays no more than is necessary;

    Reliability – all consumers are able to afford to pay for the essential use of electricity needed to participate in contemporary society, and that people will not be cut off from electricity supply for essential purposes because it is too expensive;

    Quality – prices as well as supply are stable with no supply-side or bill shocks;

    Safety – a person’s house is not at risk of burning down because they cannot use electricity for safe lighting, cooking or heating or other unsafe practises considered in the absence of supply.

69    Energy Consumers Australia (‘ECA’) also directed some of their submissions to considering the appropriate interpretation of LTIC. ECA submitted that the question for interpretation is whether “efficiency” is the goal, and LTIC is just a signpost to balancing items, or whether LTIC is the goal, and efficiency is the means to achieve it. In developing its interpretation, ECA submitted that the objective of the regulatory regime is to deliver the outcomes that are equivalent to those that would be delivered if there were competition achieved in the electricity market. ECA identified such outcomes as the consumer interests, price, quality, safety, reliable and secure supply services – in essence, the elements of the NEO, through which the LTIC must be achieved.

Impact on SAPN

70    Some of the participants submitted that SAPN’s Regulatory Proposal would ultimately work against SAPN’s interests as it would force a significant proportion of consumers off the grid through increased prices. This would thereby reduce SAPN’s customer base and impact on its ability to run its business. A number of participants noted that they were investigating alternative energy sources to run their businesses such as through diesel generation and solar energy.

Taxation and gamma

71    An issue raised by a number of the participants at the consultation was the amount of tax paid by SAPN. Many consumers called for an explanation for why consumers were paying for an income tax equivalent of “over $400 million” when, as the participants submitted, it did not appear to reflect the amount of tax SAPN was actually paying. As UnitingCare submitted, the Consumer Challenge Panel (established by the AER to assist with the AER’s regulatory determinations) noted that SAPN has historically paid minimal tax to the Australian government despite increases in the regulatory revenue allowance from 2011. UnitingCare emphasised that it was not suggesting anything improper on the part of SAPN. However, it raised the issue to ensure consumers were not paying more as a result of a “more-than-generous” credit arrangement in favour of SAPN.

72    In seeking clarity and transparency around the amount of tax paid by SAPN, the participants also called for an indication of taxation allowances for privately owned distributors to determine whether this properly represented their actual tax costs. In support of this submission, National Irrigators’ Council referred to the Minister’s criticism of SAPN for making “super profits” from its customers.

73    Consumer Utilities Advocacy Centre (‘CUAC’) considered the relationship between gamma and the LTIC in two respects. First, submitting that the value for gamma proposed by SAPN “relies on a single methodology that produces statistically unstable results because it is highly sensitive to underlying assumptions”, CUAC urged the Tribunal to consider

whether it is in the best interests of consumers for the regulator to rely on this single measure and methodology, rather than considering a wider range of evidence drawn from different methodologies to produce a more robust value for gamma, as has been done by the AER in their guideline paper.

74    Secondly, CUAC noted the recent changes in the NEL, as stated in the second reading speech of the Statues Amendment (National Electricity and Gas Laws – Limited Merits Review) Bill 2013, which sought to ensure

consumers do not pay more than necessary for the quality, safety, reliability and security of supply of electricity and natural gas under the national energy laws.

75    However, CUAC queried how this would be reflected and implemented if SAPN’s own estimate of their proposed gamma value would cost consumers $85.2 million over the 2015-2020 period.

76    Central Irrigation Trust submitted that the effect of the AER’s determination in the Final Decision in respect of gamma was viewed to be a transfer of money from their own business to SAPN. Renmark Irrigation highlighted that even small changes in the value of gamma would mean “big returns” for SAPN. It was considered to be in the best interests of consumers for the AER’s gamma methodology to be ratified by the Tribunal.

77    Major Energy Users (‘MEU’) considered gamma in the context of the incentive nature of the regime. MEU submitted that not only have companies been willing to invest in SAPN with a gamma of 0.5 or 0.4 over many years, but it was also submitted that the current owner of SAPN was prepared to invest in the distributor with a gamma of 0.4. It followed, on MEU’s submission, that there was no reason to change it as a disincentive could not be identified.

Capex and opex

78    In raising issues about the specific topics for review, a number of participants voiced concerns about capex and opex generally.

79    Business SA submitted that a broad concern was that substantial increases in the AER’s determination both in relation to capex and opex will lock in significant additional electricity costs over the next five years and beyond, as they would be incorporated into the regulatory asset base and opex used in future calculations. It was further submitted by Business SA that it considered it to be “inconceivable” for capex and opex to be increasing as it submitted that the demand on the electricity network was static, if not in decline. More fundamentally, SACOSS expressed concern that SAPN was attempting to expand the meaning of its regulatory obligations to justify additional expenditure in relation to capex and opex.

80    Drawing a connection with the concern around consumers abandoning the network in favour of alternative energy sources, Riverland Energy Association posed the question of who would be paying for existing and new assets proposed by SAPN if future customers would be seeking alternative energy sources and deserting the traditional network supply base.

Forecast bushfire safety capex

81    One of the main issues raised by participants at the consultation in relation to the bushfire mitigation capex was the absence of legislative changes by the South Australian government to warrant or require SAPN’s increase in spending on bushfire mitigation over the relevant regulatory period. SACOSS submitted that the AER was correct to conclude that the regulatory obligations imposed by Electricity Act 1996 (SA) (‘Electricity Act’) and Work Health and Safety Act 2012 (SA) (‘WHS Act’) in relation to the bushfire mitigation measures had not been expanded by the outcome of the Victorian Bushfire Royal Commission and Powerline Bushfire Safety Taskforce, as had been suggested by SAPN. Further, SACOSS submitted that, in any event, SAPN had failed to demonstrate how the additional expenditure proposed in their Regulatory Proposal was prudent and efficient. Similarly, MEU emphasised that the amount allowed for bushfire mitigation capex should not simply be the cost of service but what is efficient.

82    Business SA submitted that as SAPN’s proposal on bushfire mitigation capex had been made independent of comprehensive policy consideration, it did not represent optimal public policy outcomes. Participants warned against allowing SAPN to become an “arbiter” on what constitutes good industry practice in relation to bushfire mitigation, and disregarding the AER’s views on the basis that is “merely” an economic regulator. MEU submitted that such an approach would risk shifting from an incentive-based regime, which was specifically required by the legislation, to a regime driven by cost of service.

83    A contrasting perspective was provided by the South Australian Chapter of the Electricity Energy Society of Australia (‘EESA’). EESA acknowledged that SAPN had been achieving good results in bushfire mitigation controls. However, EESA submitted that SAPN, with whom lay a large portion of the responsibility for protecting the community during bushfires, had to demonstrate continuous improvement to its practices to address “reasonably foreseeable” events causing bushfires. Otherwise, as EESA submitted, such organisations would become vulnerable to litigation, particularly in light of the findings and changes brought about in this sphere following the Victorian Bushfires Royal Commission. It was submitted that the cost of litigation in circumstances where negligence and the “reasonable foreseeability” test were made out, would be in excess of $12 million. In light of this, and the public expectation that EESA considered existed for such measures to be implemented, EESA contended for the program of bushfire mitigation proposed by SAPN to be approved.

Forecast labour cost escalation

84    Forecast labour cost escalation received significant attention during the consultation. Participants submitted that SAPN’s proposal to increase the terms of the enterprise bargaining agreement (‘EA’ or ‘EBA’) governing SAPN’s employees was at odds commercially with other EAs across Australia, particularly in light of what were considered to be already favourable EA terms. A number of participants queried why increases in allowances for labour costs would be required where there had been no indication of increases in productivity. As CUAC submitted:

It is questionable whether any increases in real labour price growth are appropriate – and in the long term interests of consumers – without a corresponding increase in productivity growth. As stated by in the [Victorian Energy Consumer and User Alliance] submission, “productivity and labour price increases are inextricably linked”. This is particularly pertinent given that the AER has accepted the networks in both states have proposed zero productivity growth (with the exception of Jemena in Victoria), despite the significant capital expenditure over the previous period. CUAC notes that SAPN proposed that the AER adopt negative productivity growth forecasts in its alternative estimate of opex.

85    Business SA submitted that while the shareholders in SAPN were free to remunerate their labour force as they saw fit, electricity consumers would not be willing to absorb any costs considered above the efficient rate in relation to market conditions in South Australia.

86    SACOSS noted its support for the AER’s decision to use the wage price index to escalate cost rather than basing the escalation on SAPN’s EA. MEU identified the risk with allowing SAPN to determine the escalation, rather than using an external benchmark: there would be no competitive pressure on SAPN to negotiate an EA on terms most favourable to consumers. This is particularly the case given that, as MEU submitted, without an external competitor, the unions, with whom SAPN would be negotiating, could argue that the business could just pass on its costs to clients.

87    SACOSS further submitted that, if an error in the AER’s decision were identified, the Tribunal should still consider whether, in any event, it would be a materially preferable NEO decision to reverse the AER’s decision when the reviewable regulatory decision was considered as a whole.

Consumer engagement

88    The 2012 rules changes required DNSPs to improve their engagement with electricity consumers to ensure that capex and opex forecasts include expenditure to address consumers’ concerns. The consumer engagement process (‘CEP’) conducted by SAPN received some criticism from participants. Riverland Energy Association contended that the “cornerstone” of SAPN’s Review Application was based on the AER failing to giving significant weighting to the views of the consumer expressed in SAPN’s CEP. However, Riverland Energy Association identified the methods adopted by SAPN for determining the consumers’ priorities to be self-fulfilling “in the extreme”. SACOSS claimed, both at the consultation and in written submissions filed with the Tribunal subsequently, that as the CEP undertaken by SAPN was conducted in respect of areas not the subject of review before the Tribunal, it should be not be relied upon by SAPN to demonstrate the LTIC, nor be given any weight in the Tribunal’s considerations in this review. SACOSS submitted that its only relevance was to highlighting the consumer concern about potential cost increases, and directed the Tribunal’s attention to SAPN’s research findings relating to the consumers’ willingness to pay, namely that:

there is significant community concern about potential cost increases. Half of the customers surveyed are very concerned about the prospect of rising electricity costs.

89    MEU also identified problems with consumer engagement in the regulated energy distribution sphere more generally. For example, it was claimed that there was insufficient time to engage with consumers, consumers had inadequate knowledge to be able to make an informed contribution or recommendation, and the majority of time spent engaging with consumers was explaining the role of a network distributor, and distinguishing it from a network retailer. MEU considered this to be detracting from a discussion of the substantive issues.

Role of the Tribunal and the regulatory process

90    Participants also addressed the Tribunal on their perceptions of the regulatory process and what they considered to be the role of the Tribunal in the process.

91    A number of participants articulated that they saw the Tribunal as a “final backstop for consumers.” One participant “challenged” the Tribunal to bring electricity prices in South Australia back to the bottom quartile of prices amongst countries forming the Organisation for Economic Co-operation and Development (‘OECD’). SAFCA submitted that for the Tribunal to achieve the materially preferable NEO decision, it is required to take into account real-life experiences of large numbers of low and modest income households.

92    Some comments were made about the regulatory process more generally. Some participants submitted that they felt let down by the regulatory process given the sharp rise in SAPN tariffs over the last regulatory period. UnitingCare submitted that consumer behaviour over recent years put paid to the notion that regulation of network monopolies successfully functioned as a proxy for competitive and efficient markets. UnitingCare cited a number of statistics to support this argument, including that:

    31,666 South Australian households had an electricity debt, with the average debt being $758;

    3,174 South Australian small businesses had an electricity debt, with the average debt being $1,559; and

    South Australia has highest rate of energy customers on hardship programs.

93    BRIG also expressed their dissatisfaction with the regulatory process:

The network price setting process is deeply flawed. In BRIG’s experience, it has been a one sided process that has no regard for impact on consumers or the wider economy. If the AER’s increased powers are greatly reduced by the decisions of this Tribunal, we fear this will continue to be the case.

94    In relation to the regulatory review process, participants submitted that they considered it to be adversarial, intimidating and complex. Others noted that as it took significant time to understand the technical and complex concepts such as theta, gamma, and regulatory asset bases, the process itself became costly and time consuming for businesses to participate in it effectively.

95    The incentive-focus of the regulatory regime was also raised as an issue. MEU contended that the AEMC “cleverly” characterised the LTIC as being achieved by ensuring there is incentive for investors to invest in the networks, and disincentive when investment is not required. MEU considered that this emphasis on investment and investors took the focus away from the interests of the consumers who were required to pay for the network services and the impact that electricity prices had on their daily lives.

96    ECA devoted a significant portion of their submissions to the new limited merits review process, addressing what they considered to be intended from the introduction of the process. For example, ECA submitted that while SAPN does have a responsibility to meet its regulatory and other obligations, the question before the Tribunal is whether SAPN’s Regulatory Proposal represents the most efficient way of meeting such responsibilities and obligations. In this respect, ECA referred to the aim of the energy market reforms, identified by the (former) Standing Council on Energy and Resources (now COAG Energy Council), to “restore the focus of the electricity market on serving the long term interests of consumers”.

97    ECA also contrasted this case before the Tribunal with that of a NSW and ACT proceeding with respect to the absence of a consumer advocacy group as a party to the proceeding, noting that in Applications by Public Interest Advocacy Centre Ltd and Ausgrid [2016] ACompT 1 (‘Ausgrid’), the Tribunal held at [64]:

Given the role of PIAC, and the relevance of its submissions to the Tribunal’s functions and responsibilities under the legislation, the Tribunal has not needed in these matters to address separately the matters which emerged in the course of the consultation process. The role and submissions of PIAC have encompassed those matters.

98    As such, ECA submitted that it was particularly important for the Tribunal to carefully consider and place weight on the submissions of SACOSS put forward in the consultation.

99    Relevantly, in its written submissions, SACOSS urged the Tribunal to continue to

consider its consultation obligations once the parties have had an opportunity to respond to matters raised at the public forum, and demonstrates in its reasons for decision on the review how the submissions made through the consultation process have been considered by the Tribunal.

100    Before going to this issue, the Tribunal makes mention of the submissions of the Minister. It was the Minister’s contention that the Tribunal should affirm the Final Decision. In the course of these reasons, the Tribunal traverses the issues raised by the Minister, although without specific attribution to the Minister as intervenor. These issues were effectively argued by both SAPN and the AER, and need no repetition. Nevertheless, one submission of the Minister does require specific mention relevant to the consultation process.

101    The Minister submitted that the efforts of the consumers who took the time to participate in the community consultation should be recognised and their submissions to the Tribunal must be considered as part of the review. The Tribunal accepts this submission, and acknowledges that the participants in the consultation took time out of their own businesses and life to be part of the review, representing various consumer interests.

102    In light of the ultimate decision of the Tribunal, it is unnecessary to otherwise comment on the Minister's helpful submissions.

103    It should also be observed, as the following reasons indicate, the decision of the Tribunal has been reached on the basis of the contentions made by SAPN and the AER, and not by direct reference to submissions made during the consultation process. This is not because such submissions are not relevant to the review, but because in this review the decision to affirm the Final Decision has been readily arrived at by focussing on the submission of the parties and in determining whether any error had occurred on the part of the AER. Nevertheless, the consultation process and the submissions of consumers (and the Minister) may have become particularly significant (if error had been found in the Final Decision) in the consideration of the materially preferable NEO decision. This has been unnecessary in this review as no error has been found to occur.

GROUNDS OF REVIEW

104    SAPN has identified multiple grounds of review in respect of each of the topics for review. In Application by ATCO Gas Australia Pty Ltd [2016] ACompT 10 (‘ATCO 2016’), the Tribunal outlined the nature and scope of each of the relevant grounds of review: error or errors of fact, incorrect exercise of discretion, and that the decision under review is unreasonable: see ATCO 2016 at [36]-[48]. The Tribunal considers these principles to be applicable in respect of the present application by SAPN.

105    The Tribunal now turns to consider in detail each ground of review by reference to the submissions of the parties. There was a substantial degree of repetition and overlap in the submissions dealing with each ground of review, and the Tribunal has attempted to address the significant and determinative issues to arrive at its decision.

Cost of corporate income tax (gamma)

106    Gamma is one of the required parameter inputs into the post -tax revenue model (‘PTRM’) used under the NER for the determination of allowable revenues for a DNSP. It represents the “value” of imputation (tax) credits arising from company tax payments which are distributed with dividends to shareholders. These can be expected to reduce the required return on equity of shareholders (relative to receiving returns without such tax credits attached).

107    However, in the PTRM, rather than applying an estimated value of gamma to adjust the required rate of return on equity, the estimate is instead applied to reducing the revenue required to compensate the company for corporate tax paid. Consequently the return on equity incorporated into the allowed rate of return is calculated ignoring the imputation credit component of returns to shareholders. This approach draws on a demonstration by Professor Bob Officer (The Cost of Capital of a Company under an Imputation Tax System, Accounting and Finance, May 1994) of the equivalence of several such alternative approaches to valuation of company cash flow streams.

108    Because gamma is unobservable, and also because of conflicting interpretations of the word “value” in its definition, there has been ongoing debate over the appropriate numerical value to use in the PTRM. In its Final Decision, the AER applied a value for gamma of 0.4, whereas the applicant SAPN argued for a value of 0.25. This difference has a significant negative impact on the allowable revenue for SAPN over the regulatory control period which it estimates to amount to $85.2 million. The AER argues that the effect would be less than this because of the need to adjust other parameters accordingly in the PTRM, as discussed later. SAPN argues that the AER made errors in determining to use a value of 0.4 for gamma, and contends that it should have accepted the proposed value of 0.25.

109    This same issue arose in the 2015 determinations by the AER for the NSW DNSPs, which was also appealed to the Tribunal by the affected parties (arguing for a gamma value of 0.25 rather than the AER’s use of 0.4). That (differently constituted) Tribunal ruled on 26 February 2016 in the Ausgrid decision in favour of the applicants, and remitted the matter back to the AER to recalculate allowable revenues using a gamma value of 0.25. The AER has, in turn, appealed that decision to the Full Court of the Federal Court. No determination has yet been made by the Full Court.

110    This Tribunal determined that despite the existence of that appeal, it was appropriate for it to hear this review rather than leave the matter to be determined conditional on the outcome of the hearing of the Full Court of the Federal Court. The Tribunal has a legislative responsibility to hear and determine the review (within a statutorily delineated period of time), and should proceed accordingly.

111    It was also contended by SAPN that this Tribunal should follow the Ausgrid decision, or alternatively, treat it as highly persuasive. Undoubtedly, each differently constituted Tribunal should consider the importance of consistency between Tribunal decisions, but this is not the sole determinative factor nor is consistency an unqualified value. Consistency may lead to arbitrariness of decision-making, and may not produce the correct legal and just result in the particular case before the Tribunal. Each Tribunal, considering the application before it, and dealing with the relevant parties, must in accordance with the law, the issues before it, and the evidence, consider and determine the matters raised before the Tribunal.

112    It is to be recalled the Tribunal is an administrative decision-maker, not a court of law. The Tribunal cannot conclusively decide questions of law. The general statements of principle the Tribunal articulates will not and cannot have the same force as the development of principles of general law determined by the courts. The function of the Tribunal is a reviewer of decisions, and is not a primary decision-maker. The Tribunal has a responsibility to determine individual cases based upon the evidence and arguments put before it. This Tribunal proceeds accordingly.

The role of gamma in the PTRM

113    The NER (r 6.3) specify that the DNSP’s revenue requirement is to be calculated using the “building block” approach. Rule 6.4 provides detail on the building block components. Two components are relevant in this matter. One is the allowed rate of return on capital which is to be calculated in accordance with cl 6.5.2. The NER (cl 6.5.2(d)(2)) specifies that the allowed rate of return “must be … determined on a nominal vanilla basis that is consistent with the estimate of the value of imputation credits referred to in clause 6.5.3.” The second component (cl 6.4.3(b)(4)) is that “the estimated cost of corporate income tax is determined in accordance with clause 6.5.3”.

114    Clause 6.5.3 of the NER states that:

The estimated cost of corporate income tax of a Distribution Network Service Provider for each regulatory year (ETCt) must be estimated in accordance with the following formula:

ETCt = (ETIt × rt) (1 – γ)

where:

ETIt is an estimate of the taxable income for that regulatory year that would be earned by a benchmark efficient entity as a result of the provision of standard control services if such an entity, rather than the Distribution Network Service Provider, operated the business of the Distribution Network Service Provider, such estimate being determined in accordance with the post-tax revenue model;

rt is the expected statutory income tax rate for that regulatory year as determined by the AER; and

γ is the value of imputation credits.

115    Thus, the value of imputation credits provided to shareholders is deducted from that part of the allowed cash flows required to meet tax obligations. Under the “vanilla” WACC approach, those tax obligations are calculated after allowing for the effect of the tax deductibility of interest on debt. Consistency of approach means that the allowed rate of return on the regulatory asset base (another component of allowed cash flows) is calculated as a “vanilla” WACC. This is a weighted average of a cost of debt (pre company tax – ie ignoring the tax deductibility of interest at the company level) and the cost of equity (ignoring returns to shareholders in the form of imputation credits). This “vanilla” WACC is also used to convert cash flows over, and the regulatory asset base at the end of, the regulatory control period to a present value to achieve a zero NPV condition for a smoothed set of revenues or prices over the regulatory control period.

116    This “vanilla” WACC approach, for use under an imputation tax system, has its genesis in the aforementioned paper by Professor Officer, in which he demonstrated the equivalence of a zero NPV condition for a set of future cash flows under alternative approaches. One was the “vanilla WACC” approach in which net cash flows to be valued have company tax payments (allowing for tax deductibility of interest) deducted. This involved company tax being defined as net of the value of imputation credits distributed to shareholders. For consistency, required rates of return (the discount rates) ignore personal tax consequences, including the effects of imputation. Among the alternative approaches was a dividend imputation version of the standard textbook approach. In that, cash flows are calculated, after company tax, “as if” the company were unlevered, such that there is no tax deduction for debt interest allowed. Rather, the required return on debt is the after-company-tax cost of debt. Moreover, the company tax figure ignores imputation credit implications, and the required return on equity is lowered by some amount related to the value of imputation credits.

117    Two important considerations flow from this. One is whether the nature of the Officer approach has explicit implications for the interpretation of the term “value of imputation credits” used in the PTRM and the NER. The second is how a “value” of imputation credits should be estimated given the interpretation which is (or should be) adopted. These are the two issues at the heart of the dispute between SAPN (and other DNSPs) and the AER.

Regulatory background

118    The dispute over gamma needs to be placed in the context of changes made to the NER in November 2012 which changed the terminology used to describe gamma. Prior to that change, gamma was defined as “the assumed utilisation of franking credits”. Subsequently it has been defined (NER cl 6.5.3) as “the value of imputation credits”. No specific definition of what that term means, or how it is to be estimated, is given, other than that it needs to be consistent with the “vanilla WACC” approach.

119    Those rule changes also gave increased flexibility to the regulator in determination of the value to be chosen for gamma. “The current prescription of the gamma value of 0.5 in clause 6A.6.4 has also been removed to allow the regulator the ability to estimate an appropriate value that reflects the best available evidence at the time of a decision and would therefore result in a rate of return that meets the overall objective.” (AEMC, Rule Determination: National Electricity Amendment (Economic Regulation of Network Service Providers) Rule 2012, p 68).

120    Other than this statement, there appears to be no other explanation given for the change in terminology from “assumed utilisation” to “value”, although the latter term (even though not specifically defined) is more consistent with the flexibility given to the regulator and requirement (NER cl 6.5.2(e)) that:

In determining the allowed rate of return, regard must be had to:

(1)    relevant estimation methods, financial models, market data and other evidence;

(2)    the desirability of using an approach that leads to the consistent application of any estimates of financial parameters that are relevant to the estimates of, and that are common to, the return on equity and the return on debt; and

(3)     any interrelationships between estimates of financial parameters that are relevant to the estimates of the return on equity and the return on debt.

The regulatory determination process.

121    In its initial proposal, SAPN proposed a value for gamma of 0.25, calculated as the product of a distribution rate of 0.7 and a “theta” value (“the value of distributed imputation credits to investors who receive them”: SAPN, Regulatory Proposal, p 320) of 0.35. It noted that the latter figure (0.35) was different to that contained in the AER Rate of Return Guideline (‘ROR Guideline’) (where 0.7 is proposed) and also argued that the interpretation adopted by the AER of the “value of imputation credits” in those guidelines was incorrect.

122    It argued that the AER had inappropriately discarded a correct (in SAPN’s view) prior approach which interpreted theta as the “value” of imputation credits, and which led to emphasis being given in its estimation to “market value studies” (such as dividend drop-off studies). Instead it had adopted an approach based on a “utilisation” rate. SAPN argued that rather than seeking to estimate the value of distributed imputation credits, the AER instead seeks to estimate what it refers to as “the before-personal-tax reduction in company tax per one dollar of imputation credits that the representative investor receives”. Elsewhere in the Explanatory Statement, the AER refers to its conceptual definition of theta as “the expected ability of equity holders to use the imputation credits they receive to reduce their personal tax”. (The references are to AER, Explanatory Statement: Rate of Return Guideline, December 2013, pp 165 and 174 respectively). By incorrectly (in SAPN’s view) adopting this approach, the AER placed more weight on information from equity ownership and tax statistics than market value studies.

123    It is worth referring back to the Explanatory Statement of the ROR Guideline (pp 166-7) from which those statements have been extracted, since fuller reading of the document makes clear both the AER’s approach and conceptualisation of theta, and also its perspective on earlier interpretations of gamma and theta by the Tribunal and others:

We consider the relationship between the representative investor in the market and the implied representative investor from estimation methods such as tax studies and dividend drop off studies). We consider this relationship is critical in assessing what we are estimating and which estimation methods are fit for purpose.

To answer the question of the appropriate representative investor, we considered afresh:

* the Sharpe-Lintner CAPM framework under imputation as derived in Officer, Monkhouse, Lally and Van Zijl, and Lally

* analysis of this conceptual framework by academic experts

* the construction of the corporate tax building block in the rules and how this interacts with the Officer framework used within the rate of return.

Our analysis of these issues is set out in section 9.3.1, and further in appendix H. Having undertaken this analysis, we conclude that we did not fully adopt or address important aspects of this analysis during the 2009 WACC review. As a result, the Tribunal review focused only on the particular suitability of tax value studies and dividend drop off studies. This was with an incomplete conceptual framework. The Tribunal acknowledged this incomplete framework at several points in its reasons.

We conclude that the representative investor:

* Is the weighted average of investors within the defined market, where the weightings reflect market participation (equity ownership value) and risk aversion.

* In this context, the defined market is investors in Australian equity, either domestic or foreign.

* Is the representative investor at any hypothetical point during a trading year – that is, it does not disproportionately reflect an investor or set of investors at a particular point in time. This is because investors may invest at any point during the year. If a benchmark parameter is set using data from a short period in systematically different trading circumstances to the rest of the year, it produces an estimate that is only relevant to those circumstances.

Having reached this view, we consider it has important implications for the practical task of estimating the value of imputation credits. The most important implication of this relationship is that the source of evidence the Tribunal adopted for the utilisation rate (a dividend drop off study) does not produce an estimate for the representative investor. This is because dividend drop-off studies give the value weighted investor’s valuation of imputation credits:

* Based on the combined package of imputation credits, dividends, and other entitlements (unless adjusted for). That is, a value for imputation credits is not available via simple observation of the dividend drop off in these studies. The implied values for the franking credit and the cash component must be econometrically separated, which is difficult to do reliably. We discuss this further in appendix H.

* For trades around the time of dividend distribution – that is, these studies only reflect trading around the cum-dividend and ex-dividend dates.

124    In its preliminary decision, the AER rejected SAPN’s proposal and determined that the value attributed to gamma should be 0.4, which was a departure from its ROR Guideline figure of 0.5. It explained that change as reflecting the available current information, and being based primarily on data from equity ownership and taxation statistics. These indicate (respectively) potential and actual utilisation of imputation credits. The AER argued that little weight should be given to market value studies due to a range of factors potentially affecting the robustness and interpretation of their results. The AER also departed from its ROR Guideline in dropping from the list of alternative approaches the “conceptual guidelines” approach. The preliminary decision incorporated an extensive evaluation of alternative approaches to estimation of gamma and arguments provided by a range of experts and related evidence.

125    In its response to the preliminary decision and revised proposal, SAPN maintained its position that the value attributed to gamma should be 0.25, and provided additional expert reports in support of its position (and critical of the AER preliminary decision). It noted (SAPN, Revised Proposal, pp 370-1) that

There are two fundamental differences of view that explain how SA Power Networks’ approach differs so substantially from that of the AER:

1)    the first fundamental difference of approach concerns what is meant by the term ‘value of imputation credits’ in rule 6.5.3. SA Power Networks and its advisors have consistently contended that this term must mean the valuation revealed in openly traded equity markets. The reasons for this view are explained at section (a) below. By contrast, in a range of regulatory documents published over the past five years the AER and its consultants have advanced one, two, three and even more formulations of argument and explanation that seek to bridge the gap between the reference in the Rules to a ‘value’, on the one hand, and the AER’s preferred measure of the redemption rate …

2)    The second fundamental difference of approach concerns the set of comparator businesses that should be used when establishing a benchmark distribution rate. There are two key differences. First, the AER takes the view that the data for the distribution rate and the data for the valuation of imputation credits need to be drawn from the same set of firms. … Secondly, even if it were appropriate to look at a limited group of ‘comparator firms’ the AER’s approach to selecting that subset is inconsistent over time and inconsistent with other aspects of its approach to rate of return regulation.

126    These fundamental differences underpin the current dispute between SAPN and the AER (as well as being those considered by a previous Tribunal in the Ausgrid case).

127    In its Final Decision, the AER maintained its position on the appropriate value of gamma being 0.4. The AER (Final Decision, Attachment 4, pp 6-7) notes “[t]here is no consensus among experts on the appropriate value or estimation techniques to use. Further, with each estimation technique there are often a number of ways these may be applied resulting in different outcomes.” The AER also noted the commonality of proposals from a wide range of service providers and that the value of gamma was one issue currently (at that time) subject to appeal by other service providers to the Tribunal. It concluded at p 18 that “[o]verall, the evidence suggests that a reasonable estimate of the value of imputation credits is within the range 0.3 to 0.5. From within this range, we choose a value of 0.4”, and justified that choice by the greater weight it placed on data from equity ownership and taxation statistics than on market value studies (such as dividend drop-off studies). The AER did not provide specific individual estimates for the two components of gamma (ie the distribution rate (F) and the value of imputation credits (theta)).

128    In reaching its decision, the AER adopted a clear position on the interpretation of the term “value of imputation credits”. The AER based its decision on the view that “the value of imputation credits in NER cl 6.5.3 is the value of imputation credits to investors which, within the Officer framework, is the proportion of company tax returned to investors through the utilisation of imputation credits” (AER’s Outline of Submissions, para 35).

129    In implementing that approach, the AER adopted a method for calculating gamma which involves multiplying the distribution rate (or payout ratio, “F”) by the utilisation rate (“θ” or theta). There is disagreement over how best to calculate these two numbers, and whether “θ” should be interpreted as a utilisation rate or an implied market price of imputation credits. In its Outline of Submissions (at para 35), the AER:

(c)    defined the distribution rate as the proportion of imputation credits generated by the benchmark efficient entity that is distributed to investors;

(d)    defined the utilisation rate as the value to investors in the market per dollar of imputation credits distributed, which reflects the extent to which investors can utilise the imputation credits they receive to reduce their tax or obtain a refund …

(e)    estimated the distribution rate using the “cumulative payout ratio approach”, which uses data from the ATO on the accounts used by companies to track their stocks of imputation credits (franking account balances);

(f)    estimated the utilisation rate by giving varying degrees of weight to the different available estimation methods identified in the ROR Guidelines. …

130    In estimating the utilisation rate the AER gave significant weight to equity ownership information, less weight to taxation statistics and lesser weight to market value studies (such as dividend drop-off studies).

Grounds of review

131    SAPN argues that the AER made multiple errors of fact, was unreasonable, or made an incorrect exercise of its discretion in determining that the appropriate value for gamma is 0.4. These commence at paragraph 37 of its Amended Application for leave, and run for several pages. Since they are, in many cases duplicative, interrelated, and overlapping, it is not helpful for the Tribunal to reproduce them verbatim, but rather to summarize them in an order which facilitates a logical approach to analysis.

132    First, SAPN contends that the AER erred in misconstruing the correct interpretation of “value of imputation credits” through focusing on utilisation rather than implied market value.

133    Secondly, this misconstruction led the AER to incorrectly give greater weight to utilisation measures of imputation credit other than market based studies and particularly dividend drop-off studies.

134    Thirdly, SAPN contends that the AER erred in not recognising that investors would incur personal costs in extracting value from receipt of imputation credits which would reduce the value ascribed to them.

135    Fourthly, in calculating the value of imputation credits from equity ownership and taxation information, the AER used inappropriate data.

136    Fifthly, the AER in interpreting equity ownership and taxation figures failed to recognise that these are maximum possible values.

137    Sixthly, the AER did not accord sufficient weight to the dividend drop-off study of SFG Consulting (‘SFG’) due to misconceptions about the merits and reliability of drop-off studies.

138    The Tribunal notes that in relation to the concept of the “value of imputation credits” referred to in the first of the issues listed above, different theoretical models, all of which are simplifications of reality, with different strengths and weaknesses, and with different degrees of support among experts, may suggest differing approaches. Judgement about the weight to be given to alternative approaches would then be required, with resulting consequences for judgements about the subsequent issues.

Consistency of approach with the PTRM framework

139    In this section, the Tribunal addresses the first of SAPN’s challenges – that the AER erred in misconstruing the correct interpretation of “value of imputation credits” through focusing on utilisation rather than implied market value.

140    Fundamental to the dispute over gamma is the correct interpretation of the word “value” in the NER which relates to one of the two components of gamma. It is accepted by all parties, and defined in the NER (cl 6.5.3) that gamma is the product of two components. Gamma is the value of franking credits generated by the company tax paid by the entity. In turn gamma is defined as the product of the “F” (the distribution rate) and “θ” (theta – variously described as the utilisation rate or the market value (price) of distributed credits).

141    The AER places particular emphasis on the statement in cl 6.5.3 of the NER that the estimated cost of corporate income tax is to be “determined in accordance with the post-tax revenue model”. That model (abbreviated as PTRM) is based on the derivation by Professor Officer of equivalence relationships between various formulations of the calculation of the NPV of future cash flows under an imputation tax system. In particular, Officer showed that the NPV=0 criterion could be described in a number of different ways, each involving different future cash flow concepts (with differences arising from different treatment of taxation) and correspondingly consistent interpretations of a discount rate (WACC). Officer’s approach assumed a 100% distribution rate, which has been generalised in the common approach adopted by the AER and regulated entities of defining gamma as the product of “F” and “θ” as described above. Consequently, debate over the value of gamma involves debate over the values of both “F” and “θ”, with the latter being the more contentious and fundamental to the interpretation of the PTRM model. Dispute over the value of “F” relates more to the choice of empirical data for its estimation rather than fundamental theoretical issues.

142    Officer observed (at page 4 of the aforementioned article) that:

A proportion (γ) of the tax collected from the company will be rebated against personal tax and, therefore, is not really company tax but rather is a collection of personal tax at the company level. Therefore, if we wish to define the effective company tax collection, we need to reduce T by the proportion γ. …

Thus γ is the proportion of tax collected from the company which gives rise to the tax credit associated with a franked dividend. This franking credit can be utilized as tax credit against the personal tax liabilities of the shareholder. γ can be interpreted as the value of a dollar of tax credit to the shareholder.

143    Officer also observed (p 4, fn 5) that:

Where there is a market for tax credits one could use the market price to estimate the value of γ for the marginal shareholder. i.e. the shareholder who implicitly sets the price of the shares and the price of γ and the company’s cost of capital at the margin, but where there is only a covert market, estimates can only be made through dividend drop-off rates.

144    There are (at least) two possible interpretations of these comments, as has been pointed out in submissions and expert evidence. First, the derivation by Officer places emphasis on gamma as the amount of tax credits which are utilised to reduce personal tax liabilities, and this is the interpretation used by the AER. It also leads to an interpretation of “effective” company tax payments as being T(1 – γ) where T is the amount of tax collected at the company level. Secondly, Officer asserts that the market value of gamma will be set by the marginal investor, and that this can only be indirectly estimated by dividend drop-off rates. This, or at least the emphasis on drop-off studies, is the interpretation focused upon by SAPN and other DNSPs.

145    The Tribunal notes that the AER, in its reasoning in the ROR Guideline quoted earlier, did not interpret the drop-off evidence as relating to a marginal investor, but rather to some average investor prevailing at that specific time. The Tribunal also notes that Officer’s interpretation of the role of a “marginal investor” in price determination is one commonly found in academic literature, as also is the notion of an average investor.

146    It is also worth noting the emphasis given by Officer to the “marginal shareholder” in the determination of share prices and the implicit price of γ. This underscores, and potentially calls into question, a commonly-heard assertion which is also made by SAPN. The marginal investor is not the same as the average investor. The proportion of tax credits used in aggregate (ie the average utilisation) provides no information about the value of tax credits to the marginal investor. Hence, contrary to the arguments advanced, the usage of tax credits is not an upper bound on the market value of tax credits – if that is set by some “marginal investor”. Alternatively, if the market value is set by some “average” investor, an estimate for the average investor of the tax payment consequences of imputation credits distributed has relevance. As argued by SAPN, the value estimated in this way may be an upper bound due to a number of value-reducing factors. This is considered later.

147    This raises a fundamental consideration in the approach to the determination of a market value for imputation credits and implications for the cost of capital. Officer’s approach, and consequent adoption of the “vanilla WACC” approach in the PTRM does not imply anything about the precise value of γ nor how it is determined. Officer does argue that the required return on equity will be lower than would otherwise be the case (in the absence of imputation) by the value of franking credits (γ) received with dividends. And while he uses an “imputation adjusted CAPM” to illustrate, the determination of the value of γ involved in that is not specified.

148    As noted above, there are, at least, two alternative theoretical approaches to considering the determination of γ. One relies on an average investor perspective, the other on a marginal investor. On this matter, the AER (Final Decision, Attachment 4, p 60) quotes from a recent academic paper (A. Ainsworth, G. Partington and G. Warren, Do Franking Credits Matter? Exploring the Financial Implications of Dividend Imputation, May 2015):

Indeed, whether prices are set by a marginal investor, or by aggregation across investors, is an open question …

149    The average investor perspective is reflected in versions of the CAPM which allow for the imputation system such as those developed by P. Monkhouse and by Professors M. Lally and T. van Zijl, and considered by the AER. These models give rise to a utilisation rate which “in equilibrium is equal to the weighted average, by wealth and risk aversion, of the individual utilisation rates of investors in the market” (Final Decision Attachment 4, p 52). In those models, this utilisation rate is also the value of distributed imputation credits in the sense that the equilibrium required rate of return on equity (ignoring the franking credit component received) is reduced by that figure. For example, if an investor required a 10% return on a stock offering returns in a form involving no receipt of imputation credits, then if an otherwise equivalent stock provided an imputation credit yield of 3% which was valued at 50% of face value, the required return ignoring imputation credits would be 8.5%. This would imply that in the PTRM where a vanilla WACC is used and the effect of imputation incorporated into the required tax revenue component of cash flows, the utilisation rate is an appropriate estimate of theta (as an input into gamma) which can be used.

150    SAPN challenges the validity of use of these tax adjusted CAPM models, and the implication that the utilisation rate (estimated using either shareholder nationality or taxation statistics) can be used to estimate gamma. The principal critique is one of the logical consistency of using a “domestic” CAPM model when an international perspective is required. Interpreting utilisation as reflecting the proportion of domestic investors, for whom imputation credits are valuable, implies that the remaining investors are foreigners. But, in that case, it is argued, the CAPM market portfolio should be the world portfolio, and those foreigner investors would have a wealth many multiples of the amount they have invested in the local market (and of domestic investors). In the limit, the role of domestic investors in domestic stock valuation and implied value of imputation credits would be zero. The AER (Final Report, Attachment 4 p 64) considers this critique and on the advice of its expert (Associate Professor J. Handley) rejects it. Underpinning that rejection is the argument that the CAPM assumes only that there is specified group of investors who hold the specified set (market) of assets being considered. The equilibrium individual asset prices and expected returns are derived by reference to their systematic risk relative to that specified market. In that framework, which takes as given the wealth invested by foreigners in the domestic market, the weights implied in the determination of gamma would reflect relative wealth shares of the domestic market, not total wealth.

151    Such a framework may, or may not be reasonable, but there is no one generally accepted model for dealing with asset pricing in a domestic market where foreigners subject to different tax considerations participate. As the AER notes (p 64), an “international” CAPM (of which there are a number of variants) could be adopted but, as well as not being advocated by SAPN, this would require fundamental re-estimation of betas and cost of equity, as well as valuation of imputation credits.

152    On this matter, the Tribunal recognises the fact that models are, at best, simplifications of and approximations to reality. There has been, and likely always will be, debate over how to assess the relative merits of alternative models. One perspective is that models should be evaluated by explanatory performance, rather than by the reality of their assumptions. On this criterion, the possible lack of realism of the Monkhouse and Lally and van Zijl models, argued by SFG (Estimating Gamma for Regulatory Purposes, February 2015) for SAPN, is not necessarily a reason for rejecting them, provided that their predictions are not rejected empirically. Unfortunately, in terms of resolving the debate over these matters, the relevant empirical evidence is quite disparate.

153    An alternative approach to considering the valuation of imputation credits is based on assuming that it is the “marginal investor” who determines security prices and thus sets the value of imputation credits impounded into that price. As in standard price theory, investors may have different valuations for a stock (reflecting different information or personal considerations including tax), and specific amounts they demand at various market prices. Limits to arbitrage, risk, wealth constraints, or other factors may lead to their demands not being infinitely elastic at some price. Aggregated, these investor demands generate a downward sloping total demand curve for the stock. The market clearing price will be that at which the marginal investor’s valuation equals that price, with some infra-marginal investors placing a valuation on their holding above that price. In theory, at least, the marginal investor is either a domestic investor for whom theta = 1 or a foreign investor for whom theta = 0. In practice, determination of share prices (from which implied market value estimates of imputation credit value might be derived) is more complex, reflecting market and individual characteristics (including limits to arbitrage).

154    There is no guarantee (nor reason to necessarily expect) that the marginal, price-setting investor will be the same type of investor at all points in time. For example, institutional features of the stock market mean that returns only have an explicit dividend (and franking credit) component (rather than comprising only capital gains or losses) if stocks are held on the date of record. Moreover, the franking component of dividends can only be useable if holdings satisfy the 45-day holding requirement around the ex-div date. There are also constraints on franking credit usage if the stock holding is hedged against price risk (such as by use of derivatives).

155    The marginal investor perspective has several implications for assessing dividend drop-off studies, which are based on prices being determined by arbitrage by marginal investors (or by average investors active in the market at that time). First, if drop-off studies are estimating valuation of a marginal investor, the imputation credit usage figure, being an average, is not a relevant upper bound for that valuation. It may be so, although not necessarily, for the alternative approach where valuation is based on the average investor. Secondly, for reliable information to be extracted from drop-off studies, constancy of the marginal investor pre and post the dividend ex-date is required. Otherwise the drop-off figures may reflect nothing more than a short-term change in the composition of investors, and not information about the role of imputation credits in determining the required return of longer-term investors.

156    In conclusion, the Tribunal is of the view, reflected in the diversity of expert opinion, that there is no generally accepted theoretical model for explaining the valuation of imputation credits. There is broad agreement across experts that the existence of the imputation system lowers, to some degree, the cost of equity capital to Australian, domestically operating, companies, relative to a classical tax system. That is, required returns ignoring imputation credits (ie expected cash dividends plus capital gains) will be lower if imputation credits are expected to be attached to cash dividends.

157    Within the framework of the PTRM and “vanilla WACC” requirement of the NER, this translates into a reduction in the revenue allowance for corporate tax payments. Within the Officer framework underpinning that approach, that revenue reduction could be interpreted as reflecting that some part of company tax payments is a pre-payment (withholding) of personal taxes. Alternatively, the interpretation could be that the revenue reduction is the value accorded by the market to that pre-payment of personal taxes and reflected in the (non-vanilla) cost of equity capital. Under the “average investor” approach these should, in theory, coincide. Under the marginal investor approach, the relationship is indeterminate – unless specific assumptions about the identity (and constancy) of the marginal investor are made.

158    Unfortunately, the available empirical evidence is inadequate to enable confident discrimination between these alternative perspectives. There are a range of studies, reviewed in the AER’s Final Decision, using market prices which attempt to estimate the extent to which imputation credits are capitalised into stock prices and thus their market valuation. There are a range of results, and experts are divided on the merits of the various approaches and techniques.

159    Consequently, the Tribunal is of the view that the AER did not err, nor was unreasonable, in giving most weight to the “utilisation” approach. It considered the range of alternative approaches, recognised the diversity of views of experts on their merits (both theoretical and empirical), and made a judgement call. In doing so, it demonstrated responsiveness to the empirical evidence in lowering its estimate of gamma from 0.5 as proposed in its ROR Guidelines to a value of 0.4. The Tribunal recognises that this decision is the converse of that made by a differently constituted Tribunal in the Ausgrid case. The reason for this difference is twofold. First, submissions in this hearing gave greater attention to the theoretical underpinnings of the PTRM and “vanilla WACC” framework. Secondly, and as discussed immediately below, this Tribunal is of the view that the dividend drop-off evidence should be viewed in the context of the theoretical model underpinning it, and that there are significant uncertainties associated with extracting reliable evidence about tax-related parameters (such as gamma) from such studies.

Dividend drop-off evidence

160    The Tribunal now considers the second and sixth issues noted above which both assert that the AER should have given more (or complete) weight to evidence from dividend drop-off studies (and in particular the SFG study).

161    SAPN (and other DNSPs in other cases) argue that, in estimating the value of distributed imputation credits, most if not complete weight should be given to results from dividend drop-off studies. In particular, it is argued that the primary piece of evidence should be the results from a 2013 study by SFG commissioned by the Energy Networks Association. The latest version of this study updates an earlier 2011 study by SFG which had been undertaken in response to directions from the Tribunal. In contrast the AER argues that evidence from this type of approach should be given least weight.

162    The 2011 study concluded that “the appropriate estimate of theta from the dividend drop-off analysis that we have performed is 0.35 and that this estimate is paired with an estimate of the value of cash dividends in the range of 0.85 to 0.90.” The 2013 study which uses data from 2001 to October 2012 generates similar results.

163    The Tribunal notes that the SFG study is very clear about the data used and econometric techniques employed. Different specifications (reflecting statistical considerations required to achieve unbiased, efficient estimates) of the basic relationship estimated generate similar results. That basic relationship links the fall in stock price on the ex-dividend date (the drop-off) to the amount of the cash dividend and the amount of the franking (imputation) credit. Because the study includes dividend events which may involve no, partial, or full franking, it is able to estimate the sensitivity of the drop-off to both the size of dividend and the size of the franking credit in a regression relationship.

164    However, the Tribunal also notes that interpreting the drop-off sensitivity as evidence about the implicit value of a tax parameter such as the “value” of a franking credit (gamma) requires a number of further, quite strong, assumptions. In its Final Decision (Attachment 4, Section A.14) the AER gave explicit attention to issues involved in estimation of gamma from dividend drop-off studies. It considered the view of a number of experts including SFG (Professor S. Gray), Associate Professor J. Handley, Associate Professor M. Lally, Professor M. McKenzie and Associate Professor G. Partington, and examined a range of academic studies.

165    A number of issues regarding reliability of dividend drop-off studies were considered by the AER, but only one is substantive in this context, given the Tribunal’s view that the methodology and approach of the SFG study is generally acceptable (or “state of the art” as claimed by the DNSPs). The Tribunal recognises that some researchers may prefer use of alternative approaches or choices of specific data in analysing drop-off rates, but offers no comment on those empirical matters. That is the question of whether the sensitivities of the drop-off rate to dividend characteristics as estimated in such a study can be reliably interpreted as reflecting specific tax-related factors (such as the value of gamma). On this there is marked disagreement between the experts, with SFG (Professor S. Gray) asserting that the value of imputation credits can be reliably inferred, and the others referred to above, advancing a range of reasons in support of an opposing position.

166    In its Final Decision (Attachment 4, p 48) the AER notes that:

•    There is no market for imputation credits and therefore there is no directly observable market price. …

•    The value of imputation credits as estimated through a dividend drop off study:

•    is not necessarily a correct post company tax value before personal taxes and personal transaction costs (particularly demonstrated where the study estimates a value for cash dividends materially below their face value as the SFG study does) …

167    In coming to this conclusion, the AER liberally quotes criticisms of dividend drop-off studies from experts Associate Professor J. Handley, Associate Professor M. Lally, Professor M. McKenzie and Associate Professor G. Partington, and other authors. There are a range of criticisms, with the most significant being that of whether dividend drop-off information can provide information about tax parameters, such as the value of imputation credits. In this regard the AER quotes from a paper by K. Siau, S. Sault and G. Warren (2013) which references a range of other well-known published studies which have identified the problem that drop-off rates can reflect a range of other influences other than taxation:

A key methodological issue is that price movements around ex-dividend events encapsulate not only the tax differential effect, but may also reflect the presence of traders seeking to arbitrage dividends and noise associated with trading activity around ex-dividend dates. Drop-off ratios can be distorted by the need to compensate traders for transaction costs (Eades, Hess, and Kim ,1984); Lakonishok and Vermaelen, 1986; Karpoff and Walkling, 1988, 1990; Bali and Francis, 2011); or the risk involved (Fedenia and Grammatikos, 1993; Grammatikos, 1989; Heath and Jarrow, 1988; Michaely and Vila, 1995). Transaction costs may be substantial, and can drive the drop-off ratio below one (Kalay 1982, 1984; Boyd and Jagannathan, 1994). Market microstructure effects may also complicate estimation of market value, as discrete tick sizes can bias drop-off ratios downwards (Dubofsky, 1992; Bali and Hite, 1998).

A further key methodological issue is the difficulty in attributing the observed drop-off value between cash dividends and imputation credits.

168    The Tribunal also notes that the typical argument used to motivate use of drop-off studies is one of arbitrage based on differences between the cum-div price (and entitlement to dividend) and the ex-div price (with no entitlement to dividend). The “marginal investor(s)” would, in the absence of substantive transactions costs and risk, adopt a trading strategy to capitalise upon tax consequences. This behaviour would cause the price drop-off to reflect characteristics of the tax code, such as differences in capital gains tax rates and dividend tax rates, possibly allowing inferences to be made about the tax arrangements facing the marginal investor(s). But as noted above, evidence of the price drop-off differing systematically from the dividend amount could reflect a number of other factors in addition to, or other than, tax implications. These also include the “marginal investor(s)” (with differing tax considerations) being different on the cum-div and ex-div dates, or different to the marginal investor throughout the rest of the year.

169    The Tribunal also notes that such a rationale for drop-off studies appears to make attempting to estimate the market value of franking credits from such a study highly problematic. The reason is that the operation of the 45-day rule impacts upon the consequences of such trading strategies. For example, purchase of a stock cum-div and sale on the ex-div day (as implied by an arbitrage argument where the marginal investor anticipates a sufficiently low drop-off) would generally void the ability of the investor to use franking credits attached to the dividend due to the 45-day rule. This type of arbitrage strategy would thus tend to imply an equilibrium expected price drop-off (where no arbitrage gains result) equal to the cash value of the dividend, regardless of how much imputation credits were valued by longer-term investors. The situation is obviously more complex than this since equilibrium requires analysis of capital gain/loss tax consequences from trading strategies involving sales by existing holders or short sales and associated stock borrowing costs – including requirements to compensate stock lenders for the equivalent of any franked dividends received.

170    The Tribunal is thus of the view that while dividend drop-off studies may convey some information about tax parameters and valuation of their consequences for the set(s) of investors determining stock prices around the ex-div date, there are too many other confounding factors to place sole, or even, major weight on such studies for the estimation of the value of franking credits in the context of the PTRM.

171    The AER makes a number of other criticisms of drop-off studies. A general one is that there have been a range of results, sometimes implausible, from different drop-off studies. The Tribunal accepts those observations as a matter of fact, but does not regard them as a valid reason to reject the statistical results of a particular study (such as that by SFG). However, as mentioned above, the fundamental issue is whether valid tax related valuation parameters can be reliably inferred from such statistical results. Because of the weight of expert evidence questioning that such inferences can be reliably drawn, and the AER reliance on that evidence in forming a judgement, the Tribunal does not believe it needs to address those other criticisms. The uncertainty associated with drawing conclusions about the value of imputation credits from any existing drop-off study (no matter how well specified and conducted) was sufficient for the AER to make a judgement to accord limited weight to this type of evidence. Consequently, the Tribunal does not agree that the AER erred or was unreasonable in placing less weight on dividend drop-off studies in the estimation of the value of gamma.

The role of personal costs

172    In this section the Tribunal considers SAPN’s third and fifth contentions. SAPN argues that (a) there are a range of personal costs faced by investors in obtaining “value” from imputation credits which would reduce the value below that implied by their face value, and (b) this would reduce the market valuation of imputation credits relative to estimates based on the shareholder nationality or taxation statistics. Thus, the required return on equity would be higher (gamma would be lower) than implied by using a gamma value based on the shareholder nationality or taxation statistics estimates.

173    Three types of such costs are emphasised. One is a time-value of money effect. It is argued that the delays investors experience in receiving tax offsets or rebates from using franking credits reduces their value relative to face value. A second one is the consequences of the 45-day rule, precluding investors from using tax credits unless they have held the stock unhedged for that length of time around the ex-div date. Some Australian investors may have violated this rule and thus the value obtained, in terms of tax reduction or rebate, from receipt of franking credits may be less than the face value received. The third argument is that investors who aim to maximise imputation credit receipts will suffer from reduced portfolio diversification which imposes a cost, in the form of increased risk, upon them.

174    There can be little doubt that the investor-level (or personal) costs involved in making and managing investments will affect the price they are willing to pay for (or sell) an asset. For some investors, certain of those costs might be high and for others inconsequential. How they impact upon market prices, and the valuation of certain asset characteristics such as imputation credit yield, is not something well determined by theory. In practice, market prices might incorporate the personal costs of a marginal investor for whom those costs are trivial (or otherwise), or reflect some average of such costs across investors. Moreover, the effects of such costs on asset demand will be intermingled with a range of other asset price determinants such as differential views on fundamental values and liquidity needs. There would thus, it seems, need to be a very strong case to identify some such general investor costs and attribute to them some systematic effect on asset prices (and valuation of imputation credits). This is particularly so when the argument advanced is that these factors imply that predictions drawn from a theoretical model (in this case a tax-adjusted CAPM implying a price-setting role for some “average investor”) which ignores such factors, will be biased in a particular direction. Such factors could, for example, already be reflected in the composition of investors such that this group comprises those to whom such costs are of minor significance. It is not clear to the Tribunal that such a case has been substantiated.

175    SAPN contends that the delay in accessing tax benefits from receipt of imputation credits will reduce their present value to below face value. That may be the case for some investors, but would appear to imply logically that such investors will also discount the tax costs associated with any taxes yet to be paid on dividends. The net effect is unclear, and introduces the confounding effect of investor discounting of future tax liabilities from actual (or as yet unrealised) capital gains or losses which are otherwise ignored. More generally, it is not obvious that a fund manager who uses imputation credits to offset other future known or expected tax liabilities on their portfolio would find merit in applying such discounting. As a practical matter, the relatively short time lags and currently low discount rate which would likely be applied suggest that this effect is likely to be small. Because there may well be investors important to price determination who do not discount such tax-benefit delay consequences, the Tribunal is not convinced of the materiality of this argument.

176    SAPN contends that the 45-day rule will reduce the ability of domestic investors to utilise all imputation credits received. Evidence presented to the Tribunal about the reliability of tax statistics suggested that the materiality of this point is hard to judge. More generally, while this effect may be relevant, its effect on the valuation of imputation credits is far from clear. It is, for example, an institutional factor which is outside the formal structure of models used for analysis, such that the consequences with regard to results of such models which ignore it in their determination of asset prices is unclear. As noted earlier, the Tribunal also questions whether the existence of the 45-day rule (which voids imputation credits from short-term trading around the ex-div date) impedes the ability of dividend drop-off studies to inform on the value of imputation credits.

177    SAPN contends that domestic investors who structure their portfolios to acquire imputation credits will incur costs due to loss of diversification benefits, and that this will reduce the implicit price they are willing to pay for imputation credits to something below face value. Given that there is a well-documented “home-bias” in investor portfolios (found internationally generally regardless of tax systems), implying incomplete diversification benefits, the extent to which this is an additional factor of significant materiality is unclear.

178    The Tribunal is of the view that while some investors do experience investor level (personal) costs in dealing in equities, these can vary substantially across investor groups. It is thus not clear what effect such costs will have on equity market prices or on the need to adjust estimates for implied values of franking credits drawn from shareholder distribution or tax statistics.

Estimates using shareholder nationality, company type, and tax statistics

179    SAPN’s contention that the AER erred in adopting a wrong conception of “value” of imputation credits and consequently placing more weight on shareholder and tax statistics then on dividend drop-off results for estimation purposes has been dealt with earlier. The Tribunal determined that there was no error involved.

180    This leaves as a matter for further contention the appropriate statistics to be used which forms the basis of the fourth contention of SAPN. SAPN contend that the AER erred in three ways in data choices in estimating gamma. However, because the AER only gave a range of figures for each of the inputs to the gamma calculation and not explicit figures used in reaching its final estimate of gamma, the issue before the Tribunal is whether the net result of its decision based on these ranges for inputs is unreasonable or in error. To assess that, however, it is necessary to determine whether the ranges adopted were reasonable.

181    The following Table (extracted from Table 1 of SAPN’s Outline of Submissions) provides a concise summary of the AER analysis and decisions which are points of contention.

The distribution rate

182    First, the AER considered the possibility of a different value for the Distribution Rate (F) to the figure of 0.7 based on all equity, which was adopted in the ROR Guideline. An alternative figure of 0.77 was considered based on listed equity only. This raises the question of whether the distribution rate of the hypothetical benchmark efficient entity (‘BEE’) is better proxied by the average of any company or listed companies only. SAPN asserts that by having regard to the distribution rate of listed companies the AER was in error. The principal reason advanced is that the BEE is assumed to have only domestic earnings, whereas many large listed companies have foreign earnings which do not generate imputation credits. Then, if the dividend payout ratio (dividends/earnings) is the same as for the BEE, the distribution rate of imputation credits (credits distributed/credits generated) will be higher. That argument, does not, however, allow for the possibility that such companies with foreign earnings have a lower dividend payout ratio.

183    More generally, dividend payout ratios and distribution rates can be expected to vary between companies based on ownership characteristics and need/preferences for internally generated capital. Unlisted companies vary markedly. At one extreme there are small companies owned by individuals on high marginal tax rates who may prefer earnings retention to generate concessionally-taxed long-term capital gains or to defer the additional tax which would need to be paid on franked dividends. At the other extreme, large foreign-owned Australian registered companies may also prefer retention and reinvestment of earnings rather than distribution of dividends and attached franking credits which would be wasted.

184    The conclusion the Tribunal reaches is that there is no compelling reason which has been advanced to believe that the “average” unlisted company is any better or worse than the “average” listed company as a proxy for the BEE. Consequently the Tribunal does not believe that the AER made an error or was unreasonable or incorrectly exercised its discretion in considering estimates of distribution rates for listed entities.

Equity ownership data

185    The AER placed most reliance on estimates of the domestic ownership rate for estimating theta. This is consistent with its conceptual interpretation of the role of gamma in the PTRM and the NER as considered earlier in these reasons, and with which this Tribunal does not find error.

186    SAPN contends that the AER then erred in two regards in use of ownership statistics. First, it argues that the AER did not take account of the fact that the domestic ownership percentage is at best an upper bound on the value of theta because of the inability of some investors to use imputation credits. As this Tribunal has noted previously, that argument is incorrect if it is believed that stock prices (and thus imputation credit value) are determined by some marginal investor. Even if the average investor perspective is taken, there may be other relevant factors, not adequately captured in theoretical models, which preclude an interpretation as an upper bound.

187    Secondly, SAPN contends that the actual equity ownership figures used by the AER are essentially not fit for purpose. The principal argument advanced here is that the AER uses data since 2000 in determining the potential range of relevant values for equity ownership. Instead, SAPN argues, it should be the current rate of domestic equity ownership which should be used. It points to the AER’s own analysis (Final Decision, Attachment 4, p 96, Figure 4-3) showing that the domestic ownership share at the end of 2014 was 0.6 for total equity and 0.45 for listed equity. These were both marginally below the midpoint of the ranges of 0.56-0.68 and 0.38-0.55 respectively in post-2000 data considered by the AER.

188    The AER’s counterargument is that (a) there is considerable temporal variability in ownership shares; (b) there is a degree of uncertainty about the precision of available information; and (c) the regulatory control period spans five years such that it is necessary to predict ownership rates over the whole of that period (AER’s Outline of Submissions, para 96). The Tribunal is unable to see the validity of the last point in the context of estimating the value of imputation credits as at the start of the regulatory period. It relates to an interpretation of theta as purely a utilisation rate over a five-year period. While the Tribunal recognises the need for analysis of historical data to reduce uncertainty surrounding current figures, it would expect that sound reasons would be provided for using figures significantly different from the current value given that this is close to the historical average. As discussed below, the implied value for theta used by the AER in adjusting the MRP is 0.6 which is not inconsistent with the 2014 domestic ownership share of total equity.

189    SAPN also asserts (Outline of Submissions, para 127) that the listed company, domestic equity ownership, figure will provide a better proxy (than the total figure which includes unlisted companies) for that applicable to a BEE. This argument appears at variance with SAPN’s argument in the context of distribution rates (discussed above) that unlisted equity would provide a better proxy than listed equity for the BEE. The Tribunal is however willing to accept that there might be different proxies better suited to estimation of different characteristics of the hypothetical BEE. Nevertheless, the Tribunal has not been presented with convincing evidence that the listed equity data should not have been considered by the AER.

190    The Tribunal thus does not find that the AER erred or that its decision was unreasonable in considering historical data on domestic equity ownership shares for both listed and all companies. The issue of whether it was unreasonable in forming a judgement about gamma, based on weighing a range of evidence, including the implied choice of values from the ranges examined, is considered below.

The use of taxation statistics

191    SAPN contends that the AER erred, or made an incorrect exercise of discretion and was unreasonable in its use of taxation statistics by failing to recognise that the credit redemption rate of 0.45 from such statistics (which number is not in dispute) (a) did not indicate the value of imputation credits to investors and (b) and was an upper bound on the value of theta.

192    The AER gave less weight to the taxation statistics than to equity ownership information (and more weight than to market value (dividend drop-off studies). It argued for giving greater weight to equity ownership data on the grounds of concerns about the reliability of the tax statistics.

193    The Tribunal was presented with a range of information which indicates that there is a considerable degree of uncertainty about the reliability of the tax statistics regarding the redemption rate of imputation credits. That, of itself, implies that the redemption rate figure from tax statistics is not an upper bound for the utilisation rate, but rather a noisy estimate. Whether, ignoring that fact, it would be an upper bound for the market valuation of distributed credits has been considered earlier. As explained there, the answer to that question depends upon the unsettled issue, on which experts disagree, of how stock market prices are determined. In particular, if a marginal price-setting investor perspective is taken, the average utilisation rate implies nothing about valuation by a marginal investor. If an average investor perspective is taken, then the figure is potentially more relevant to being a noisy upper bound on valuation.

194    The Tribunal does note, however, that the arguments advance in the context of the shareholder-nationality approach for a market value less than implied by those statistics do not all apply in this case. In particular, the argument relating to the 45-day rule would be irrelevant, since any effect of that would already be reflected in the utilisation rate being less than 100%.

195    The argument that the redemption rate does not reflect the value of imputation credits to investors has also been considered earlier. The Tribunal has noted that experts are divided on this issue, and has found no reason to accept that the interpretation by the AER is incorrect.

Conclusion

196    In the face of significant uncertainty, the approach by the AER of considering a range of approaches to estimating gamma and applying different weights to those approaches is, the Tribunal believes, appropriate. It is clear that some experts would apply different weights to the alternative types of evidence, and that some support the AER’s relative ranking while others disagree. In particular, some would accord much higher weight to results of dividend drop-off studies. The Tribunal has noted the arguments about the problems of deriving reliable tax-related parameters such as investor valuation of imputation credits from drop-off parameters, and is of the view that the AER did not err in forming the judgement it did regarding weight to give to different forms of evidence.

197    The Tribunal also feels it worth commenting on the interrelation between the value of gamma and the market risk premium (‘MRP’) estimate, since this was raised in submissions. Specifically, the AER noted that in deriving a value for the MRP based on historical data from both pre and post imputation periods, an adjustment incorporating gamma is required. It appears, from evidence submitted, that in doing so the AER has implicitly adopted a higher gamma for that adjustment than used explicitly in the matters considered here. Doing so means that the MRP and allowed rate of return on equity is higher than would be the case if a lower value was used. The AER states (Final Decision, Attachment 3, p 398) that “[o]ur adjustment is based on investors valuing distributed franking credits at 60 per cent of their face value”.

198    At the risk of repetition of earlier remarks, but to elaborate, this arises because the PTRM is a post-company tax model in which company tax payments are measured net of the implied value of personal tax pre-payments resulting from distribution of imputation credits. The “vanilla WACC” rate of return on equity component is before investor tax consequences arising from, or costs incurred in, deriving that return. It is the rate of return the company would need to provide in the form of capital gains and cash dividends to meet shareholder requirements assuming that no imputation credits were distributed. Consequently, the allowed rate of return in the revenue determination is increased, relative to that (capital gains and cash dividends return) observed in the market, by the value of imputation credits distributed. In adjusting historical data on the MRP to reflect the introduction of imputation in 1987, the MRP data post that date is increased by an assumed value of imputation credits received by investors. This is offset (to some degree) by the subtraction of the value of distributed imputation credits from the allowed revenues for meeting tax obligations. Thus, a lower assumed valuation of gamma would have two partially offsetting effects on the allowed revenue. The rate of return on capital component would be reduced (since there would be a smaller upward adjustment of the rate of return) while the tax component would be increased (due to reduced deduction of a lower value of distributed credits).

199    In theory, if not in practice, there are circumstances where these effects could be exactly offsetting. One could be where the company was fully equity financed, had a CAPM beta of unity, distributed all cash flows, and the required return involved grossing up the observed market returns and MRP under imputation by a gamma factor which was used also in determining the tax revenues. In that case, it would appear that the actual value of gamma might be irrelevant to the allowed revenue determination – it would simply affect the allocation of revenue between the rate of return and tax components. In practice, the effects are less clear due to complications such as (a) estimation of a market rate of return using both historical pre-imputation and post-imputation observations; (b) tax code features such as differences in accounting and tax treatment of depreciation; (c) use of debt financing; and (d) equity betas different from unity.

200    In the current context, the possible use of a higher theta value in adjusting the historical MRP than in determining the tax revenue allowance (although the AER does not provide information on the precise value used for this latter figure) could involve some inconsistency which would work in favour of SAPN. Alternatively, if a theta value of 0.6 was also applied in the determination of gamma, a result for gamma of 0.4 suggests use of a distribution rate of slightly below 0.7. That would make redundant the debate over whether AER’s consideration of the distribution rate of 0.77 of listed companies is preferable to a consideration of all companies (with a distribution rate of 0.7).

201    The Tribunal simply notes that (a) consistency in the adjustment of the MRP for the value of franking credits distributed (theta) and in the value assumed for determining the net corporate tax revenue component is desirable; and (b) assumption of a higher or lower value for theta (and thus, ceteris paribus) gamma will have partially offsetting effects on the rate of return and taxation components of allowable cash flows.

202    The grounds of review regarding the value of gamma are rejected.

Return on debt

203    The issue regarding the return on debt is the choice of transition arrangements between the previously used approach for determining the cost of debt of a benchmark efficient DNSP and a new approach to be fully implemented by the end of a 10-year transition process. The previous approach is referred to as an “on-the-day” approach in which the cost of debt for each year of the regulatory period was determined as that prevailing at the start of the regulatory period. The new approach is referred to as an “historical trailing average” approach and was facilitated by a change to the NER described below. In that approach the cost of debt in each year of the regulatory period is calculated as the average cost of the debt portfolio of a benchmark efficient DNSP at the start of that year, resulting from assuming that it had raised equal amounts of 10-year maturity debt over the past 10 years. It can thus vary in each year of the regulatory control period.

204    There is no dispute on the decision to change to the new approach, nor on the benefits of some form of gradual transition to the new approach. The issue is purely about the specific transition arrangements to be used. Moreover, it relates to the transition of only one of the two components of the cost of debt, namely the debt risk premium (‘DRP’). Both parties are in agreement on the merits of, and on the form of, a gradual transition arrangement for the risk-free component of the cost of debt.

205    This transition of the risk-free component involves a determination in the first year of the regulatory period using the previous “on-the-day” approach – which takes the rate prevailing in the market at (or just prior to) the start of the regulatory control period. In each subsequent year the weight given to the “on-the-day” figure (ie the rate at the start of the regulatory control period) in calculating the return on debt for that year would decline by 10% per year. The risk-free rate (calculated as an average over some previously agreed averaging period) just prior to the start of each subsequent year would be given a weight of 10% and would remain part of the calculation with that same weight for 10 years. Thus, the on-the-day rate would have 90% weight in the second year and the average rate determined for the start of that second year would have 10% weight. In the third year, the on-the-day weight would decline to 80% with the rate determined for the start of that third year being given a weight of 10% – and the rate determined for the start of the second year remaining in the calculation with a weight of 10%. After 10 years of this process full adjustment to the new “historical trailing average” approach will have occurred.

206    In the Final Decision the AER considered four options for transitioning to the new approach. The dispute relates to only two of those options. One of the options starts with an on-the-day rate for the first regulatory year of the regulatory control period and then transitions to the average approach over 10 years in the manner described above (‘Option 2’). Option 2 was the option adopted by the AER in the Final Decision and it is the same transition arrangement as outlined by the AER in its ROR Guideline – although there was no (nor needed to be no) separation of borrowing costs there into a risk-free and DRP component made there. The other option uses a hybrid approach. Under this approach the risk-free rate would transition like Option 2, however the DRP would transition immediately (‘Option 3’). Option 3 is the transition approach ultimately advanced by SAPN and rejected by the AER. SAPN challenges the decision of the AER that there should also be gradual transition arrangements for the DRP component.

207    SAPN argues that the correct approach, which it proposed, should instead be an immediate shift to Option 3.

208    The DRP prevailing at the commencement of the regulatory control period was lower (at 1.95% per annum) than its 10-year trailing average which SAPN argued was 2.37% per annum at that date, giving a gap of 0.42% per annum for the first year of the regulatory control period (SAPN’s Submissions in Reply, Figure 1). Because of the past history of the DRP (with much higher values than 1.95 occurring in the wake of the global financial crisis), SAPN argue that the gap between the Option 2 and Option 3 allowed DRPs would be significantly higher in subsequent years of the regulatory control period. That is, the Option 3 DRP would be more than 42 basis points above the Option 2 DRP in later years of the regulatory control period.

209    It is possible to calculate the effect of the AER’s decision, compared to SAPN’s preferred approach, on the allowed revenue over the regulatory period. Future values of the DRP prevailing at the start of each year (and thus the allowable DRP) are, by definition, unknown at the start of the regulatory period. But the effect of a particular outcome on the allowed DRP is the same under both Options 2 and 3. The reason is that the DRP prevailing at the start of a future year enters into the calculation of the allowable DRP for that year (and each of the subsequent 10 years) with a weight of 0.1 in both cases. While the difference in the allowed DRP between the two options will change year-by-year due to the “rolling off” of the DRP from 10 years prior in Option 3, the newly included current year DRP will affect the average calculated by both options equivalently.

210    SAPN argues that in comparing Options 2 and 3 there is a material reduction in allowed revenue under the AER’s choice of Option 2, which would be in excess of $40 million over the five-year period. The precise figure would depend upon what numbers would be used in Option 3 for past values of the DRP for a BEE. In oral submission, the AER indicated that their calculations of the effect would be $22 million over the five years: Transcript, p 157. There is debate about how to determine the appropriate numbers, and while this is relevant to the AER’s stated reasons for its preference for Option 2, the appropriate method of determining such numbers is not a matter before this Tribunal.

211    SAPN asserts that the Final Decision of the AER involved error or errors of fact, an incorrect exercise of discretion, or was unreasonable. The errors asserted relate to each of the AER’s considerations in choosing a transition method. These were given in the Final Decision (Attachment 3, pp 170-1) as:

    the impact on promoting efficient financing practices consistent with the principles of incentive based regulation;

    the impact on a BEE’s opportunity to recover at least its efficient financing costs over the life of its assets;

    matching the allowed rate of return with efficient financing cashflows over a single regulatory period, and the potential conflict between this consideration and providing a BEE with a reasonable opportunity to recover efficient financing costs over the life of its assets;

    avoiding a potential bias in regulatory decision-making that can arise from choosing an approach that uses historical data after the results of that historical data is already known;

    avoiding the practical difficulties in the use of historical data to calculate the allowed return on debt, particularly during the global financial crisis.

Background

212    The NER (cl 6.5.2) provide rules for how the return on debt, as one component of the allowed rate of return, for a DNSP is to be determined and requires it must contribute to the allowed rate of return objective (‘ARORO’) which is defined in NER (cl 6.5.2.(c)) as:

The allowed rate of return objective is that the rate of return for a Distribution Network Service Provider is to be commensurate with the efficient financing costs of a benchmark efficient entity with a similar degree of risk as that which applies to the Distribution Network Service Provider in respect of the provision of standard control services (the allowed rate of return objective).

213    Prior to rule changes in November 2012, the NER required that the return on debt must be the same figure for each year of the regulatory control period. In practice this was implemented through use of the “on-the-day” approach in which the return on debt for all years was determined as the prevailing cost of debt finance at the start of the regulatory control period (using an average over some agreed, preceding, number of days).

214    The rule changes allowed for the return on debt to be determined to be the same for each year of a regulatory control period, or to be different for each year (NER cl 6.5.2(i)) and, subject to contributing to the ARORO, NER cl 6.5.2(j) allowed for a methodology which resulted in the return on debt reflecting:

(a)    the return that would be required by debt investors in a BEE if it raised debt at the time or shortly before the making of the distribution determination for the regulatory control period; or

(b)    the average return that would have been required by debt investors in a BEE if it raised debt over an historical period prior to the commencement of a regulatory year in the regulatory control period; or

(c)    some combination of the returns referred to in subparagraphs (a) and (b).

215    The 2012 changes included NER cl 6.5.2(k)(4) which required the AER to take into account any impacts on a BEE that could arise from a change in the debt estimation methodology. Clause 6.5.2(k) states that:

In estimating the return on debt under paragraph (h), regard must be had to the following factors:

(1)    the desirability of minimising any difference between the return on debt and the return on debt of a benchmark efficient entity referred to in the allowed rate of return objective;

(2)    the interrelationship between the return on equity and the return on debt;

(3)    the incentives that the return on debt may provide in relation to capital expenditure over the regulatory control period, including as to the timing of any capital expenditure; and

(4)    any impacts (including in relation to the costs of servicing debt across regulatory control periods) on a benchmark efficient entity referred to in the allowed rate of return objective that could arise as a result of changing the methodology that is used to estimate the return on debt from one regulatory control period to the next.

216    It is noteworthy that while these requirements refer to a “benchmark efficient entity” there is no precise guidance given on the debt financing policy assumed to be followed by such an entity. Indeed NER cl 6.5.2(j) quoted above allows for virtually any pattern of financing of such an entity. The AEMC’s rule determination (AEMC, Rule Determination: National Electricity Amendment (Economic Regulation of Network Service Providers) Rule 2012) noted at 7.4.1 (p 84) that “efficient benchmarking service providers may have different efficient debt management strategies”. Moreover, NER cl 6.5.2(k) implies that it in adopting a change in methodology, consideration is to be given to the effect on a hypothetical “benchmark efficient entity”, not the DNSP which is the subject of the determination. In practice, of course, it is possible that the DNSP could be such an entity.

217    Also NER cl 6.5.2(k)(1) and (4) are not specific to the time periods to which regard needs to be given. Thus, in subcl (k)(1), the minimisation between allowed and actual cost of debt (for a BEE) could refer to the regulatory control period, or to the life of the assets being (partially) financed by that debt. Similarly, subcl (k)(4) could be interpreted as referring to only the two regulatory control periods spanning the change in methodology, or to a longer period such as the life of assets. One component of the dispute between SAPN and the AER effectively relates to different interpretations of these rules, with the AER tending to focus on effects over the life of the assets and SAPN focusing on the current regulatory control period. In the Final Decision (Attachment 3, p 153) the AER states that “[w]e discuss impacts that occur across regulatory control periods, such as over the life of a benchmark efficient entity’s regulated assets. We consider the rules require us to do so.” It is the view of this Tribunal that the conclusion of whether the AER committed errors or was unreasonable in making its decision on the transition process hinges upon the validity of this interpretation of the rules.

218    As noted above, the AER’s decision to adopt Option 2 is consistent with its ROR Guideline published in December 2013, following a draft document in August 2013. There, the AER stated that it would henceforth use a trailing average portfolio approach for total debt costs following completion of a transitional period of 10 years (of the same form as that described for the risk-free component above). It argued that this approach would be adopted for all service providers and that doing so would contribute to confidence in the regulatory process. DNSPs would therefore have been aware of the intended change some 18 months before it was due to come into effect, enabling them to make some adjustments to debt portfolio financing arrangements in anticipation.

219    In its Explanatory Statement of the ROR Guideline (p 120) the AER referenced the AEMC commentary regarding the motivation for the 2012 changes allowing for the possibility of a transition period:

The purpose … is for the regulator to have regard to the impacts of changes in the methodology for estimating the return on debt from one regulatory control period to another. Consideration should be given to the potential for consumers and service providers to face significant and unexpected change in costs or prices that may have negative effects on confidence in the predictability of the regulatory arrangements. …

Its purpose is to allow consideration of transitional strategies so that any significant costs and practical difficulties in moving from one approach to another is taken into account.

220    In the ROR Guideline at 6.1, it is stated that:

The allowed return on debt must be estimated such that it contributes to the achievement of the allowed rate of return objective. It should therefore provide compensation to a service provider for the debt financing cost which is commensurate with the efficient financing costs of a benchmark efficient entity with a similar degree of risk.

The regulatory decision process

221    SAPN provided its initial access proposal to the AER in October 2014, in which it proposed transitioning to the new trailing average approach for the cost of debt (SAPN, Regulatory Proposal, pp 338-41). In that proposal the suggested approach involved direct reference to the cost of debt for a BBB rated entity rather than separate examination of risk-free and DRP components. This was, in effect, the Option 2 approach chosen by the AER (although it did not separately identify the risk-free and DRP components of borrowing costs), and was consistent with the AER’s ROR Guideline. On 6 February 2015, before the AER’s draft decision in this case, SAPN made a submission to the AER following the AER’s November 2014 draft determinations for NSW DNSPs, which involved the adoption of Option 2. In this submission it recommended use of the hybrid (Option 3) approach.

222    In its Preliminary Decision in April 2015, the AER determined that Option 2, consistent with SAPN’s original proposal, should be adopted. It commented that it was not clear that the NER allowed for SAPN to submit a changed proposal such as via its 6 February 2015 submission (SAPN Preliminary Decision, Attachment 3, p 129).

223     Subsequently in the Final Decision (Attachment 3, p 12) the AER noted:

In the preliminary decision we accepted SA Power Networks’ initial proposal to adopt a transition applied to both the base rate and debt risk premium components of the return on debt. Under the normal decision making process, this means SA Power Networks could not change its position in the revised proposal. However, it appears the drafting of the transitional arrangements in the NER leads to an outcome which makes it possible for service providers to depart from their proposal after the AER has accepted it, and allows them to introduce a new position after the preliminary decision stage. Such an outcome raises concern on the relevance of the preliminary decision process.

224    In its Revised Proposal submitted in July 2015, SAPN explained at p 380 that the significant fall in the DRP since submission of its initial proposal, and the implication of the high weight given to the “on-the-day” approach in the Option 2 transition process thus depressing the allowed rate of return was one reason for its changed position. It also argued that the hybrid Option 3 approach was more consistent with the efficient financing processes of a benchmark firm.

225    In its Final Decision, the AER reaffirmed its decision to use Option 2 as the transition method. Other issues where there were differences in method for determining the cost of debt between the AER decision and SAPN’s Revised Proposal included adoption of a BBB+ rating rather than BBB rating for the BEE, and sources of, and required adjustments to, data for determining the DRP for 10-year debt of a BBB+ rated entity. These are not matters under review and thus do not need consideration here.

Debt financing practices and regulation

226    It is generally accepted that the debt financing practices of a regulated access provider will be influenced by the specific features of regulation. This is because the regulatory approach will determine the allowed cost of debt, and resulting revenue implications, based on its assumed financing method. Risks to the service provider of actual debt costs differing from the allowed cost (and thus revenue allowed) will arise from differences between the financing pattern adopted and that assumed in the regulatory process.

227    Under the previous “on-the-day” approach, an allowed return on debt was determined at the start of the regulatory control period and fixed for each of the years in that period. Under the building block approach, the allowed revenue over the period thus included an amount sufficient to meet interest payments on the debt financed component of the regulatory asset base in each year, calculated using that specified cost of debt figure (and specified leverage). If the entity had an actual cost of debt above or below that figure it would thus be, respectively, under-compensated or overcompensated for its debt financing costs during that regulatory control period.

228    To avoid that risk, one option would be for the DNSP to arrange its debt finance portfolio such that all debt would mature just prior to the commencement of a regulatory control period. It would then raise new debt at the same time – when the allowed cost of debt is determined. Doing so would, in principle, mean that its actual cost of debt would match the allowed cost in the new regulatory control period (or differ only to the extent that its debt costs were different to those of a BEE).

229    In practice, a number of complications mean that such an approach is unlikely to be consistent with efficient debt financing, and possibly infeasible. Particularly relevant to the current matter is the potential risks to the DNSP of attempting to refinance its entire debt portfolio at a specific point in time. It would then be subject to risk of adverse movements in its borrowing costs which are not compensated under the regulatory regime, which bases allowed cost of debt on that applicable to a BEE. Such risks could arise either from random shocks to the DRP of the issuer, unwillingness of investors to absorb a large volume of new debt from the issuer, or opportunistic behaviour by lenders aware of the issuer’s debt rollover requirements. While regulatory agreement to use some private, pre-agreed, averaging period for calculating the cost of debt can mitigate some of this risk, the concentration of debt rollover within a relatively short period is not generally accepted as good risk management practice.

230    Another complication was that, in the on-the-day approach, the AER based the cost of debt on an assumed 10-year maturity for debt, even though the regulatory control period was five years. Except in the highly unlikely event of five-year and 10-year debt having the same cost at the start of the regulatory control period, issuing five-year debt to match the five-year resetting of debt costs which occurs at the start of the next control, and thus hedge debt financing cost risks, would lead to actual debt costs differing from allowed debt costs. In general, given an historical tendency for an upward sloping yield curve (10-year rates greater than five-year rates), the DNSP issuing five-year debt would tend to have debt costs less than the corresponding allowable revenue, which might, to some degree, compensate for the rollover risks in such a strategy.