COPYRIGHT TRIBUNAL OF AUSTRALIA

Phonographic Performance Company of Australia Limited under s 154 of the Copyright Act 1968 (Cth) [2016] ACopyT 2

ORDERS

THE TRIBUNAL DIRECTS THAT:

1.    The parties are to confer and are to inform the Tribunal, by a single agreed email communication, of the further steps necessary to finalise the matter, within 28 days of today’s date.

REASONS FOR DECISION

THE TRIBUNAL:

Background

1    On 28 August 2015, we delivered reasons for our decision that the schemes referred to us by the parties were not reasonable in the circumstances in accordance with s 154(4) of the Copyright Act 1968 (Cth) (the Copyright Act) (Phonographic Performance Company of Australia Limited under s 154 of the Copyright Act 1968 (Cth) [2015] ACopyT 3). We refer to these reasons in this decision as our “principal reasons”. These reasons are to be read in conjunction with our principal reasons. Abbreviations used in these reasons are explained in our principal reasons.

2    Although our principal reasons need to be read as a whole it is convenient, for the purpose of these reasons, to summarise the conclusions we reached. These conclusions form the basis for subsequent events and assist in understanding the further conclusions we have reached.

3    After a hearing extending over some 14 days between March and June 2015 we concluded that:

Generally:

(1)    None of the schemes referred to us could be considered to be reasonable in the circumstances because PPCA’s scheme required payments on a per stream basis and CRA’s scheme(s) required payment on a percentage of revenue basis. In particular, we said at [264] that “we accept CRA’s submission that it would not be reasonable to force radio stations to pay for simulcasting on a per stream (and no other) basis”. Despite an express invitation to consider a scheme in which radio stations were given the option of paying on either basis, no scheme to that effect was proffered by the parties during the first tranche of the hearing.

(2)    Another reason that none of schemes referred to us could be considered to be reasonable in the circumstances was because the percentage of revenue rates CRA proposed were too low and the per stream rates PPCA proposed were too high. Because the parties refused to engage with the case of the other party in any meaningful way, we were left without a rational foundation in the case as put for either party for determining an upper range for a percentage of revenue rate and a lower range for a per stream rate.

(3)    A scheme which permitted radio stations to decide whether to pay on a percentage of revenue basis or a per stream basis (a hybrid scheme) would be reasonable. In this regard we said at [123] that “the advantages of per stream payments for simulcasting are such that it would be reasonable to try to encourage radio stations to implement such a process”. At [125] we added this caveat:

We do not suggest that it would be appropriate for any incentive to be provided by increasing the percentage of revenue rate that would otherwise be applied. Such an approach, in our view, would not be reasonable. The incentive relates to the implementation of the per stream payment scheme and, accordingly, it is reasonable only to reduce that rate below what it would otherwise be for a period.

(4)    A scheme which accepted that, as a starting point, the sound recording right has at least the same value as the musical works right could be reasonable.

As to reasonable percentage of revenue rates:

(5)    The value of the musical works right may be ascertained from the APRA-CRA Agreement, which involves a percentage of revenue payment.

(6)    The value of the simulcast right for musical works was embedded in the overall value of the musical works right in the APRA-CRA Agreement. Provided a rational basis was used, we were prepared to try to disaggregate the value of the simulcast right from the overall value indicated by the musical works right. The first proxy for value should be the average simulcast audience share. It would also be rational to recognise that audience share does not capture certain advantages of simulcasting over broadcasting so the proxy of simulcast audience share should be increased to ensure that the percentage of revenue rates determined would reflect the added value attributable to these benefits and our view that the simulcast audience share was likely to increase over time. In valuing the advantages of simulcasting over broadcasting we noted that it must be recognised that, unlike radio stations in the US, radio stations in Australia had not yet developed effective monetisation strategies for simulcasting with the consequence that part of what was being valued were potential benefits. We noted also, however, that we found it “difficult to imagine that these capacities will not be exploited in the foreseeable future” (at [148]) but that the “present value of a potential is necessarily less than the present value of something which, in part at least, is actually being realised” (at [273]). We also noted that “we cannot predict the future for listening via the simulcast” (at [144]).

(7)    If the value of the potential of simulcasting over and above that of broadcasting and the likely future growth of simulcasting was estimated then that value could be added to the disaggregated rates for simulcasting in the APRA-CRA Agreement to ascertain percentage of revenue rates for the simulcasting of sound recordings which should be reasonable in the circumstances.

As to reasonable per stream rates:

(8)    All of the agreements on which PPCA relied involved a per stream rate that was too high for the circumstances applying to radio stations in Australia. Even the lowest rates in the PPCA webcasting agreements entered into in 2013 and 2014 disclosed “per stream rates higher than what would be achieved in a notional bargain between PPCA (or the record companies) and radio stations” (at [305]). We considered that in any notional bargain “radio stations are likely to do substantially better than the webcasters PPCA did deals with” (at [309]).

(9)    While we could estimate a reasonable rate on this basis, another alternative was to convert the percentage of revenue rates that would be payable under the percentage of revenue payment option (determined as set out above) into per stream rates, an exercise analogous to that undertaken in Exhibit 94. This exercise would reflect our conclusion that “the per stream rate should be generally equivalent to the percentage of revenue rate we consider reasonable” (at [332(7)(b)]), albeit that perhaps for the first year the per stream rate should be discounted to give radio stations an incentive to pay on a per stream basis (at [332(7)(c)]).

Otherwise:

(10)    In the circumstances, procedural fairness required the parties to be given a further opportunity to be heard.

4    The parties have proposed amended schemes based in part on further statements of Dr Eisenach, Dr Epstein and Mr Samuel. The principal areas of dispute to which these amended schemes give rise are the rates to be selected on both a percentage of revenue basis and per stream basis, the measure of revenue against which the selected percentages are to be calculated, PPCA’s proposal to default to a per stream payment only from 2020 onwards, and Dr Eisenach’s proposed increases for selection bias as between payments on either a percentage of revenue or per stream basis.

5    A number of these issues may be disposed of in short order.

Default to per stream only payments

6    PPCA proposes that radio stations be required to pay on a per stream basis from 1 July 2020. According to PPCA, this requirement gives effect to the Tribunal’s objectives of fixing a value based on the APRA-CRA Agreement, encourages radio stations to adopt per stream reporting, and reflects sound economic logic as articulated by Dr Eisenach.

7    We are unable to accept that this proposal is reasonable in the circumstances. We consider it clear from our principal reasons that we believe that for so long as radio stations must pay for the broadcast right on a percentage of revenue basis, they should be able to pay for the simulcast right on that basis if they so wish. Our reasons in support of this position are set out in full in our principal reasons and it is unnecessary to repeat them here.

8    To the extent anything more needs to be said, we also consider it to be clear from our principal reasons that while we accept it is worthwhile to encourage radio stations to adopt a per stream payment model for simulcasting, this is to be achieved, if at all, by giving radio stations an incentive to do so, such as by giving a discount on the per stream rate compared to the percentage of revenue rate for some period. It is not to be forced on radio stations, at least not while they are required to pay for the broadcasting right on a percentage of revenue basis.

9    It follows that PPCA’s submissions to the effect that such an incentive is unlikely to work and will cease to work as soon as the incentive is removed are not a persuasive reason for adopting PPCA’s proposal. If the incentive is unlikely to work, then radio stations will either pay on a percentage of revenue basis because that is what they do for the broadcast right or will pay on whichever basis is cheaper for them – which is what we intended by proposing a hybrid scheme (discussed further below in the context of selection bias).

10    Accordingly, we consider that PPCA’s scheme must be varied by removing the progression to a per stream only payment system from 1 July 2020. This conclusion means that the concerns which CRA expressed about PPCA’s proposal for a flat fee of $5000 for certain regional stations after the expiry of the election period need not be considered.

The definition of gross revenue

11    PPCA’s position is that the definition of “gross revenue” in the APRA-CRA Agreement is out of date and amenable to “gaming” by radio stations, and thus an inappropriate model for the Tribunal to employ. The APRA-CRA Agreement captures revenue from the source of advertising during broadcasts and other revenue “directly referable” to “on air activity” (that is, broadcasting) and the communication of music “over the Website” (“Website” being defined by reference to a station’s particular website). In contrast, PPCA proposes a definition that captures all revenue and then excludes categories of revenue that have no connection to broadcasting or simulcasting (such as the lease of space on transmission towers or interest earned from investments).

12    As a matter of principle, we prefer a definition that identifies what is included in revenue as opposed to excluded from it. This is the approach in the APRA-CRA Agreement.

13    Nevertheless, it is now six years on from the negotiation of the APRA-CRA Agreement and we consider the definition in it to be a product of its times. While CRA submitted that the definition in the APRA-CRA Agreement was “technologically neutral”, that is not apparent from the face of the Agreement (irrespective of how the parties to that Agreement might conduct themselves under its provisions). In particular, the definition of “Website” is fixed by reference to an identified site in a schedule to the Agreement. Given that digital innovations occur rapidly and radio stations do not merely have a website, but have mobile applications and conduct activities such as providing their simulcasts through iHeart Radio (an aggregated simulcasting service), it would be inappropriate to repeat the terms of the deal done in 2010 in the present scheme. The synergies between online activities and simulcasting also seem not to have been apparent as at 2010, given the use of the concept of “directly referable” to the communication of music over the “Website” which appears in the APRA-CRA Agreement.

14    What then should be included in the definition of revenue? Leaving aside revenue derived from advertising during the broadcast, Dr Eisenach, Dr Epstein and Mr Samuel agreed that the definition of revenue should capture revenue which stations derive from, or which is reasonably related to, simulcasting. Irrespective of the approach taken by APRA and CRA (which is a matter for those parties and not this Tribunal), we are concerned that the test of “directly referable” in the APRA-CRA Agreement is too restrictive and liable to give rise to disputes between these parties. Accordingly, while it is not appropriate for the Tribunal to undertake a drafting exercise on behalf of the parties, we propose to identify those categories of revenue we consider should be included for the purposes of this scheme. In so doing, we accept that this probably means that this scheme involves a larger revenue base than the APRA-CRA Agreement. As explained below, we are not concerned by this probability.

15    Consistent with our reasoning below in respect of selection bias, we do not consider that an approach that attempts to result in exact equivalence of payments under this scheme between the two possible payment methods or with the APRA-CRA Agreement is necessary or desirable. Such an approach involves an unjustifiable and inappropriate assumption that the process in which we are involved embodies values of accuracy, precision and exactitude. We consider it counter-productive to the interests of the parties to adopt any suggestion to this effect. The fact is that, of its nature, the task we must perform leaves open a wide range within which any selected value might be considered reasonable. If the parties and experts called by them do not accept the existence of this wide range then, as the course of this matter shows, the capacity for sensible debate about the range itself, and the factors that might assist in selecting from the range, is undermined. A commercial resolution outside the context of this Tribunal then becomes impossible. Instead, as occurred in this case, enormous expense is incurred in attempts to unearth every possible factor that might be relevant to value, no matter how insignificant. Equally of concern, no doubt also at great expense and with good intentions, experts are permitted to create economic models which, when all is said and done, merely translate a series of broad-brush evaluative assessments into a mathematical form. Such an approach tends to suggest that there is one answer to the problem, rather than a range of reasonable outcomes based on an evaluative assessment. In case it was suggested to the contrary, we wish to say that we do not consider it inappropriate that experts and parties expressly accept and adopt a range within which they believe the value should fall. To the contrary, we think that doing so would be beneficial. It would work against the artifice involved in the expression of singular values and the adverse consequences of such artifice – namely, the tendency to lose sight of the artifice and for a party to lock itself into a single point value from which no deviation may be contemplated.

16    As a general matter, these concerns suggest that the Tribunal should be taking a more active role than it does at present about the nature and extent of the evidence with which it wishes to be presented. To enable this to occur, a different management method which enables the Tribunal to understand the substance (and not merely the existence) of the issues in dispute at an early stage, before any expert or other evidence is prepared, may be necessary. Consideration might also have to be given to the Tribunal approving the experts each party may call, the instructions to be given to any expert who is proposed to give evidence, the use of joint experts, and/or the use of a single expert appointed by the Tribunal. In particular, it might be useful for the Tribunal to be informed, before any expert report is prepared, of the various methods available to assist in valuing the relevant right and the kinds of information and exercises that would need to be undertaken to deploy these methods, before any expert is permitted to prepare an expert report. The Tribunal might then be able to assist the parties to avoid unnecessary expenditure associated with the preparation of expert evidence which the Tribunal does not consider to be useful.

17    As to the present case, acceptance of the true nature of the task involved is sufficient justification for not being overly concerned by adoption of a definition of “gross revenue” which means that this scheme has a greater potential revenue base than that in the APRA-CRA Agreement. The concept of “at least equivalent value” to the musical work rights which we reached in our principal reasons is merely the first step in this process, and itself involves a contestable evaluation. Every step that must follow is of the same character. The result will never produce true equivalence, either externally (with the APRA-CRA Agreement) or internally (as between the methods of payment). As such, any attempt to impose otherwise sound economic theories or to undertake otherwise appropriate statistical corrections (such as for selection bias) on this process seems to us to involve an undesirable pretence that what we are involved in is something other than what it is; an estimation of a range of reasonableness followed by the selection of outcomes within the range which must necessarily be understood as one of any number of possibly reasonable outcomes. Importantly, both the range and the individual outcomes are based on a weighing all of the factors we think are significant, not the application of economic theory which, to be useful, would require information of a different kind and quality from that practically available in cases such as the present.

18    We return now to those categories of revenue that we consider should be included for the purposes of this scheme. We note that there is no dispute that:

(1)    Revenue from the sale of advertisements which are broadcast should be included;

(2)    Revenue from the sale of broadcast programs which contain music should be included;

(3)    Any other revenue which is derived from any aspect of the broadcast activity itself should be included; and

(4)    Revenue from the sale of advertisements which are simulcast, whether this be before, during or after a simulcast should be included. Further, this should be so whether the simulcast is accessible from the website of the radio station, a mobile application or any other digital application.

19    We also consider that:

(1)    If an advertisement or any other revenue raising exercise appears in that part of a website or other digital application from which the simulcast can be accessed (that is, the notion of physical proximity to the simulcast application), the revenue should be included. This must include the station’s advertising (for example) on any website or digital application of a third party from which the station’s simulcast can be accessed (such as iHeart Radio).

(2)    If an advertisement or any other revenue-raising exercise appears anywhere on a website or other digital application because the simulcast has been accessed (that is, the notion of a causal relationship with the simulcast application), the revenue should be included. Again, this must include advertising (etc) on any website or digital application of a third party from which the station’s simulcast can be accessed (such as iHeart Radio).

20    We say this because we are satisfied that these revenues are derived from or so closely related to the simulcasting activity that 100% of them should be included in the relevant gross revenue.

21    A contentious issue arises where the advertising or other revenue raising exercise is part of a website or digital application, but the connection (physical or causal) with simulcasting is unclear. The APRA-CRA definition could exclude such revenues (we say no more because this is not a matter for us to decide). PPCA submits that 100% of these revenues should be included. CRA opposes this position. The difficulty with PPCA’s proposition is the fact that people might access a website or digital application for a variety of reasons and, realistically, their reasons might change during the access. This means that some people will have no interest in the simulcast when they access the website or digital application, and others listening via the simulcast will spend time viewing advertisements, playing games, entering competitions and the like. Again, we do not consider this issue capable of resolution other than by reference to broad considerations.

22    First, we find it impossible to believe that any person accessing the website or digital application for the purpose of listening to the simulcast will not be exposed to the advertisements and other revenue raising opportunities available.

23    Second, we find it hard to accept that a person who is not currently listening to a station’s broadcast would have a station website or digital application open and yet elect to listen to the broadcast rather than the simulcast, where both mediums were suitable.

24    Third, the evidence indicated that radio stations are focused on achieving listener loyalty. The broadcast, the simulcast, and other entertainment activities available on websites and digital applications are all part of a strategy to facilitate listener loyalty.

25    These factors indicate that a material proportion of these revenues should be included in the relevant revenue base. The selection of any such proportion is necessarily arbitrary. Given the synergies between simulcasting and online applications, which do not exist for broadcasting other than by way of cross-promotion, we consider that 50% of all other online revenues (that is, other than revenues where 100% should be included, as set out above) should be included within the revenue base.

26    We accept also that revenue from related entities should be included if it is revenue that, had it been received by the licensee, should be within the definition of gross revenue.

27    We accept too that “revenue” includes non-cash consideration such as credits, which seems to be common ground (cl 1(b) of the CRA definition and cl 1.2 in sch 4 of the PPCA definition).

28    We accept that a bad debt which is written off and then paid should be included in gross revenue.

29    We accept that, for the avoidance of doubt, it would be useful to identify revenues which are definitely to be excluded, generally as PPCA proposes, in cl 1.5 (provided that the drafting makes it clear that other revenues not expressly excluded are not included merely by reason of omission). The excluded list should include the agreed matters such as revenue from licensing tower space and from interest earned.

Selection bias

30    It will be apparent from what we have said above that we consider a specific adjustment to rates to account for selection bias involves the artifice of putative accuracy with which we do not wish to engage.

31    Selection bias is the effect that occurs if each station chooses to pay under the method by which it will pay the least amount. If this occurs, the overall amount received under the hybrid scheme will be less than the overall amount that would be received under a unitary scheme, all other factors remaining equal.

32    There are a number of reasons why we are untroubled by selection bias in this case.

33    The APRA-CRA Agreement involves a single payment for all identified rights in the musical works. The scheme with which we are concerned involves a payment for the simulcast right alone. While the rates are set by reference to the estimated value of the simulcast right for musical works embedded in the total values which appear in the APRA-CRA Agreement, the object is not to ensure that the amount PPCA obtains under this licence scheme is the same as the value for simulcasting derived by APRA under the APRA-CRA Agreement. This is consistent with our approach to gross revenue which includes more revenue sources than the APRA-CRA Agreement.

34    It is also unrealistic to assume that there is any way in which PPCA is guaranteed to be paid the same overall amount, irrespective of the choice of payment method made by radio stations. This is a necessary consequence of a hybrid scheme. When we referred at [332[7](b)] of our principal reasons to the per stream rate being “generally equivalent to the percentage of revenue rate”, we did not intend to suggest that it was possible or even appropriate to attempt to achieve absolute equivalence in overall payments under either method.

35    We are not concerned that the overall payments may be different depending on the choices which radio stations make. Without reprising our principal reasons, we consider that radio stations must be able to pay based on a percentage of revenue basis for the simulcast right because they already have to do so for the broadcast right. If a radio station decides it is worthwhile to calculate what it would have to pay on a per stream basis that will almost necessarily be because the station believes it may save money to do so. We would consider this a positive consequence, as there are multiple reasons for wishing to encourage payments on a per stream basis. Thus, if it is correct that low music use percentage (or MUP) stations will have an incentive to pay on a per stream basis, this is desirable. If it is also correct that high MUP stations will have an incentive to pay on a percentage of revenue basis, then that consequence is unavoidable.

36    For these reasons, we are satisfied that no specific adjustment for selection bias should be made. This does not mean, however, that the fact that radio stations get a choice and PPCA does not is irrelevant to the rate setting exercise (an issue to which we return below).

37    These conclusions also mean that, from a practical perspective, we do not need to consider further any proposed discount to the per stream rates to encourage radio stations to pay on that basis.

Rates

38    Although we have said it before, it is necessary to repeat that the task which the Tribunal is required to perform is to estimate what proportion of the total percentage of revenue rates derived from the APRA-CRA Agreement, which provide the benchmark for the total value of the sound recording right, should be attributed to simulcasting.

39    Dr Eisenach’s original opinion was that this task could not be performed. We accept that his most recent efforts represent a good faith attempt to perform the “impossible” by a process of economic modelling, but we do not believe the effort should have been made. The undesirable consequence has been to introduce a large number of new issues incapable of resolution other than by evaluative processes and, again, to yield outcomes of spurious precision and accuracy.

40    It would be possible, but not desirable, to undertake a step-by-step analysis of Dr Eisenach’s methodology for the purpose of explaining why we do not accept his conclusions. This kind of analysis, although carried out by Mr Samuel and Dr Epstein, is undesirable because it is apt to mask the nature of the task involved. The reality, which we thought apparent from our principal reasons, is this. We know that the percentage of revenue rates in the APRA-CRA Agreement include the simulcast right. We have said that we accept that the simulcast right has more than a merely nominal value. We have said that part of the value of the simulcast right is represented by simulcasting audience share. We have said that simulcasting audience share does not take account of other benefits of simulcasting, as compared to broadcasting, so that some additional value above a merely proportional audience share should be recognised. Some of those values (which will contribute to listener loyalty) exist at present, but others reflect unrealised potential. Given these parameters, seeking to disaggregate the value of the simulcast right from the overall percentage of revenue rates in the APRA-CRA Agreement by a mathematical modelling exercise strikes us as a quixotic endeavour.

41    Accordingly, we do not accept Dr Eisenach’s overall approach.

42    Dr Eisenach’s outcomes are also inconsistent with certain findings in the principal reasons.

43    For example, Dr Eisenach notes that we accepted at [116] of our principal reasons that “if an income-based valuation method is used, the issue is what the profits would have been assuming the activity sought to be valued could not be carried out”. While this is correct, as our principal reasons also disclose, we considered that the best available evidence of value is the APRA-CRA Agreement and the best proxy for the current value of the simulcast right is audience share together with an estimate for the existing advantages of simulcasting over and above broadcasting as well as the potential of simulcasting to grow and offer increased synergies with the online presence of radio stations. Given this, we are unable to see any justification for a model which starts on the basis of an estimate of profits attributable to simulcasting.

44    Even if this approach is valid, as we have said, it introduces a range of further judgments which are required about issues such as allocation of costs, the relationship between revenues and profits, and the allocation of profits based on audience share. Dr Eisenach’s most recent analysis engages with all of those issues and more. In respect of each issue, he reaches a conclusion with which Dr Epstein and Mr Samuel disagree.

45    Dr Eisenach then uses the average percentage of revenue rate which he has calculated by this method to determine equivalent per stream rates. Given the above, it is necessarily the case that his conclusions about the appropriate per stream rate starts from a position we consider invalid.

46    In this regard, Dr Eisenach noted that the per stream rates he calculated were higher than the rates in the PPCA webcasting agreements. Given our finding that radio stations would be able to negotiate better rates than those in the PPCA webcasting agreements, Dr Eisenach explained his reasons for concluding to the contrary. Based on this evidence, PPCA contends that the Tribunal’s conclusion in the principal reasons that the radio stations have greater bargaining power than the webcasters is incorrect. In particular, PPCA notes that two of the webcasters are major broadcasters (ARN and Nova), the services are essentially passive listening (like radio), and ARN and Nova negotiated rates the same as the other webcasters. The nature of the services involved is in dispute but we consider it unnecessary to resolve this issue. Assuming PPCA’s characterisation of the services as radio-like to be correct, we nevertheless do not find the analysis persuasive. The analysis fails to recognise the reality that, irrespective of the identity of the webcasters, webcasting in Australia is an immature market (in contrast to the commercial broadcasting market, which is mature and stable). Further, unlike the webcasters which are or were parties to the PPCA webcasting agreements (including Nova and ARN), all radio stations interested in simulcasting act as a collective entity for the purpose of this notional bargain. These two factors informed our conclusions in the principal reasons and continue to do so. Accordingly, the per stream rates in the PPCA webcasting agreements are too high to apply to radio stations. It follows that percentage of revenue rates which equate to these per stream rates are also too high.

47    It is also relevant that Dr Eisenach acknowledges that the added value we are attempting to capture is not “readily quantifiable”. We agree. We are quantifying the essentially unquantifiable because we are satisfied it would be unreasonable not to do so.

48    We also do not accept that the decision in the US Webcaster IV proceedings, in which the Copyright Royalty Board reduced the per stream rate for simulcasting from USD 0.0025 per stream to USD 0.0017 provides support for Dr Eisenach’s analysis. We explained in our principal reasons that there are substantial differences between the simulcasting markets in the US and Australia (at [272]-[284]). Those differences, particularly the size of the market and the well-developed strategies for the monetisation of simulcasting in the US, continue to support the conclusion which we reached in our principal reasons that the US per stream rates would represent a substantial over-value of the simulcast rights in Australia. Those conclusions remain unaffected by our acknowledgment of the fact that radio stations compete with online music service providers (at [146]) and the observations in the Webcaster IV proceedings to similar effect upon which PPCA relies.

49    It is not necessary to say much about the further calculations carried out by Mr Samuel. His main point is that we proposed that the range of 1% to 10% be applied to the percentages in the APRA-CRA Agreement and not the percentage simulcast rates in column 3 of the table at [255] of our principal reasons. We agree that the necessary consequence is that the simulcast right is allocated a proportionally greater value than the broadcast right, and that applying any increase to the rates in APRA-CRA Agreement will magnify the effect of the proportional difference. The range of 1% to 10% was expressed as an assumption (at [254]). Whether it is applied to the rates in the APRA-CRA Agreement or the column 3 rates is immaterial. The relevant issue is setting an amount of increase above the value proxy of the simulcast audience share percentage which is justified by reason of the characteristics and potentialities of simulcasting not applicable to broadcasting. We identified the matters we considered bear upon this inquiry at [130] to [156] of our principal reasons. They need not be repeated.

50    CRA made a number of points in support of its position that any adjustment to increase the value of simulcasting compared to broadcasting over and above the proxy of audience share should be small. CRA pointed out that ratings, which determine advertising rates, are based on audiences within the station’s licence area and that for regional stations, which are not subject to ratings, advertisements tend to be for local services and products. We accept this is true. However, as CRA also correctly acknowledged, the relevance of the reach of simulcasting is the capacity it provides to ensure listener loyalty. CRA pointed out that despite the influx of online music providers over recent years, the audience for radio has remained steady, which suggests that these providers are not a substantial threat to the profitability of radio. Again, we accept the accuracy of the statement about stable audience share, but consider that it further supports the existence of an additional value of simulcasting compared to broadcasting given that a listener can listen to the simulcast from any location, thus increasing listener loyalty. CRA pointed out that we had accepted that Australia was well behind the US in the development of strategies to monetise simulcasting, including by targeted advertising, which it said suggests that “a critical mass for simulcasting” to support monetisation strategies appears to be some time away. In accordance with our principal reasons, we accept that the additional benefits which will enable the direct monetisation of simulcasting are largely unrealised as yet in Australia. We also remain of the view, however, that this is likely to change in the not too distant future.

51    Dr Epstein referred to the rate increases between the APRA-CRA Agreements in 1999 and 2010 as providing support for his view that the additional benefits provided by simulcasting should be seen to be of low value. He noted that the increase for a station at 61% MUP was 0.17% (2.25% in 1999 to 2.42% in 2010). However, as we noted in our principal reasons, the increase covered the new rights incorporated by the APRA-CRA Agreement in 2010 (the webcasting, podcasting and the ACMOS rights). Further, it cannot be assumed that the broadcast right has maintained its value given the increased audience share of simulcasting. As we said previously at [238], any increase in simulcast audience, to some extent at least, involves a decrease in the broadcast audience (that is, not every new simulcast listener is a new listener; some listeners will have switched from broadcast to simulcast listening). It follows that while the overall increase in the rates payable under the APRA-CRA Agreement as between 1999 and 2010 is relatively low (an average of 7.6%), that fact discloses little about the relative value of simulcasting at present. The audience share for simulcasting has effectively doubled for metropolitan stations as between 2010 and 2015, and the potential benefits offered by simulcasting must be more apparent now than in 2010.

52    Given these conclusions, the question remains – where to from here? Step one is to evaluate a reasonable starting point based on the proxy of current simulcast audience share.

53    In our principal reasons we adopted the percentage of 10.7% for the current simulcast audience share because it was accepted as applicable at that time by Mr Samuel and Dr Epstein despite their contention that, as a result of that share being based on metropolitan stations, it was likely to over-estimate the total simulcast audience share for all radio stations. Mr Samuel, in particular, has looked at the most recent information for metropolitan and regional stations and concluded that 10% is more than sufficient to account for the current audience share. Dr Epstein also uses 10% as his starting point.

54    We are aware that the ratings information does not identify people who listen exclusively by simulcast. We are aware also that if the simulcast was not available not every listener would cease listening in every circumstance. Depending on circumstances, nearly all or very few listeners might revert to the broadcast. These are all imponderable issues. None of them suggest that the current simulcast audience share is an inappropriate starting point or that to use this as a proxy for value would over-estimate the value of simulcasting. As we have said, in an industry where listener numbers are the key driver of revenue, a capacity to reach listeners over the internet is an inherently valuable right. This suggests that current audience simulcasting share is a reasonable proxy for current simulcast value leaving aside the advantages of simulcasting over broadcasting, the potential to further monetise simulcasting, and the likely growth of simulcasting.

55    What we can say from the most recent survey information is this:

(1)    The data for regional stations is not as reliable as the data for metropolitan stations. This is because only three regional areas are rated, the absolute numbers involved are smaller and thus liable to greater distortion, and many regional stations ceased simulcasting as discussed in our principal reasons – a phenomenon which must skew the data set. All that can be said with any degree of confidence is that, at present, a greater proportion of people in metropolitan areas listen via simulcasts than in regional areas.

(2)    The data for metropolitan stations is the best we have. Every metropolitan station is rated. The ratings occur eight times each year. Data exists over a number of years. The absolute numbers involved seem large enough to give confidence in using them. However one might wish to describe the trend, listening by simulcast has increased over the past five years.

56    For these reasons, we consider that the current audience share for simulcasting should be based on the data for the metropolitan stations. While this may well over-state the total simulcast audience share across Australia, the data available from the regional stations is too unreliable to warrant a different approach.

57    The average simulcast audience share was 10.5% in 2014 and 10.1% in 2015. The average over 2014 and 2015 is thus 10.3%. We will use 10.3% as our starting figure.

58    One of the aspects of value we are trying to capture is the future growth of simulcasting (but this is not the only aspect, given the present advantages simulcasting offers compared to broadcasting as identified in our principal reasons). We have said that we consider simulcasting is likely to grow. In this regard, Dr Eisenach has carried out regression analyses in an attempt to prove that the simulcast audience share is likely to nearly double again within the next five years. We do not find this analysis either persuasive or helpful. The assumption of linear growth is unjustifiable. The growth trend predicted by Dr Eisenach’s model is demonstrably incorrect for periods where data is now available. The spurious gloss of accuracy which this analysis involves is no better than any prediction we might make.

59    Conversely, Dr Epstein contends that as ownership of smart phones and computers is now nearly universal in Australia simulcast audience share is unlikely to grow as Dr Eisenach predicts unless listening habits change. As our principal reasons disclose, we consider that listening habits are likely to change over time and thus simulcasting is likely to grow over the long-term. The slight decrease in audience share between 2014 and 2015 is not sufficient to persuade us to the contrary.

60    For these reasons, given the predictive nature of the exercise, we suggested to the parties that they might consider a scheme under which, if the average simulcasting audience share increases to a defined point over a sufficient number of years (say, three years), the rate would increase by an equivalent amount. In principle, CRA was prepared to adopt this approach. PPCA was not. Having regard to the fact that for such a model to work simulcast ratings would have to continue indefinitely and on the same basis as presently, and PPCA would have to be able to access this data, we suspect our suggestion might give rise to yet more disputes between these parties. As a result, we do not propose to pursue this option further, which leaves us with the task of prediction of growth as well as quantifying the unquantifiable.

61    Dr Epstein is content to add 1% to a current 10% audience share proxy based on his overall opinion that the additional benefits offered by simulcasting should be valued at the “low end of the indicated range”. Mr Samuel’s approach, based as it is on a percentage of the values of the simulcast audience share rather than the aggregated values of all rights in the APRA-CRA Agreement, yields lower additional values than Dr Epstein’s flat rate of a 1% “uplift”.

62    We consider that these outcomes are insufficient to reflect the present advantages and future potential of simulcasting which we identified in our principal reasons. As such, we would characterise Dr Epstein’s 11% as below the low end of the potential range of reasonable values (with Mr Samuel’s being lower still). It will be apparent from what we have said above that we consider the outcome of Dr Eisenach’s analysis – a 28% value for simulcasting – well beyond the high end of the reasonable range.

63    CRA submitted that:

While the Tribunal’s task is to estimate such benefits, the difficulty in predicting whether, and if so, when, those benefits may materialise…in itself tends to caution against attributing a relatively large value to [them]….

64    However, we have already found that: (i) there are current advantages to simulcasting which are not (and never can be) available to broadcasting, (ii) simulcasting is likely to grow, although to what extent and at what pace we cannot predict, and (iii) the current advantages of simulcasting will become increasingly important in a digital world. In other words, we are satisfied that the value of simulcasting will increase over the life of the scheme. This suggests that undue caution will unreasonably under-value the simulcast right.

65    Other factors also suggest that undue caution is inappropriate.

66    First, under both methods a radio station with a higher MUP will pay more than a radio station with a lower MUP. The radio station controls its music use and thus can control its simulcasting costs as it sees fit. PPCA has no control over music use by any station. It follows that radio stations have capacity to manage the risks of a material over-estimate but PPCA (and those it represents) have no capacity to manage the risks of a material under-estimate.

67    Second, if it is worthwhile for a station to calculate the best payment method for it from year to year, then it should be expected the radio station would do so and pay the least amount possible. Again, PPCA has no control over the choice a radio station makes.

68    Third, given propositions (ii) and (iii) as set out at [64] above we should try to ensure that the rate set reflects the estimated value of simulcasting over the life of the scheme. It is proposed that the scheme continue until 2020/2021. However, experience has shown that once rates are set in this context they tend to continue for years after a scheme or agreement has reached the end of its term. This is the case with the APRA-CRA Agreement, which is rolling over from year to year. The extraordinary effort to which the parties have gone in respect of this matter supports the conclusion that revisiting the rates in 2020/2021 is unlikely bar some material unforeseen contingency. This indicates that the scheme is likely to continue for more than five years. It follows that the estimated value we are trying to reflect is not the current value of the simulcast right but the average value over the life of the scheme until 2020/2021, recognising also that it is likely the rates will continue to apply for some time thereafter. The consequence for this case is that the rate we select must be higher than what we perceive to be the current value of the right (because we consider the value will increase, not decrease, over the life of the scheme).

69    Having regard to the matters set out above, we consider that the range of possible reasonable values for the simulcast right over the life of the scheme is between a low of around 12% of the total value for the sound recording right to a high of around 20%. The average of this range is 16%. Given the gross quality of all of the estimates involved in the process we see no reason not to adopt 16% of the total value for the simulcast sound recording right as set out below:

70    Mr Samuel and Dr Epstein both converted the percentage of revenue rate they calculated at 61% MUP (the average music use percentage across all stations) into a per stream rate using the method which we applied to Exhibit 94 in our principal reasons (at [315] to [319]), albeit updating the information in that exhibit with information now available.

71    PPCA suggested that a better figure to use than the 61% MUP rate is the proportion of the actual amount paid by CRA members to APRA compared to the total revenue base averaged over the last two years. In 2014, the percentage was 2.76%. In 2015, the percentage was 2.71%. The average over the last two years is thus 2.735%. This indicates that the average of payments overall under the APRA-CRA Agreement are within the 65-69.99% band rather than the 60-64.99%, despite the average MUP of all stations being 61%. Doing the same exercise we did in [315] of our principal reasons yields the following:

(1)    An overall average by adding all of the rates in the simulcasting @16% column and dividing by 12:

0.316%

(2)    The simulcasting @ 16% rate calculated for the overall average music use percentage of all radio stations (61%):

0.387%

72    The average of these figures is 0.35%, which we propose to adopt. As set out in [316] of our principal reasons, the per stream rates should accordingly result in radio stations paying around 0.35% of revenue for the simulcast sound recording right.

73    If the actual income figures are also used, rather than the estimates in Exhibit 94, the percentage used in [319] of our principal reasons (1.35%) should be read as 1.32%. On that basis, if the process used in our principal reasons is adopted, the calculation of the per stream rate is:

(1)    Multiply the 16% simulcast rate by the average percentage of revenue rate paid under the APRA-CRA Agreement over the last two years (2.735%). The result of this is 0.438%.

(2)    Multiply 0.438% (0.438/100) by the PPCA rate as proposed at the time Exhibit 94 was prepared ($0.0018) and divide by 1.32%. The result is $0.00059 (0.059 cents) per stream.

74    For the sake of completeness only we note that the effect of using Dr Eisenach’s new method for converting the percentage of revenue rate into a per stream rate, if applied to our inputs rather than those of Dr Eisenach, results in a per stream rate of $0.00048 (assuming the spreadsheet provided by the parties is correct and being used correctly – both matters which need not be resolved). In other words, and perhaps without the parties realising it, if the inputs are constant the per stream rate calculation method CRA advanced (consistent with the principal reasons) yields a higher rate and the per stream rate calculation method PPCA advanced (Dr Eisenach’s latest method) yields a lower rate. We prefer to proceed as we did in our principal reasons. First, the approach we used based on Exhibit 94 was exposed to the parties at the time of the principal reasons and any concerns about it could have been raised at the time or thereafter. Second, we see no reason to revisit the method so late in the life of this matter. Third, there is no reason to assume that the estimates in Dr Eisenach’s latest work are more reliable than those in his earlier work or that, if they are, the overall effect of the different inputs is desirable given what we have identified as the evaluative nature of the exercise overall.

75    The per stream rate of $0.00059, we note, is:

    about double CRA’s latest proposal ($0.00029) based on Mr Samuel’s evidence;

    about 40% above Dr Epstein’s calculation ($0.00036);

    about 30% below PPCA’s latest proposal leaving aside selection bias based on Eisenach’s evidence; and

    about 43% below PPCA’s latest proposal adjusted for selection bias based on Eisenach’s evidence.

76    Importantly, from our perspective, the per stream rate of $0.00059 is materially below the range of the PPCA webcasting agreements which, while by far the lowest rates of any identified by PPCA, we nevertheless found would be too high for the simulcast right. The fact that the per stream rate of $0.00059 is materially below the rate from the PPCA webcasting agreements, having regard to our findings about the difference between the webcasting and commercial radio industries and the bargaining power of individual webcasters and the collective of radio stations, indicates to us that the percentage of revenue rates and the per stream rate we have selected are within the (wide) bounds of reasonableness.

77    We also consider that our conclusions are able to be embodied in a variation of the amended schemes proposed by both PPCA and CRA for the purpose of s 154(4) of the Copyright Act. We will give the parties an opportunity to consider these reasons and to propose further steps required to finalise this matter as appropriate.

Associate:    

Dated:        22 April 2016