Copyright Board Canada

Commission du droit d’auteur Canada

 

Date	2012-07-06
Citation	File: Public Performance of Sound Recordings
Regime	Collective Administration of Performing Rights and of Communication Rights Copyright Act, subsection 68(3)
Members	Mr. Justice William J. Vancise Mr. Claude Majeau Mrs. Jacinthe Théberge
Proposed Tariffs Considered	Re:Sound Tariff No. 6.B – Use of Recorded Music to Accompany Physical Activities, 2008-2012

statement of royalties to be collected for the performance in public or the communication to the public by telecommunication, in canada, of published sound recordings embodying musical works and performers’ performances of such works

Reasons for decision

I. INTRODUCTION 1

II. POSITION OF THE PARTIES 1

III. EVIDENCE 2

A. Re:Sound 2

B. FIC and Goodlife 6

i. Evidence related to music use and repertoire use in fitness centres 6

ii. Evidence rebutting the case presented by Re:Sound 6

iii. Evidence in support of the Objectors’ counter-proposal 9

iv. Evidence on ability to pay 11

IV. LEGAL ISSUES 12

A. Which sound recordings are eligible to remuneration? 12

B. Does Re:Sound represent all eligible sound recordings or only those in its repertoire? 13

C. How does a sound recording join Re:Sound repertoire? 14

D. Who is entitled to collect royalties from Re:Sound when the right to remuneration has been assigned? 14

E. The relevance of the previous conclusions in these proceedings 15

V. ANALYSIS 17

A. The Re:Sound Proposal 17

i. Assumptions in the Economic Models 18

ii. Problems With Data Collection 19

iii. The Accuracy of the Calculations 20

iv. A General Problem With the Result 21

v. Conclusion 21

B. The Objectors’ Proposal 21

i. Introduction 21

ii. Re:Sound Satellite and SOCAN 22.A as Proxies 22

iii. Re:Sound 3 as Proxy 23

iv. The Proper Use of Re:Sound 3 24

v. SOCAN 19 as Proxy for Music in Classes 25

C. Other Options Considered and Rejected 27

i. A Tariff Based on the Amount of Royalties of SOCAN 19 28

ii. A Tariff Based on the Australian Tariff 28

iii. A Tariff Based on SOCAN Tariff 7 (Skating Rinks) 29

iv. A Tariff Based on SOCAN Tariff 18 (Recorded Music for Dancing) 29

v. No Tariff 30

D. The Re:Sound Tariff for Fitness Classes: A Transitional Measure 30

E. Dance Instruction, Skating and Other Physical Activities 32

VI. TARIFF WORDING 33

VII. TRANSITIONAL PROVISIONS 33

I. INTRODUCTION

[1] On March 30, 2007, pursuant to section 67.1(2) of the Copyright Act [1] (the “Act”), Re:Sound Music Licensing Company (Re:Sound) [2] filed its first proposed statement of royalties for the use of recorded music to accompany dance and fitness (Tariff 6). We issued our decision on Tariff 6.A (Use of Recorded Music to Accompany Dance) on July 15, 2011. [3] We now deal with the remaining part, Tariff 6.B (Use of Recorded Music to Accompany Physical Activities).

[2] Of those who objected to Tariff 6, only the Fitness Industry Council of Canada (FIC) and Goodlife Fitness Centres Inc. (Goodlife) (together, the “Objectors”) took a position with respect to the activities now covered by Tariff 6.B.

[3] The proposed tariff targets most forms of physical activity. However, the only evidence filed in these proceedings concerned venues, referred to hereafter as fitness centres or clubs, offering their members access to fitness facilities such as pools, courts or workout areas, with or without fitness classes. Consequently, the analysis that follows deal only with fitness centres and fitness classes. Other forms of physical activity targeted in the tariff, such as dance instruction and skating, will be addressed at the end of these reasons.

II. POSITION OF THE PARTIES

[4] Re:Sound initially proposed separate rates for fitness venues and fitness classes. For fitness venues, Re:Sound proposed a rate of 5 per cent of gross receipts, with a yearly minimum of $100. For fitness classes, it proposed a fixed fee of $3 per class.

[5] In its statement of case, Re:Sound proposed to collapse the two rates into one of $1.55 per member per month. Based on the information available, this would yield royalties of roughly $86 million per year.

[6] Re:Sound justifies this amount as follows. Fitness centres make foreground uses of sound recordings. They rely on music extensively; in group class settings, it is an essential component of the physical activity. Members are willing to pay significantly more for their memberships to have their club play recorded music. This willingness to pay translates into increased club profits. Rights holders are entitled to their fair share of these gains.

[7] The Objectors were represented by separate counsel but presented evidence jointly. They opposed Re:Sound’s position on three main grounds. First, the fitness industry could not afford to pay the proposed royalties given its precarious financial position; the retroactive application of the tariff would only add to this difficulty. Second, the proposed tariff is not based on sound economic and statistical evidence. Third, the collective overestimates the centres’ actual use of Re:Sound’s repertoire and underestimates their ability to play sound recordings for which Re:Sound is not entitled to collect royalties.

[8] The Objectors proposed alternative rates of $0.78 annually multiplied by the average weekly number of participants in fitness classes and $0.15 annually multiplied by the average weekly number of participants in workout areas. Based on the information available, this would yield royalties in the neighbourhood of $3 million per year. To arrive at this amount, the Objectors used several tariffs certified by the Board as benchmarks.

III. EVIDENCE

A. Re:Sound

[9] Dr. Costas Karageorghis, Deputy Head (Research), School of Sport and Education, Brunel University (London, U.K.) commented on the effects of music use before, during and after physical activity on measurable outcomes. Before, music optimizes the arousal level of the person readying for exercise. Music used during exercise, whether synchronous (exerciser consciously moves in time with a musical beat) or asynchronous, exerts ergogenic and psychological effects that increase the probability of achieving intended outcomes; indeed, synchronous music forms the core element of fitness classes. Post-exercise music helps recuperation. Overall, music has a consistent and measurable effect on the behaviour and psychological state of exercisers and positively influences performance. The impact increases when music is strategically selected. To fitness centres, music is both integral and essential.

[10] Dr. Elaine Popp, Program Head for the University of Guelph-Humber’s Kinesiology degree program and Associate Dean for the School of Hospitality, Recreation and Tourism, Humber Institute of Technology and Advanced Learning, provided an overview of the fitness industry and of the use and importance of recorded music to that industry. She explained how choreographing group class movements to carefully selected music energizes and motivates participants. She offered the view that only selected music can help achieve such results; simply tuning to radio will not do.

[11] Doris Tay, Director of Distribution for Re:Sound was asked to estimate Re:Sound’s share of music played in fitness centres. She analyzed a list of 173 sound recordings played at one particular Goodlife centre during a period of one day (“the Goodlife list”), supplied as part of the interrogatory process. Listed recordings were first checked against the collective’s Log Processing Distribution System, a database containing information on over one million records of sound recordings. If a recording was not listed in the database, Google searches were performed using available information. The collectives acting for makers of sound recordings, Audio-Visual Licensing Agency (AVLA) and the Société de gestion collective des droits des producteurs de phonogrammes et vidéogrammes du Québec (SOPROQ), also helped to identify the sound recordings.

[12] Ms. Tay concluded that 67 sound recordings were eligible to receive equitable remuneration and 59 were not; she was unable to determine the status of the remaining 47 recordings. Of the recordings identified, 53 per cent were determined to be eligible. Ms. Tay allocated the remaining 47 according to the same proportion, implying that undetermined recordings were as likely to be eligible as those recordings whose eligibility could be determined conclusively.

[13] In rebuttal, Ms. Tay identified what she considered to be logical flaws in the eligibility analysis performed by Mr. Black of Les Mills, and discussed below. First, he had not considered the possibility that a performer, instead of the maker, may have brought a sound recording into the repertoire. Second, he did not take into account existing arrangements among record labels to collect royalties internationally. She also gave evidence relating to several large companies that produce music for fitness, in an attempt to demonstrate that even though some claimed to offer royalty-free covers (a cover version is any recording of a song other than the first), these were in fact eligible sound recordings within the meaning of section 19 of the Act.

[14] Alan Mak, Principal, Rosen & Associates Limited testified about the accounting practices of Goodlife and the interpretation of its financial statements.

[15] The core of Re:Sound’s case rested on two reports: one by Dr. Adriana Bernardino, Executive Vice-President, Research, Advanis Inc., and one by Dr. John McHale, Established Professor and Head of Economics at the J.E. Cairnes School of Business and Economics, National University of Ireland.

[16] Dr. Bernardino’s report describes the two information gathering exercises she conducted to determine the value that clients of fitness centres attribute to recorded music to accompany fitness: a qualitative exercise – focus groups – and a quantitative exercise – a survey. She first conducted focus groups of fitness club members. In these groups, participants were asked questions about their club’s characteristics and their preferences – what was important to them in the decision to join the club. The information from the focus groups was used to design the survey.

[17] Participants in the survey were recruited by telephone using standard random sampling techniques. People who qualified for participating in the survey and agreed to participate were sent an email with a link to the online survey itself.

[18] In addition to the usual demographic information, the survey asked questions about the respondent’s current club. The responses were then used to create two sets of choice tasks. In the first, respondents had to choose between two hypothetical clubs the one they preferred. The clubs differed in their characteristics, including how crowded they were, how experienced the instructors, and (most importantly for our purposes) the nature and extent of the use of recorded music. In the second, respondents were required to assume that their current club would no longer use recorded music. They were then asked questions aimed at finding out how much the price of their membership would have to drop for them to remain at their current, music-less club instead of leaving for the alternative, music-using club.

[19] Dr. Bernardino estimated a standard discrete-choice model using the data collected from the survey. For each individual, she computed the likelihood of remaining at their current club. After averaging these likelihoods across the respondents, she computed the percentage drop in membership fee required to bring the average likelihood back to the point where it would have been had recorded music been played. In other words, the percentage measures the amount that respondents would have required as a refund in order to remain at their current, music-less club instead of moving to the alternative, music-using club. Her estimate was 32 per cent of membership fees, or approximately $15 of the average monthly fee of $45. According to Dr. Bernardino, this is a measure of the value of recorded music to fitness centre users.

[20] Dr. Joffre Dan Swait Jr., Full Professor of the Department of Marketing, Business Economics and Law in the Faculty of Business, University of Alberta, was consulted in the design stage of Dr. Bernardino’s research. Having reviewed her report, he concluded that both her selection of choice modeling for the estimation of consumer willingness to pay for the use of recorded music and her particular choice model were appropriate, and that the results of her research were valid and reliable.

[21] Dr. McHale’s report outlined some general principles of valuation and drew on Dr. Bernardino’s calculations to derive what he considered to be a fair tariff. Dr. McHale proposed a model of the fitness industry based on monopolistic competition. Monopolistic competition is a type of market structure where a large number of firms produce somewhat differentiated products. Because of the differentiation, each firm enjoys a certain amount of monopoly power. But because the products are sufficiently similar to constitute a single market, and because there is free entry and exit into the industry over time, this monopoly power is constrained by the presence of close substitutes, leading to a competitive-like market outcome.

[22] Dr. McHale described the supply-side and demand-side of the market for fitness club memberships. He assumed that the supply-side exhibits constant marginal cost: the cost to any given club of each additional member is constant. Dr. McHale further assumed that the demand curve faced by clubs exhibits constant elasticity: the number of members that would leave if fees increased by a given percentage would be the same, irrespective of the amount.

[23] Dr. McHale’s final assumption was that fitness centres and music collectives would engage in Nash bargaining over the surplus value of music. In 1950, John Nash described the axiomatic foundation of a very simple bargaining rule. If the (two) parties to a bargain are equal in all respects (preferences, patience, and bargaining power), the amount to be shared should be divided equally. In Dr. McHale’s model, the amount to be shared is the difference between the value to members of clubs using recorded music and the value to members of clubs not using recorded music. This value, which would be recorded as revenues of the club in an accounting sense, is proposed to be shared by paying half of it to the music rights holders before adjusting for repertoire.

[24] Dr. McHale used the data from the interrogatories served on Goodlife, FIC and some of its member clubs, together with Dr. Bernardino’s 32 per cent to calibrate his model. His conclusion was that the value of recorded music to fitness clubs is 16 per cent of club total [4] revenue. Re:Sound then adjusted this figure in two respects.

[25] The first adjustment reflects the fact than when recorded music is played publicly, royalties are payable for both the recording (Re:Sound) and the music (SOCAN). Since the Board has consistently held these two rights to be equal in value, Re:Sound assumed that it was entitled to no more than 8 per cent of revenues.

[26] The 8 per cent figure presumes that SOCAN and Re:Sound each control essentially all the music played in fitness centres. This can be true of SOCAN but not of Re:Sound: there are far more non-eligible recordings than there are non-eligible musical works. A repertoire use adjustment is required. Re:Sound multiplied the 8 per cent by the 53 per cent obtained by Ms. Tay in her repertoire use analysis to arrive at a tariff of 4.24 per cent of revenues.

[27] Re:Sound hypothesized that it would be challenging to require fitness centres to produce financial statements on an ongoing basis. To alleviate this, Re:Sound proposed to convert the revenue figure to a figure related to membership. It proposed using the 2008 International Health, Racquet and Sportsclub Association (IHRSA) Global Report. The report states that there are 5,047 fitness clubs in Canada, which collectively have 4,635,000 members, producing annual revenues of 1.89 billion US dollars.

[28] Re:Sound converted the 1.89 billion US dollars to 2.03 billion Canadian dollars, multiplied by 4.24 per cent and divided by 4,635,000 members. This produces a rate of $18.59 per member per year. Dividing by 12 yields the proposed Re:Sound rate of $1.55 per member per month.

B. FIC and Goodlife

i. Evidence related to music use and repertoire use in fitness centres

[29] Maureen Hagan, Vice-President of Operations, Goodlife testified about the role of fitness instructors, their use of music and the importance of creativity in delivering fitness classes.

[30] Michele Nesbitt, Group Fitness Coordinator (Edmonton) for Goodlife testified about her experience in group classes including the use and preference of various types of music in group fitness classes, the implementation of the Body Training Systems (BTS) programs and its use of cover recordings that, according to her, are royalty-free.

[31] Richard Boggs, CEO of The Step Company, testified about music use in the company’s BTS division and explained what leads him to conclude that it purchases royalty-free music from its suppliers.

[32] Malcolm Black, Managing Director, Les Mills Music Licensing Limited (LMML) and Les Mills Media Limited explained how LMML acquires the music it supplies to Les Mills International Limited (Les Mills) for use in its workout programs. Les Mills uses sound recordings from three sources. Some are used pursuant to a blanket license with Sony Music and Warner Music. Some are licensed from independent labels. Some are commissioned by LMML, who owns the rights in these recordings. [5] Mr. Black called into question Re:Sound’s analysis of 11 specific tracks on the Goodlife list. He also testified that Zumba, another company that makes fitness-oriented products, represents itself as using royalty-free music.

[33] Dan Brodbeck, Recording Engineer, Ontario Institute of Audio Recording Technology, testified on the creation and use of covers. He explained that, depending on the intention of its creator, a cover can be anything between clearly distinguishable from, and virtually identical to, the original recording.

ii. Evidence rebutting the case presented by Re:Sound

[34] The Objectors relied on three expert witnesses to counter the testimony of Drs. Bernardino and McHale – Dr. Ruth Corbin of CorbinPartners Inc., Dr. Michael Hanemann of the University of California at Berkeley, and Dr. David Reitman, Vice-President, Charles River Associates. Some of their testimony covers the same ground. For brevity, we outline each argument made by these witnesses only once, regardless of how many of them raised the point.

[35] Dr. Corbin critiqued Dr. Bernardino’s use of focus groups. Moderators repeatedly prompted participants to speak about music. Music was mentioned without prompting only once during the focus groups, yet music figured prominently in the survey design. This prominent placement introduced order bias: survey respondents tend to pay more attention to items that appear higher on a page or computer screen.

[36] Dr. Corbin considered that the sampling for the survey was not random, noting that less than 0.1 per cent of potential respondents chose to participate in the survey. No one was asked to complete the survey who did not engage in at least one activity where recorded music was played. That is, people who attended a club to play tennis or swim laps were excluded from the target population.

[37] Dr. Corbin emphasized the high likelihood of guessing at various questions. One reason for this was that respondents who said that they do not engage in a particular activity (e.g., taking cardio classes) were asked if music is played in that activity. The survey did not contain a “don’t know” option; neither did it allow participants to skip a particular question.

[38] Dr. Corbin was concerned that no attempt was made to ensure that the survey respondent and the initial contact were the same person. She argued that the methodology for determining outliers (respondents whose responses are unusual as compared to the rest of the respondents) and straightliners (respondents whose responses correspond to a straight line down the computer screen or printed page) was arbitrary and not replicable. She also took issue with the way in which the survey questions changed after the survey had begun.

[39] Dr. Corbin considered the underlying random utility model used to generate the survey to be unrealistic, since it requires decisions to be fully rational. Finally, she noted that since the Bernardino model actually measures willingness to accept, it probably overestimates willingness to pay, which survey economists and statisticians generally agree is less.

[40] Dr. Hanemann emphasized that anything that can be purchased has potential substitutes. If those substitutes are perfect, the value of the two must be identical. If the substitutability is imperfect, then each one may have some additional value relative to the other. In doing an exercise such as the one done by Dr. Bernardino, it is essential to use the next best alternative in computing value.

[41] Dr. Bernardino assumed that no music use is the next best alternative to the use of Re:Sound music. Yet Re:Sound’s own conclusion that 47 per cent of music used in fitness centres is not eligible to remuneration tends to show that the next best alternative to music in the Re:Sound repertoire is non-Re:Sound music. Incorrectly choosing “no music” as the next best alternative, where other substitutes involving music exist, necessarily overstates the economic value of Re:Sound music to club members.

[42] Dr. Hanemann critiqued the survey design as containing a forced choice task: respondents had to select one of the two fitness centres presented to them in the survey. This is both artificial and unrealistic. The test employs language that repeatedly emphasizes the hypothetical nature of the choice. It does not allow respondents to say they would select neither location, a choice always open to all. It ignores the existence of other fitness centres, including the participant’s own. Dr. Hanemann explained that forced choice is generally not used in surveys for these reasons. While the other choice task was open-ended, Dr. Hanemann rejected the idea that fusing the data from the two choice tasks removed the flaws of each. The open-ended task suffered from vagueness: the respondents were asked if they would remain at their current fitness club or “go somewhere else.” It was even possible that some people would go to no fitness club whatsoever, but it was not possible to discern this choice from the survey.

[43] Dr. Hanemann critiqued the manner in which fitness centre attributes were presented to respondents. Critical, non-essential attributes (location, availability of television sets, babysitting services) were omitted; the absence of such attributes will tend to inflate artificially the value of those attributes that are presented. At the same time, the number of attributes that were present in the two fitness centres from which the respondents had to choose should be kept under control. Having too many attributes for consideration can drive respondents to read only part of the set of attributes when responding, thus biasing the survey. Finally, giving a prominent location to music would tend to inflate its value.

[44] Dr. Hanemann drew attention to Dr. Bernardino’s treatment of the price variable in the survey in the context of the regression analysis she performed. Different clubs invoice on different schedules, and members may pay on different frequencies. Dr. Bernardino sought to avoid the problems associated with this by defining the change in price as a percentage. However, this means that the value of music is also a percentage. So a person who pays $50 per month and is willing to accept a 10 per cent membership fee increase values the playing of recorded music at $5 per month. Another respondent who also is willing to accept a 10 per cent increase in membership fees but pays $100 per month values the playing of recorded music at $10 per month. This is, as Dr. Hanemann explained, non-standard treatment of a price variable in a regression. Furthermore, it is inconsistent with the treatment of all the other variables in the survey, which were measured in absolute terms, not relative terms.

[45] Dr. Reitman explained that Dr. McHale’s model was, in fact, a model of one fitness centre, not a model of the fitness industry. The confusion is not uncommon: models of monopolistic competition are often presented this way in economics textbooks.

[46] Dr. Reitman clarified that Dr. Bernardino’s figure of 32 per cent for the value of music assumes that no other fitness club plays recorded music. The switching effects on demand are much smaller if all clubs were banned from playing recorded music.

[47] Dr. Reitman noted that Dr. McHale’s figure of 16 per cent was simply the equal split applied to Dr. Bernardino’s calculated value of music of 32 per cent. While there are substantial algebraic calculations in Dr. McHale’s report, these reduce to the equal split assumed by Dr. McHale.

[48] Dr. Reitman further noted that while music may have a positive value, the marginal value of music to a club is zero when the fitness industry is in long-run equilibrium. This is because the long-run equilibrium of an industry structured as monopolistic competition always has zero economic profits. Thus, the incremental contribution of any input to the profits of that industry must be zero. Dr. Reitman did not provide this explanation to argue that the equitable tariff should be zero; rather, he used it to argue that any model based on monopolistic competition was inappropriate for these proceedings.

iii. Evidence in support of the Objectors’ counter-proposal

[49] Dr. Reitman presented a counter-proposal, based on two assumptions. First, the value to members of music played by fitness centres is the same for all those who listen to it, whether in a workout area or in a fitness class. Second, members who use a workout area and do not listen to club-provided music still value that music, but as background music. This led him to propose one rate (Part A) for fitness classes and one (Part B) for workout areas. In turn, the value of music in the workout areas would be calculated differently for those who listen to the music being played by the centre and for those who listen to music from some other source (television set, personal audio device) or not at all.

[50] Dr. Reitman offered three possible proxies for Part A. The first is SOCAN Tariff 19 (Fitness Activities and Dance Instruction) (hereafter “SOCAN 19”): $2.14 per average weekly participant. Dr. Reitman multiplied this figure by 0.8, which is his estimate of the ratio of the value of the sound recording to the musical work in that market. The Board has consistently set this ratio at 1. Dr. Reitman offered the opinion that the ratio must be lower here. For example, the ratio of 1 has been used for commercial radio. A radio station will not attempt to pass a cover recording for the original; to that extent, the cover and the original are not substitutes. In the fitness context, the cover and the original are substitutes. Insofar as a substitution effect tends to reduce the relative value of the right to perform the sound recording, the relative value of Re:Sound’s rights is less than 100 per cent of SOCAN’s. This yields $1.71, before repertoire adjustment.

[51] The second proxy is the rate set in SOCAN Tariff 22.A for limited downloads of music (hereafter “SOCAN 22.A”). Limited downloads use technology that causes the file to become unusable when a subscription ends. Dr. Reitman drew the analogy between someone choosing sound recordings for limited download in order to optimize performance in a private workout and an instructor choosing sound recordings to play in a fitness class setting.

[52] The SOCAN 22.A rate is 6.3 per cent of revenues. At the time of the hearing Napster To Go offered limited download subscriptions at a rate of $14.95 per month. Dr. Reitman then assumed that the typical Napster user listens to music for 60 hours per month, and that the typical workout at a fitness centre lasts one hour. These assumptions yielded an hourly subscription price of $0.25. This was multiplied by 52, since the Part A rate was to be expressed as an annual rate based on one weekly visit; this yielded $12.96. Applying the SOCAN rate of 6.3 per cent to the annual subscription price yields $0.82.

[53] The third proxy is the Re:Sound satellite radio tariff (hereafter “Re:Sound Satellite”), undiscounted for repertoire, of 4.39 per cent of revenues. Sirius offered annual subscriptions at a monthly price of $13.74. Dr. Reitman then assumed that Sirius subscribers listen to 40 hours of music per month. The assumptions about workout patterns remain the same. The hourly subscription price is $0.34 and the annual price is $17.86. Applying the Re:Sound rate of 4.39 per cent to this amount yields $0.78.

[54] The rates obtained using SOCAN 19 was more than twice as high as what the other two proxies yielded. Dr. Reitman surmised that this may reflect the additional value of synchronous music uses, since SOCAN 19 targets fitness classes. From this, he concluded that the Part A rate was best calculated by reference to SOCAN 19.

[55] For Part B music, Dr. Reitman performed two calculations. To reflect his assumption that members who use a workout area and do not listen to club-provided music value that music as background music, he used the per attendee background music rate from Re:Sound Tariff 3 (hereafter “Re:Sound 3”) ($0.0008) after removing the adjustment for repertoire: $0.0019 per person per day. This was multiplied by 52, since the Part B rate was also to be expressed as an annual rate based on one weekly visit. This yielded $0.0988. Dr. Bernardino’s survey having concluded that 67 per cent of those attending fitness centres do not listen to club-provided music in workout areas, he applied that correction, to yield a rate of $0.066 per year per member.

[56] The second calculation concerned the 33 per cent of clients who listen to club-provided music in workout areas. The starting assumption was that these clients should attract Part A royalties. Yet having used SOCAN 19 to calculate Part A royalties, he then concluded that the SOCAN 22.A and Re:Sound satellite radio tariffs, being more likely to reflect the value of asynchronous uses of music, would be more appropriate to calculate the rate for music listened to in workout areas. The Re:Sound Satellite proxy yields a Part B rate of $0.32; with the SOCAN 22.A proxy, the rate is $0.34, Dr. Reitman proposed using the simple average as Part B rate: $0.33.

[57] To adjust the rate for repertoire use, Dr. Reitman referred to Dr. Bernardino’s survey, in which 53 per cent of participants declared that most music played in fitness centres is recorded music, 20.2 declare satellite radio and 26.8 per cent declared either commercial radio or no music. He disregarded the third category, applied the repertoire adjustment used to set the satellite radio tariff to the second and applied the repertoire adjustment derived from the Goodlife list to the first. This yielded a repertoire adjustment of 45.82 per cent. Applying this repertoire adjustment to the non-adjusted rates of $1.71 and $0.33 yielded final rates of $0.78 for part A and $0.15 for Part B. These represent the FIC and Goodlife counter-proposal.

iv. Evidence on ability to pay

[58] David Hardy, president of FIC and co-owner of several fitness clubs in the Edmonton area, Janice Renshaw, owner of Body Waves Fitness and Wellness Centre in Caledonia, Ontario and Mike Raymond, president of Curves International testified before the Board. Kevin Galbraith, owner of the Fitness Firm in Burlington, Ontario, filed a witness statement. These witnesses testified that their clubs would have difficulty passing on the cost of the Re:Sound tariff to their members. Mr. Raymond offered two reasons for this. First, it is not possible to raise the prices for existing members because existing members have signed contracts which maintain the price at a given level during the term of the contract. Second, while it is possible to raise prices for new members, industry practice is to offer new members discounts in order to get them to join. Thus, it is likely that even if the posted price passed on the cost of the tariff, the effective price would not be able to do so.

[59] All four witnesses also stated that their ability to pay the tariff was worsened by reason that they had not made provisions to pay the tariff retroactively; payments would come out of current revenues. In fact, Mr. Raymond estimated that about 20 per cent of Curves franchises would be at risk of closing if the proposed tariff was certified.

[60] James L. Horvath Managing Director, ValuQuest Limited, used IHRSA data to analyze the impact of the proposed tariff. He assumed that the state of the Canadian fitness industry is similar to that of the American one (from which the IHRSA report samples) and that the sampling is representative. Based on his analysis, he concluded that the proposed tariff would have a detrimental impact on the fitness industry in Canada, causing many clubs to cease operation.

[61] Armand Capisciolto, National Accounting Standards Partner for BDO Canada testified as a rebuttal witness to Mr. Mak.

[62] Mr. John Muszak, Vice-President Marketing, Goodlife commented on recorded music as value driver for getting, keeping and growing membership and testified about the possible impact of the proposed tariff on the industry.

[63] Dr. Jan Schroeder, Associate Professor of Physical Education at California State University, testified about the demographics of the North American fitness clubs, the influence of age demographics on group exercise activities and the influence of these trends on the use of music in such activities.

IV. LEGAL ISSUES

[64] Considerable confusion surfaced concerning a number of legal issues. Most concerned the extent of the repertoire for which Re:Sound is entitled to collect royalties. In a nutshell, Re:Sound claims to represent all eligible recordings. The Objectors argued that Re:Sound can collect only on account of those titles that are in its repertoire, and only if producers have not waived the right to collect royalties.

[65] The approach we use to set the tariff makes it unnecessary to deal at length with most legal issues addressed in the course of these proceedings. Nevertheless, outlining certain basic principles may help the reader understand some of our conclusions (and uncertainties) concerning the evidence.

A. Which sound recordings are eligible to remuneration?

[66] Section 20 of the Act provides that a sound recording can trigger equitable remuneration pursuant to section 19 (i.e. is eligible) if the maker resides in a Rome convention country (residence-based eligibility) or if all fixations were made in such a country (location-based eligibility). For the purposes of these proceedings, this is important in at least three respects.

[67] First, recordings made in a non-Rome country by a maker who is not a Rome resident are not eligible. This includes the vast majority of recordings made in the United States. [6] Users can avoid paying Re:Sound royalties by playing only sound recordings that are not eligible. This task becomes easier if music is provided, for example, by a supplier of exercise music or videos (e.g., Zumba), if the supplier only uses non-eligible recordings.

[68] Second, who performers are, where they live and whether they are members of a collective are irrelevant to the issue of sound recording eligibility; this depends entirely on the status of the maker or the location of the fixation.

[69] Third, residence-based eligibility is determined by looking at the maker, not its parent company or its Canadian agent. A recording made in a non-Rome country by a non-Rome maker is ineligible even if the maker is a wholly owned subsidiary of a Rome resident or if the maker’s Canadian agent is a Rome maker. Contractual arrangements between the maker and its parent company or agent cannot make eligible a recording that is not.

B. Does Re:Sound represent all eligible sound recordings or only those in its repertoire?

[70] Re:Sound claims royalties in respect of the entire eligible repertoire. Its argument can be outlined as follows. Section 20 of the Act determines eligibility. Subparagraph 68(2)(a)(i) requires that remuneration flow only to recordings that section 20 declares to be eligible. If remuneration must flow only to eligible recordings, it must flow to all such recordings. This is dictated also by the subsection 19(3) requirement that equitable remuneration always be shared equally between performers and makers. We disagree, for several reasons.

[71] First, in most regimes the Board administers, a clear distinction exists between what is eligible and what attracts royalties pursuant to a tariff. Generally speaking, a collective can only collect royalties for what its repertoire contains. The retransmission and private copying regimes are the only exceptions to this rule. In these regimes, copyright owners who have not joined a collective (sometimes referred to as “orphans”) can claim their share from a collective society the Board designates. For that reason, the amount of royalties is set at a level sufficient to compensate all eligible copyright subject matters.

[72] Second, the interpretation Re:Sound proposes is incompatible with subsection 67.1(4) of the Act, the relevant parts of which read as follows:

Where a proposed tariff is not filed with respect to the [...] sound recording in question, no action may be commenced, without the written consent of the Minister, for

[...]

(b) the recovery of royalties referred to in section 19.

[73] Section 67 provides that the only section 19 collectives targeted in sections 67 to 68.2 are those that collect the remuneration for sound recordings of musical works. If, as Re:Sound argues, it represents all eligible sound recordings of musical works, then the words “with respect to the [...] sound recording in question” (note the singular) are redundant. The eligible repertoire would be either “all in” or “all out”.

[74] Third, the 68(2)(a)(i) requirement that remuneration flow only to eligible recordings does not require that remuneration flow to all such recordings. It is just as reasonable to conclude that the requirement exists to ensure that no remuneration flow to non-eligible recordings, whether or not remuneration flows to all eligible recordings.

[75] Re:Sound is entitled to collect equitable remuneration only in regard to recordings that were brought into its repertoire. Given the wording of subsection 67.1(4) of the Act, users are entitled to use eligible recordings that are not in Re:Sound’s repertoire for free, unless the Minister responsible for the Act authorizes an action for recovery of royalties.

C. How does a sound recording join Re:Sound repertoire?

[76] Either the maker or the performer can bring an eligible recording into the Re:Sound repertoire. Furthermore, when this occurs, all the rights in the recording become part of the repertoire, not just those of the person who brought the recording into the repertoire. [7]

[77] As a result, when the recording is in the repertoire, Re:Sound is entitled to collect royalties for the benefit of both performers and makers, even if some have not assigned their rights to Re:Sound. Re:Sound is in turn required to distribute equal shares to performers and makers, whether or not they are members of a Re:Sound collective. Hence, the requirement in subsection 19(3) of the Act that equitable remuneration be equally shared between the performer and maker is satisfied.

D. Who is entitled to collect royalties from Re:Sound when the right to remuneration has been assigned?

[78] The section 19 remuneration right clearly is one of “the rights conferred by this Part” for which section 25 of the Act specifies that they are transferable. Subsection 19(3) requires that royalties be divided so that performers and makers each receive half. No mention is made of any successor in title to the remuneration right. Re:Sound relies on the underlined words to argue that it must pay royalties to the original makers or performers, irrespective of who may be entitled to that money in the end. We disagree: Re:Sound must pay the royalties to those who are actually entitled to them.

[79] Re:Sound’s own distribution practices are incompatible with its position. Its member collectives, to whom royalties are distributed, are not original makers or performers.

[80] Re:Sound’s proposition is illogical to the extent that the remuneration right often survives the demise of its original owner. Successors in title cannot cash a cheque made to the benefit of a dead performer or a non-existing label. The ability to devolve must exist. So must the ability to assign.

[81] Finally, Re:Sound’s interpretation is incompatible with section 58.1 of the Copyright Amendment Act. [8] That provision limits the ambit of any assignment of a right that would be a right to remuneration under the Act made in an agreement concluded before April 25, 1996. A provision that limits assignments is not required where no assignment is possible.

[82] Sometimes, we were left with the impression that, in Re:Sound’s opinion, eligibility to collect equitable remuneration is determined “at the time the relevant tariff is certified”. [9] This cannot be correct. Of necessity, the Board makes decisions on repertoire use based on the evidence available to it at the time of the hearing. Yet tariffs are inherently prospective, and Re:Sound’s repertoire changes every day. Of necessity, its royalty distributions will be based on the repertoire it represents when a decision to distribute is made, not on some historical snapshot of a period past. To remunerate in 2012 a recording made in 1960 and identified as eligible in a 2007 repertoire analysis would make no sense: the recording ceased to be eligible on January 1, 2011. [10] It would be just as illogical to deny remuneration in 2012 to a recording for the sole reason that it did not exist in 2007.

E. The relevance of the previous conclusions in these proceedings

[83] Re:Sound’s claim to represent 53 per cent of the repertoire is necessarily too high. First, Re:Sound claimed to represent all eligible recordings. Re:Sound is not entitled to receive royalties for recordings in the public domain, non-eligible recordings, eligible recordings that are not in Re:Sound’s repertoire and (possibly) eligible recordings in repertoire for which the rights have been otherwise cleared.

[84] Second, it is highly probable that the category of titles whose eligibility could not be determined contains a smaller proportion of Re:Sound repertoire than the known repertoire category. Re:Sound determines eligibility based on declarations by AVLA, SOPROQ or their members; it is highly improbable that a music label will file eligibility declarations concerning recordings that it does not own.

[85] Third, a cover recording is a separate sound recording, however similar it may be to the original. Many covers of eligible recordings are not themselves eligible. Not all eligible covers are in the repertoire of Re:Sound; if LMML takes full advantage of the work-for-hire provisions of the New Zealand Copyright Act, the sound recordings that LMML commissions are not in the repertoire even if the commissioned maker’s own recordings are both eligible and in the repertoire.

[86] That being said, the extent to which Zumba, Les Mills and others have succeeded in supplying recordings that do not attract Re:Sound royalties depends on a number of factors for which we have insufficient evidence. This includes the countries in which fixations are made, the nationality of the makers, the existence and applicability of work-for-hire provisions, the regulation of the relationship between eligible makers and collective societies in their country of origin and the terms of the contracts between makers and performers.

[87] For the reasons just stated, Re:Sound’s claim that it represents 53 per cent of the repertoire is clearly exaggerated. The Objectors offered the figure of 33 per cent, based on Mr. Black’s analysis of the Goodlife list. That figure probably is too low, for the reasons stated in the previous paragraph. That being said, since the analyses of Ms. Tay and Mr. Black are the only evidence that is even remotely reliable which are available to us to determine repertoire use, we will use these as best we can, based on the following principles.

[88] We start with Ms. Tay’s analysis of the 173 recordings found on the Goodlife list, [11] which she classified as eligible (Yes: 67), not eligible (No: 59) and eligibility undetermined (Maybe: 47).

[89] If the Objectors do not challenge Ms. Tay’s classification, we accept it. Not all eligible recordings are in Re:Sound’s repertoire; however, based on past experience, we come to the conclusion that the probability of this occurrence is low, especially in view of the other principles we apply.

[90] If the Objectors challenge Ms. Tay’s classification and she refutes the challenge with credible evidence, we accept Ms. Tay’s classification. This includes evidence to the effect that the performer, but not the maker, is a member of one of Re:Sound’s member collectives. Contractual arrangements between makers and performers may prevent member performers from bringing a recording into the repertoire; however, since Ms. Tay’s evidence is the only one we have, we accept it.

[91] Where the Objectors challenge Ms. Tay’s classification and she does not refute the challenge, we accept the Objectors’ position if we are satisfied that local legislation allows services such as LMML or Zumba to clear the rights in a way that may prevent the recording from being made part of Re:Sound’s repertoire. Since Re:Sound is best equipped to determine what is in its repertoire and what is not, we treat the absence of refutation as evidence to the effect that the title has remained outside of the repertoire.

[92] Our analysis yields the following.

[93] Of the YES recordings, 55 are unchallenged; they remain YES. Two are recordings for which LMML holds a licence but that Mr. Black conceded are or may be AVLA titles; they remain YES. Five are recordings for which LMML holds a licence and for which Ms. Tay provided evidence of performer membership; they remain YES. Five are recordings for which LMML holds a licence and for which Ms. Tay did not provide evidence of performer membership; they become NO.

[94] The 59 NO recordings were unchallenged; they remain NO.

[95] Of the MAYBE recordings, one is a recording for which LMML holds a licence and for which Ms. Tay provided evidence of performer membership; it becomes YES. Thirty are recordings that LMML commissioned. Ms. Tay did not provide evidence in their respect. New Zealand copyright legislation provides that the first owner of a commissioned recording is the person who commissioned it. They become NO. Fifteen are recordings found in Zumba products. Ms. Tay did not provide evidence in their respect. Australian copyright legislation provides that the first owner of a commissioned recording is the person who commissioned it. They become NO. The single remaining MAYBE is removed from the calculation entirely, since it does not significantly influence the final result.

[96] Based on the preceding analysis, we find that 63 titles are in Re:Sound’s repertoire while 109 are not. This yields a repertoire use of 63 / (63 + 109) = 36.6 per cent. The manner in which we arrive at this number is far from ideal; however, it is the only rational one we have, given the evidence made available to us and the assumptions made by the parties.

[97] We have already mentioned various factors that may justify the number we use being higher or lower. There are others which we need not list here. Suffice it to say that on the whole, these factors tend to neutralize one another and that none would lead us to conclude that the number should be adjust upwards or downwards. Consequently, 36.6 per cent is the repertoire adjustment we will apply where required.

V. ANALYSIS

A. The Re:Sound Proposal

[98] The reliability of the Re:Sound proposal depends on the reasonableness of the assumptions in the economic models of Drs. Bernardino and McHale, the reliability of the data collected (i.e. the survey methodology) and the accurateness of the calculations to produce final numbers. We enumerate four difficulties with the assumptions, six difficulties with the data collection, and a difficulty with the accurateness of the calculations, as well as an overall problem with the result.

i. Assumptions in the Economic Models

[99] First, Dr. Bernardino’s analysis assumes that the behaviour of fitness centre members can be modeled using a random utility model. Economic psychology literature suggests that people use heuristics (short-cuts) to make complex decisions; indeed, the pre-test showed that people completing the survey did so. The random utility model assumes that the decision maker chooses the alternative that maximises his utility; using heuristics violate that assumption by increasing the chances of not selecting that alternative. As a result, the estimated random utility model is a poor approximation to the choice behaviour displayed in the survey. Attempts to control for heuristics using cut-offs were never sufficiently explained; as a result, we are unable to determine whether it solved the problems of heuristics or not. Furthermore, cut-offs can only control for binary heuristic behaviour, not interactive heuristics.

[100] Second, Dr. Bernardino assumed that demand for fitness centres exhibits constant elasticity, that is, the number of people that would leave their fitness centre is the same if the monthly membership fee doubled from $300 to $600 than if it doubled from $10 to $20. This assumption does not seem reasonable to us: income is earned in dollars, not percentages. Dr. Bernardino relied on this assumption to take several short-cuts in calculating the value of recorded music to accompany fitness. If the assumption is invalid, so are these short-cuts and, by implication, the calculated value of recorded music.

[101] Third, Dr. McHale assumed that there is a representative fitness centre exhibiting a constant marginal cost for each additional member. Even if we admitted the possibility of a constant marginal cost, the likelihood of a representative fitness centre in the realm of marginal cost is extremely remote in a market where monopolistic competition exists. It strains credibility that the ongoing costs of adding a member would be the same for a centre that only offers circuit training as for a full-service club. Dr. McHale’s assumption of homogeneity is standard in the literature on monopolistic competition, but not supported by the evidence of the heterogeneous offers at fitness clubs.

[102] Fourth, we cannot accept that a classical Nash bargaining model is appropriate in this instance. First, a negotiation between Re:Sound and an association with signing authority for all the industry is too much of a construct: the industry is too heterogeneous to speak with one voice. Second, the existence of alternatives to the Re:Sound repertoire other than not playing any music at all will shift power significantly to the centres. A related, third point is that there is reason to believe that in this specific market, the person who should be bargaining for all or part of the clubs’ uses may not be the clubs themselves or their representative, but a third party provider such as LMML. Fourth, though this is a minor point, the fact that (at least from a legal perspective) the outcome of any negotiation must be filed with the Board for certification will tend to shift bargaining power in ways that it is not possible to assess.

ii. Problems With Data Collection

[103] Even if we were comfortable with the assumptions in the economic models, there are several problems with the data collection process. The fact that many of the biases in data collection are difficult to quantify complicates things further. If we could quantify them, we could decide which ones could be ignored and which ones merit corrections. In the present case, we find serious issues with the reliability of the data.

[104] First, biases are introduced by not including a “don’t know” option in the survey. This implies that respondents are forced to guess a response to a question for which they do not know the answer. From a statistical perspective, these answers may be viewed as random. The problem is that the random answers cannot be separated from the genuine answers.

[105] Second, biases are introduced by asking respondents to value attributes they do not use. Respondents do place value on such attributes, based on the possibility that they will use them at some point in the future. [12] However, the calculation of non-use values is necessarily more inaccurate than use values. Yet Dr. Bernardino assigned equal weight to the non-use values as to use values.

[106] Third, biases are introduced by highlighting music over other attributes in the survey instrument. As explained in paragraph 35, this was done to ensure that attention was paid to the very reason for which the survey was conducted: to value music. The extent or impact of the order bias caused by the particular positioning of music was discussed at length at the hearing. We are more concerned by the fact that the bias was introduced deliberately [13] and that it concerned the very attribute for which the survey was conducted.

[107] Fourth, biases are introduced by asking survey participants about one “given” but not others. Focus groups led Dr. Bernardino to conclude music is a “given”: users cannot do without it, but will not think of it unless they are deprived of it. To ensure that music, the object of the survey, received the required attention, the final survey left aside some non-music features and repositioned music to focus more attention on alternative music attributes. Introducing a single given among other critical but non essential characteristics artificially inflates the value of the given: the sum of the individually assessed values of all givens in a product is liable to be several times the value of the product itself. Yet another way to look at this is to compare the relative values survey participants ascribed to critical but non essential characteristics as compared to music. Convenient location, welcoming environment and flexible contract term options all were rated as more important than music. Yet this is logically impossible if music represents one third of total value to participants, as Dr. Bernardino concluded. We are not in a position to say by how much the value of music is inflated; nevertheless, the omission of so many givens, combined with the fact that music was not mentioned organically in the focus groups, gives credence to the hypothesis that recorded music is worth less than Re:Sound suggests.

[108] Fifth, we agree with the Objectors that the next best alternative to the Re:Sound repertoire is non-Re:Sound music, not the absence of music. The fact that listeners cannot distinguish Re:Sound repertoire from other sound recordings, far from casting doubt on this conclusion, serves to reinforce it: if listeners cannot hear the difference, substitutability is enhanced. Nor do we find the argument appealing that music must first be valued in the abstract, and only then adjusted for Re:Sound’s repertoire share: what is not in the repertoire is a substitute and the existence of a substitute means that no music is not the best alternative to using the Re:Sound repertoire.

[109] It is possible to set a Re:Sound tariff by simply adjusting an existing SOCAN tariff to repertoire share. The Board has done so frequently. However, this methodology fails in the discrete-choice context, where the next best alternative must be specified precisely.

[110] Sixth, the design of the survey is fatiguing. Respondents are asked to read and assimilate a considerable amount of information about two different fitness centres and choose between them four separate times. Fatiguing surveys often lead to potential straight-line bias. There are two ways to deal with straightlining. One is to remove the straightlined responses. The other is to remove every response by someone who straightlined one or more questions. Dr. Bernardino chose the former. In our view, the latter is more appropriate, since every response by a straightliner is suspect.

iii. The Accuracy of the Calculations

[111] Dr. Bernardino calculates a measure of compensating variation, which is generally the amount of compensation someone has to receive in exchange for that person to feel equally well off after something has been taken away. In this case, she calculates how much less someone would be willing to pay for a monthly membership if the fitness centre ceased to use recorded music. Biases are introduced by calculating the variation measure as an average, rather than individually. It is possible to compute compensating variation at the individual level (which is meaningful) and discuss the portions of the distribution that would leave their fitness centres under different pricing assumptions. What was done here was to compute a compensating variation measure between two average prices and determine the cut-off price as if it could be applied to an average centre member. Compensating variation between averages is a non-linear calculation and as such, meaningless: Jensen’s inequality [14] states that reversing the order of calculating compensating variation and its average is not possible. We cannot revert to the individual computations Dr. Bernardino stated she performed, since they were not supplied to us. We find it important to raise this point, even though the Objectors did not.

iv. A General Problem With the Result

[112] The quantum proposed by Re:Sound, $86 million, is at least five times as much as what Re:Sound receives from commercial radio stations. This offends common sense: all things being equal, the value of sound recordings to commercial radio must be much more than for fitness centres.

v. Conclusion

[113] For the reasons stated above, we reject the approach Re:Sound proposed to value the use of music in fitness centres.

B. The Objectors’ Proposal

i. Introduction

[114] Before analyzing the Objectors’ proposal, we find it helpful to comment on Dr. Reitman’s discussion of Ramsey pricing. [15] We agree that a purely optimal tariff in the sense of minimizing the distance of the tariff-induced from the Pareto-efficient allocation must be consistent with the solution of a Ramsey problem. [16] We also agree that the use of a benchmark (or proxy market) for which the elasticity of demand is proportional to the market under consideration simplifies the Ramsey problem.

[115] That being said, Dr. Reitman may have misunderstood our mandate. The Board is asked to establish fair and equitable royalties. At no time has the Board equated fair and equitable with efficient; indeed, at some times, the Board has applied policy criteria to distance itself from what otherwise appeared to be efficient. [17] At no time has the Board discussed a measure of efficiency that minimizes the distance between a Pareto-efficient allocation and the allocation created by the tariff.

[116] Dr. Reitman’s discussion is not without interest. He notes the well-known shortcut: if one price is a Ramsey price, setting a second price as a ratio to the first ensures that the second price is also a Ramsey price. This gives theoretical gravitas to benchmarking and the use of proxies. We believe that benchmarks or proxies are frequently a reasonable way to proceed, for a variety of reasons stated in numerous Board decisions. Having said this, we do not foresee basing our use of proxies on Ramsey pricing. Ramsey pricing requires setting all tariffs at the same time; the Board sets one or at most several tariffs at once.

[117] We next turn to the proposal itself. The proposal is in effect two rates. Part A is based on SOCAN 19. Part B is based on Re:Sound 3, Re:Sound Satellite and SOCAN 22.A.

[118] For reasons that will soon become clear, we start our analysis with Part B. We find that SOCAN 22.A and Re:Sound Satellite are not appropriate benchmarks and that Dr. Reitman uses Re:Sound 3 improperly.

ii. Re:Sound Satellite and SOCAN 22.A as Proxies

[119] The use Dr. Reitman makes of SOCAN 22.A and Re:Sound Satellite is both confusing and problematic. By using these tariffs as he does, Dr. Reitman contradicts his assumption that, when listened to, music used in workout areas is as valuable as music used in fitness classes: he sets the rate for that workout music using a proxy that is not SOCAN 19, which he uses to set the rate for music in classes. He does so without re-framing his starting assumptions.

[120] Also, we find it difficult to relate limited downloads or satellite radio to music played in fitness centres. For limited downloads, Dr. Reitman invokes a hypothetical subscriber who uses Napster To Go to create the exact mix that person finds optimal for enhancing the workout experience. We do not find this analogy useful. A user can arrive at the same mix by using other, more prevalent music offerings, including private copying and permanent downloads; we see no reason to favour the use as benchmark of a more marginal segment of the market.

[121] It is as difficult to see how satellite radio is related to music to accompany fitness, especially given the way he envisages that music being used while training: on a portable satellite radio device. Such devices exist and are available in Canada. We have no evidence of their market penetration [18] or of the extent to which they can be used in clubs. [19]

[122] In any event, the way in which Dr. Reitman proposes using the satellite radio and limited download tariffs as benchmarks involves too many untested assumptions.

[123] One is that the typical user of the workout area visits the fitness centre once a week for a period of one hour. The evidence in Dr. Bernardino’s survey shows that respondents attend their club 3.04 times per week. [20] This figure is broadly consistent with the published literature on club attendance. We prefer Dr. Bernardino’s evidence over Dr. Reitman’s hypothesis.

[124] Another assumption is that subscribers listen to their limited download service an average of 60 hours per month, but that subscribers listen to their satellite radio service 40 hours per month. There is no evidence for either figure, and no explanation why one would be different than the other.

[125] A third is that the value of sound recordings is 80 per cent of the value of musical works, since a cover version of the original song will do as well for fitness purposes. The Board has always treated sound recordings and musical works as having the same value, all other things being equal.

iii. Re:Sound 3 as Proxy

[126] The assumption that the value of music in workout areas is different according to whether it is listened to or not led Dr. Reitman to perform separate calculations for each. For music not listened to, Dr. Reitman uses Re:Sound 3 as starting point. He then performs a series of relatively complicated calculations to derive a figure that can then be multiplied by the average number of weekly participants, as in SOCAN 19. For reasons explained below, that rate base is currently unworkable. These calculations also require some of the same assumptions as we have just discarded with respect to Dr. Reitman’s other proxies.

[127] We reject using Re:Sound 3 in the manner Dr. Reitman proposes.

iv. The Proper Use of Re:Sound 3

[128] We agree that music in classes and in workout areas should be valued separately. We see no point in valuing workout area music differently based on declarations made by participants that they listen or not to the music in the course of a survey we reject.

[129] All workout area music is background music, pure and simple. Whether listened to or not, it serves a single purpose: to enhance the customer’s experience. [21] ^{, [22]} There is no reason to believe that background music has a different value for the fitness industry than for retail stores.

[130] The structure of Re:Sound 3 is especially suited to this industry: users pay according to attendance if possible, and if not according to floor area, then capacity. Fitness centres generally maintain a tight control over attendance; most should be able to report on that basis. Therefore, we conclude that Re:Sound 3 should apply to the use of music in workout areas.

[131] We do not remove from the calculation those who come to a club to participate in classes, for two reasons. First, someone who comes to the club to attend a class may work out or stretch before or after the class. During this time, the attendee would hear background music. Second, the evidence suggests that background music is heard in the change rooms and lobby; class attendees spend some time in those areas as well. There is no need to remove from the calculation those who work out while listening to their own music. We have no evidence that would allow us to compare the frequency with which club users listen to their own music as compared to users of other premises where background music is played.

[132] There is no need to perform a repertoire adjustment. Nothing indicates that fitness centres use background music differently than other premises. We estimate the amount to be generated by this part of the tariff to be $406,000.

[133] Subsection 3(2) of Re:Sound 3 expressly provides that the tariff does not apply to the use of music to accompany fitness activities. The tariff continues to apply on an interim basis pursuant to section 68.2(3) of the Act. This apparent difficulty is readily overcome: the use of recorded music everywhere in a fitness centre other than in classes will be made subject to Re:Sound 3 by incorporating it by reference into tariff 6.B.

v. SOCAN 19 as Proxy for Music in Classes

[134] Dr. Reitman proposed using SOCAN 19 as proxy for Part A of the tariff. On surface, this is indeed the most logical place to begin, since this is a tariff for “a licence to perform […] works in SOCAN’s repertoire, in conjunction with physical exercises”. The tariff sets, for each room in which performances take place, an annual fee of “$2.14 multiplied by the average number of participants per week in the room, with a minimum annual fee of $64.”

[135] Re:Sound argued that we should not use SOCAN 19 as a proxy:

The structure and royalty rates for SOCAN Tariff 19 and the predecessor PROCAN and CAPAC tariffs have never been given any substantive consideration. Neither the Board nor the Appeal Board have determined the value of the use of musical works by a fitness centre. Furthermore, inflation has significantly eroded the royalty rate for SOCAN’s Tariff 19 since it was last increased for 1992. [23]

[136] The argument has substantial merit. A review of the decisions of the Board and its predecessor confirms that the tariff has never been the subject of even a cursory examination. Nevertheless, as a tariff that has prevailed for almost 30 years in various forms, it is a reality in the marketplace; thus, we must consider the extent to which it could be useful in the present proceeding.

[137] The starting point of our analysis was the royalties collected by SOCAN in 2007 [24] under this tariff. SOCAN collected $786,676 from 4,994 licensees. It reported two categories of licensees: licensees with fewer than 60 average weekly participants collectively sought 2,814 licenses for which revenue of $150,638 was generated; and licensees with 60 or more average weekly participants collectively sought 2,180 licenses for which revenue of $636,038 was generated.

[138] Simple division indicates that small users paid an average of $53.53 per license, whereas others paid an average of $291.76. Each of these figures is prima facie suspect, for different reasons. Small users are paying less on average than the minimum fee of $64. [25] Large users pay an amount which implies average weekly attendance of 136 people. Since the average fitness centre in Canada has membership of 918 people, [26] large users should have an even larger membership than this. Even if the membership of the large users were the same as the average centre, payments of this magnitude would imply that only 15 per cent of members attend their gym in an average week, and only once at that. It seems implausible to us that so few people would attend the gym as implied by these figures.

[139] One may examine these figures in an alternative, possibly more dramatic, way. Some members (we call them “ghosts”) pay their dues and do not attend the club at all. About a third of the other members attend the club between two and three times per week. There is some evidence that the portion of ghosts is between 20 and 40 per cent of members. Applying these figures to the average fitness centre and multiplying by the number of fitness centres in Canada yields estimated SOCAN royalties of between $4 million and $8 million per year. We do not have sufficiently precise evidence on the figures used in this calculation to narrow the range. Nevertheless, the fact that SOCAN collected less than $1 million in 2007 led us to ask questions to SOCAN. From this, we learned the following.

[140] First, there exist many interpretations of the expression “average number of participants per week in the room”. The logical (and probably correct) interpretation is to count all people in each room where fitness activities take place and music is played every day of the year and to divide by 52, assuming the club is open every day of the year. Yet some licensees use the average number of attendees per class, irrespective of the number of classes in a week. Others use the number of paid memberships, treat many rooms as if they were one, or use a “representative” week.

[141] An interpretation is biased if it tends to systematically overestimate or underestimate royalties. It is incorrect if it does not follow the text of the tariff but rather uses shortcuts to approximate the amounts owed. The interpretation based on the number of attendees per class is biased. The interpretations based on paid memberships, combining rooms or using a representative week are unbiased but incorrect.

[142] The misinterpretation of the tariff is worrisome; if users accidentally or deliberately misinterpret our tariffs in a biased way, the amounts generated are much lower than they would otherwise be.

[143] Second, there may be a tariff enforcement problem. There are about 5,000 licensees and about 5,000 fitness centres in Canada. [27] At first blush, this implies that each fitness centre holds a single license. However, this interpretation is incorrect. First, SOCAN 19 applies to dance instruction as well as fitness. We have reason to believe that the number of dance instruction licences is significant. Second, licensees are supposed to hold a license and pay for each room in which fitness activities take place and SOCAN repertoire is played. It would seem therefore that the tariff is not fully enforced.

[144] SOCAN shed some additional light on the under-enforcement/misinterpretation issue:

[…] since Tariff 19 came into force in its present form, SOCAN has faced a significant challenge in administering this tariff. In light of the fact that Tariff 19 is a self assessing tariff, and given the nature of the businesses that need a Tariff 19 licence, it is generally impossible for SOCAN to monitor the use of its music under this tariff. Unless SOCAN is a member of the gym or is signed up to take a dance class, it is not given access to these events. Thus, it is virtually impossible for SOCAN to verify the information reported. Moreover, given the relatively low tariff rates and royalties generated by the application of Tariff 19 by comparison to other tariffs, SOCAN must balance its Tariff 19 licensing efforts with the realities underlying the expenditure of time and resources required to do so. [28]

[145] Third, SOCAN has conducted very few audits relative to Tariff 19.

[146] Fourth, SOCAN collects 31 per cent of its revenue “from Tariff 19” pursuant to agreements between it and users of the tariff, copies of which were filed with us at our request. The agreements make up a significant portion of the marketplace and include some of Canada’s largest fitness centres and dance class providers.

[147] We conclude that SOCAN Tariff 19 as it stands is not usable as a proxy for setting royalties for the use of music in fitness classes. The language of the tariff is apparently confusing, to the point where users misunderstand, either deliberately or accidentally, how to compute payments under the tariff. SOCAN does not devote substantial resources to monitoring or auditing Tariff 19, to the point where the enforcement can be called into question. The fact that SOCAN enters into agreements that account for a significant portion of its revenues related to this tariff is further evidence that the tariff is unenforceable, unenforced, misunderstood or some combination of these.

C. Other Options Considered and Rejected

[148] We considered several options other than SOCAN 19 to set royalties for the use of music in fitness classes. Despite their initial attractiveness, we reject them all save one, which we certify. To justify that decision, it is necessary that we review the options we rejected.

i. A Tariff Based on the Amount of Royalties of SOCAN 19

[149] We considered setting the amount payable for all uses of music in fitness classes throughout Canada using the amount collected under SOCAN 19, adjusted for repertoire. The surface attractiveness of this proposal is evident. Regardless of how the tariff is interpreted or enforced, the amounts collected by SOCAN are a market reality.

[150] There are however several problems with this option. First, it preserves any existing under-enforcement and misinterpretations of the tariff. It would be ideal to acknowledge these and work to fix them. We cannot do so because SOCAN 19 is not before us and we do not have enough data to fix these problems on the present record.

[151] Second, it would make Re:Sound functionally subservient to SOCAN in relation to this tariff. That is, this would have the same effect as if Re:Sound cedes all responsibility for administration and enforcement of its tariff to SOCAN. This would go considerably further than setting a Re:Sound tariff on the assumption that sound recordings have the same value as musical works, before adjusting for repertoire, as the Board often does.

[152] Third, royalties collected pursuant to SOCAN 19 come from both fitness centres and dance schools. The information available does not allow us to segregate one from the other.

[153] Finally, the fact that the amounts said to be collected pursuant to the tariff are a mixture of amounts paid pursuant to the tariff and those paid pursuant to agreements makes this option murky. These three reasons make this option unworkable and we reject it as a basis for setting the tariff for fitness class music uses.

ii. A Tariff Based on the Australian Tariff

[154] On May 17, 2010, the Copyright Tribunal of Australia issued a decision relating to music in fitness centres. [29] Re:Sound urged us to draw inspiration from this decision into account. The decision cannot be used as a benchmark, if only because the Federal Court of Australia has set it aside and remanded the case for further proceedings. [30]

iii. A Tariff Based on SOCAN Tariff 7 (Skating Rinks)

[155] Skating is a form of physical exercise; as such it is comparable to the activities that take place at fitness centres. SOCAN Tariff 7 covers music played at roller and skating rinks. Individual skaters exercise to the music played both synchronously and asynchronously.

[156] It would be very simple to apply SOCAN Tariff 7 directly; it is 1.2 per cent of revenues. Applying the calculations used by Re:Sound in its Statement of Case, 1.2 per cent of revenues is equivalent to $0.42 per member per month. Applying the 36.6 per cent repertoire correction yields a rate of $0.15 per member per month.

[157] Tariff 7 is not an appropriate proxy to set a tariff for fitness centres. The business models are largely different: a substantial number of licensees under Tariff 7 pay the minimum fee, since they do not charge for admission to the skating rinks. The economic importance of the two activities is extremely different.

iv. A Tariff Based on SOCAN Tariff 18 (Recorded Music for Dancing)

[158] Re:Sound argued that music is at least as intrinsic to dance as it is to fitness; as such, the value attributed to music for Tariff 6.B could be used to set the rate in Tariff 6.A. We recently rejected that argument:

[25] Re:Sound implicitly claims a substantial similarity between the dance and fitness sectors. In our opinion, there are several key differences. First, the business models are different. In the fitness sector, the largest portion of revenue comes from monthly membership and (where these are charged separately) class fees. In the dance sector, the largest portion of revenue comes from admission fees and food and beverage sales. Fitness clubs have members whereas dance clubs have patrons.

[26] Second, the use of music is different. In a dance venue, the use of music is both ubiquitous and necessary. In a fitness centre, the use of music is partial and optional. It is partial since there may be some rooms in which music is not played, such as the class room when a yoga class is given, or the change rooms. It is optional by reason that if the fitness centre does not offer any classes and has a workout area, it could satisfy the demands of some members by having television screens or docking stations for personal music players. [31]

[159] The option under consideration here is a reversal of Re:Sound’s argument. That is, we consider the possibility of using dance as a proxy for fitness, not fitness as a proxy for dance.

[160] Ignoring the fact that we have already explained the differences between these industries, there is an additional source of problems we encounter in adapting SOCAN 18 to Re:Sound 6.B. The structure of SOCAN 18 requires payments which relate to the number of days per year the venue is open. Most fitness centres are open all year round, seven days per week, thus making the nuances in SOCAN 18 irrelevant. For both of these reasons, we find SOCAN 18 inappropriate as a proxy for Re:Sound 6.B.

v. No Tariff

[161] In an earlier decision pertaining to SOCAN’s Tariffs with respect to other uses of music on the Internet (Tariffs 22.B to 22.G, 1996-2006), the Board declined to certify a SOCAN Internet tariff for certain uses in the absence of proper and reliable evidence. [32] The Federal Court of Appeal upheld this decision (the “Other Sites” decision). [33] Do the facts in Re:Sound 6.B sufficiently match those in the other sites decision to justify declining to certify a tariff in this instance?

[162] The fact that evidence exists in the matter under consideration does not mean it is reliable. The Board has rejected the expert evidence of Re:Sound and the Objectors as unreliable. The Board has not rejected the factual evidence filed by the parties. As such, this makes it difficult to claim that there is no evidence whatsoever, a necessary criterion in the Other Sites decision.

[163] There are several other differences relative to the Other Sites decision. First, it cannot be readily argued that the amounts involved would necessarily be modest. According to Re:Sound, the Canadian fitness club industry annual revenues are estimated at $2 billion. Second, the tariff period has not even expired. The vast majority of the companies engaged in the fitness business have been doing so since at least 2008. Third, the Board knows who the tariff is targeting.

[164] The facts in this instance do not fit those of the Other Sites decision cited by reason that there is some evidence on which we can base a tariff and accordingly, we must certify a tariff.

D. The Re:Sound Tariff for Fitness Classes: A Transitional Measure

[165] We are certain of the following: that Re:Sound is entitled to a tariff for the use of sound recordings in fitness classes; that we are required to certify one; that we do not have any usable evidence relating to the value of that tariff; and, that our most logical and obvious proxy, SOCAN 19, is unusable for the reasons stated.

[166] Our preferred course of action is to set a flat fee for all users. The advantage of the flat fee is that compliance monitoring becomes easy; if a fitness centre is using Re:Sound repertoire, it must pay the flat fee.

[167] There is an obvious objection to flat fees: they may overcharge some users and undercharge some users. We note, by contrast, that the Board’s minimum fees are always flat fees. Furthermore, we believe that this tariff will be transitional. We expect Re:Sound to submit better evidence next time the Board hears the matter, allowing the panel latitude to set a tariff more permanent in nature. Nevertheless, the tariff we certify is fair under the present circumstances.

[168] It remains to fix the flat fee. Based on some of the agreements we mentioned in paragraph 146, we compute an average of the amounts paid to SOCAN. The Board sometimes hesitates to use agreements as the basis of a certified tariff, however, in this case the agreements entered into by SOCAN are the best indication of the current state of affairs in the market. Without commenting on whether or not the agreements are fair, we note that they are a fact of life for industry participants.

[169] The average obtained from the agreements we use is $288.90. [34] We adjust this figure for Re:Sound’s repertoire share, 36.6 per cent, and obtain the flat fee, $105.74. We certify Re:Sound Tariff 6.B as a flat fee of $105.74 per venue using recorded music in fitness classes. This is in addition to the royalties payable under Re:Sound 3 for the use of background music in workout areas.

[170] We note that some centres, particularly those paying the minimum fee of $64 to SOCAN, will pay more to Re:Sound than to SOCAN. We also note that some large users may pay more for their use of background music than for foreground music. We consider both of these situations somewhat problematic, but the tariff we certify is the best we can do under the circumstances.

[171] A tariff requiring payment to Re:Sound to use its sound recordings in excess of the required payments to SOCAN to use the musical works fixed in the same sound recordings is not per se unfair. This can emerge naturally if Re:Sound proposes a higher rate than SOCAN collects at the time and presents strong economic evidence to support its position. In such a situation, a tariff that prevented users from paying more to Re:Sound than to SOCAN would be unfair to Re:Sound’s rights holders.

[172] We note that our decision comes after a lengthy and difficult process made no easier by the lack of reliable and relevant evidence. Our preference is to set tariffs based on strong economic evidence presented by the parties. In this case we were not able to do so. As stated above, in our opinion the 2008-2012 tariff for fitness classes is transitional. The next time we examine Re:Sound Tariff 6.B, it would be preferable to also have SOCAN 19 before us. We would then be able to consider all of the relevant economic evidence and set both of these tariffs simultaneously.

E. Dance Instruction, Skating and Other Physical Activities

[173] The proposed tariff targets other forms of physical activity than fitness. However, little or no attention was paid to anything else during the proceedings. To set royalties for skating, dance instruction and other physical activities, we use the best information available to us.

[174] For the reasons already stated at paragraphs 165 and 166, we set the royalties for dance instruction in the same way we set the royalties for fitness classes.

[175] According to the agreements we mentioned in paragraph 146, the royalties paid for dance instruction are essentially the minimum set out in SOCAN 19. This is what we use as starting point. Adjusting the SOCAN 19 minimum of $64.00 to Re:Sound’s repertoire share of 36.6 per cent yields $23.42. This is what we certify as the flat yearly fee, per venue, for dance instruction. This rate will also apply to all other physical activities targeted in the tariff for which no specific rate is set: given the complete absence of evidence as to the music uses that may occur in these instances, we conclude that their value is minimal.

[176] SOCAN Tariff 7 targets skating, nothing else. The Board has found repeatedly that, absent any evidence to the contrary and all other things being equal, a sound recording should trigger the same royalties as a musical work. We agree: consequently, SOCAN Tariff 7 is a suitable proxy for setting the Re:Sound royalties for skating activities. The royalties in SOCAN tariff 7 are 1.2 per cent of gross receipts from admissions, exclusive of sales and amusement taxes, subject to a minimum annual fee of $104.31. An adjustment to reflect Re:Sound’s repertoire share of 36.6 per cent yields a rate of 0.44 per cent and a minimum of $38.18. These are the royalties we certify.

[177] The tariff expressly provides that background uses of recorded music are subject to Re:Sound 3 only with respect to fitness centres. We know that background music is played in these centres; we have no reason to believe that this occurs in other fitness venues. We leave it to Re:Sound to provide the required evidence when either this tariff or Re:Sound 3 is under review.

[178] To the extent possible, the Board tries to include an estimate of the quantum of royalties for its tariffs. In the present case, however, we would need to make too many assumptions, and would only be able to arrive at a fairly inaccurate estimate. We prefer to refrain from providing such an estimate.

VI. TARIFF WORDING

[179] The content of the tariff is largely explained above. Only two comments on its wording are warranted.

[180] Paragraph 3(2)(a) provides that fitness centres that pay royalties pursuant to Re:Sound 3 will file a single report and pay all royalties at the same date, thereby reducing somewhat the administrative burden of the tariff.

[181] Paragraph 8(2)(a) allows Re:Sound to share information received pursuant to the tariff with SOCAN in connection with the collection of royalties or the enforcement of a tariff. The reasons for this are found in a recent decision of the Board. [35] There is no need to repeat them here.

[182] Otherwise, the wording reflects boilerplate accounting and reporting provisions found in other Re:Sound tariff.

VII. TRANSITIONAL PROVISIONS

[183] The tariff contains transitional provisions made necessary because the tariff takes effect on January 1, 2008, while it is being certified much later.

[184] A table sets out multiplying interest factors to be used on sums retroactively owed in a given year or quarter. As we explained recently:

[131] In our opinion, the practice of using interest factors should be generalized. This tariff is being certified more than four years after it takes effect. It increases the royalties payable. Collectives cannot distribute funds they have not collected. Rights holders cannot spend or invest royalties they have not received. These are opportunity costs; their loss must be accounted for. This reasoning applies equally to tariffs of first impression and to variations to existing tariffs. [36]

Gilles McDougall

Secretary General