PUBLIC VERSION Electronically Filed Docket: 19-CRB-0005-WR (2021-2025) Filing Date: 01/14/2020 07:53:38 PM EST
Before the UNITED STATES COPYRIGHT ROYALTY BOARD LIBRARY OF CONGRESS Washington, D.C.
In the Matter of:
Determination of Rates and Terms for Docket No. 19-CRB-0005-WR Digital Performance of Sound Recordings (2021-2025) and Making of Ephemeral Copies to Facilitate those Performances (Web V)
WRITTEN REBUTTAL STATEMENT OF THE NATIONAL ASSOCIATION OF BROADCASTERS
Sarang Vijay Damle (D.C. Bar No. 1619619) [email protected] 555 Eleventh Street, NW, Suite 1000 Washington, D.C. 20004-1304 T: (202) 637-2200 F: (202) 637-2201 Joseph R. Wetzel (CA Bar No. 238008) [email protected] Andrew M. Gass (CA Bar No. 259694) [email protected] 505 Montgomery Street, Suite 2000 San Francisco, CA 94111-6538 T: (415) 391-0600 F: (415) 395-8095
Counsel for the National Association of Broadcasters
January 10, 2020 PUBLIC VERSION
Table of Contents for the Written Rebuttal Statement of The National Association of Broadcasters
A. Introductory Memorandum to the Written Rebuttal Statement
B. Index of Rebuttal Witness Testimony
C. Written Rebuttal Testimony of Gregory K. Leonard and Appendices
D. Written Rebuttal Testimony of John R. Hauser and Appendices
E. Written Rebuttal Testimony of Joseph Ritz
F. Confidentiality Declaration and Certification
G. Certificate of Service
NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25) PUBLIC VERSION
7$% $ PUBLIC VERSION
Before the UNITED STATES COPYRIGHT ROYALTY BOARD LIBRARY OF CONGRESS Washington, D.C.
In the Matter of:
Determination of Rates and Terms for Digital Performance of Sound Recordings Docket No. 19-CRB-0005-WR and Making of Ephemeral Copies to (2021-2025) Facilitate those Performances (Web V)
INTRODUCTORY MEMORANDUM TO THE WRITTEN REBUTTAL STATEMENT OF THE NATIONAL ASSOCIATION OF BROADCASTERS
The National Association of Broadcasters (“NAB”) presented evidence and expert opinion in its written direct case that digital simulcasts of over-the-air radio broadcasts are fundamentally different from other non-interactive services, and as a result, the Judges should establish a lower royalty rate for simulcasts. It should not have come as a surprise to SoundExchange that NAB would present such a case; NAB proposed just such a differential in Web IV, and SoundExchange’s own members have entered into numerous direct licenses applying materially different economic terms to simulcasts than to custom radio offerings. SoundExchange had every opportunity in its written direct case to demonstrate why simulcast transmissions should be treated the same as custom radio transmissions, or to address simulcasts at all.1 Instead, SoundExchange’s written direct case ignored completely the important distinctions between simulcasts and other non- interactive services. SoundExchange’s entire direct case simply assumes that simulcasters and
1 As noted in the NAB’s written direct case, although there are some non-custom, non- simulcast streaming services in the marketplace, they form a fairly small share of the market. See Written Direct Testimony of Gregory K. Leonard, ¶ 35. 1 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
custom radio services occupy the same market and that consumers view them interchangeably.
But all of the available evidence demonstrates otherwise.
Based on the record to date, NAB has established that a separate, lower rate for simulcasts
is appropriate. In virtually every agreement between record labels and services that offer both
simulcasts and custom radio, the labels have agreed to license simulcasts at a substantially lower
rate than custom radio. So too for other music license agreements, such as those executed by
“performing rights organizations” like ASCAP and BMI. Moreover, NAB conducted a consumer
survey that shows that, if simulcasts were unavailable, almost a third of simulcast listeners would
switch to over-the-air broadcasts and not to any royalty-generating service. Using these survey
results, NAB’s economic expert Dr. Gregory Leonard performed an opportunity cost analysis that
shows that, in a hypothetical competitive marketplace, willing buyers and willing sellers of
licenses to stream sound recordings would agree to rates for simulcasts far below the current
statutory rate. Because the plurality of simulcast listeners would switch to over-the-air broadcasts
if they could not listen to simulcasts, sellers would recognize that the cost of not doing a deal
would be a substantial loss of accretive revenue, resulting in a lower royalty rate. This result makes sense: as NAB’s fact witnesses explained, people listen to simulcasts for reasons that are very different from the reasons they listen to music-centered services, including the on-air personalities, content about local events and emergencies, and human connection.
Nothing in SoundExchange’s written direct case is contrary to—or even addresses—any of these points. Indeed, without any support for doing so, SoundExchange lumps together simulcast and custom radio and proposes a single higher royalty rate for all non-interactive services. That rate is based on Dr. Robert Willig’s opportunity cost calculations and Mr. Jonathan
Orszag’s benchmarking analysis. Neither expert provides any reason that simulcasts and custom
2 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
radio services should pay the same royalty rate, and both reports suffer from serious flaws that
inappropriately inflate SoundExchange’s rate proposal. To make matters worse, one of
SoundExchange’s core arguments for a rate increase is the alleged convergence between custom
radio services and interactive services—a rationale with no application whatsoever to simulcasts.
Dr. Willig’s opportunity cost calculations are unreliable because they are based on a flawed
consumer survey conducted by Professor Gal Zauberman, which purports to determine how
listeners of “streaming radio services” would listen to music if those “streaming radio services”
were no longer available. As an initial matter, this survey sheds no light at all on the switching behavior of simulcast listeners in particular, since the switching hypothetical asked respondents to
imagine that both simulcasts and custom radio were no longer available. Moreover, as NAB’s
survey expert Dr. John Hauser explains, flaws in the design of Professor Zauberman’s survey lead
to overestimates of switching to new, paid music subscriptions. For example, if a respondent
indicated that he or she had not listened to an existing paid on-demand streaming service in the
last thirty days, the respondent was only given the option to subscribe to a paid on-demand service;
there was no option to switch to the respondent’s existing paid on-demand service, even though
there are many reasons why someone might not have listened to (or remembered listening to) his
or her existing paid service in the prior thirty days. Likewise, the switching question—which
asked respondents which music-listening options they would choose—improperly encouraged
respondents to select music over non-music activities. In addition, the survey failed to distinguish between making new purchases of physical or digital records, and listening to an existing physical
or digital music collection. Finally, a basic flaw in the survey’s design means that the survey
cannot accurately measure how listeners would allocate their time across substitute options. Any
one of these flaws, standing alone, would render Professor Zauberman’s survey an unreliable tool
3 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
to estimate switching to royalty-generating activities. Taken together, these flaws are so serious
that the survey should be ignored in its entirety.
Moreover, as Dr. Leonard explains, the flaws in Professor Zauberman’s survey infect
Dr. Willig’s opportunity cost calculations, and therefore his Shapley Value and Nash-in-Nash
analyses. By assuming that any respondent who did not report listening to a subscription service
in the prior thirty days is not already a subscriber and would pay a new subscription fee in the
absence of non-subscription non-interactive services, Dr. Willig overstates the opportunity cost.
Similarly, Dr. Willig inflates the opportunity cost by assuming that each respondent who said he
or she would switch to a CD or digital music collection would purchase new CDs or digital music rather than listen to an existing collection. Further, Dr. Willig implausibly assumes that, where a respondent indicated he or she would switch to more than one subscription service, that respondent would actually subscribe to multiple services instead of just one—again resulting in a magnified opportunity cost. Finally, even setting aside the problems with Dr. Willig’s opportunity cost inputs, his Shapley Value and Nash-in-Nash frameworks are not specific to simulcast, offer no justification for using Pandora’s financials as a proxy for simulcasters, and make a number of improper simplifying assumptions that may result in an overstated royalty calculation.
Mr. Orszag’s benchmarking analysis is an equally unsound basis for SoundExchange’s proposed rate. As a threshold matter, Mr. Orszag entirely ignores the most relevant non-interactive benchmarks: iHeartMedia Inc.’s (“iHeart”) renewed direct deals with independent record labels.
As Dr. Leonard explained in his Written Direct Testimony, direct licenses that have been renewed with the bilateral assent of both the buyer (the radio station) and the seller (the record label) are the best evidence of a willing-buyer willing-seller transaction at the effective per-play royalty rate that predated the renewal. Because of the complexity of the financial terms of sound recording
4 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
license agreements, when they initially agree to a license, the radio station and the record label can only predict how those terms will play out. But once that license is up for renewal, the radio station and the record label know the actual financial performance of the deal. If the deal is renewed, both parties must have deemed it beneficial, both ex post as well as ex ante. Dr. Leonard thus relies for his benchmarking analysis on the fifteen instances in which independent record labels that originally entered into direct deals with iHeart have chosen to renew their agreements. Mr. Orszag overlooks these agreements completely.
Instead, Mr. Orszag bases his proposed royalty rate on the effective royalty rates as a percentage of revenue paid by subscription interactive services. He claims that subscription interactive rates are appropriate benchmarks because the functionality between interactive and non-interactive services has converged, particularly with respect to subscribers’ use of playlists.
But this alleged convergence has zero relevance to simulcasts, which do not allow users to have any control over what they hear. Indeed, simulcasts provide the least amount of interactivity of any non-interactive service, and cannot offer the core feature (and primary driver of demand) of an interactive service: the ability to listen to any song in the service’s library at any time.
Moreover, Mr. Orszag’s benchmark agreements are the product of the undisputed market power of the major record labels, yet his rate proposal fails to adjust for that power. Mr. Orszag admits—as he must, given the deposition testimony of Warner and Universal witnesses—that the major record labels are “must-haves” for the interactive services, at least in the long term. Indeed, nothing has changed since the Judges’ consistent findings in Web IV, SDARS III, and
Phonorecords III that the majors have complementary oligopoly power: an interactive service still needs the full catalog of each major to be successful; an interactive service cannot induce price
5 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
competition by steering; and there is no evidence of price competition among the major labels to be included in the libraries of interactive services.
Nonetheless, Mr. Orszag claims that the interactive services have recently gained bargaining power that offsets the major labels’ complementary oligopoly power. He relies in
substantial part on the alleged ability of the interactive services to steer by placing songs on playlists or in other prominent locations on the service. But he ignores [
. In addition, the evidence developed in discovery contradicts Mr. Orszag’s
claim that .
For example, Universal’s Aaron Harrison testified that
Indeed, both publicly available materials and documents produced by the labels establish that the interactive services are
engaged in intense competition with one another for end users, which weakens their bargaining power in negotiations with the majors. The substitutability of Spotify and Apple Music, and thus
the likelihood that a user unsatisfied with one service could simply switch to the other, make it
imperative that the services continue to offer the core functionality of on-demand access to all
major content and maintain the longstanding $9.99 per month price point while offering discounted
family and student plans. Accordingly, if the Judges rely on the interactive benchmarks (which
they should not, for all the reasons explained in NAB’s submission), they must be adjusted for the
complementary oligopoly power of the major record labels.
Finally, SoundExchange’s proposed changes to the minimum fee and audit-related terms
of the statutory license are unworkable. With respect to minimum fees, Dr. Leonard explains why
6 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
SoundExchange has not offered evidence sufficient to support the doubling it proposes. With respect to audits, iHeart’s Joe Ritz discusses how an audit is a complicated, detailed, and sometimes contentious process. Auditors frequently seek extensive raw data and related information that is several years old, requiring the licensee to task engineers, finance team members, and legal personnel with piecing together and interpreting data from multiple systems created by employees no longer working at the company. It simply is not possible to respond to every data request within sixty days, and allowing audits of multiple time periods in the same year—especially by an auditor that is not independent—would only slow the process down. Not surprisingly, SoundExchange’s proposals are inconsistent with the audit provisions of license agreements reached in the market.
Summary of the Written Rebuttal Testimony of NAB’s Witnesses
NAB’s Written Rebuttal Statement includes testimony from the following expert and fact witnesses.
Expert Witnesses
Dr. Gregory K. Leonard is an economist and Vice President at Charles River Associates.
In his Written Direct Testimony, Dr. Leonard explained that simulcast is differentiated from other non-subscription non-interactive services (such as custom radio) in economically important ways that dictate a lower per-play royalty rate for simulcasts. Further, Dr. Leonard reviewed the available direct license agreements and determined that iHeart’s renewed licenses with independent record labels provide a useful benchmark for both (i) the need for separate rates for simulcast and custom radio and (ii) what those rates should be. Dr. Leonard’s benchmarking analysis demonstrated that competitive rates for simulcasts and custom radio would fall somewhere between [ ] and [ ], and [ ] and [ ], respectively.
7 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
Additionally, Dr. Leonard used consumer survey data provided by Dr. John Hauser (discussed below) to conduct an opportunity cost analysis to determine the net loss in royalties a label would
expect from licensing simulcasts. Using that approach, Dr. Leonard calculated the opportunity
cost of licensing simulcasts, and thus the appropriate per-play royalty rate for simulcasts, as falling
somewhere between $0.00065 and $0.00108.
Dr. Leonard submits Written Rebuttal Testimony in response to the Written Direct
Testimony of Dr. Robert Willig, Mr. Jonathan Orszag, and Dr. Catherine Tucker. Dr. Leonard
explains that SoundExchange’s experts assume—with no justification—that the royalty for
simulcast and the royalty for custom radio should be the same, despite economically significant
distinctions between those services. He further explains that Dr. Willig’s royalty proposal is
flawed because his improper use of SoundExchange’s consumer survey results in overstated
opportunity cost calculations. Dr. Leonard addresses various problems with Mr. Orszag’s benchmarking analysis, including that (i) he ignores non-interactive benchmarks that are more
comparable than his interactive benchmarks; (ii) his subscription interactive benchmark is not
appropriate for simulcasts, which have not “converged” with interactive services; and (iii) his
adjustment for interactivity based on ratio equivalency is insufficient. Moreover, Dr. Leonard
explains that Mr. Orszag is wrong in arguing that the complementary oligopoly problem posed by
the major sound recording companies no longer exists: Mr. Orszag admits that the major record
labels are “must-haves”; the interactive services still have no ability to steer; [
]; and the interactive services are engaged in intense competition for end
users that weakens their bargaining power vis-à-vis the major record labels. Dr. Leonard also
rebuts Dr. Tucker’s claim that the ability to pay higher rates somehow justifies higher rates and
8 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
explains that SoundExchange’s proposed increase in the minimum fee is untethered from its
incremental administrative costs. Finally, in rebuttal to SoundExchange’s unitary rate proposal,
Dr. Leonard provides an update to his opinions on workably competitive royalty rates based on
the CRB’s recent announcement of the 2020 webcasting rates, and discusses additional evidence
of the promotional value of radio.
Dr. John Hauser is the Kirin Professor of Marketing at the Massachusetts Institute of
Technology (“MIT”) Sloan School of Management. He is an expert in survey design, demand
forecasting, product confusion, product feature valuation, and measurement of consumer preferences, beliefs, and willingness to pay. In his Written Direct Testimony, Dr. Hauser
conducted a consumer survey to measure what consumers would do in place of listening to Internet
simulcasts of terrestrial commercial radio if such simulcasts were not available. The survey also
determined consumers’ listening behavior with respect to Internet simulcasts of terrestrial
commercial radio including (1) how much they listen, (2) what content they listen to, and (3) the
importance of that content to consumers. The results of Dr. Hauser’s survey confirm that non-
music aspects of NAB members’ simulcasts drive listening behavior, and show that simulcast
listeners tend not to be people who would otherwise listen to content that would result in higher
royalties for record companies.
On behalf of NAB and Sirius XM/Pandora, Dr. Hauser submits Written Rebuttal
Testimony in response to the Written Direct Testimony of Professor Gal Zauberman, who designed
and administered a survey attempting to measure the music-listening behavior of listeners of music
streaming services and to determine how those listeners would listen to music if those music
streaming services were not available. In his rebuttal testimony, Dr. Hauser explains why
Professor Zauberman’s survey data cannot be used to reliably estimate such switching behavior,
9 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
particularly with respect to simulcast listeners. Dr. Hauser also explains that flaws in the design
of Professor Zauberman’s survey likely lead to overestimates of switching to new, paid music
subscriptions. Further, he shows that the design of Professor Zauberman’s survey prevents it from
reliably estimating the switching behavior of listeners of Internet simulcasts of terrestrial
commercial radio and listeners of Sirius XM over the Internet in particular. As a result of the flaws
in the design of Professor Zauberman’s survey, and other differences between Dr. Hauser’s survey
and Professor Zauberman’s survey, only Dr. Hauser’s survey can be used to measure the switching behavior of listeners to Internet simulcasts of terrestrial radio.
Fact Witness
Joseph Ritz is Vice President, Finance, Digital Networks at iHeart. He submits Written
Rebuttal Testimony in response to SoundExchange’s proposed changes to the audit-related terms
of the statutory license. Mr. Ritz explains that, contrary to the testimony of SoundExchange’s
Chief Operating Officer Jonathan Bender, iHeart has strong business incentives to ensure its
royalty payments are timely and accurate. He discusses how iHeart takes its sound recording
royalty obligations seriously in order to foster and maintain its valued relationships with artists
and direct deal partners. He also explains the burdens that audits typically impose on iHeart,
especially when they require iHeart personnel to take substantial time away from their jobs to
reconstruct extensive raw data from many years prior. In particular, Mr. Ritz describes iHeart’s
ongoing 2011-2013 SoundExchange audit and highlights several logistical issues and delays on
the part of the auditors, including multiple requests for redelivery of already available files and a
15-month delay in receiving the draft audit report. Mr. Ritz explains why audits of multiple time periods in the same year, strict deadlines for responding to auditor requests and cost-shifting for
failure to comply, and SoundExchange’s participation in the audit are unworkable proposals. He
10 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
testifies that these proposed changes will only lead to longer delays and more costs for both the licensee and SoundExchange. Lastly, Mr. Ritz supports NAB’s proposal to amend the interest rate for underpayments to the rate recently adopted in SDARS III.
Respectfully submitted,
LATHAM & WATKINS LLP
Sarang Vijay Damle (D.C. Bar No. 1619619) [email protected] 555 Eleventh Street, NW, Suite 1000 Washington, D.C. 20004-1304 T: (202) 637-2200 F: (202) 637-2201
Joseph R. Wetzel (CA Bar No. 238008) [email protected] Andrew M. Gass (CA Bar No. 259694) [email protected] 505 Montgomery Street, Suite 2000 San Francisco, CA 94111-6538 T: (415) 391-0600 F: (415) 395-8095
Counsel for the National Association of Broadcasters
11 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
7$% % PUBLIC VERSION PUBLIC VERSION
7$% & PUBLIC VERSION
Before the UNITED STATES COPYRIGHT ROYALTY BOARD LIBRARY OF CONGRESS Washington, D.C.
In the Matter of:
Determination of Rates and Terms for Docket No. 19-CRB-0005-WR Digital Performance of Sound Recordings (2021-2025) and Making of Ephemeral Copies to Facilitate those Performances (Web V)
EXPERT REBUTTAL WITNESS STATEMENT OF DR. GREGORY K. LEONARD
January 10, 2020
NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
TABLE OF CONTENTS
I. QUALIFICATIONS ...... 1
II. ASSIGNMENT AND SUMMARY OF OPINIONS ...... 1
III. SOUNDEXCHANGE’S EXPERTS DID NO ANALYSIS OF THE APPROPRIATE ROYALTY FOR SIMULCAST, BUT INSTEAD ASSUMED (WITH NO ATTEMPT AT JUSTIFICATION) THAT THE ROYALTY FOR SIMULCAST SHOULD BE THE SAME AS THAT FOR CUSTOM RADIO ...... 3
IV. DR. WILLIG’S ROYALTY PROPOSAL IS FLAWED ...... 7
A. SoundExchange’s Consumer Survey Provides an Unreliable Basis for Calculating the Opportunity Cost of Licensing Simulcast ...... 7
B. Dr. Willig’s Flawed Use of the SoundExchange Survey Results in Overstated Opportunity Cost Calculations ...... 11
C. The “Share of Ear” Survey Also Provides an Unreliable Basis for Dr. Willig’s Opportunity Cost Calculations ...... 17
D. Dr. Willig’s Use of the Shapley Value and Nash-in-Nash Frameworks Is Flawed ...... 22
V. MR. ORSZAG’S ROYALTY PROPOSAL IS FLAWED ...... 25
A. Mr. Orszag’s Subscription Interactive Streaming Services Benchmark Is Not Appropriate for Non-Subscription Non-Interactive Streaming Services ...... 25
B. Mr. Orszag’s Adjustment for Interactivity Based on Ratio Equivalency Is Insufficient ...... 35
C. Mr. Orszag’s Claim That the Complementary Oligopoly Problem No Longer Exists Is Incorrect ...... 40
1. The Major Labels Still Have Complementary Oligopoly Power ...... 40
2. The Interactive Services Have No Ability to Steer ...... 43
3. The Evidence Shows the Major Labels Recognized Their Strong Bargaining Positions Vis-à-vis the Services ...... 49
4. The Interactive Services Are Engaged in Intense Competition With Each Other, Which Weakens Their Bargaining Positions Vis-à-vis the Major Labels ...... 51 i NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
VI. DR. TUCKER’S OPINIONS DO NOT SUPPORT A HIGHER ROYALTY RATE FOR SIMULCAST ...... 63
VII. SOUNDEXCHANGE HAS FAILED TO DEMONSTRATE THAT THE MINIMUM FEE SHOULD BE INCREASED ...... 65
VIII. UPDATED SIMULCAST ROYALTY RATE BASED ON AN OPPORTUNITY COST FRAMEWORK ...... 68
IX. PROMOTIONAL VALUE OF RADIO ...... 69
ii NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
I. QUALIFICATIONS
1. My qualifications are discussed in my opening report.
II. ASSIGNMENT AND SUMMARY OF OPINIONS
2. I have been asked by counsel for the National Association of Broadcasters to review and respond to the opinions expressed by SoundExchange’s witnesses in their written direct testimony. My opinions are based on the information and data cited herein. I reserve the right to update my analysis and opinions if more information and data become available through discovery. I also reserve the right to respond to any further opinions and analysis offered by other experts in this case.
3. Based on my analysis, as described in greater detail below, I have reached the following opinions:
• SoundExchange’s experts did not perform any analysis of simulcast. Indeed, simulcast is barely mentioned in their reports. Instead, they calculated a royalty for custom radio and then assume, without justification, that this royalty should also apply to simulcast. • A separate analysis for simulcast is necessary because simulcast is an economically distinct service from custom radio and because the distinction has been recognized in market transactions that [ ]. • The SoundExchange user survey and Dr. Willig’s use of the survey results are unreliable for the purpose of analyzing the royalty for simulcast because: o By virtue of its design, the survey does not measure how simulcast users would respond to the absence of simulcast. o The survey shows different diversion patterns than the Hauser survey, which is designed to study simulcast user behavior. o The survey does not provide a reliable means to identify existing subscribers to paid services or to identify diversion to CDs/Vinyl/Digital Downloads. o The survey shows diversions of questionably large size to alternatives requiring a payment, suggesting the hypothetical nature of the exercise may have induced an overstatement of willingness to make a payment. o Dr. Willig assumes that respondents who chose multiple paid alternatives in an initial question would actually pay for each of those alternatives, despite many
1 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION PUBLIC VERSION PUBLIC VERSION
survey instrument did not allow respondents to provide different answers concerning their likely
behavior in the absence of simulcast versus the absence of other “free Streaming Radio”
services. As discussed further below, the evidence suggests that survey respondents do, in fact,
provide different answers for simulcast and custom radio, if allowed by the survey instrument to
do so.
6. Dr. Willig based his Shapley Value and Nash-in-Nash calculations for non-subscription
non-interactive services on financial information from Pandora.2 He did not do a version of the
calculations based on the financial information of any simulcaster, nor did he demonstrate that
the financial situation for any simulcaster is the same as that of Pandora.
7. Mr. Orszag uses subscription interactive service royalties as benchmarks for non-
interactive service royalties.3 In doing so, he makes no distinction between simulcast and custom radio. In attempting to justify subscription interactive services as a valid benchmark for non- interactive services, Mr. Orszag’s discussion is almost entirely directed at custom radio. Indeed,
many of his claimed justifications do not even apply to simulcast. For example, Mr. Orszag
claims that the subscription interactive benchmarks include the same sellers (i.e., record
companies) and similar buyers; however, in his application of ratio equivalency to non-
subscription non-interactive streaming services Mr. Orszag uses effective revenue per play
information for Pandora’s and iHeart’s ad-supported custom radio services and does not use any
information pertaining to simulcast. Thus, the “buyers” in his benchmark analysis are
Artists, American Association of Independent Music, Sony Music Entertainment, UMG Recordings, Inc., Warner Music Group Corp., and Jagjaguwar Inc., p. 10; see generally Written Direct Testimony of Gal Zauberman, September 20, 2019 (“Zauberman Report”), Appendix D. 2 Written Direct Testimony of Robert Willig, September 23, 2019 (“Willig Report”), p. 20-21, 24-26. 3 See Sections III and IV of the Written Direct Testimony of Jonathan Orszag, September 23, 2019 (“Orszag Report”).
4 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
exclusively custom radio services and do not include simulcasters. Additionally, Mr. Orszag claims that the subscription interactive benchmarks are relevant because of the convergence between interactive services’ and non-interactive services’ functionality over time, and that this alleged convergence stems primarily from the growing use of playlists by interactive service subscribers. However, none of Mr. Orszag’s supposed evidence of increased “lean-back listening” by subscribers of interactive streaming services applies to simulcast at all, and Mr.
Orszag makes no attempt to establish convergence between interactive services’ and simulcasting’s functionality over time. Nor could he, because there is no such convergence to speak of.
8. In short, neither Dr. Willig nor Mr. Orszag has performed any actual analysis of simulcast. Indeed, the word “simulcast” appears in their reports a total of two times combined.
Rather than analyzing simulcast, Dr. Willig and Mr. Orszag assume that the conclusions they reach based on what are essentially analyses of custom radio apply to simulcast as well and that therefore the royalty for simulcast should be equal to the royalty for custom radio. However, they provide no justification for this assumption.
9. As I discussed in my opening report, this assumption is invalid because simulcast differs in economically important respects from custom radio.4 For example, Dr. Willig argues that ad- supported non-interactive services should be viewed as a part of a service’s broader business model in which the ad-supported service is used as a way to funnel users into the service’s more profitable subscription services.5 However, it is not even possible for many simulcasters to use
4 Written Direct Testimony of Dr. Gregory K. Leonard, September 23, 2019 (“Leonard Report”), p. 20. 5 Willig Report, pp. 27-31.
5 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION PUBLIC VERSION
11. In short, neither Dr. Willig nor Mr. Orszag has provided any reliable analysis of what the
royalty rate for simulcast should be. They have instead assumed, without providing any
justification and contrary to actual willing-buyer willing-seller (“WBWS”) agreements, that the
royalty rate for simulcast should be the same as that for custom radio.
IV. DR. WILLIG’S ROYALTY PROPOSAL IS FLAWED
A. SoundExchange’s Consumer Survey Provides an Unreliable Basis for Calculating the Opportunity Cost of Licensing Simulcast
12. A key input into Dr. Willig’s opportunity cost calculations is a consumer survey
commissioned by SoundExchange.10 Respondents to the survey were asked a screening question
about the entertainment services that they use. Those who indicated they used a music streaming
service continued the survey and were asked which music-listening options they had used in the
past 30 days.11 Next, each respondent was randomly assigned to either the scenario where (1) free streaming radio was no longer available or (2) paid streaming radio was no longer available.
If the service assumed to be no longer available in the scenario to which the respondent was assigned was a service the respondent had previously indicated he or she had used within the past
30 days, the respondent moved forward in the survey.12 The survey then directed respondents to
select how they would fill their music-listening time in the assigned scenario. The options listed
were paid on-demand streaming; free on-demand streaming; paid streaming radio service (in the
no free scenario); free streaming radio service (in the no paid scenario); satellite receiver in the
car; AM/FM radio on a traditional radio receiver; CDs, vinyl records, or MP3 files; and do
10 Willig Report, pp. 21-22. 11 Zauberman Report, pp. 13-14, 16. 12 Zauberman Report, pp. 14-15.
7 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
something other than listen to music.13 A subset of respondents was further asked how they would allocate their time among the options they had picked.14
13. I understand that NAB’s survey expert Dr. John Hauser is providing a critique of the
SoundExchange survey from a survey perspective. Here, I only discuss the economic
implications of the survey’s flaws.
14. First, as explained in the Hauser Rebuttal Report, the SoundExchange survey is biased
against respondents expressing preferences for non-music alternatives, as Q2 asks respondents
“which of the following music-listening option(s)” they would use as substitutes for listening to music with a free (or paid, depending on the scenario) streaming radio service.15 This has the effect of overstating the opportunity cost that Dr. Willig ultimately calculates because non-music alternatives do not generate any royalties for SoundExchange. In contrast, the Hauser survey listed specific non-music alternatives such as “non-music digital content” and “print or online content.”16 Absent the survey’s bias, more users would have expressed a preference for non-
music alternatives, thereby resulting in a lower opportunity cost.
15. Second, as discussed above, the SoundExchange survey was not separately implemented
for simulcast. Instead, all non-subscription non-interactive services were lumped together.
Specifically, the SoundExchange survey asked respondents to “Imagine that all FREE [PAID]
streaming radio services were no longer available”; it did not ask respondents to imagine that
only simulcasts were no longer available.17 Even if the survey were otherwise free of flaws, it
13 Zauberman Report, pp. 16-19. 14 Zauberman Report, pp. 19-22. 15 Written Rebuttal Testimony of John R. Hauser, January 10, 2020 (“Hauser Rebuttal Report”), p. 3. 16 Written Direct Testimony of John R. Hauser, September 23, 2019 (“Hauser Report”), p. R-2. 17 Zauberman Report, pp. 16-17.
8 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
would not be appropriate to apply the SoundExchange survey results to simulcast unless the
distribution of survey responses would have been the same for a simulcast-only version of the
survey. A comparison of the SoundExchange survey results (all non-subscription non-interactive
services lumped together) to the Hauser survey results (simulcast only) provides a rough test of
whether the distribution of survey responses is the same for simulcast as for other non-
subscription non-interactive services.18 Table 1 below compares the diversion ratios from non- subscription non-interactive services to other options derived from the results of the
SoundExchange and Hauser surveys.19 There are substantial differences between the two sets of
diversion ratios. Specifically, the SoundExchange survey produced uniformly higher diversion ratios for royalty-bearing alternatives than did the Hauser survey. This difference suggests that simulcast users would have a lower opportunity cost to SoundExchange than would users of other non-subscription non-interactive services.20 With a lower opportunity cost for simulcast,
all else equal, Dr. Willig’s Shapley Value and Nash-in-Nash calculations would have yielded a
lower royalty for simulcast than, e.g., custom radio.
18 It is a “rough” test because the SoundExchange survey is flawed in other ways that could have led to biased and unreliable results. 19 Note that the Hauser survey asked respondents about a hypothetical scenario which did not include simulcasting but did include custom radio. This means that the respondents were allowed to “switch” to custom radio in the absence of simulcast in the Hauser survey, but not in the SoundExchange survey. To account for this difference in posed hypothetical scenarios, I have calculated the percentage of the subset of respondents who did not choose custom radio who said they would switch to each of the other alternatives. 20 Another possibility, of course, is that the flaws in the SoundExchange survey render it unreliable for any purpose.
9 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
Table 1: SoundExchange and Hauser Survey Diversion Ratios
SoundExchange Hauser
Subscription On-Demand 9.1% 1.5% SiriusXM 5.5% 3.4% Ad-supported On-Demand 16.4% 10.3% CDs/Vinyl/Digital Downloads 14.8% 1.9%
16. One particularly notable difference between the two surveys is with respect to the
diversion ratio from non-subscription non-interactive services (SoundExchange)/simulcast
(Hauser) to purchases of CDs/Vinyl/Digital Downloads. The SoundExchange survey finds a diversion ratio for this alternative of 14.8%, while the Hauser survey finds a diversion ratio of only 1.9%. The SoundExchange survey’s 14.8% figure is implausibly high given the sharp decline in purchases of CDs/Vinyl/Digital Downloads in recent years. From 2014-2018, the
RIAA reports that [
].22
Furthermore, Universal documents demonstrate [
].23 Research from MusicWatch demonstrates that [
21 RIAA 2018 Year-End Music Industry Revenue Report and RIAA 2014 Year-End Industry Shipment and Revenue Statistics. Vinyl is classified as LP/EP and both CD and vinyl sales excludes singles. 22 RIAA 2018 Year-End Music Industry Revenue Report and RIAA 2014 Year-End Industry Shipment and Revenue Statistics. 23 SOUNDEX_W5_000046466 (Universal Music Group U.S. Recorded Music, 2012 – 2019 YTD Revenue Sources).
10 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
].24 It is unlikely that such a large
percentage of simulcast users (for whom the service is zero-fee and who likely have relatively
low willingness to pay (“WTP”) for music) would switch to the small, declining, and costly
alternative of purchasing CDs/Vinyl/Digital Downloads in the absence of simulcast. Moreover,
the SoundExchange survey’s finding of a large diversion to CDs/Vinyl/Digital Downloads is
inconsistent with evidence in the record that shows that [
].25 The implausible results of the SoundExchange survey may be the result of having a
single response option for switching to CDs/Vinyl/Digital Downloads that the respondent
already owns or would purchase, as discussed in more detail below.
B. Dr. Willig’s Flawed Use of the SoundExchange Survey Results in Overstated Opportunity Cost Calculations
17. In incorporating the SoundExchange survey results into his calculations, Dr. Willig makes several flawed assumptions that result in an overstatement of his calculation of opportunity cost and, ultimately, the royalty.
18. First, Dr. Willig effectively identifies a survey respondent as a subscriber or non- subscriber to a given service depending on whether the respondent reported listening to the
service in the last 30 days.26 Any respondent who reported not listening to a subscription service in the last 30 days is assumed by Dr. Willig not to be a subscriber to that service. However, not all subscribers of a service use the service within any given 30-day period. For example, in the
24 SOUNDEX_W5_000045448 (Research Snapshot Alexa, Streaming & Social, May 2018), at SOUNDEX_W5_000045455. 25 SOUNDEX_W5_000186861 at SOUNDEX_W5_000186861 and SOUNDEX_W5_000186883. 26 Willig Report, Appendix E, p. E-1.
11 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
third quarter of 2018, Spotify had 5 million inactive premium subscribers, which is 6% of total
premium subscribers and up from 1% earlier the same year.27 As explained in the Hauser
Rebuttal Report, a respondent may not have used his or her subscription service in the past 30
days because the respondent instead listened to the option that the SoundExchange survey
instructed was no longer available, or for some other reason.28 Moreover, as discussed in the
Hauser Rebuttal Report, many respondents may not accurately recall the services they used in
the past 30 days.29 Thus, some of the survey respondents Dr. Willig has classified as non-
subscribers may in fact be subscribers. In that case, Dr. Willig overstates the opportunity cost
associated with such survey respondents, because he assumes they would pay a subscription fee
in the absence of non-subscription non-interactive services when in fact they had already paid the subscription fee.
19. Second, the SoundExchange survey asks each respondent to indicate which music- listening options he or she would use in the absence of non-subscription non-interactive services.30 Some of the survey respondents indicated they would use more than one subscription
service.31 Dr. Willig assumes that each such respondent would, in fact, subscribe to multiple
services, paying multiple subscription fees, rather than subscribing to just one.32 For example,
27 Inactive users are defined as users who did not listen to any music on Spotify in the last thirty days of a given quarter. See “1 Potentially Troubling Trend for Spotify,” The Motley Fool, November 6, 2018, https://www fool.com/investing/2018/11/06/1-potentially-troubling-trend-for-spotify.aspx. Performing the same calculation from this article using data from Spotify’s SEC filings from October 2019, the number of inactive users has increased to around 6 million. See Spotify Form 6-K, October 2019. 28 Hauser Rebuttal Report, p. 10-14. 29 Hauser Rebuttal Report, pp. 11-13. 30 Zauberman Report, pp. 16-17. In the case of the “no paid radio streaming” hypothetical, each respondent was asked to indicate which music-listening options he or she would use in the absence of subscription non- interactive services. 31 SOUNDEX_W5_000044743 (Zauberman Survey Data) 32 Willig Report, Appendix E.
12 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION PUBLIC VERSION
Table 2: SoundExchange Actual vs. Hypothetical Music Listening Last 30 Days Hypothetical Difference Paid On-Demand 34.0% 40.2% 6.3% Paid Noninteractive 23.0% 35.1% 12.1% Sirius XM 28.8% 32.6% 3.7% CDs, Vinyl Records, or MP3 Files 56.9% 68.4% 11.4%
Ad-Supported Noninteractive 100.0% n/a n/a Ad-Supported On-Demand 73.1% 84.6% 11.5% AM/FM Radio on a receiver 84.7% 89.0% 4.2%
The largest gain in respondents in the hypothetical scenario is in paid non-interactive services, with ad-supported on-demand services and CDs/Vinyl/MP3 approximately tied for second place.36 In total, the sum of gains across the paid alternatives is more than twice the sum of gains across the free alternatives. If the survey is to be believed, respondents are quite willing to incur additional expenses in the hypothetical scenario even though free services are available. There is further evidence of a large stated hypothetical willingness to make a payment. Of respondents with no listening to subscription services in the last 30 days, 14% stated they would add a subscription service category (i.e., paid on-demand, paid non-interactive, or Sirius XM) and, of these, a third (5%) stated they would add two or all three subscription service categories in response to the absence of free non-interactive services. Of respondents who reported to listening to two of the subscription service categories in the last 30 days, 19% stated that they would add the third category in response to the absence of free non-interactive services. While some greater willingness to make a payment might be expected given the (hypothetical)
36 Another anomaly in the SoundExchange survey results is that, as shown in Table 2, 59.9% of respondents indicated listening to CDs, Vinyl Records, or MP3 Files in the last 30 days. Other evidence suggests this figure is quite high. A study prepared for the RIAA found that [ ]. SOUNDEX_W5_000187914 at SOUNDEX_W5_000187919.
14 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
elimination of the free non-interactive alternative (although this would be tempered by the
continued availability of free on-demand alternatives), the question is whether the observed
hypothetical stated willingness to make a payment is economically plausible. I find the
respondents’ stated willingness to make payments implausible, particularly given other evidence
that users of free services have low willingness to pay and given that the stated diversion to paid
alternatives outstrips diversion to free alternatives. The academic literature recognizes that the
hypothetical nature of the “payment” in this type of survey can lead respondents to overstate their true willingness to pay, as they do not actually have to make the payment that they say they
would make.37
21. Another notable result of the SoundExchange survey is that, of the subset of respondents
who were asked to allocate their time among the options they had picked, 37% of them said they
would spend zero additional time on at least one of the choices they had previously indicated.38
Indeed, of the 376 respondents who said they would turn to at least one paid alternative and were subsequently asked to allocate their time, 101 zeroed out one of the paid alternatives they had chosen previously. This demonstrates that when pressed to further think about their time, many of the respondents narrowed their choices and often eliminated a paid alternative. However, because this question was only asked to a subset of the respondents -- namely those who chose more than one option, and among those at least one free on-demand service39 -- it is not possible to observe the time allocation of anyone that picked more than one subscription service, but not a
free service. It is likely that some of the respondents who were not asked about their time
37 See, e.g., F. Voelckner, “An Empirical Comparison of Methods for Measuring Consumers’ Willingness to Pay,” Marketing Letters, 2006; J. Murphy, et al., “A Meta-analysis of Hypothetical Bias in Stated Preference Valuation,” Environmental and Resource Economics, 2005. 38 Calculated using SOUNDEX_W5_000044743 (Zauberman Survey Data). 39 Zauberman Report, p. 19-22.
15 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
allocation and who said they would “switch” to more than one subscription service would have
allocated zero time to at least one of the subscriptions if also pressed. To the extent this is the case, Dr. Willig’s opportunity cost is again overstated.
22. Third, a sizable percentage of the SoundExchange survey respondents said that they would listen to CDs/Vinyl/Digital Downloads in the absence of non-subscription non-interactive services.40 However, the survey did not ask respondents to specify whether the
CDs/Vinyl/Digital Downloads to which they would listen would be from their existing collection
or new purchases; instead, this listening option was described as “Listen to CDs, vinyl records,
or MP3 files that you currently own or would purchase.”41 The distinction is important because
a simulcast user who states he or she would listen more to their existing CD/Vinyl/Digital
Download collection in the absence of simulcast would not represent an opportunity cost to
SoundExchange. Dr. Willig, however, assumes that each respondent who said that he or she
would switch to CDs/Vinyl/Digital Downloads and who said that he or she had not listened to
CDs in the last 30 days would purchase new CDs/Vinyl/Digital Downloads rather than listen to
an existing collection.42 To the extent any of these respondents would listen to their existing
collections, rather than purchase new CDs/Vinyl/Digital Downloads, Dr. Willig’s opportunity
cost calculation is again overstated.
40 SOUNDEX_W5_000044743 (Zauberman Survey Data). 41 Zauberman Report, Appendix D, p. 60 (emphasis added). 42 Willig Report, Appendix E, p. E-6.
16 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
C. The “Share of Ear” Survey Also Provides an Unreliable Basis for Dr. Willig’s Opportunity Cost Calculations
23. As an alternative to the SoundExchange survey, Dr. Willig uses Edison Research’s Q2
2019 “Share of Ear” survey.43 This survey asks respondents to track time spent listening to
music in various channels and then calculates percentage of listening time spent on each channel.
Dr. Willig assumes that the diversion ratio from non-subscription non-interactive services to an
alternative is equal to the listening share of that alternative divided by one minus the listening
share of non-subscription non-interactive services.44
24. Dr. Willig’s use of the “share of ear” survey is flawed for a number of reasons. First, Dr.
Willig’s Shapley Value framework using the “share of ear” survey again combines simulcast
with other ad-supported non-interactive services rather than analyzing simulcast separately.45
25. Second, because the “share of ear” survey attributes no share to non-music alternatives,
Dr. Willig’s approach does not account for such alternatives at all. That is, in his approach based on the “share of ear” survey, Dr. Willig assumes zero diversion to non-music alternatives.46
Given that both the SoundExchange and Hauser surveys find an appreciable percentage of respondents saying they would turn to non-music alternatives, Dr. Willig has no justification for making this assumption.47 Moreover, this assumption necessarily results in an overstatement of
43 Willig Report, p. 32, Appendix F. Edison Research, “Share of Ear,” Q2 2019 (SoundExchange Exhibit 44). 44 Willig Report, Appendix F, pp. F-6 to F-11. 45 Willig Report, pp. 32-34. In fact, Dr. Willig’s assumption discussed further below that diversion ratios are proportional to shares means that Dr. Willig assumes that the diversion ratio from simulcast to, say, subscription interactive services is equal to the diversion ratio from custom radio to subscription interactive services. This assumption makes no economic sense given the economic evidence that degree of interactivity affects substitution, which implies that the diversion ratio from custom radio (greater interactivity) to subscription interactive services should be greater than the diversion ratio from simulcast (lesser interactivity) to subscription interactive services (see, e.g., SOUNDEX_W5_000186861 at SOUNDEX_W5_000186880). 46 Willig Report, pp. 21-23, 32-33, Appendix F, pp. F-8 to F-11. 47 Of the SoundExchange survey respondents, 39% selected non-music alternatives as one of the alternatives they would use in the hypothetical scenario. Zauberman Survey Data. Of the Hauser survey respondents, 19% 17 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
the opportunity cost because non-music alternatives represent no opportunity cost to
SoundExchange. Recalculating Dr. Willig’s opportunity cost, assuming 19% diversion to non-
music alternatives, reduces it from [ ].
26. Third, Dr. Willig assumes that the “share of ear” associated with CDs is indicative of the percentage of non-subscription non-interactive plays that would be diverted to the purchase of
new CDs in the absence of the non-subscription non-interactive services.48 However, the CD-
listening that makes up the reported “share of ear” for CDs is based on owned CDs and does not
specify how recently these were purchased, meaning that both new and existing CDs would be
included.49 By Dr. Willig’s own logic that listener diversion is accurately represented by the various alternatives’ “share of ear,” he should have assumed that listeners would have diverted plays to new CDs only in proportion to new CD share of ear (rather than total CD share of ear,
which is by definition greater than new CD share of ear), or, put another way, that some of the
total diversion to CDs would have been to existing CDs, which do not represent an opportunity
cost to SoundExchange. By assuming that the entire diversion of plays to CDs represents
diversion to new CDs, Dr. Willig overstates the number of plays diverted to CDs that would
represent an opportunity cost to SoundExchange. Dr. Willig makes a similar incorrect
assumption regarding diversion of plays to the subscription services. He assumes that the
selected non-music alternatives as the alternative they would use in the hypothetical scenario. Hauser Report, Appendix R, items 16-19, 21-24. Note that, as discussed above, the SoundExchange survey respondents could choose multiple alternatives. A subset was asked how they would allocate time among the alternatives identified in the first step, and many respondents assigned zero time to one or more alternatives previously identified. If a respondent was asked how they would allocate their time between their chosen options, and they allocated zero time to the “Do something else” option, this respondent is not included in the percentage of people who chose to “Do something else.” 48 Willig Report, Appendix F. 49 Edison Research, “Share of Ear,” Q2 2019 (SoundExchange Exhibit 44); “‘Share of Ear’ Documents the Transition from Owned to Rented Music,” Edison Research, February 1, 2019, https://www.edisonresearch.com/share-of-ear-documents-the-transition-from-owned-to-rented-music/.
18 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
entirety of the diversion of non-subscription non-interactive plays to a given subscription service would invoke the payment of a royalty to SoundExchange. However, based on the results of the
SoundExchange survey, a substantial fraction of users of non-subscription non-interactive services are already subscribers to other subscription services. Diversion of plays by these users to their existing subscription services would not invoke the payment of any additional royalty to
SoundExchange. Thus, by ignoring the fact that some non-subscription non-interactive users are already subscribed to subscription services, Dr. Willig again overstates the number of diverted plays that represent an opportunity cost to SoundExchange.
27. Fourth, economic literature undermines Dr. Willig’s assumption that shares provide reliable estimates of diversion ratios in this type of differentiated product market.50 For example, in a market with “premium” and “economy” segments, the premium products likely have higher diversion ratios with each other than they do with the economy products, regardless of what the market shares would suggest. In other words, users of a premium product are likely to view
50 Dr. Willig’s assumption that diversion ratios are proportional to shares imposes the “independence of irrelevant alternatives” (IIA) property, which underlies, for example, the “logit” model of demand. The IIA assumption places strong restrictions on substitution patterns between products (i.e., the own- and cross-price elasticities of demand and related diversion ratios). Because of IIA’s restrictiveness regarding substitution patterns, the literature warns about imposition of the IIA property and associated use of the logit model of demand in circumstances when the nature of substitution patterns is a central question of interest. See, e.g., D. McFadden, “Econometric Models of Probabilistic Choice,” in Structural Analysis of Discrete Data with Econometric Applications, 1981, pp. 222-223 (“…models satisfying [IIA] yield implausible conclusions when there are strong contrasts in the similarity of the alternatives”); S. Berry, “Estimating Discrete Choice Models of Product Differentiation,” RAND Journal of Economics, 1994, p. 250 (“[t]he logit model products unreasonable substitution patterns”); J. Hausman and G. Leonard, “Economic Analysis of Differentiated Products Mergers Using Real World Data,” George Mason Law Review, 1997, p. 322 (“…the [IIA] assumption…implicitly restricts the demand structure by constraining the pattern of demand substitution between products”); D. Brownstone and K. Train, “Forecasting New Product Penetration with Flexible Substitution Patterns,” Journal of Econometrics, 1999, p. 110 (“…identification of the correct substitution patterns is an empirical issue, and the IIA property…imposes a particular substitution pattern rather than allowing the data analysis to find and reflect whatever substitution pattern actually occurs”); A. Nevo, “Mergers with Differentiated Products: The Case of the Ready-to-Eat Cereal Industry,” RAND Journal of Economics, 2000, p. 402 (“the logit model greatly restricts the own- and cross-price elasiticities”); P. Davis and E. Garces, Quantitative Techniques for Competition and Antitrust Analysis, 2010, pp. 477-478 (“…the logit model imposes severe limitations on own- and cross-price elasticities…we recommend strongly against using [logit] models in situations where we must learn something about substitution patterns…”).
19 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
services. Yet, this is inconsistent with the economic evidence that degree of interactivity affects
substitution, which implies that the diversion ratio from custom radio (greater interactivity) to subscription interactive services should be greater than the diversion ratio from simulcast (lesser interactivity) to subscription interactive services.55
29. The Hauser survey results provide support for the conclusion that listening shares provide poor estimates of diversion ratios from simulcast to other alternatives. Table 3 compares the
diversion ratios implied by the listening shares to the diversion ratios from the Hauser survey.
As with the SoundExchange survey, there are substantial differences between the two sets of
diversion ratios. Specifically, the “share of ear” survey (along with the assumption that diversion
ratios are proportional to shares) produces uniformly higher diversion ratios for royalty-bearing
alternatives than does the Hauser survey. This difference suggests that either (1) the “share of
ear” survey and assumption that diversion ratios are proportional to shares produces unreliable
diversion ratios, or (2) simulcast users have a different and lower opportunity cost to
SoundExchange than users of other non-subscription non-interactive services. With lower
diversion ratios, all else equal, Dr. Willig’s Shapley Value calculations would have yielded a
lower royalty for simulcast than, e.g., custom radio.
Table 3: Share of Ear and Hauser Diversion Ratios
Share of Ear Hauser
Subscription On Demand 1.5% SiriusXM 3.4% Ad-supported On Demand 10.3% CDs/Vinyl/Digital Downloads 1.9%
55 See, e.g., SOUNDEX_W5_000186861 at SOUNDEX_W5_000186880.
21 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
D. Dr. Willig’s Use of the Shapley Value and Nash-in-Nash Frameworks Is Flawed
30. As inputs into the Shapley Value and Nash-in-Nash frameworks that he uses to calculate
his proposed royalty, Dr. Willig uses (i) the opportunity cost information based on the
SoundExchange survey and his calculation of royalty rates for alternatives to non-subscription
non-interactive services and (ii) financial information from Pandora.56 These frameworks are
flawed not only because they rely on the unreliable survey information and assumptions
addressed above, but also for additional reasons.
31. As noted above, the inputs for Dr. Willig’s Shapley Value and Nash-in-Nash frameworks
are not specific to simulcast, and Dr. Willig provides no justification for assuming that the values
of the inputs are accurate for simulcast. Of particular note is Dr. Willig’s use of Pandora’s
financial information as a stand-in for the financial situation of every non-subscription non-
interactive service.57 Given the substantial difference between Pandora and simulcast in terms of business model, there is no reason to believe that Pandora is representative of the simulcast services provided by, e.g., Mel Wheeler, Inc. Indeed, as discussed in my opening report, the economic evidence suggests that [ ].58 Thus, there is no
basis to conclude that the royalties that emerge from Dr. Willig’s Shapley Value and Nash-in-
Nash frameworks are appropriate for simulcast.
32. Moreover, in constructing his Shapley Value and Nash-in-Nash frameworks, Dr. Willig makes a number of simplifying assumptions. While it is common in economic modeling to make
simplifying assumptions, one must be careful lest the omission of an important factor renders the
56 Willig Report, pp. 16-26, 34-37, Appendices D, E, G. 57 Willig Report, Appendix D. 58 Leonard Report, pp. 27-29.
22 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
model misleading as applied to a real-world situation. There are several respects in which
Dr. Willig’s frameworks make simplifying assumptions that may result in an overstatement in
the calculation of the royalty.
33. First, Dr. Willig assumes that the royalties for musical compositions are set exogenously
outside of his frameworks.59 Yet, for Pandora, historically the royalties for musical compositions have been closely and contractually linked to sound recording royalties.60 Thus, in the real
world, if sound recording royalties were increased by the amount that Dr. Willig proposes based
on his models, music composition royalties would be expected to increase as well. But, in that
event, Dr. Willig’s frameworks incorrectly model the music composition royalties as exogenous
and, moreover, assume a level of musical composition royalties that is too low. If Dr. Willig’s
frameworks were made more realistic to allow musical composition royalties to be endogenous
(for example, set to be positively related to sound recording royalties), the frameworks would
generate lower royalties than they do with the unrealistic assumption that musical composition royalties are exogenously determined.61
34. Second, and similarly, Dr. Willig’s frameworks implicitly assume that any royalty for
sound recordings would be assessed as a lump sum that would have no effect on service
providers’ decisions regarding pricing, use of music versus non-music content, and other
59 Willig Report, pp. 14-15, Appendix C. 60 PANWEBV_00005062, at PANWEBV_00005076; “Pandora Details Its Pay Raise to Publishers,” Billboard, February 12, 2016, https://www.billboard.com/articles/business/6874911/pandora-publisher-pay; “Pandora Is Now Paying Publishers 1/5th of the Money It Gives Labels,” Music Business Worldwide, February 17, 2016, https://www musicbusinessworldwide.com/pandora-is-paying-major-publishers-fifth/. 61 When the musical composition royalties are assumed to be a positive function of the sound recording royalties, an increase in the sound recording royalties decreases the surplus available to split among the labels and services because the musical composition copyright holders capture more of the surplus.
23 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
characteristics of their services.62 In reality, however, [
] and thus represent a marginal cost to services that, under
economic principles, would be expected to affect service providers’ decisions regarding pricing
and product characteristics. Indeed, Dr. Willig even converts his calculated lump sum royalty to
a per-play royalty that he argues should be imposed in the future.63 If this per-play royalty were
imposed in the real world, however, it would affect services’ decision-making, and the services’
economic outcomes (revenue, plays, and profitability) would be different than the levels assumed
as inputs in Dr. Willig’s frameworks. That means the inputs to Dr. Willig’s frameworks that led
to his per-play royalty proposal are incorrect. The result is that the royalty Dr. Willig proposes
may give SoundExchange an excessive share of the surplus.64
35. Third, a flaw specific to Dr. Willig’s Nash-in-Nash framework is his assumption that each pair of negotiating parties would split the surplus 50/50. Dr. Willig seems to imply that a
50/50 split is a necessary implication of the Nash Bargaining Solution (“NBS”).65 However, a
50/50 split necessarily emerges from the NBS only if the negotiating parties are assumed to have
the same utility function. In general, if the parties have different utility functions (e.g., different
levels of risk-aversion), the split of the surplus will not be 50/50.66 Dr. Willig provides no
62 Willig Report, p. 16. A lump sum royalty would not affect a service’s marginal costs and thus would not affect decisions on the margin regarding pricing and product characteristics. However, a lump sum royalty may affect a service’s decision as to whether to participate in the market. If the lump sum royalty makes participation unprofitable, the service would decide not to participate. 63 Willig Report, pp. 27, 31-32. 64 Specifically, the surplus available to split among labels and services would be less than Dr. Willig assumes. This, in turn, means that the royalty Dr. Willig proposes (which assumes the incorrect higher level of surplus) gives too much of the correct level of surplus to the labels. 65 Willig Report, pp. 35-36. 66 For example, see M. Osborne and A. Rubinstein, A Course in Game Theory, 1994, p. 303 (“A main goal of Nash’s theory is to provide a relationship between the players’ attitudes towards risk and the outcome of the bargaining...We compare two bargaining problems that differ only in that one player’s preference relation in one 24 NAB Written Rebuttal Statement Dkt No 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
justification for the assumption that labels and services all have the same utility function. Absent
such an assumption, Dr. Willig would have to determine the split between each pair of
negotiating parties. He has not shown how that could be done given available information.
36. Finally, there is a question as to whether the Shapley Value framework is appropriate given the task in this proceeding, which I understand to be modeling the outcome of a WBWS negotiation assuming effective competition. Shapley Values provide a specific split of surplus among contributing factors that possess certain properties. Shapely Values are not meant to mimic a WBWS outcome and, indeed, a WBWS outcome may well differ from the outcome implied by Shapley Values, depending on the institutional structure.67 This point is underscored
by the fact that Dr. Willig’s Nash-in-Nash framework gives a different answer than his Shapley
Value framework.68 In contrast to the Shapley Value framework, the Nash-in-Nash framework
is an attempt to model WBWS outcomes—but Dr. Willig’s Nash-in-Nash framework here
remains flawed for all the reasons described above.
V. MR. ORSZAG’S ROYALTY PROPOSAL IS FLAWED
A. Mr. Orszag’s Subscription Interactive Streaming Services Benchmark Is Not Appropriate for Non-Subscription Non-Interactive Streaming Services
37. Mr. Orszag’s proposed royalty rate for non-subscription non-interactive streaming services of $0.0025 per play is based on the effective royalty rates as a percentage of revenue paid by subscription interactive streaming services. Specifically, Mr. Orszag’s recommended
of the problems is more risk-averse than it is in the other; we verify that the outcome of the former problem is worse for the player than that of the latter.”). 67 See, e.g., F. Gul, “Bargaining Foundations of Shapley Value,” Econometrica, 1989. In an experiment described in C. Moellers, et al., “Communication in Vertical Markets: Experimental Evidence,” International Journal of Industrial Organization, 2017, the Shapley Value framework failed to predict “market” outcomes accurately. 68 Willig Report, p. 36.
25 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
rate is equal to the effective royalty rate as a percentage of revenue paid by Spotify to the major
record labels and Merlin for its subscription interactive streaming service for the year ending
April 2019 of [ ], multiplied by the effective revenue per play paid for Pandora and
iHeart’s non-subscription non-interactive services for the year ending April 2019 of ].69
Mr. Orszag’s rationales for using subscription interactive streaming services as a benchmark for subscription and non-subscription non-interactive streaming services are: the subscription interactive benchmarks are market-based and were used in prior webcasting proceedings; the subscription interactive benchmarks include the same sellers (i.e., record companies) and similar buyers; and there has been a convergence between interactive services’ and non-interactive services’ functionality over time.70 Each of Mr. Orszag’s rationales is flawed.
38. First, the complementary oligopoly power of the major record labels in the interactive
market renders this market less than effectively competitive. This is the conclusion that was
reached by the Judges in Web IV, where the Judges determined that an effective competition
adjustment was necessary to account for major record labels’ complementary oligopoly power.71
Mr. Orszag incorrectly believes that such an effective competition adjustment is no longer necessary because the major record labels’ bargaining power is now offset by the interactive services’ supposed bargaining power resulting from their alleged must-have status vis-à-vis the major labels.72 As I discuss in Section V.C below, however, the complementary oligopoly power
of the major record labels is not offset by the supposed bargaining power of the subscription
interactive streaming services. Thus, even though the interactive streaming services benchmarks
69 Orszag Report, pp. 41-43. 70 Orszag Report, pp. 19-29. 71 Web IV Determination, 81 Fed. Reg. 26316, 26343-44 (CRB May 2, 2016). 72 Orszag Report, Section V.
26 NAB Written Rebuttal Statement Dkt. No 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
are market-based, they do not reflect effective competition and therefore still need to be adjusted
if they are to be used as benchmarks in this proceeding. Mr. Orszag did not perform an effective
competition adjustment, which inflates his recommended royalty rate for non-subscription non-
interactive streaming services. As discussed in my opening report, to ensure that the royalty
rates for non-subscription non-interactive services (including simulcast) reflected effective
competition, in Web IV, the Judges used as benchmarks royalties from actual license agreements
between labels and services that contained steering provisions.73 Steering is a means by which a
service can create competition among labels, mitigating the complementary oligopoly problem.
The steering discount applied in Web IV was 12%.74 Applying this discount to Mr. Orszag’s
$0.0025 per-play rate lowers his recommended rate for all webcasting (custom radio and
simulcast combined) to $0.0022.
39. Second, although the interactive services benchmarks used by Mr. Orszag include the
same sellers—i.e., labels—they do not include similar buyers. Mr. Orszag’s per-play rate is
based solely on the effective royalty rate as a percentage of revenue of that one buyer,
Spotify, paid to major record labels and Merlin for its subscription interactive streaming service.
As I explain in the following paragraphs, that service is a fundamentally different service from non-subscription non-interactive streaming services, especially simulcasts.
40. Third, and relatedly, Mr. Orszag’s assessment of the increased comparability between interactive and non-interactive streaming services is incorrect as a general matter, and has no application at all to simulcasts specifically. Mr. Orszag argues that there has been a convergence between interactive services’ and non-interactive services’ functionality over time, which makes
73 Web IV Determination, 81 Fed. Reg. 26316, 26365-69 (CRB May 2, 2016). 74 Id. at 26405.
27 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
them comparable to each other. This alleged convergence stems primarily from the growing use
of playlists by subscribers of interactive streaming services.75 In other words, Mr. Orszag claims
that increased “lean-back listening” by subscribers of interactive streaming services has made
their user experience more similar to the user experience of a non-interactive streaming service.
This claim is flawed for several reasons.
41. Most importantly, the evidence of “convergence” that Mr. Orszag cites does not apply to
simulcast. For example, Mr. Orszag classifies various types of activities as “lean-back listening” and claims that “as of 2017, up to 68 percent of listening on Spotify is lean-back.”76 Mr. Orszag
further claims:
Apple’s and Spotify’s increasing focus on the lean-back listening experience illustrates the convergence between interactive and noninteractive services in terms of user behavior and the type of services they offer; in the noninteractive world, the stations are effectively playlists. The playlists on interactive services and the stations on noninteractive services are not necessarily identical, but they attempt to achieve the same goals, which is to provide listeners with a lean-back listening experience and sometimes music discovery options, which attract and retain users on the platform. A person unfamiliar with the services simply listening to the music programming would not necessarily know the difference between lean-back listening with interactive service-generated playlists and lean-back listening with noninteractive service stations.77
But while 68% of Spotify listening is lean-back—i.e., playlists—only 15% of monthly content
hours are from editorially curated playlists, which is the closest type of playlist to what one
would get listening to simulcast.78 The rest of the Spotify activities that Mr. Orszag classifies as
lean-back listening are algorithmically-curated playlists created by Spotify for each user (17%)
75 Orszag Report, pp. 65-67. 76 Orszag Report, pp. 23-24. 77 Orszag Report, p. 26. 78 Orszag Report, pp. 23-24.
28 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
and a user’s own-generated playlists (36%).79 Neither of these types of playlists are available on
a simulcast. SoundExchange witnesses Aaron Harrison, UMG Recordings, Inc.’s Senior Vice
President, Business & Legal Affairs, Digital, and Reni Adadevoh, Warner’s Vice President of
Legal and Business Affairs, [
].80 Thus, there is, in fact, little to no evidence of overlap and thus convergence between interactive services and simulcast.81 With regard to custom radio, the addition of lean-back features to
interactive services is an effort to provide an additional functionality to a subset of interactive
users who may value such functionality. Many interactive users do not use the functionality (and
thus do not value it significantly). Moreover, even among the subset of interactive users that use
the functionality, its existence does not necessarily create “convergence.” Two products sharing
a feature does not necessarily make them substitutes. For example, collision avoidance features
are available on both a Honda Civic or a Mercedes E-series sedan, but these two car models are
hardly close substitutes.
42. Moreover, interactive services provide other playlist-related features that are valuable to
users but are not even possible on simulcast (and, indeed, are not permitted under the statutory
license for non-subscription non-interactive services). For example, on an interactive service a
user can view the entire playlist, including upcoming songs, and can skip songs and move
79 Orszag Report, pp. 23-24. 80 Harrison Dep., pp. 105, 118-119; Adadevoh Dep., pp. 116-117. 81 There is also relatively little overlap between ad-supported interactive services and custom radio given that custom radio users cannot listen to user-generated playlists. Mr. Orszag ignores the distinctions between the playlist categories.
29 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
PUBLIC VERSION
functionality demonstrates that interactive service users are subscribing to take advantage of on- demand functionality, exercising specific music choices as opposed to listening to pre- programmed simulcasts.
47. Simulcasts contain limited, pre-selected music content and provide the listener with no in-platform on-demand functionality. As long as such on-demand features have a significant value to users (as they currently do, as demonstrated, for example, by the higher subscription price for interactive than non-interactive services90 and established user behavior on these platforms), economically meaningful divergence exists between interactive services and simulcast.
48. SoundExchange expert Dr. Tucker highlights the value of on-demand features to users in her report, which is contrary to Mr. Orszag’s claims about convergence. Dr. Tucker discusses how on-demand services have gained popularity in recent years, and that “[i]ndustry analysts generally agree that paid subscriptions to services that offer interactive functionalities (such as
Spotify Premium) are the future of music.”91 If there is “convergence,” as Mr. Orszag claims, there would be no basis for Dr. Tucker to distinguish interactive services. In fact, the increasing demand for on-demand services demonstrates that they are significantly differentiated from services that do not offer on-demand functionality.
90 Pandora’s non-interactive subscription service, Pandora Plus, is priced monthly at $4.99, whereas its interactive subscription service, Pandora Premium, is priced monthly at $9.99. See “Upgrade to Pandora Plus or Premium,” Pandora Help, https://help.pandora.com/s/article/Upgrade-to-Pandora-Plus-or-Pandora-Premium- 1519949306612?language=en_US. iHeartRadio also prices its non-interactive subscription service, iHeartRadio Plus, and its interactive subscription service similarly at $4.99 and $9.99 per month respectively. See “How much is iHeartRadio All Access? What can I use to pay for it?” iHeartRadio, https://help.iheart.com/hc/en- us/articles/235721027-How-much-is-iHeartRadio-All-Access-What-can-I-use-to-pay-for-it- and “How much is iHeartRadio Plus? What can I use to pay for it?” iHeartRadio, https://help.iheart.com/hc/en- us/articles/235720747--How-much-is-iHeartRadio-Plus-What-can-I-use-to-pay-for-it-. 91 See Written Direct Testimony of Catherine Tucker, September 23, 2019 (“Tucker Report”), pp. 7-8.
32 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
49. While Mr. Orszag has failed to demonstrate convergence of interactive services with
custom radio, he has not even addressed convergence of interactive services with simulcast. In
fact, interactive services have not converged toward simulcast, and simulcast has not converged toward interactive services, both because it is not possible given the nature of simulcast and because it is not allowed under the statutory license. Convergence of functionality would require moving outside of the statutory license and, therefore, is not relevant for determining the rate
within the statutory license. Accordingly, Mr. Orszag’s justification for using subscription
interactive services as a benchmark for simulcast is not valid.
50. Fourth, Mr. Orszag ignores non-interactive benchmarks that are more comparable than
his interactive benchmarks. As discussed extensively in my opening report, iHeart’s direct deals
with the renewal indies are the most relevant benchmarks for establishing the statutory rate for
non-subscription non-interactive streaming services including custom radio and simulcast.
These deals represent the best benchmarks for several reasons, including that they are market
transactions consistent with Section 114’s WBWS standard and (i) cover the licensing of rights
for the public performance of sound recordings; (ii) cover custom radio and simulcast services;
and (iii) include the same sellers (i.e., record companies) and same buyers (i.e., non-subscription
non-interactive streaming services) that would be involved in the statutory license. Moreover, by limiting my analysis to iHeart’s agreements with the renewal indies and calculating iHeart’s effective per-play royalty rates paid to these renewal indies during the pre-renewal period, I have focused on rates that must have been acceptable to both parties because both parties agreed to renew the direct deal at the same contractual rate with knowledge of the actual per-play rate.92
Further, the rates in these agreements are relatively free of complementary oligopoly power due
92 Leonard Report, pp. 31-37.
33 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
to the fact that the renewal indies likely have less complementary oligopoly power than the major record labels. Finally, any argument that these non-interactive benchmarks are in the shadow of regulation ignores the fact that when the renewal indies agreed to the lower effective rates than the statutory rate those rates were freely negotiated. The renewal indies must have understood that iHeart could and perhaps would steer toward the labels’ recordings or that licensing iHeart offers other (e.g., promotional) benefits.93 Indeed, I have reviewed evidence
produced in discovery that [
].94
51. Finally, Mr. Orszag also ignores ad-supported interactive services as a potential
benchmark. Such services share all of the characteristics of subscription interactive services that
Mr. Orszag pointed to as support for their use as benchmarks, plus they are also ad-supported,
which makes them more comparable to non-subscription non-interactive services. Spotify’s ad-
supported interactive service had a per-play royalty of [ ] and a percentage-of-revenue
royalty of [ ] from January through April of 2019.95 Both of these figures are well below the respective figures ($0.0025 per play and [ ] of revenue) that Mr. Orszag derives from his subscription interactive benchmark. I note that, if the Spotify ad-supported interactive
service rate were to be used as a benchmark for non-subscription non-interactive services, adjustments would need to be made to account for effective competition and lack of interactivity.
93 Leonard Report, pp. 34-35. 94 See, e.g., NAB00003944; NAB00003948; NAB00003969; NAB00004024. 95 Calculated using Orszag data.
34 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
have no willingness to pay.”98 However, Mr. Orszag argues that users of non-subscription non- interactive services do pay “with their time spent listening to advertisements,”99 and, therefore,
“a benchmark analysis that uses subscription interactive services is appropriate if the interactivity
adjustment is calculated using the revenue of noninteractive ad-supported services (thus
reflecting their particular willingness to pay) rather than the revenue of noninteractive
subscription services.”100 Mr. Orszag concludes that his “approach to this analysis appears to
have been tacitly endorsed by the Judges in Web IV, who suggested that any attempt to apply
ratio equivalency to the ad-supported market would have to consider the revenue earned by the
services in that market.”101 He also concludes that the subscription interactive benchmark is now
more applicable to the non-subscription non-interactive market because “consumers increasingly
view subscription interactive services as a viable (and sometimes preferred) alternative to an ad-
supported noninteractive service.”102 Mr. Orszag’s interactivity adjustment as it pertains to non-
subscription non-interactive streaming services, and in particular simulcast, is flawed for several
reasons.
54. First, Mr. Orszag misses an important distinction between subscription and ad-supported
services that undermines his argument regarding the applicability of ratio equivalency, namely
that the relationship between interactivity and revenue generation differs between the two types
of services. Subscription services generate revenues through subscription fees paid by users. To
the extent some users value interactivity, a subscription service may be able to charge a higher
98 Orszag Report, p. 38. 99 Orszag Report, pp. 38-39. 100 Orszag Report, p. 39. 101 Orszag Report, p. 39. 102 Orszag Report, p. 41.
36 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
price for an interactive service than a non-interactive service. An ad-supported service, in
contrast, generates revenue through placing advertisements. Advertisers have no reason to prefer
advertising on a service with greater interactivity. In particular, they would not be willing to pay
more per impression for an ad on an interactive service than a non-interactive service (unless an
interactive service attracts more valuable impressions than a non-interactive service, a
proposition for which I have seen no evidence). In any event, the relationship between revenue
generation and interactivity is substantially different for ad-supported than for subscription
services. Moreover, there is no reason to think that the elasticities of demand of the buyers for
the respective types of services (advertisers for ad-supported and users for subscription) are the
same or that the cost structures are the same (e.g., ad-supported services may have costs and
investments associated with an advertising platform). These key economic differences cause Mr.
Orszag’s rationale for ratio equivalency between subscription and ad-supported services to fail.
Relatedly, Mr. Orszag overstates the case on user WTP. While it is true that users of ad-
supported services “pay” with the time spent with ads, they had the choice to pay for a
subscription service and avoid ads. Thus, by revealed preference, their WTP to avoid ads (and
by implication their WTP for more music in place of ads) is less than that of subscribers to paid
services (moreover, as discussed above, the ad-supported service receives revenue from the
advertiser, not based on the user’s time). This distinction is the one the Judges appear to have
been making,103 and it is correct.
55. Second, Mr. Orszag’s application of ratio equivalency as it relates to simulcast is completely unsupported. In fact, he treats simulcast the same as all other non-subscription non- interactive streaming services without providing any support for such an assumption. For
103 Web IV Determination, 81 Fed. Reg. 26316, 23650, n. 110 (CRB May 2, 2016).
37 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
example, Mr. Orszag’s application of ratio equivalency to non-subscription non-interactive
streaming services uses the effective revenue per play for Pandora and iHeart for their ad-
supported custom radio services for the year ending April 2019 of ].104 Thus,
Mr. Orszag’s analysis relies exclusively on the effective revenue per play for custom radio and
assumes with no support that simulcast should be treated the same as custom radio. As
previously discussed, simulcast differs in economically important ways from custom radio and,
therefore, Mr. Orszag’s assumption is unsupported. A more appropriate (but still incorrect)
application of ratio equivalency pertaining to simulcast would be to calculate the effective
revenue per play for simulcasters. Mr. Orszag has not done this. As discussed in my opening report, the revenue per play for many simulcasters is likely low due to difficulties in generating
revenues on the simulcast. Even using the simulcaster revenue per play would be incorrect,
however. As discussed in my opening report, many simulcasts, unlike interactive services or
custom radio, contain significant non-music content, and thus some of the simulcast revenue
should be attributed to this content rather than music.105 Dividing revenue by music plays
without any adjustment for non-music content would fail to account for the value of the non-
music content.
56. Third, Mr. Orszag wrongly states that there is no connection in the marketplace between
percentage-of-revenue rates and the degree of interactivity offered. Mr. Orszag claims: “We do
104 Revenue per play for Pandora is ]
enue per play. He does not explain how his proposal makes economic sense given this context. Any claim that iHeart and Pandora are perfect substitutes for users is unsupported. 105 Leonard Report, pp. 19-24. Additionally, Mr. Orszag ignores that Mr. Harrison testified that [ ]. Harrison Dep., pp. 138-139; Written Direct Testimony of Aaron Harrison, September 22, 2019 (“Harrison Report”), pp. 24-25. However, simulcast is not a close substitute for interactive services.
38 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
not see in the marketplace [
.]”106 This statement is incorrect. As previously discussed, marketplace evidence provided by music copyright licensing agreements have specified different percentage- of-revenue royalty rates for simulcast and custom radio, which have different degrees of interactivity. This evidence includes iHeart’s direct deals with independent record labels and the recent PRO licenses (and offered licenses) for rights to publicly perform musical compositions.
Specifically, iHeart’s direct deals with several renewal indies specify [
].107
57. Finally, Mr. Orszag’s conclusion that the subscription interactive benchmark is now more applicable to the non-subscription non-interactive market because “consumers increasingly view subscription interactive services as a viable (and sometimes preferred) alternative to an ad- supported noninteractive service” is unsupported, particularly with respect to simulcast, for the reasons discussed above. Mr. Orszag also claims that the SoundExchange survey finds a cross elasticity between non-interactive ad-supported services and interactive services, with 14% of users of non-interactive ad-supported services (who did not use interactive subscription services) stating that they would start listening to interactive subscription services if the ad-supported services were not available.108 Cross (price) elasticity is not well-defined when the price charged
for the product is zero. However, even so, Mr. Orszag ignores the fact that several other
106 Orszag Report, p. 33. 107 See Leonard Report, pp. 31-36, which specifies these royalty terms for iHeart’s direct deal with Big Machine, a representative agreement for iHeart’s agreements with renewal indies. 108 Orszag Report, pp. 40-41.
39 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
services, such as ad-supported interactive services and CDs/Vinyl/Digital Downloads, are found
by the SoundExchange survey to have higher diversion ratios than subscription interactive
services109, implying they are closer substitutes for ad-supported non-interactive services than interactive services. A diversion ratio of 14% is not particularly large in any event. Moreover, as discussed above, the SoundExchange survey does not address simulcast users and likely biases responses in favor of subscription services. As seen in Table 1 above, when the Hauser
survey results are used, the diversion ratio from simulcast to subscription interactive services is
only 1.5% and lower than the diversion ratios to other alternatives.
C. Mr. Orszag’s Claim That the Complementary Oligopoly Problem No Longer Exists Is Incorrect
58. Mr. Orszag argues that the royalty rates from subscription interactive services need not be adjusted to account for any complementary oligopoly power of the major labels before using these rates as benchmarks for non-subscription non-interactive services.110 Mr. Orszag’s
argument is incorrect.
1. The Major Labels Still Have Complementary Oligopoly Power
59. As an initial point, Mr. Orszag does not even attempt to demonstrate that the
complementary oligopoly power of the majors with respect to the interactive services has
diminished. Rather, he admits that it “likely remains the case that the Majors are ‘must-haves’
for the interactive services, at least in the long run.”111 That is consistent with the testimony of
109 Willig Report, p. F-9. 110 Orszag Report, pp. 75-76. 111 Orszag Report, p. 45.
40 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
PUBLIC VERSION PUBLIC VERSION PUBLIC VERSION
64. Warner and Universal personnel testified that [
]. For example, Reni Adadevoh testified that Warner wants to ensure
] and that [
]125 Ms. Adadevoh furthered testified that [
].126 Similarly, Aaron Harrison testified that [
].127 Mr. Harrison testified that if
.]128 He further testified that [
.]”129 He explained that these provisions “[
.]”130
125 Adadevoh Dep., pp. 31-34. 126 Adadevoh Dep., p. 54. 127 Harrison Dep., p. 59. 128 Harrison Dep., pp. 63-64. 129 Harrison Dep., pp. 168-169, 188. 130 Harrison Dep., pp. 169-170.
46 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
PUBLIC VERSION PUBLIC VERSION PUBLIC VERSION
conditions must be conducted before a conclusion of market power is reached. [
].144 The remaining market share is held by other services, such as Google and Amazon. Thus, using market share as a screen would suggest the inquiry should end without further analysis, with the conclusion that Spotify and Apple have no market power.
71. Nevertheless, looking more deeply into the industry’s economic conditions confirms the conclusion of no market power. The conditions under which a firm may have market power are that its rivals face barriers to expansion and there are barriers to entry by new rivals. Otherwise, an attempt by the firm to increase unilaterally its price to buyers (or decrease unilaterally its price to input suppliers) will be met with switching by buyers (or input suppliers) to its competitors, which will have no trouble accommodating the increased demand given no substantial barriers to expansion, thereby defeating the attempted price increase. HHIs and market shares are not indicative of a firm’s market power when there are no substantial barriers to expansion and entry.145 As discussed below, these conditions hold for interactive services. Thus, Mr. Orszag’s
144 According to the Department of Justice, “[a]lthough the courts ‘have not yet identified a precise level at which monopoly power will be inferred’, they have demanded a dominant market share…A high share indicates that it is appropriate to examine other relevant factors. In this regard, if a firm has maintained a market share in excess of two-thirds for a significant period and market conditions (for example, barriers to entry) are such that the firm’s market share is unlikely to be eroded in the near future, the Department believes that such evidence ordinarily should establish a rebuttable presumption that the firm possesses monopoly power…In many decades of section 2 enforcement, we are aware of no court that has found monopoly power when defendant’s share was less than fifty percent, suggesting instances of monopoly power below such a share, even if theoretically possible, are exceedingly rare in practice.” (emphasis added). See “Competition and Monopoly: Single-Firm Conduct under Section 2 of the Sherman Act : Chapter 2,” DOJ, https://www.justice.gov/atr/competition-and- monopoly-single-firm-conduct-under-section-2-sherman-act-chapter-2. See also D. Carlton and J. Perloff, Modern Industrial Organization, 4th Edition, 2005, p. 644: “There is no agreement as to exactly what share (or change in share) is ‘high,’ but many economists regard a share in the range of 30 to 50 percent as too low to indicate significant market power in an industry with a competitive fringe comprising the remainder of the market.” 145 D. Carlton and J. Perloff, Modern Industrial Organization, 4th Edition, 2005, p. 644: “Market shares are imperfect indicators of market power… For example, if entry is easy, then the industry pricing is severely constrained regardless of whether an existing firm has a large market share.”
52 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
HHI and market share calculations are economically meaningless and, moreover, the interactive
services do not have increased market power/bargaining power in their negotiations with the
major labels.
72. It is undisputed that Spotify and Apple Music compete for users. Spotify points out in
public filings with the U.S. Securities and Exchange Commission that its “closest competitor” is
Apple Music.146 As articulated in Spotify’s 2018 filing, it faces “increasing competition for
Users from a growing variety of businesses, including other subscription music services around
the world, many of which offer services that seek to emulate our Service.”147 Interactive, on-
demand services emulate and enhance the same user-valued features to ensure user retention.148
For Spotify, user retention (i.e., competition) depends in large part on its “ability to continue to offer leading technologies and products, compelling content, superior functionality, and an engaging User experience.”149 Without these offerings, users can easily switch to Spotify’s
closest competitor, Apple Music, or any other on-demand streaming service. Third-party apps
such as Soundiiz further facilitate switching by enabling users to transfer personal playlists
between streaming services, another contributing factor to these services’ substitutability.150
73. From the point of view of users, interactive services are substitutable, and therefore, each
interactive service must continue to be competitive to retain its users. While the interactive
146 SOUNDEX_W5_000033096, at SOUNDEX_W5_000033103. 147 SOUNDEX_W5_000033096, at SOUNDEX_W5_000033131 (emphasis added). 148 SOUNDEX_W5_000033096, at SOUNDEX_W5_000033131. “Why Do We Still Pay Only $10 a Month for Music?,” Rolling Stone, December 11, 2019, https://www-rollingstone- com.cdn.ampproject.org/c/s/www.rollingstone.com/music/music-news/music-streaming-10-month-fee- 924809/amp/. 149 SOUNDEX_W5_000033096, at SOUNDEX_W5_000033117; SOUNDEX_W5_000026604, at SOUNDEX_W5_000026613. 150 “How to transfer playlists from Spotify to Apple Music,” AppleInsider, August 18, 2019, https://appleinsider.com/articles/19/08/18/how-to-transfer-playlists-from-spotify-to-apple-music.
53 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
discounts from this base price have been offered through, e.g., family plans.162 If anything, competition among interactive services has grown more intense.
76. The effects of this downstream competition flow back upstream, weakening the services’ bargaining positions in licensing negotiations with the labels. The downstream competition for users makes it crucial for an interactive service to be able to offer the core functionality of broad access to music. This need, and the inability of services to steer, means that the services have little leverage in negotiations with the major labels. From the labels’ perspective, because the services are substitutable from the point of view of users, they are also substitutable from the point of view of labels seeking to reach those users (and, of course, the labels have other channels for reaching users as well). A service could have market power over a label only to the extent that it was a conduit between the label and users for which there was no close substitute.
However, the services are substitutes for each other. Consistent with these points, none of the major interactive services has ever attempted to go without one of the major labels.
77. Mr. Orszag claims that a label would be hurt more than a service by a bargaining impasse between them. This claim is unsupported. Following a significant decrease in its library of songs, a subscription interactive service likely would have to give immediate discounts on subscription fees to disgruntled users who were not getting what they paid for, would suffer subscriber losses given the ease of switching to other services, and would sustain lasting harm to its reputation. Mr. Orszag points to pre-payment of subscriber fees, but he fails to show that the extent of pre-payment is significant, either in terms of users who do it or length of time covered.
In any event, a service would have to refund a significant portion of the pre-payments to avoid
162 See, e.g., “Best Music Streaming Services,” Consumer Reports, September 18, 2019, https://www.consumerreports.org/streaming-media/best-music-streaming-service-for-you/.
56 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
harm to reputation that would otherwise occur by providing less than was promised. The label,
in contrast, would recapture a portion of the royalties via other channels to which users turned to
hear the label’s music. Moreover, a label would have a greater ability to wait out the impasse, given that it would continue to receive royalties from other sources, whereas the service’s entire subscription revenues would potentially be at risk given that it was not able to meet the primary driver of demand for an interactive service.
78. A further important factor is that the services face no substantial barriers to expansion.
Thus, any service could easily expand to accommodate users switching from other services.
Again, Apple and Amazon have shown how easy expansion is. Apple has participated in the
U.S. interactive subscription service segment for less than five years but expanded quickly and has overtaken Spotify in paid U.S. subscribers.163 There are also no significant barriers to entry.
In fact, potential entrants may face lower entry barriers since Congress’s adoption of the Music
Modernization Act (“MMA”). Starting in January 2021—the start of the rate term being decided here—the MMA will establish a new blanket license for the musical composition rights that are necessary to launch an interactive streaming service.164 As the then-CEO of the Digital Media
Association testified to Congress, “[b]y reducing uncertainty and increasing efficiency, the
MMA will result in greater consumer choice in the on-demand music streaming space, as new
entrants will be able to launch a service with lower barriers to entry.”165 Entry by a new service faces no significant barriers. This is particularly the case for a platform with economies of scope, as in the cases of Apple, Google, and Amazon. Since 2015, several new interactive
163 “Apple Music Overtakes Spotify in Paid U.S. Subscribers,” The Wall Street Journal, April 5, 2019, https://www.wsj.com/articles/apple-music-overtakes-spotify-in-u-s-subscribers-11554475924. 164 Public Law 115-264: Music Modernization Act, 132 Stat. 3676, October 11, 2018, Title I. 165 Statement of Christopher Harrison (Chief Executive Officer Digital Media Association), at the Hearing on “Protecting and Promoting Music Creation for the 21st Century,” May 15, 2018.
57 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
services have entered in the US, including Apple Music (2015), iHeartRadio and iHeartRadio All
Access (2016), Primephonic (2017), Pandora Premium (2017), YouTube Music (2018), and
Qobuz (2019).166 Furthermore, the fact that industry revenue has been growing makes entry more attractive. Even the major labels have acquired or launched interactive streaming services to enter the market – e.g., Sony has launched its own Hi-Res interactive service in Japan.167
79. Furthermore, similar to other internet industries, interactive services participate in a segment that is dynamic and subject to rapid change, as demonstrated by the entry and subsequent growth of Apple and Amazon. Even firms with leading shares can lose that share quickly to existing or new competitors. For example, SoundCloud, despite being one of the top five interactive services as of 2018, struggled to maintain its position as indicated by its mass layoffs and rapid decline in its market valuation in 2017 due to intense competition from other services that blunted its growth.168
166 “Introducing Apple Music — All The Ways You Love Music. All in One Place,” Apple, June 8, 2015, https://www.apple.com/newsroom/2015/06/08Introducing-Apple-Music-All-The-Ways-You-Love-Music-All-in- One-Place-/ ; “iHeartMedia Launches First Ever -- and Only – Interactive on Demand Radio Services, ‘iHeartRadio Plus’ & ‘iHeartRadio All Access,’” BusinessWire, December 1, 2016, https://www.businesswire.com/news/home/20161201005685/en/iHeartMedia-Launches----%E2%80%93- Interactive-Demand-Radio/; “Primephonic launches hi-res classical-music streaming service,” Musically, June 14, 2017, https://musically.com/2017/06/14/primephonic-classical-music-streaming/; “Pandora’s on-demand music service finally arrives,” TechCrunch, March 13, 2017, https://techcrunch.com/2017/03/13/pandoras-on- demand-music-service-finally-arrives/; “YouTube Music and YouTube Premium Officially Launch in US, Canada, UK, and Other Countries,” The Verge, June 18, 2018, https://www.theverge.com/2018/6/18/17475122/youtube-music-premium-launch-us-canada-uk; “New high-res music streaming service Qobuz launches in US, takes aim at Spotify, Pandora,” USA Today, February 14, 2019, https://www.usatoday.com/story/tech/news/2019/02/14/spotify-pandora-have-new-competitor-music-streaming- service-qobuz/2860467002/. 167 “Why Has Sony Music Just Launched A Streaming Service (And Spotify Rival) In Japan?” Music Business Worldwide, December 11, 2018, https://www.musicbusinessworldwide.com/why-has-sony-just-launched-a- music-streaming-service-and-spotify-rival-in-japan/. 168 “Most popular music streaming services in the United States as of March 2018, by reach”, Statista, November 20, 2019, https://www.statista.com/statistics/798715/most-popular-us-music-streaming-services-ranked-by- reach/; “As Streaming Services Boom, SoundCloud Strives for Relevancy,” Wired, April 25, 2017, https://www.wired.com/2017/04/soundcloud-crossroads/; “SoundCloud, the ‘YouTube for audio’, cuts 173 jobs, closes San Francisco, London offices” TechCrunch, July 6, 2017, 58 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
80. Mr. Orszag claims that [
], which he argues is a further sign that the interactive services are exercising increased bargaining power in licensing negotiations with the labels.169 (Mr. Orszag, however, ignores the
labels’ ] on interactive services, which as discussed
above is a sign of the labels’ continued greater relative bargaining power over the services.) Mr.
Orszag’s claim is flawed for the following reasons.
81. First, in making this argument, Mr. Orszag only considers subscription interactive
services as a benchmark, and ignores ad-supported interactive services which are a better
benchmark for non-subscription non-interactive services (although still substantially flawed for the reasons discussed throughout this report).170 In fact, [
]. The reason
that [
], by making efforts and investment to improve the value of its advertising platform to advertisers. Indeed, according to
Spotify’s Form 20-F of 2018, its ad-supported service generates revenue “from the sale of
display, audio, and video advertising delivered through advertising impressions.”172 In order to
increase its advertising revenue, Spotify has “introduced a number of new advertising products,
https://techcrunch.com/2017/07/06/soundcloud-the-youtube-for-audio-cuts-173-jobs-closes-san-francisco- london-offices/. 169 Orszag Report, pp. 73-75. 170 Orszag Report, pp. 35-37, 42-43. 171 Calculated using Orszag data. 172 SOUNDEX_W5_000026604, at SOUNDEX_W5_000026653.
59 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
including sponsored playlists … and continue[d] to focus on analytics and measurement tools to
evaluate, demonstrate, and improve the effectiveness of advertising campaigns on [its]
platform.”173 These efforts have been made to “improve the efficiency and scalability of
[Spotify’s] advertising platform.”174 As Dr. Tucker acknowledges, Spotify launched its
“programmatic advertising” offering in 2015. She explains that such programmatic advertising
“enhances the attractiveness of digital music services to advertisers and enables digital music
services to compete better with large players in online advertising such as Google and
Facebook.”175 Dr. Tucker also acknowledges that Spotify launched its fast-growing self-service
advertising platform in 2017, “Ad Studio.” She notes that self-service advertising platforms such
as these “platforms can help digital music services increase their advertising inventory.”176 To the extent [ ] is the result of Spotify’s efforts and investments in building its advertising platform, it would be incorrect to interpret [
] as due to increased Spotify bargaining power. Instead, the
[
] provides no support for an increase in Spotify market power – rather, it reflects Spotify’s
173 SOUNDEX_W5_000026604, at SOUNDEX_W5_000026653. 174 SOUNDEX_W5_000026604, at SOUNDEX_W5_000026658. 175 Tucker Report, pp. 21-22. 176 Tucker Report, pp. 22-23. Spotify notes that the Ad Studio is their “fastest growing advertising channel” based on quarterly revenue and advertiser growth. See “Spotify’s Ad Studio is its fastest-growing advertising channel,” Music Ally, January 30, 2019, https://musically.com/2019/01/30/spotifys-ad-studio-is-its-fastest- growing-advertising-channel/. Additionally, Spotify’s Global Head of Advertising said, “The self-serve tool Ad Studio that allows smaller businesses to activate the Spotify free audience is a big growth area for us.” See “Spotify: Our Ads Business is a real Trigger for Growth,” Music Business Worldwide, October 31, 2018, https://www musicbusinessworldwide.com/spotify-our-ads-business-is-a-real-trigger-for-growth-moving- forward/.
60 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25)
PUBLIC VERSION
continuous efforts to improve its product quality and revenue generating ability in order to
maintain its market position in a highly competitive market.
82. Moreover, Mr. Orszag is mistaken as a matter of economics in his claim that a decrease
in the royalty as a percentage of revenue implies increasing market power on the part of the
interactive services. Mr. Orszag’s fundamental assumption—that the royalty should be the same
percentage of revenue, not only across services (which is wrong for the reasons discussed
above), but also across time, even as the downstream revenue per play fluctuates—is simply
wrong as a matter of economics. Indeed, if the interactive services are competitive in the
upstream licensing market and the downstream demand curve each service faces flattens (which
would occur, for example, with increased downstream competitiveness among them), the
percentage-of-revenue royalty would be expected to fall, [ ].
Thus, Mr. Orszag has no basis in economics to claim that the [
] is an indication that those services have
increased market power or bargaining power in upstream licensing negotiations.
83. This can be illustrated in a simplified, but instructive example. Suppose a monopolist
label of sound recordings sells a license to plays to a perfectly competitive intermediary industry
(i.e., services) at a royalty of r per unit. Each unit of the input is transformed by the
intermediaries into one unit of (homogeneous) output. Assume the intermediaries are perfectly
competitive and exhibit constant returns to scale with cost c per unit to transform the input.
Suppose downstream demand is Q(p), where p is the downstream unit price. Under these assumptions, downstream price equals to marginal cost177 (i.e., = + ) and the label chooses
177 In the competitive equilibrium, a competitive firm’s price equals marginal cost. See D. Carlton and J. Perloff, Modern Industrial Organization, 4th Edition, 2005, pp. 58-59.
61 NAB Written Rebuttal Statement Dkt. No. 19-CRB-0005-WR (2021-25) PUBLIC VERSION
r to maximize its revenue that is equal to ∙ ( + ). The royalty as a percentage of revenue is