PUBLIC VERSION

Before the UNITED STATES COPYRIGHT ROYALTY JUDGES Library of Congress Washington, D.C.

In the Matter of:

Determination of Royalty Rates and. Docket No. 16—CRB—0001—SR/PSSR Terms for Transmission of Sound. (2018-2022) Recordings by Satellite Radio and "Preexisting" Subscription Services (SDARS III)

VOLUME I

MUSIC CHOICE'S WRITTEN DIRECT STATEMENT

WRITTEN DIRECT TESTIMONY OF DAVID J. DKL BKCCARO WRITTEN DIRECT TESTIMONY OF DAMON WILLIAMS

WRITTEN DIRECT TESTIMONY OF GREGORY S. CRAWFORD PhD Music Choice's Written Direct Statement

1000000000000000000000000000000000010000000 0 Before the UNITED STATES COPYRIGHT ROYALTY JUDGES Library of Congress Washington, D.C. 0 In the Matter of: Docket No. 16-CRB-0001—SR/PSSR Determination of Royalty Rates (2018-2022) 0 and Terms for Transmission of Sound Recordings by Satellite Radio and "Preexisting" Subscription Services (SDARS III)

MUSIC CHOICE'S WRITTEN DIRECT STATEMENT

Pursuant to 87 C.F.R. $ 851.4 and the Judges'otice of Participants, Order dated 0 Commencement ofVoluntary Negotiation Period, and Case Scheduling March 14, 2016 and the Judges'rder for Further Proceedings and Case Scheduling

dated June 28, 2016 (the "Judges'rders"), Music Choice respectfully submits its written direct statement. CONTENTS OF MUSIC CHOICE'S WRITTEN DIRECT CASE 0 VOLUME I: WITNESS TESTIMONY Music Choice '~ Pursuant to 87 C.F.R. $ 851.4(b)(1) and. the Judges'rders, submits written direct testimony from the following expert and. fact witnesses, ,~ included in Volume I: Music (1) David J. Del Beccaro, the President and Chief Executive OKcer of 0 Choice, presents testimony concerning the business operations ofMusic Choice. He ("PSS") also testi6es concerning the history of the Preexisting Subscription Services royalty rate, including how it was initially set too high based upon an erroneous "PRO estimate of the aggregate musical composition performance royalty rate (the subsequently increased in a Rate") by the original CARP, and how that rate was series of settlements only to avoid. the expense of litigation. Mr. Del Beccaro also

explains that the Judges'mposition of a further rate increase in the SDARS II

proceeding was based solely upon the assumption of a future channel increase that

Music Choice never actually implemented and was therefore (and for other reasons)

erroneous.

Mx. Del Beccaro then testi6es to various changes in the PSS marketplace that

have led to a signi6cant change in Music Choice's business since the time of the

SDARS II proceeding, and. demonstrates the impact of the implicit 7.5% rate from

the SDARS II determination on Music Choice's business during the next rate period. In particular, addxessing each of the Section 801(b) policy objectives, he demonstrates the adverse impact the implicit 7.5% rate will have on those

objectives, and why the change in circumstances since the SDARS II proceeding warrant a reduction in the PSS rate to a rate no higher than 5.6% of revenue. Mr.

Del Beccaxo will further testify concerning the lack of any true marketplace value

for the Section 112 ephemeral license royalty rate and. the lack of any need. to

change the long-established terms of the PSS license. and (2) Damon Williams, the Senior Vice President of Programming Strategy Partnerships at Music Choice, presents testimony demonstrating the various ways in which Music Choice provides its own original, creative content to its residential 0 subscribers, including its unique approach to cuxation of the music channels 0 themselves, as well as various on-screen, social media, and. other visual elements related to the service, which increase subscriber engagement and. increase the

promotional impact of the service. Mr. Williams also testi6es speci6cally about that

promotional impact, particularly the many ways in which Music Choice increases

record. sales and. on-demand streaming, and. promotes artists and genres that are therefore promoting the continued. release of a 0 underserved, by other music services, 0 wider variety of music to the public. 0 Dr. Greeorv Crawford. Ph.D., Professor of Economics, Director of Graduate 0 (3) Studies at the University of Zurich Department of Economics, provides his expert 0 opinion as to the proper rates to be set in this proceeding. After observing the Judges'rior ruling that there is no usable marketplace benchmark for the PSS, he 0 relies upon well-established economic principles to develop an economic framework to determine the likely outcome of a relevant hypothetical marketplace negotiation 0 of a license for the sound. recording performance right between a record. company and a PSS. He uses the Asymmetric Nash Bargaining Framework, examuung the

0 actual marketplace dynamics a8ecting Music Choice and the record. companies and

analyzing the factors of agreement pro6ts, threat points and baxgaining power that

would be applicable in the hypothetical negotiation. Professor Crawford then

quanti6es those inputs using empirical data, including the cuxrent and. short-term projections of future performance ofMusic Choice's business if it were limited to the residential audio service covered by the PSS license.

Using this methodology, Professor Crawford. demonstrates that in the relevant

hypothetical marketplace negotiation, even with various conservative assumptions

such as allocating no value to the promotional impact of Music Choice and allocating

signi6cantly less value to Music Choice's bundled video offerings than the value

supported by empirical evidence, the PSS rate would most likely be 8.5% of gross

revenue, but in any event no higher than 5.6% of revenue. He also considers the

Section 801(b)(l) policy objectives and. concludes that, on balance, they indicate that

the PSS rate should be set below the high end. ofthe range yielded. by the

hypothetical marketplace negotiation. VOLUME II 4 VOLUME III: MUSIC CHOICE EXHIBITS

Pursuant to 87 C.F.R. $ 851.4(b)(1) and the Judges'rders, Music Choice

submits the following exhibits, included in Volumes II L III: Exhibit Sponsoring Description Version No. Witness David J. Del Speaking Engagements Public Beccaro MC1 David J. Del Channel Lineups Public MC2 Beccaro David J. Del Register of Copyrights'006 PSS Decision Restricted MC8 Beccaro David J. Del 1993 Sony License Restricted MC4 Beccaro David J. Del 1994 EMI License Restricted MC5 Beccaro David J. Del 1993 WMGLicense Restricted MC6 Beccaro 0

Exhibit Sponsoring Description Version No. Witness Del ASCAP License Restricted ' David J. MC 7 Beccaro David J. Del BMI License Restricted 0 MC8 Beccaro ,0 David J. Del SESAC License Restricted 0 MC9 Beccaro David J. Del Music Choice Residential Audio Only Restricted MC 10 Beccaro Financials 0 David J. Del Music Choice Residential Consolidated Restricted MC 11 Beccaro Financials David J. Del News and Notes on 2016 Mid-Year Public MC 12 Beccaro RIAA Music Shi ment and Revenue Statistics David J. Del News and Notes on 2015 RIAA Shipment and Public Beccaro Revenue Statistics DamonWllhams Damon Williams Speaking Engagements Public Restricted Damon Wllllams July 2016 study by IPSOS OTX MediaCT Damon Williams Music Choice Channel Programming Restricted Plllloso ines Damon Williams Music Choice On-Screen Dis la Public Damon Williams Comparison of Music Choice Country Channel Restricted versus Radio Station Listenersbi Punch Album Premiere Restricted Damon Williams Five Finger Death MC 19 Reca Pubhc MC 20 DamonWilliams Brand New This Week Pro am Presentation Damon Williams Brand New This Week Artist Social Media Pubhc Back Your Music, The F- Restricted DamonWllllams Materials from Take MC 22 Word, MC Icons and Primed Pro rams Damon Williams PR that Music Choice Has Done for Artists Public Damon Williams Sampling of Other Emails Thanlang Music Restricted Choice Damon Williams Representative Plaques Recognizing Music Public MC 25 Choice Restricted MC 26 DamonWilliams Primed Reca Deck Restricted MC 27 DamonWilliams Music Choice Promotional Pro ram Deck Restricted 0 MC 28 Damon Williams Artist Post on Music Choice Blo DamonWilliams Record Label Promotional Emails Listing Public 0 MC 29 A'a on Music Choice Damon Williams Sample Email Re: Record Label Recognition Restricted MC 80 ofPromotional Im act Damon Williams Sample Email Re: Recognition Of Promotional Restricted MC 31 Im act 0 Exhibit Sponsoring Description Version No. Witness Damon Williams Recognition Received &om Rich Homie Quan Restricted and Promotional Materials Created for Rich MC 82 Homie Quan DamonWilliams Presentation Summarizing Rich Homie Quan Restricted MC 83 Promotion DamonWilliams List ofArtist Studio Visits and On Location Restricted MC 84 Promotions DamonWilliams A Case Study ofArtist Promotion on Music Restricted MC 85 Choice Damon Williams Presentation summarizing the Mainland Primed Restricted MC 36 program DamonWilliams Materials related to the Chris Brown Week Restricted MC 37 promotion DamonWilliams Case study ofMusic Choice Bugatti Beez Restricted MC 88 promotion MC 89 Damon Williams Neilsen Music 360 Report 2015 Highlights Public MC 40 Damon Williams Infinite Dial 2014 study Public DamonWilliams Bmails from labels and artists asking for Music Restricted MC 41 Choice's support

MUSIC CHOICE'S RATE PROPOSAL

Section Pursuant to 87 C.F.R. $ 851.4(b)(8), Music Choice proposes that the

114 sound. recording performance license rate for Music Choice be reduced. to no higher than 5.6 percent of gross revenues as that term is currently de6ned in the applicable regulations. Because the ephemeral copies made by Music Choice have no independent economic value, Music Choice proposes that the Section 112 ephemeral license fee be included within the performance royalty rate noted. above. Music

Choice proposes that no changes be made to the terms and. other regulations

applicable to the PSS license. 0 0 0 0

, ~ 0 DATED: October 19„2016

By ,0 Paul M. Fakler ,0 John Sullivan Margaret Wheeler-Frothingham 0 1675 Broadway 0 New York, NY 10019-5874 Tel: (212} 484-3900 Fax: (212) 484-3990 0 Email: [email protected] 0 [email protected] ' [email protected]

, ~ ' Jackson D. Toof ,0 1717 K Street., N.W. Washington, D.C. 20006-5344 Tel: (202} 857-6000 '0 Fax: (202) 857-6395 Email: [email protected] 0 Counsel to Music Choice 0 0 0 0

0 0 0

0 Written Direct Testimony of David J. Del Beccaro

~ 0000000000000000000000000000000000000000000 PUBLIC VERSION

Before the UNITED STATES COPYRIGHT ROYALTY JUDGES Library of Congress Washington, D.C.

In the Matter of:

Determination of Royalty Rates and Docket No. 16—CRB—0001—SR/PSSR Terms for Transmission of Sound. (2018-2022) Recordings by Satellite Radio and. "Preexisting" Subscription Services (SDARS III)

WRITTEN DIRECT TESTIMONY OF DAVID J. DEL BECCARO (On behalf of Music Choice)

My name is David. J. Del Beccaro and. I am the Presid.ent and. CEO of

Music Choice. I have overseen all aspects of Music Choice's business since the company's inception in 1987. I submit this testimony in connection with the above-captioned proceeding, in which the Copyright Royalty Judges ("CRJ") will adjust the rates for the statutory license used by Music Choice for the public performance of sound recordings as a preexisting subscription service pursuant to 17 U.S.C. $ 114(f)(l) and. for the related. ephemeral copying of those sound. recordings pursuant to 17 U.S.C. $ 112(e).

My Background

I helped commercialize Music Choice (formerly named Digital Cable Radio

Associates) beginning in 1987, when I served. as Vice President, Business

Development for Jerrold Communications ("Jerrold."), a division of General PUBLIC VERSION

Instrument Corporation ("GI"). After approximately four years of product development and. market testing within Jerrold., I helped secure financing for the digital music service concept through a partnership of major cable and music companies, beginning in 1991 when the company launched as a stand-alone entity.

Between 1991 and 2000, a number of major companies became investors in the venture through various predecessors and. af51iates. Those companies are now:

ARRIS Group, Inc.; Comcast Corporation; Cox Communications, Inc.; EMI

Music Publishing North America Holdings, Inc.; Microsoft Corporation; Sony

Corporation of America; Spectrum Management Holding Company, LLC; and

Time Warner Inc. Notably, these partners include both a company that owns record. labels, a music publisher, and. companies that own cable companies. All of the deals we have done with those record. labels and. cable companies owned by our partners have been conducted at arm's length, with no favoritism shown to Music Choice. Indeed., some of the most liKcult and. protracted. negotiations have been those with our cable company partners. Nor are we in any position to show favoritism to cable or record. companies owned by our partners as we have several partners with no record label or cable affiliates. Further, no single partner hold.s a controlling interest in Music Choice and., in~et, as a group the non-cable partners hold. a signi6cantly greater ownership interest (by a two to one margin) in Music Choice than our cable partners do as a group.

In my capacity as President and. CEO of Music Choice over the past twenty-five years, I have devised., implemented. and. overseen various changes in PUBLIC VERSION

the company's offerings and technologies, as the company has had to adapt to

an increasingly difEicult and competitive market for cable music services. In this

time, I have become intimately familiar with various facets of the music

ind.ustry, including the production and promotion of sound. recordings, artist

promotion, and the many forms of broad.casting, television networks, and. digital

music services available in the marketplace. I have been quoted. in The New

York Times, Los Angeles Times, Associated Press, Reuters, MultiChannel

News, and Billboard., among other national publications on various topics

related. to music and technology. A list of my recent speaking engagements on

these topics is submitted as Exhibit MC l.

Prior to holding my current position, I served as the Vice President of

New Business Development at Jerrold., as noted above. Before joining Jerrold., I

held. various marketing and. financial positions at GI and Ford Motor Company.

I have B.S. and M.S. degrees in Industrial Engineering from Stanford.

University and a B.A. in Management Engineering from Claremont McKenna

College.

I am familiar with the operations of Music Choice and. its relationships

with copyright owners and. their representatives, including the America~

Society of Composers, Authors and Publishers ("ASCAP"), Broadcast Music, Inc.

("BMI"), SESAC, Inc. ("SESAC"), the Recording Industry Association ofAmerica

("RIAA"), various individual record companies (including all of the major labels), SoundExchange and. other licensing entities. As part of my PUBLIC VERSION responsibilities as President and CEO of Music Choice, I also keep myself generally apprised of the copyright licensing costs faced by similar businesses.

The following testimony is based upon my personal knowledge and information available to me in the course of performing my duties as President and CEO of

Music Choice. The Music Choice Service Music Choice's residential audio service that is the subject of this proceeding is a music service comprised of 50 channels of diverse audio programming delivered to subscribers'elevisions, with an additional 25 channels made available to authenticated television subscribers through Music

Choice's website and a mobile app. A list of Music Choice's current channel lineup is submitted as Exhibit MC 2. Each channel provides a distinct musical genre or sub-genre to the listener. Our service is delivered to customers by cable operators and. other multi-channel video programming distributors ("MVPDs") as part of a package of television offerings to customers in the home (e.g., the

Music Choice service is included by MVPDs as part, of their digital basic cable service). We typically enter into long-term contracts with each of our MVPD affiliates. At this time, themeighted average term of our affiliate agreements is ..'.. Virtually all customers receive all 50 of our television channels through our residential service. Our programming currently reaches almost 50 million residential customers across the United States.

Music Choice provides its audio service to residential customers under the statutory license for the public performance of sound recordings by a PUBI I C VERSION

"preexisting subscription service" ("PSS"), as that term is defined in Section

114(j)(11) of the Copyright Act. We fully comply with the sound recording

performance complement, as required by the statutory license. Accordingly, we 0 do not play more than three different selections of sound recordings from any one phonorecord within a three hour period on any of our channels. We do not 0 consecutively play more than four different selections of sound recordings by the 0 same artist or from a compilation set of phonorecords within a three hour period 0 on any of our channels. We do not pre-announce our play list. We make regular 0 reports of our programming, and regularly remit the required license fees to 0 SoundExchange. Since the statutory license was enacted in 1995, Music Choice has paid the record companies and recording artists over 0 royalties. The Applicable PSS Rate-Setting Standard As a PSS, the Music Choice residential audio service is subject to a

special standard under Sections 114{f)(l) and 801(b){1) of the Copyright Act. As

, ~ the Register of Copyrights confirmed in connection with the 2006 PSS rate ' proceeding (SDABS this standard. is designed to protect the business 0 I), 0 expectancies of the few pioneering digital music services that were making digital performances of sound recordings prior to 1995, when the limited digital

performance right for. sound recordings first came into existence, and 1998, 0 when the rate standard for certain statutory licenses were modified to a 0 market-based standard for newer services that first entered. the market

0 PUBLIC VERSION

thereafter. A copy of the Register's decision, discussing the history of the PSS

rate-setting standard is submitted as Exhibit MC 3. 0 As noted. above, I began commercializing the Music Choice service in 1987. At that time, there was no public performance right of any kind for sound

recordings. When the service was actually launched. to customers nationally in 0 1991, after signi6cant investment by Music Choice in new technology and. 0 systems, there still was no public performance right for sound. recordings. 0 Because of this, Music Choice had. no reason to believe that it would ever have 0 to pay a penny for the sound. recording performance right (just like its primary 0 competitor at the time, terrestrial radio). The legal land.scape started to change 0 in the mid.-90s, and, sound. recording owners were granted. a very limited. digital performance right in 1996. Partly in order to protect the few existing digital '0, music services, including Music Choice, at the same time Congress created the new performance right it also createl a compulsory license. Congress provided. that the rate for this license would. be set, not as a pure market rate, but rather 0 by evaluation of four policy-based. factors now codified in Section 801(b), which , ~ require consideration of, among other things, the licensees'inancial conditions,

marketplace realities and. the avoid.ance of disruption to the services'arkets.

' The legal landscape changed. yet again in 1998, with the passage of the

Digital Millennium Copyright Act. In that legislation, Congress changed the

rate-setting stand.ard for new digital music services to a market-based, willing

buyer / willing seller stand.ard. In d.oing so, however, Congress went out of its

0 PUBLIC VERSION way to allow what were then the three preexisting subscription services operating in the residential cable radio market, Music Choice, Muzak, and

DMX, to keep the non-market, policy-based rate-setting standard. As the

Register of Copyrights noted in her decision, Congress grandfathered the standard for the PSS to protect their business expectancies, in recognition that they had launched. new businesses and invested significant sums under a very different set of legal rules than those that would be applied to newer services.

Consequently, and. unlike most other statutory licenses for the sound. recording digital performance right, the PSS license standard provides for a

"reasonable royalty" that is set without reference to a market rate, other than having a market rate as its absolute upper boundary. According to the statute, the reasonable royalty rate is set based upon evaluation of the following policy objectives:

(A) To maximize the availability of creative works to the public.

(8) To afford the copyright owner a fair return for his or her creative work and the copyright user a fair income under existing economic conditions.

(C) To re6ect the relative roles of the copyright owner and the copyright user with respect to relative creative contribution, technological contribution, capital investment, cost, risk, and contribution to the opening of new markets for creative expression and media for their communication.

(D) To minimize any disruptive impact on the structure of the industries involved and on. generally prevailing industry practices.

17 U.S.C. $ 801(b)(1). 0 0 PUBLIC VERSION 0 0 History of PSS Rates As noted above, Music Choice had no expectation that it would. have to

pay any sound. recording royalties at the time the service was developed and.

launched.. In 1998 and 1994, shortly before the creation of the digital 0 performance right, certain of the major music companies, namely Warner Music 0 Group, Sony Music and EMI Music, invested in Music Choice. In connection 0 with that investment, and. for the purpose of establishing that there was a 0 specif'ic, recognized. value to the performance of sound recordings, the record. 0 labels insisted. that Music Choice agree to pay them a license fee. At the time, Warner, Sony and. EMI collectively controlled over two-thirds of the market for

sound recordings and therefore had. signi6cant bargaining power. Also, Music

0 Choice was desperate to add partners that would contribute capital to the

and the music companies were willing to d.o so — partially in trad.e for , ~ company 0 Music Choice agreeing to a royalty rate at the requested. amount. The royalty 0 rate was two percent of revenue, adjusted. for the percentage of each record 0 company's music played. on the Music Choice service so that two percent would cover the entire record industry. Copies of these licenses are submitted. as 0 Exhibits MC 4-6. Although these royalty payments were agreed to before there was an independent statutory obligation to pay for the performance of sound.

recordings, I believe based. upon my experience negotiating these deals that the

0 two percent rate (allocated among the whole record industry) represented. the 0 value the record. labels hoped. to place on the sound. recording performance right at that time. 0 PUBLIC VERSION

During the first PSS rate-setting proceeding, commenced. in 1996, the

Copyright Arbitration Royalty Panel ("CARP") initially set the royalty rate at 5 percent of gross domestic revenue from each licensed. residential service. On appeal, in 1998, the Librarian of Congress adjusted. the rate to 6.5 percent. First, the Librarian considered the PSS'usical composition performance licenses with the Performing Rights Organizations ("PROs") as a marketplace referent. In particular, he used the high-end. estimate of the sum of the services'icense rates for the digital performance of the underlying musical compositions paid to ASCAP, BMI, and SESAC to establish the highest possible reasonable rate for the equivalent digital sound. record.ing performance license.

The Librarian relied upon the CARP's estimate of ten percent of revenue as the sum of the maximum likely aggregate PRO license rate. That f'igure was estimated because certain of the licenses were in a period. of negotiation and litigation, and. therefore had. not yet been Gnally set. Nor had. there been a history of final rates during prior periods because the services were new at the time. As it turned. out, however, the CARP's estimate was too high. When Music

Choice's rates to be paid to the three PROs were Gnalized., the aggregate rate was approximately MW3] of gross revenue from our residential audio service—roughly M-"&~ of the estimate upon which the original rate was based.. That aggregate rate has not risen since that time and remains approximately ~~~&. Music Choice's current licenses with the PROs are submitted. as Exhibits MC 7-9. PUBLIC VERSION

Next, the Librarian considered. the various evidence relevant to the policy objectives contained. in the statute, and used. those considerations to set the reasonable royalty rate at 6.5 percent, or 65 percent of the highest estimate of the aggregate PRO rate. Changes in circumstances during the intervening twenty years, however, have only strengthened. Music Choice's case under each of the policy factors, as discussed in more detail below.

On further appeal, the Court ofAppeals for the District of Columbia

Circuit approved the stand.ard. and. method. the Librarian used., and. aKrmed the

6.5~/o rate.

In twenty years, the original 6.5/o royalty rate has been adjusted. only once through the CARP (now CRQ process, in the SOBS II proceeding Bve years ago. Prior to that, Music Choice twice settled. with SoundExchange solely to avoid. the prohibitive costs associated. with the rate litigation process. Due to an inherent inequality in bargaining power, discussed further below, and. because the starting point for negotiations was the 6.5 percent rate established by the Librarian on the basis of inQated PRO royalty estimates, Music Choice had. no real choice but to accept a rate increase to 7 percent for 2002 to 2003 and

7.25 percent for 2004 through 2007. This increase in no way reGected any of th' statutory policy objectives relevant to this proceeding, nor was it even a true willing buyer / willing seller, marketplace transaction reQecting fair market value. The increase was simply a result of the fact that Music Choice could. not, from a business perspective, justify the expense in money and. sta8 resources for

10 PUBLIC VERSION another proceeding so soon after the conclusion of the appeal process of the Grst proceeding. Indeed the sum of the additional payments caused. by the rate increase over the term of the royalty period was less than the millions of dollars it would. have cost Music Choice to litigate a red.uction in that rate.

Despite repeated. efforts by Music Choice to engage in settlement discussions, when the royalty rate came up for adjustment for the next rate period., Sound.Exchange did not negotiate a settlement until directed to by the

Jud.ges d.uring the direct trial opening statements of the SOBS I proceeding in

June 2007. By the time they complied., Music Choice had. already incurred over a million d.ollars in legal fees and again was forced to agree to another slight increase, to 7.6 percent only in the Gnal rate period. year of 2012, solely to avoid. incurring millions more in legal fees. This pattern repeated. itself yet again with respect to the subsequent rate period., adjudicated. in SOBS II. It was Music Choice's hope that having been taken to task by the Judges in SDABS I, SoundExchange would. be more willing to negotiate a reasonable settlement the next time around.. Unfortunately, that was not the case. Despite the fact that Music Choice reached out to

Sound.E~hange many months before the SDABS II voluntary negotiation period. began, SoundExchange did. not engage in settlement discussions until after the negotiation period was over. Even then, and after we shared our financial d.ata (without any reciprocity), we received. the same old demands that our rate be signi6cantly increased, with no serious economic justi6cation given

11 PUBLIC VERSION other than the fact that the cost of a proceeding would. be far more burd.ensome on Music Choice than on Sound.Exchange.

In this proceeding, Music Choice reached. out to SoundExchange yet again, in January 2016, to attempt settlement solely to avoid. the costs of litigation. SoundExchange once again failed. to negotiate, and. did. not even respond to Music Choice's offer until July. The terms SoundExchange then finally deigned. to offer were again unreasonable, and. even included

As the Jud.ges may be aware, these proceedings cost millions of dollars in legal and. expert fees for each party. For a company the size of Music Choice, the cost of a rate-setting proceeding creates enormous pressure to settle, even at unfair rates. That pressure is not felt evenly by Music Choice and

Sound.Exchange. Notably, unlike the record. labels and. artists, Music Choice's interests are not represented by a large trade organization. Indeed., because

DMX had. to file for bankruptcy and. sell omits assets, and. Muzak's participation in the market is insignificant—as evidenced. by its withdrawal from every one of these proceedings—music Choice must directly bear the entire burd.en and costs of PSS rate-setting itself. Music Choice's declining revenue and. current financial situation, discussed. in more detail below, make this burden even more difficult to bear; the expense of participating in this proceeding is contributing

12 PUBLIC VERSION to Music Choice's

SoundExchange, on the other hand., has no independent business to protect or to "distract" it from hyper-litigation of these proceedings. Moreover, all of its litigation expenses are merely deducted from the royalties it collects from licensees like Music Choice. Those expenses are then effectively spread. among SoundExchange's thousands of members, in proportion to the revenues they receive from SoundExchange.

Consequently, SoundExchange has no incentive whatsoever to enter into reasonable settlements and. has every incentive to force Music Choice into these proceedings every 6ve years (as it has done the last three times — in 2006, 2011, and now) unless Music Choice agrees to a substantial increase in its royalty rate. This dynamic is worsened by the fact that the PSS rates are adjusted. in a joint proceeding with Sirius XM Radio, a service with a completely different market structure, fundamentally different demand characteristics, and. incomparably higher revenues than Music Choice. Prom SoundExchange's point of view, once it is already committed to a proceeding against Sirius XM, it does not cost much more incrementally to put on a marginal case against Music

Choice. But SoundExchange is very cognizant of the severe impact the costs will have on Music Choice. Finally, SoundExchange does not view the proceeding as carrying any risk. Whether Music Choice's rate is lowered or increased (even by a vast multiple), the difference in resulting royalties to any of the record labels

13 PUBLIC VERSION

or artists is insigni6cant. The risk to Music Choice, on the other hand., is

severe. Small percentage point royalty rate adjustments can have outsized.

effects on Music Choice and. its viability as a business, especially now, as

discussed. below.

Faced. with SoundExchange's incentives and. pattern of conduct, Music

Choice had no choice but to litigate the SOBS II proceed.ing. The one-way

ratchet of ever-increasing and. unsustainable rates demanded by

SoundExchange, propped, up by the original CARP's signi6cant overestimation, risked putting Music Choice out of business. Music Choice therefore sought a

corrected. and. reasonable rate, set pursuant to the applicable policy factors.

The Prior Proceeding (SDARS I1) The existing PSS rate is the product of a determination by the Jud.ges in SOBS II, but also a product of the erroneously-inQated. rate Rom the first PSS

CARP proceeding and. the intervening rate increases extracted. by

SounlExchange solely based upon leverage from the expense of these rate proceedings. In SEAS II, the Judges set a PSS sound recording royalty rate

of 8% for 2013, rising to 8.5% for 2014-2017 (versus the prior 7.5% rate). The

Judges first found. they had. no usable marketplace benchmarks due to unique

characteristics of the PSS market that "signi6cantly dim[med] the possibility of market comparators." Notwithstanding that all of these unique characteristics

14 PUELIC VERSION

are replicated. in Music Choice's PRO licenses for its resid.ential audio service,

the Judges rejected. those licenses as benchmarks.i

Having found no satisfactory marketplace benchmark, the Jud.ges began

their analysis with the then-existing 7.5% rate—which was the prod.uct, as

discussed. above, of a gross overestimate of the PRO license rate and. a series of

forced. settlements between Music Choice and. Sound.Exchange. In starting with

7.5%, the Judges did. not address the problematic genesis of that rate.

Reviewing the record. evidence, the Judges found. that the 7.5% rate adequately

furthered all four of the Section 801(b) policy factors based upon Music Choice's

business at that time, but found that the rate would. not be adequate during the

future rate period. because it did. not account for the fact that Music Choice was

then planning to expand. from 46 to 300 channels, resulting in a greater "usage"

of music. The Jud.ges held. that this planned. channel increase, alone, justi6ed an

increase in the rate from 7.5% to 8.5%, to compensate for the resulting

increased "usage" of music.

Of note, no participant had. advanced this "usage"-based argument, and.

Music Choice therefore had. no opportunity to respond to it during the SDARS II proceeding. As the Jud.ges recently acknowledged, the Court of Appeals (and the

'onetheless, the PRO licenses were used to set the high end ofthe range ofreasonable rates in the first PSS CARP proceeding, and I still believe that those licenses present the best marketplace comparator for the PSS license. In an ancillary market for the record companies such as the PSS license, particularly where the Music Choice service has never had any substitutional impact on the record companies'rimary sales and on-demand streaming revenues, there is no reason why the sound recording rights should be any more expensive than the music publishing rights.

15 PUBLIC VERSION

Register of Copyrights) have ruled that the Jud.ges should not adjust rates based on theories or arguments that participants did not raise. See In re

Determination ofRoyalty Rates and Terms for Ephemeral Recording and

Webcasting Digital Performance of Sound Recordings (Web IV), 81 Fed.. Reg.

26,816, at 26,819 (May 2, 2016) ("The Register appears to interpret [case law] as

barring the Jud.ges from relying on theories 'Grst presented in the and. not advanced by any participant."'). Indeed, if MusicJudges'etermination Choice had. had. the opportunity to address this "usage" theory before the prior panel applied. it, I believe the Judges would have realized. it was ill-founded. As a preliminary matter, the PSS royalties are not paid for broad. "usage" of sound recordings. The license allows the public performance of sound recordings.

Merely making additional channels available does not necessarily lead to a greater number of performances. Although a channel increase would. give

Music Choice subscribers a greater variety of channels to choose from, they can still only listen to one channel at a time. Ind.eed., in the entire time since Music

Choice added 25 channels to its app and internet platforms, the best month of listening to those channels added. approximately

The principle that the Section 114 license royalties should be paid based only on performances, not some broader concept of "usage," is reflected in the webcasting license rates—royalties do not go up based merely on additional channels offered by a webcaster, but only on additional performances, de6ned

16 PUBLIC VERSION

(similar to fundamental principles of copyright law) as occurring only when a

song is actually received. by a listener.2 And setting all that aside, to the extent

ad.ditional channels made our service more valuable to consumers, that

increased value would. be reflected in a relative increase in our revenue. As I

explain below, we charge as much as we can given the market pressures

affecting us and. the entire cable industry~. Any increased. value, includ.ing if any

value would. have been created. by additional channels, would help us resist decreases or at least slow the decline in rates, giving us more revenue than we

would. have had without providing that additional value. This relative increase

in revenue would. lead to relatively more royalties paid. to SoundExchange because our rate is calculated based. upon our revenue. If, on the other hand., an increase in channels did. not create any incremental value to Music Choice or its

subscribers, I do not believe that Music Choice should have to pay a greater percentage of its revenue for sound. recording rights. This principle is also reflected. in various PRO licenses for Music Choice

and. other licensees, includ.ing terrestrial radio broad.casters, which are set on a percentage of revenue where that percentage rate does not increase or decrease based. on the number of channels broadcasted.. Indeed., nobody would. argue that a broadcaster like CBS Radio should. have to pay a higher percentage of its revenue just because it operates more radio stations than another broadcaster.

Music Choice is not able to track the actual number ofperformances to enable a per- performance rate, so a percentage ofrevenue rate must therefore be used.

17 PUBLIC VERSION

There is no reason that, standing alone, a greater number of performances

should also mean a higher percentage of revenue royalty rate.

More important, even if the reasoning behind the rate increase had been

sound, the factual premise upon which the increase was based turned out to be

false. Music Choice never went through with its planned expansion to 300

channels.

Even und.er the

prior panel's reasoning, then, Music Choice has been overpaying for the entire

past rate period, when the rate should have been kept at 7.5% of revenue. Music Choice's Rate Proposal for the Upcoming Period Music Choice proposes that the PSS rate be reduced to a rate no higher

than 5.6% of gross revenue during the next rate period. As explained above, the rate increase implemented in SOBS II was premised solely on a planned

future expansion of Music Choice's channels that never actually occurred.

Accordingly, even under the reasoning of that decision, 7.5% was the

appropriate rate for the current perrod under the statutory policy factors.

However, because of the recent significant decrease in Music Choice's revenues

from its residential audio service, which is projected. to continue through the

next rate period., Music Choice's rate reduction must be greater than the 1% erroneous increase, because that increase was applied to years with

significantly higher revenue. Accordingly, to make Music Choice whole—

18 PUBLIC VERSION assuming no further loss of revenue or adjustments based upon the statutory policy factors — its rate must be reduced to 5.7% for the upcoming rate period. But there are additional reasons supporting our proposed decrease from the 7.5% implicit in the SDARS II determination to a rate no higher than 5.6%, based upon significant changes to Music Choice's business since the time of the last proceeding, which materially impact the applicable Section 801(b) policy factor analysis.

I. Affording a Fair Return and Pair Income Under Existing Economic Conditions

One of the four statutory policy objectives is "to afford the copyright owner a fair return for his or her creative work and the copyright user a fair income under existing economic conditions." 17 U.S.C. $ 801(b)(1)(B). Music Choice has recently experienced a significant deterioration of the financial condition of its residential audio service, which will certainly continue through the next rate period and necessitates rate relief for that period.

A. Music Choice's Historical Financial Performance

Music Choice was the world's first digital music service. As such, in order to begin providing our cable music channels to subscribers in 1988, we had to invent much of the technology necessary to get high-quality digital music programming to subscribers'elevisions. At the same time, we had to create a market for subscribers to pay for a radio-type service that they were used to getting for free from terrestrial radio.

19 PUBLIC VERSION

The road to financial viability upon which Music Choice has traveled- and continues to travel — is neither straight nor short. For almost 80 years,

Music Choice has struggled to launch, sustain, and grow our domestic residential services in a highly competitive and rapidly changing marketplace.

To date, our partners have invested over of capital to fuel our continuing operations. Significantly, Music Choice has dedicated this capital to deploy domestic residential services that provide invaluable promotional benefits to the record industry.

Despite our best efforts, we have not yet recouped accumulated losses from our domestic residential operations,

. In the first CARP proceeding, we submitted a proposed Gve-year budget in which we projected domestic residential operations of Music Choice would achieve over ~ in total gains between. 1996 and 2001. Unfortunately, our domestic residential operations actually incurred additional losses of ~'. during that period.

On a cumulative basis, accounting for the years of losses experienced during our history, the Music Choice residential business has still not become profitable. As of the end of 2015, Music Choice had a cumulative loss

20 0 0 PUBLI C VERSION 0 0 of 0 0 ~ Any improvement in our future cumulative loss position is dependent upon 0 significant increases in 0 0 . Further, any success in these 0 of labels 0 will inure to the benefit the record 0 0 0 Even taking Music Choice's entire, consolidated business into account, 0 and allowing the company's ancillary commercial background music business

0 line to subsidize the losses from our residential service, the 0 return on their capital investment has been partners'nnualized 0 0 0 B. The Financial Performance of Music Choice's Residential Service Has 0 Significantly Deteriorated Since SDARS II Due To Changes in the 0 MVPD Marketplace 0 At the time of SDARS II, Music Choice's residential music service had 0 reached a period of modest (if inconsistent) profitability after years of losses. Recently, however, our residential services'inancial performance has suffered a 0 significant deterioration due to the dynamics of the MVPD marketplace in which we 0 operate. Consolidation among MVPDs combined with shrinking margins in the 0 0 cable industry, along with continued competitive pressure, have led to 5+~ 0 In 0 0 particular, these dynamics have resulted in 0 21 PUBLIC VERSION

All of these fundamental changes to

Music Choice's existing economic conditions since the last proceeding, which will continue through the next rate period, necessitate a rate reduction.

When Music Choice first launched., it was sold as a premium, a la carte, service to cable subscribers and we charged a price of $2.50. After a short period. in which a relatively small number of early adopters and. highly-engaged. music lovers subscribed. to the stand.-alone service, market penetration and. revenues plateaued, and. we had. to pursue a different strategy. Once Music Choice migrated away from our unsuccessful premium service mod.el, we started. pricing our domestic residential services as a basic digital offering, to be bundled. with subscribers'asic cable packages.

Music Choice has experienced. — and continues to experience — significant pricing pressure when we renegotiate licensing fee arrangements with cable af61iates, due in part to increased consolidation in the cable industry,

. This, combined with continued competitive pressure from various

22 PUBLIC VERSION 0 market entrants seeking to undercut our pricing over the years, has resulted in reduced license fees. By 1996, Music Choice averaged only 0 customer per month for its residential audio service. By the time of the SDARS this number was down to per customer. Since 0 II proceeding, SDARS II, this downward pressure had continued. The average price for the largest MVPDs, which. make up ~ ot'our total subscriber base, is uow 0 per customer. Just three years ago, the average price for Qg~ of our subscriber base was per subscriber. That is a /gn~ reduction in just the past three years, In addition to the dynamics described. above, rate pressure has also further intensified due to recent changes in. market conditions affecting the 5DTDs. Even though they have continued to consolid.ate, increasing their. ability to negotiate lower rates from Music Choice, MVPDs'wn margins are shrinking. This is due to several factors, including increasing consolidation among media programming 0 networks and other content providers, the impact of cord-cutters and cord-nevers (who either cancel or never subscribe to cable television in the first place, relying 0 instead on broadband), and constraints on MVPDs'bility to raise consumer prices 0 due the economy and cable prices that are already very high. These shrinking 0 0 margins have led the MVPDs to take even tougher negotiating positions than in the 0 past, even with significant content providers. 0 0 It is important to remember that Music Choice is a television content provider, and in negotiations with MVPDs we compete and negotiate in a similar 0 23 0 PUBLIC VERSION manner to other television content providers. It used. to be that once a content provider was carried on a cable system, it would never be dropped.. Those days are over. %Me MVPDs always had. signi6cant leverage over new and marginal content providers seeking carriage on their systems, in the past couple of years the MVPDs, including mid-size operators whose margins are even more affected given their relative lack of bargaining leverage, have begun aggressively dropping even name- brand. networks in a way that has never before happened on a sustained. level. For example, Suddenlink and. Cable One dropped at least 24 networks and. 15 networks, respectively, Rom their services in 2014, and GCI dropped 18 networks last year.

Altice, another major MVPD, has announced. that it intends to "aggressively" cut costs in the near term, including by dropping networks. Very popular networks, such as AMC (which has the successful Walking Dead series, among other highly- viewed, programming) have been dropped. by smaller operators. Virtually every independent network has had. to take reduced. fees from the MVPDs in this period.

Even the top-shelf networks are not immune to this pressure. Although

ESPN is too popular to be dropped in its entirety, certain MVPDs such as Comcast and Cox have dropped. its channels from lower-priced "skinny" packages, so even if the peg-subscriber price is not reduced., ESPN's overall revenues are. All of these developments have occurred in the past two years. These same pressures apply, even more so, to Music Choice and. have driven a fund.amental change in the financial outlook for our residential music service, which is reflected in our 2016 financial statements and. projections, as discussed below. PUBLI C VERSION

Since its launch, Music Choice has relied almost exclusively upon licensing fee revenues from MVPDs to sustain its operations,

. As our licensing fee rates have been driven down., Music Choice attempted to develop an advertising program and other revenue streams, but those efforts have been less successful than we had hoped. I believe that our revenues, profitability and return on investment for the residential service will continue to be limited to, and constrained by, our licensing fees. We have tightened. our belts as a company during our 25 years of operation, and attempted to invest in other product lines, hoping that diversification would. help us — but again, with limited success to date. In short, in response to increasing competition and extreme downward pressure on the rates we can charge for our service, we have run a very lean operation.

As I discuss in more detail below, Music Choice has historically invested many millions of dollars on improvements to its residential service, almost all of which also benefit record. companies and artists by increasing the promotional impact of our service. These improvements and. investments have also helped develop Music Choice into a high-value service, and have enabled us to earn while maintaining high market penetration. Because

, and given the consolidation in the cable industry, we have And because of our high existing market penetration, we have no realistic way to significantly increase our market share

25 0 PUBLIC VERSION in the near term. The only way to improve our Gnancial performance within our

control is to cut costs. The only cost that can realistically be lowered. in any

0 material way without harming the quality of our service (thereby making us

less competitive and. driving our fees and margin down even faster) is the sound.

recording performance license at issue in this proceeding, which is already too 0 high. 0 0 0 any further acceleration of the decline in our license fees. 0 C. Financial Statements for Music Choice's Residential Audio Service Schedules showing Music Choice's f'inancial results from the operation of , ~ 0 its residential audio service Rom this year and. projections through the next rate period, assuming the current, erroneous 8.5% of revenue PSS rate for 2016 and 0 2017, and. a corrected. 7.5% rate for the next rate period., are submitted. as Exhibits MC 10-11. As explained. in more d.etail below, Music Choice's costs are 0 0 presented. in two ways. First, because the Jud.ges indicated in the last 0 proceeding that they view the economic performance of the audio channels 0 covered by the license alone as most relevant, in Exhibit MC 10 I have allocated. various shared. costs and. revenues to the audio and video components of the , ~ bundled. resid.ential service to create a prof'it and. loss statement showing our

i ~ current and. projected. residential service performance if we only offered. the 0 PUBLIC VERSION audio channels. Second, because Music Choice's audio and video services are always bundled together and viewed as complementary parts of a single offering, and because we do not allocate the shared costs or revenues in the ordinary course of business, I also present a similar profit and loss statement for the combined residential service, set forth in Exhibit MC 11.

As may be seen on Exhibit MC 10, due to the recent change in marketplace dynamics described above, Music Choice's residential audio business (on a stand-alone basis)

When the overall bundled residential service (including video) is taken as a whole in Exhibit MC 11, the results in 2016 and 2017 are even worse, with slightly better results starting in 2018

: On the other hand, given past experience, I have confidence in the projections for the key driver of the negative results,

As previously noted,

27 PUBLI C VERSION

1. Methodology for creating residential audio only Gnancial analysis As noted above, Music Choice creates its financial statements on a consolidated basis in the ordinary course of its business. Although the residential audio channels covered by the PSS license are our original, core business, we also bundle VOD and/or the Music Choice Play video channel with our residential music service for some affiliates. We created these video offerings as an extension of the residential audio service, largely as a way to increase the value of the audio channels to the MVPDS to stop, or at least slow, the decay in our rates. From the time it was first offered in 2004, VOD was always bundled with our audio channels with no separate fee. When we first rolled out the Music Choice Play video channel in 2010 (which was originally called SWRV), we charged a separate per-subscriber fee. Ultimately, we were unsuccessful in getting significant market penetration while charging a separate fee, and now all new deals including Music Choice Play have it bundled with no separate fee.

Particularly because the video products are bundled with no separate fee, those products add value and. contribute to the overall residential license fee revenue. In fact, one of the ways we have effectively resisted even sharper drops in our license fees was to add the video prod.ucts to the service. Our VOD offering is particularly popular. 0 0 PUBLIC VERSION

0 0 0 0 Music Choice's video products are direct licenses with the 0 not covered. by the PSS license, and. we have to negotiate 0 record companies for those prod.ucts.s 0 The Judges mad.e clear in SDABS II, however, that the financial analysis they viewed. as relevant was an analysis limited. to the residential audio 0 channels covered. by the PSS license, and not Music Choice's other business lines. To that end., I created the financial statement submitted. in Exhibit MC 10 by working with the Music Choice accounting department to go through each category of costs reported and. projected. for our entire consolidated. business and. 0 0 determining how much of those costs would. have been spent if we were only providing the residential audio channels. This analysis included. personnel costs, where I analyzed. every employee in the company and. determined. which would. still be needed if we were only providing the residential audio channels.

As noted. above, although the video products are bundled. with our audio

channels with no separate additional fee, they ad.d. value and. account for some 0

s In addition to our residential music service, Music Choice also has a separate commercial 0 background music business line. That business, like the video products, is not covered by the PSS license, but unlike the video products is covered by a different statutory license.

29

0 PUBLIC VERSION

portion of that license fee revenue. To allocate those revenues, I considered the

Music Choice Play and. VOD products separately.

With respect to Play, when we used. to charge separately for that channel,

the price we were able to get was on average approximately per

subscriber, and even now that we no longer break out that separate charge in

the contracts, we internally allocate approximately:—

5%5$emm from the resid.ential license fees to Music Choice Play in the ordinary

course of our business, so I used. that allocation for revenue from af6liates that

get Music Choice Play. Notably, when we entered into contract renewal

negotiations with affiliates who had. been paying the separate

for Play, the starting point of the rate negotiation was the total per-subscriber

fee, including the ad.ditional amount of the separate Play fee, just not broken out separately anymore.

With respect to VOD, although we never broke out a separate fee, for

many years our VOD offering was the biggest differentiator between our service

and. that of a Canadian competitor, Stingray, that has been attempting to break

into the United. States market by und.ercutting our prices. Until 2015, Stingray

did. not provide a full VOD offering. I went back and. analyzed. all of the vari~os

contract renewals prior to 2015 where Stingray was trying to take business

away from us and. found. that over all of these negotiations, we would keep the business as long as our price reflected. the additional value added. by our VOD offering.

30 PUBLIC VERSION

Specifically, I analyzed all available d.ata from 2012 through 2014 for deals involving VOD and. potential competition from Stingray to determine how much higher we were able to charge for our bundled VOD service over the lower price offered. by Stingray. The smallest MVPDs have the least negotiating leverage, anl the d.ata reQect this. The average premium for the smallest MVPDs was —:, while mid-sized MVPDs yielded. a premium of and. large MVPDs a premium of SW~. We dil not conclude contract negotiations with any "super-large" MVPDs luring this period.; however, they clearly have the most leverage and. I therefore conservatively assumed. an average premium of SV% for those MVPDs. The weighted average of all these d.ata points shows a ~&~ premium. Based upon this analysis, I allocated. )kkMK~] of our resid.ential service license fees to VOD, for those subscribers receiving VOD. This resulted. in an adjustment of ~K~l to overall residential audio license revenue, because not all of our subscribers receive VOD.

2. Methodology for creating combined. residential Gnancial analysis For comparison, I also created a similar statement analyzing the actual and. projectel financial performance of the entire residential music business, including vid.eo, similar to the analysis I presented. in SDARS II. This analysis is submitted. as Exhibit MC ll. I used. the same methodology outlined above, of analyzing each of our categories of costs and employees and: determining how much we would. spend. if-we only provid.ed. the residential music service.

31 0 0 0 PUBLIC VERSION 0 0 Revenues were simpler to allocate because we d.o track our resid.ential revenues separately from our commercial background service revenues. 0 D. Providing a Fair Return to the Copyright Owners 0 The savings from Music Choice's proposed rate, which would amount to 0 — 0 approximately per year in the next rate period as compared to a 0 retention of the 7.5% rate implicit in the SDARS lI decision 0 This will have a significant impact on our ability to will not be felt at all the That decrease certainly by record companies or recording artists. Even the entirety of the PSS royalty, especially by the time it is divided up an. distributed to all of the record companies and artists, is less than a rounding error to any of them. Moreover, in stark contrast to Music Choice, the RIAA recently reported. that the record. industry has reached the inflection point where revenues from on- demand. streaming have more than made up for. any losses from decreases in physical and digital record sales. A copy of that report is submitted as Exhibit MC 12. Also, even during the more difficult periods in the transition from CDs to 0 downloads, and then to streaming, the record companies have always managed to generate significantly higher profits than Music Choice. J Choice's Music Choice 0 Finally, Music proposed rate is higher than the rate 0 pays for the analogous PRO licenses, and is substantially higher than the 2% rate 0 demanded by the record companies prior to the advent of the statutory license.

0 32 '00 0 0 PUBLIC VERSION

0 least the rate 0 There simply is no reasonable dispute that this factor requires at 0 decrease requested by Music Choice, 0 0 II. Maximizing the Availability of Creative Works to the Public 0 Another statutory policy objective is "to maximize the availability of creative detail 0 works to the public." 17 U.S.C. $ 801(b)(1)(A). As discussed in greater by 0 Damon Williams, Music Choice's Senior Vice President of Programming Strategy 0 0 and Partnerships, our residential audio service maximizes the availability of creative works to the public in at least two ways, First, Music Choice provides 0 valuable discovery and promotional opportunities for record companies and their 0 artists, especially for artists in genres not served by terrestrial radio, which leads 0 not only to higher sales but also a greater variety of music. Second, Music Choice

0 contributes a great deal of original, creative content to its service that is provided to 0 the public, including the very curation of its channels, the on-screen display, 0 recordings of in-studio promotional performances by artists, various social media 0 content tied to the audio channels to drive subscriber engagement, and many 0 different types of other promotional materials for artists. 0 Given the changes in the financial performance of Music Choice's residential 0 0 audio service since the time of SDARS II, the implicit 7.5% rate from that decision puts Music Choice's ability to continue these activities at risk. As further discussed

connection with the relative contributions policy factor, recent 0 below in 0 developments

0 33 0 0 PUBLIC VERSION

= the artist promotional visits discussed

by Mr. Williams. artists per month come to our

studios to create live performance recordings, interviews, and. other materials to

promote their records. Many of these materials are used on our audio channels and

social media platforms to drive additional listening and sales of those recordings by

our millions of residential subscribers.

'3WÃM l%i==~RZlBIB!IK:.-»

8RKQ

PSSIRw~KiEK3 se~35NF3ssn@,"M

34 0 PUBLIC VERSION 0 0

Importantly, the artists I often are artists

, ~ that receive virtually no support from any other major national media outlet. The 0 0 record. company executives actually responsible for selling records and. promoting 0 artists are constantly begging us to bring their artists in for these promotional 0 opportunities. Unfortunately, the lawyers on the digital rights side of the 0 companies do not understand. that by driving our rights costs up they are cutting oK 0 0 their noses to spite their faces. 0 0 Another example of our ability to maximize creative works made available to 0 the public involves our original content. ~aaa:=-. 0 0 ~~&&8!If'~ This '00 original content is important not only to maintain subscriber engagement with the audio channels and promote discovery, sales, and streaming of records, this content

0 is also key to driving our MVPD aKliates to d.o promotions with Music Choice. 0 0 These joint promotions with our MVPDs greatly extend the reach and impact of the and. artists. 0 promotional opportunities we provide the record companies ~ 0 0 0

0 35 0 PUBLI C VERSION

0 III. Relative contributions 0 The third statutory objective is "to reflect the relative roles of the copyright owner and the copyright user in the product made available to the public with

respect to relative creative contribution, technological contribution, capital

investment, cost, risk, and contribution to the opening of new markets for creative expression and media for their communication." 17 U.S.C $ 801(b)(1)(C). With respect to each of these types of relative contribution, while Music Choice continues

0 to contribute more than the record companies for all but arguably one (creative

contribution), all of Music Choice's relative contributions have been put at risk due 0 to the change in the financial performance of Music Choice's residential audio 0 service.

0 Music Choice's investments in curation,

original content, technology, and developing new markets and media have been 0- crucial to differentiating Music Choice from various competition and creating the value proposition that justifies our subscription fees. 0 0

0 will eventually make our audio service less valuable, leading to more financial stress, and At the same time, the record 36 0 0 0 PUBLIC VERSION 0 companies also Almost all of these investments enhance the promotional benefits Music Choice provides to the

' record companies and artists. These benefits are far more valuable to record

' companies and artists, because they affect their primary revenue streams of sales, 0 0 downloads, and streaming, than any potential increase or decrease in the PSS rate. 0 It is in everyone's interest to provide the requested rate relief to Music Choice. 0 ,0 A. Relative creative contribution 0 As noted above in connection with the related policy objective of maximizing 0 the availability of creative works to the public, Music Choice has devoted

0 to creative related to its residential audio service, 0 substantial resources inputs including the curation of its channels and the creation of many types of original 0 content for the service. Over the course of its operation, Music Choice has invested

' 0 in. programming and original content costs for the combined residential service. 0 0 0 0 0 While the record companies certainly make significant creative contributions 0 to the recordings subject to the statutory license, they have many more significant 0 ways of generating a return on those contributions and the PSS license is ancillary, 0 if not completely insigni6cant to such return. Those recordings are not made with the PSS in mind, or for the purpose of providing Music Choice with raw materials

for its service. And while the recordings are an important creative input to our 37 0 0 PUBLI C VERSION

residential audio service, the record companies do not contribute in any way to the

various investments Music Choice makes in what makes Music Choice's service

unique and compelling: the curation and. original content described above.

0 B. Relative technological contribution 0 0 As the world's first d.igital music service, Music choice had. to invent a 0 significant amount of the technology necessary to launch its business. Music Choice created various technological components of its system for the purpose of opening new avenues for transmitting sound recordings to a larger and. more diverse audience. This technology included technology to uplink the programming signals to satellites and. transmit them through cable services„ technology to identify the name of the sound recording and. artist during the performance, and. technology for programming, encryption and. transmission of the programming containing the sound recordings. Music Choice has made numerous additional technological contributions to its service since it launched, designed to further increase the exposure of the sound. recordings to new and. larger audiences, and specifically to enhance the promotional 0 value of the service to the record labels. Examples include improvements to the on- screen displays containing promotional information about the song and artist; the

where artists visit and record the value-added 0 creation of production studios 0 promotional recordings discussed above that are featured on the Music Choice

0 service, improvements to the digital playback system to improve the to provide deeper music lists and more interesting mixes onprogrammers'lexibility more channels; the creation of data center and disaster recovery systems necessary to 38

0 PUBLIC VERSION operate a 24/7 network and the creation of and. improvements to the Music

Choice website and. apps and technology for distribution via MVPD IPTV

(internet protocol TV) platforms, and expanding television subscribers'ccess and. use of the service through development of enhanced. audio/visual navigation and. contextual audio/video. We have also continued. to improve the satellite uplink and. other technologies necessary to put our channels on satellite for distribution, and. previously developed. technology to integrate our resid.ential audio service with various social networking platforms, such as Twitter and

Facebook, as well as texting, all to promote customer engagement. This increase in subscriber engagement, in turn, increases the promotional value of the service to the record labels and, artists.

In contrast, the record. companies have not invested. in any technology specific to Music Choice's residential audio service.

Over the course of its operation, Music Choice has invested. R$&f in technology costs for the residential service.

39 PUBLIC VERSION

C. Relative capital investments

This factor is closely related. to the prior factors, because each of the

improvements listed above required. signi6cant capital investments. While the

record labels have not made any investments in equipment or technology to

facilitate Music Choice's curation, original content, or transmission of their sound

recordings to the public, Music Choice, as noted. above, has made substantial

investments over the course of its operation for equipment and technology. From inception through 2016, Music Choice has spent ~ = =- just on capital expenditures related, to product development, product enhancement and infrastructure.

It is important to note that any future capital investments

D. Relative costs and. risks

As was recognized by the original CARP and the Librarian of Congress on

appeal in the first rate proceeding, Music Choice faces significant costs and risks

associated with its service while lowering the costs and risks of the record. companies and. artists.

40 PUBLIC VERSION

1. Music Choice lowers the costs and risks of record companies and artists

As a preliminary matter, it is important to note that all of the costs incurred by the record labels in connection with their sound recordings are sunk costs. These costs are not incurred for the purpose of providing content to Music Choice, and would not change irrespective of the rate set in the proceeding, and even if Music

Choice went out of business. The Music Choice service does not increase those costs in any way. In fact, Music Choice substantially lowers the cost to the record labels for the promotion of their sound recordings. As described in more detail in the testimony of Damon. Williams, Music Choice provides various value-added promotions to the record labels and artists. The record labels frequently thank us and recognize the promotional value they receive are provided free of charge to the record labels. lorn For example, we recently did a promotion for Chris Brown, discussed in more detail in Mr. Williams's testimony, which was very successful. Chris Brown and his label chose to do that promotion with Music Choice rather than any other music service, because they realize that Music Choice provides a superior and unmatched promotional platform. In fact, they turned down at least two offers at the time—one f '; " in order to do that promotion with

Music Choice, precisely because they realized that the greater promotional value from Music Choice was worth far more than any deals offered by

The head of urban music for Chris Brown's label specifically told us that Music Choice is responsible for more urban music sales for Sony than any other

41 0 0 PUBLIC VERSION 0 0 outlet. As noted above, however, Music Choice has recently = 0 0 0 0 The promotional impact of airplay and. other promotions on Music Choice also 0 lowers the record. companies'isks. As explained. in Mr. Williams's testimony, 0 Music Choice's promotional impact not only drives increased revenue and profits, 0 that impact is particularly strong for artists in various genres that are not promoted. by terrestrial radio, and. therefore provide the most risk to the record. companies. As he also explains, Music Choice's promotional impact is now recognized throughout the music industry, as refLected. in the tracking and reporting of our plays by the 0 BDS and Mediabase services. Plays on Music Choice are closely monitored. by the record. companies and. have become very important to the promotion of new releases. 0 " 0 I SSII-"': '::: .::-'.-:—S:"-"'"-''~:. St'ISSI SS:-:~~ 0 "":.!S':.:::,::::::::, S" ':,'::::S:.".SSSS"" I:::—!SS— ~ '.SI:-: SSSS!'":-

0 I": 'SS: .SSS-":::i,,:;,:-"'SS:-,'-~ 0 lV~|NEKiK,~E.;M~~MKF3MK, -5

'5, ~ 0 Sm~mr~== 0 0 0 42 0 0

PUBLIC VERSION 0 0

, ~ 0 2. Music Choice's costs and risks are high and increasing As noted above, Music Choice— ~, and. its core residential audio service still has not reached. profitability

on. a cumulative basis. Although at the time of SDABS II, the service had achieved

a modest profit on an annual basis, various recent changed market dynamics

0 described above have fundamentally increased the risks associated with our residential audio business. At the same time, our residential audio business is our

only business line where our costs are increasing, including the rate increase

implemented in SDARS II.

E. Relative contribution to the opening of new markets and media '0,0 As noted. above, Music Choice was the very first digital music service, and e invested significant amounts in creating an entirely new way of getting music to consumers in their home. Music Choice continued to improve its service and.

innovate, extending its service's reach through its app, website, and social media.

, ~ The record. companies have not made any similar contributions to the markets we

serve. The varioua.technology, creative content, and. other investments described. 0 above have all contributed to the opening of new markets and media for consumers 0 to experience and discover music. But again, these investments are now at risk due to changed circumstances since SDARS II. '0

43 0 0 PUBLI C VERSION

IV. To Minimize Any Market Disruption The rate set in the original PSS CARP proceeding was obviously too high, in large part due to the CARP's inaccurate estimate of the services'omposition performance right licenses. This is proven by the fact that of the three original services, only Music Choice remains an active participant in this market. DMX declared bankruptcy and sold off its assets. The company that purchased those assets eventually consolidated with Muzak. Muzak's residential service is inconsequential to its overall business, and limits its participation to only one affiliate because it found it was unable to make money when competing for affiliates. Thus, even the original rate of 6.5% has been disruptive. As I described above, this disruption has only been amplified by the nature of these rate proceedings, in which Music Choice as the sole surviving active PSS must spend millions of dollars to avoid Sound Exchange's demands for massive rate increases that would put us out of business.

As explained in detail above, after finally achieving a modest profitability at the time of SDABS IZ, Music Choice now faces a sudden and significant deterioration of its financial condition due to changes in the MVPD marketplace. The record. companies remain much larger and continue to generate much more revenue than. Music Choice, putting them in a far superior J position to absorb the impact of a lower rate. The entire in royalties Music Choice has paid the record labels over the last twenty years amounts to a mere Nggn of the record industry's total retail revenue from sound recordings in 2015 alone, which was approximately $ 7 billion. The RIAA's 44 PUBLIC VERSION

2015 year-end report is submitted as Exhibit MC 13. And unlike Music Choice, the current and future outlook for the record companies appears to be improving. According the RIAA's most recent report, the recording industry has hit the inevitable inflection point where the signi6cant increases in on-demand. streaming revenue are greater than any declines in physical or download. sales, and. overall revenues are increasing. The RIAA's 2016 mid-year report is submitted. as Exhibit MC 12. In this context, it is clear that this policy objective supports Music Choice's rate proposal. Finally, setting a lower rate for Music Choice will not have, and. has not had., any precedential value with respect to the record labels'egotiations or proceedings for other licenses. For example, in the SDARS I rate decision, the

Judges noted. that respective markets for Music Choice and. Sirius XM are so fundamentally different, including with respect to demand. characteristics, that the Music Choice rate is absolutely unusable as a benchmark for the SDARS rate. As the Jud.ges noted in SDARS II, these same fund.amental differences exist when Music Choice's market is compared. to that for any other music service.

The rate reduction proposed. by Music Choice will allow Music Choice to

This, in turn, will put us in a materially better position to survive the economic diKculties outlined in my testimony. The resulting

45 PUBLIC VERSION reduction in royalty payments will be insigni6cant to the individual record. companies and artists who receive those payments.

V. Ephemeral License The ephemeral license provided. in Section 112(e) of the Copyright Act has no independent economic value to Music Choice. These copies are not sold or distributed., and. are not used for any purpose other than to facilitate our licensed. performances. I am unaware of any marketplace context in which the record labels seek, or get, a separate payment just for ephemeral copies in ad.dition to performance royalties. For example, when Music Choice obtains licenses for music vid,eos,

on the performance side, Music Choice pays the PROs a license fee for the right to perform a song, but there is no additional fee for any ephemeral copies created. and. used. to facilitate those performances. With respect to the statutory sound. recording licenses at issue, the ephemeral rate has historically been included. within the section 114 performance rate, and a certain percentage of that performance royalty is apportioned. to the section 112 license.

Consequently, the ephemeral license fee should. be inAuded within the royalty fee for the performance license, and. attributed a small percentage of that fee.

VI. Regulations The various regulations associated with the PSS license have been in place, substantively unchanged., for many years. Both Music Choice and.

SoundExchange have ample experience operating under the current PUBLI C VERSION regulations, and there have been no problems sufficient to warrant any change to those regulations in this proceeding. Conclusion

In SDAAS II, the prior panel determined that, under Music Choice's financial condition at that time, the existing 7.5% of gross revenue rate was best suited to satisfy all of the Section 801(b) policy objectives, except that Music Choice's then- planned increase from 46 to 300 channels warranted a rate increase to 8.5%. But,

Music Choice never implemented that rate increase, and therefore has been overpaying for this entire rate period. Making Music Choice whole for that overpayment, alone, warrants a reduction to 5.7% of revenue. But changed financial circumstances, driven by fundamental changes in the MVPD market since

SDABS II, provide independently sufficient justification pursuant to the policy objectives for lowering the rate to at most 5.6%. Notably, these market dynamics would impact any PSS, and any other participant in the cable radio market. While this rate decrease will not completely solve all of the financial difficulties Music

Choice will experience during the next rate period, it will provide material relief and put Music Choice in a better position to avoid

At the same time, the rate reduction will'nave no material impact at all on. the record companies and artists, who should have every interest in allowing Music Choice to continue providing its music discovery and promotion platform to its 50 million subscribers throughout the country.

47 Before the UNITED STATES COPYMGHT ROYALTY JUDGES Library of Congress Washington, D.C.

In the Matter of: Docket No. 16—CRB-0001-SR/PSSR (2018-2022) Determination of Royalty Rates and Terms for Transmission of Sound Recordings by Satellite Radio and "Preexisting" Subscription Services (SDARS IH)

DECLA.RATION OF DAVID J. DEL BECCARO contained in I, David J. Del Beccaro, declare under penalty ofperjury that the statements correct to the best ofmy my Written Direct Testimony in the above-captioned matter are true and New knowledge, information, and belief. Executed this 19th day of October 2016 in New York,

York.

& David J. Del Beccaro Written Direct Testimony of Damon Williams

00000000000000000000000000000000000000000000 0

i ~ PUBLIC VERSION i ~ Before the UNITED STATES COPYRIGHT ROYALTY JUDGES Library of Congress Washington, D.C. I ~ 0 ' In the Matter of: Determination of Royalty Rates Docket No. 16-CRB-0001-SR/PSSR 0 and Terms for Transmission of Sound. Recordings by Satellite (2018-2022) Radio and "Preexisting" Subscription Services (SDARS III)

WRITTEN DIRECT TESTIMONY OF DAMON WILLIAMS (On behalf of Music Choice)

My name is Damon Williams. I am the Senior Vice President of i Programming Strategy and. Partnerships for Music Choice. In that capacity, I 0 develop high-level programming strategies, initiatives and partnerships to 0 impact the audio and video programming at Music Choice. In addition, I formulate, oversee and implement the business strategy for Music Choice's interface with the music industry. In this capacity, and in coordination with the record labels, I create measurable initiatives that drive tune-in to Music

Choice's network. I oversee all label relationships and work directly with

record label executives, managers, independent promoters, and artists to 0 further their promotional strategies through various aspects of the Music Choice service. 0 I submit this testimony to explain the many ways in which the Music 0 Choice residential audio service increases exposure and revenues for record. 0 0 0 0 0 PUBLIC VERSION 0 companies and recording artists, including how the Music Choice service

helps drive streaming traffic and. record sales, and leads to the national distribution of more recordings. This promotional impact provides an

important bene6t to record companies and. artists, and makes Music Choice

an important part of the music industry ecosystem. I will also describe some

of the many ways in which Music Choice contributes original, creative

content to its residential music network and. in doing so creates value we 0 independent of the musical sound. recordings play. 0 My Background I joined Music Choice in 1998, spending my 6rst years with the company as the lead programmer for its urban music channels. This led. to various roles managing Programming, Production and Content Development '0 for Music Choice. My vision for compelling promotional programming was first realized. with the introduction of "Tha'orner," the first uncut rap show 0 their new releases on Music for television, which promoted. artists and Choice's audio service. In addition, I wrote and hosted a four-part audio

0 Hip-Hop" which aired on the Music Choice Hip- 0 series called. "The History of Hop and RSB channel. For a decade and a half, I have brought both up-and-

coming and. established recording artists into Music Choice's New York City

production studio for exclusive interviews and studio sessions, in partnership

with record labels. I am a &equent speaker at important music industry

events like the Worldwide Radio Summit and South by Southwest (SXSW) 0 PUBLIC VERSION and have been featured on various cable networks like CNN to speak on behalf of Music Choice about topics related to the music industry. A list of my recent public speaking engagements on music industry issues and press in which I have been quoted is submitted as Exhibit MC 14.

When I joined Music Choice, I had over Gfteen years of broadcast programming experience as a leading music programmer in the country'

broadcast industry. My accomplishments in broadcast included d.riving the

ratings of Radio One's WHYS in Washington, D.C. to the number one spot in

only six months. Under my direction, that station earned Gavin Magazine's

"Urban Radio Station of the Year" award. In 1997, while at WKYS, I was

presented. with Impact Magazine's "Programmer of the Year" award..

Before my time at WKYS, I served. as the program director of In6nity

Broadcasting's hip-hop station WPGC-AM, where I spearheaded. the

crossover of the station's format &om hip-hop to gospel, overseeing

programming in both formats. Prior to that, beginning in 1992, I was selected.

to help Willis Broad.casting launch a new mainstream urban station 92.1

KISS-FM. During my tenure at 92.1 KISS-FM, the station was considered a

key breakout station for record sales and. new artists by many prominent

urban record. labels such as Def Jam Records, Interscope Bad Boy Records

and Jive Records.

Throughout my career, both in terrestrial radio and at Music Choice, I

have built relationships with record labels and artists by helping to promote 0 PUBLIC VERSION 0 their records and "break" new artists (i.e., identify the most promising, but 0 unknown, new artists and help them find their audience and achieve their 0 first commercial breakthrough). I am familiar with the operations of Music 0 Choice and with its relationships with record labels and artists. I am also 0 familiar with various other aspects of the music industry such as the 0 promotional effect of broadcasting — including both terrestrial radio and 0 digital broadcasting — on record sales and artist development. As part of my responsibilities as Senior Vice President of Programming , generally Strategy and Partnerships for Music Choice, I also keep myself aware of the programming activities of other broadcasters, including various promotional strategies employed. by the digital music services, and of the record labels and artists. Music Choice, I serve as Chairman of the In addition to my work for in 1988, the Rhythm 5 Rhythm 8r, Blues Foundation. Since its founding receive royalties &om Blues Foundation has assisted R&B artists who did not record labels by providing them with financial support and working to preserve the history and legacy of RRB music. The Foundation has launched

several award-winning programs to educate the public and to provide work opportunities for R8cB artists. Key Board members include Jeff Harleston 0 and Bruce Resnikoff (Universal Music Group Fund), Iris Gordy (Gordy/Fuqua Fund), Del Bryant and Charlie Feldman (BMI), Terry Stewart (Former

President, The Rock and Roll Hall of Fame), Steve Pamon (COO Parkwood 0 0 0 0 0 PUBLIC VERSION

Entertainment) with notable advisory Board members like Kenny Gamble and. Bonnie Raitt. More information on the Rhythm 5 Blues Foundation can be found at: http://www.rhythm-n-blues.org/abouthistory.php. As Chairman, I

oversee all aspects of the RSB Foundation's programs for artists, including

the administration of $ 1.5 million in grant funds from the Universal/Motown

Records Fund and Gordy/Fuqua Fund.

Although Music Choice has expanded into new areas, including Video On

Demand ('VOD") and the linear Music Choice Play video channel, the residential

audio channels remain the heart and soul of what we do. It is our core service, and

is our service that all customers have and know. Nonetheless, we at Music Choice,

as well as our cable distributors and. the record labels, all recognize that our newer

video channels are complementary to our residential audio service, particularly

with respect to artist promotion — which I will discuss in more detail below.

The following testimony is based. upon my personal knowledge and

information available to me in the course of performing my dutiee as Senior

Vice President of Programming Strategy and Partnerships for Music Choice.

I. Music Choice's Creative Contributions and. Original Content Music Choice is not a mere technological "delivery platform" for the

distribution of music, like a digital download retailer or an on-demand streaming

service. The large catalog of recordings we have access to under the statutory

license is only the starting point, or raw material, for the service we provide.

Moreover, many other music services have access to the same universe of sound. PUBLIC VERSION

differentiate recordings, including services that provide music for free. In order to Music our service and provide the kind of value that justifies our subscription fees,

Choice must add significant creative content and other inputs to our residential and. audio offering. These creative inputs include the curation of our audio channels,

various elements of our on-screen display and user interface, which increase our subscribers'ngagement with the music and increase the promotional impact of the

Music Choice service.

A. Music Choice's Curation of its Music Channels Requires Creativity, Originality, Skill, and Deep Knowledge

The Music Choice residential service is a curated, non-interactive audio

service. While every non-interactive music service may have access to the same

music, this does not mean that the quality of the curation is the same. Music

Choice's audience tunes in to our service because of the daily professional music

curation provided by our staff of music programming experts. Our audience turns

on our service because they know that they will hear a cohesive playlist of music

regardless of which channel or genre they are listening to. Our customers are not

concerned with any one song or artist, they want a prolonged and compelling overall

listening experience. This is why our listeners average of listening per day, with ~ of customers reporting satisfaction with our product according to a July 2016 study by Ipsos OTX MediaCT. See id. at slide 25. That

study is submitted as Exhibit MC 15. Consumers can use a variety of other media platforms including downloads

and interactive streaming when they want to select specific songs or artists to hear. 0 0 PUBLIC VERSION While other types of music services offer an interactive/on-demand. platform for

users to hear specific songs, or playlists generated by computers — which can lead to

0 a data-dump of songs haplessly grouped. together by algorithm-powered metadata channels the 0 strings — our appeal is that we o6'er something different. To program 0 0 way that we do takes creativity, strategy, and analysis by people who have a deep 0 knowledge of the music and culture of each and every genre. We hand select each 0 song on our service to develop our content pool, and. we then develop a set of curation rules that help us refine that content pool and. play it back to the end user produces the Music Choice listening experience. in a thoughtful manner that To help create this listening experience, Music Choice invests multiple times companies like Ipsos OTX MediaCT and Nielsen, a year in audience research by which helps us to understand the profile of cable TV customers and their listening music programmers to analyze this data habits. We then use our seasoned. team of and develop programming philosophies and programming plans. In addition to relationships with those duties, our music programmers have to develop meaningful the music industry at large, including with labels, artists, manager and industry ecosystem that fuels trades associations, all of which are part of the music industry

0 record promotion and sales. in 0 The Music Choice paradigm is to curate commercial-free audio channels tastes 0 order to create unique destinations that appeal to the wide variety of musical in the marketplace. Music Choice researches the music, identifies the biggest hit 0 songs, features core artists from each respective channel's genre, and also 0 0 PUBLIC VERSION strategically promotes new releases and new, lesser-known, artists to help consumers discover new music and drive purchasing and on-demand streaming

decisions. Our programming philosophy is based in human thought and judgment,

not just algorithmic programming based on mathematical formulas, and therefore

our programming philosophy cannot be replicated by the computers or algorithms

used by many of our competitors.

Music Choice has been able to succeed because we have developed a curation

system supported by a set of programming goals that each one of our programmers

uses as guide. This system involves: consumer experience. (1) Categorizing music to better control playback and on a set of (2) Structuring the playback of the audio channels based parameters that are evaluated by our programming team on a weekly basis. These

important parameters, like "artists separation rules" that prevent too many of the

same artist song titles from playing back to back or too often, are what help make

the Music Choice experience diBerent from services that depend on technology for

their programming. Many of the computer algorithm based services tend to consider

things like a song's "sonic attributes" instead of its relevance to a channel or its

relevance in the marketplace. Music Choice depends on it programming team to

ensure that we are DMCA compliant and aligned with the criteria used by

important music industry measurement platforms like Mediabase and. BDS, which

track airplay and provide the record labels with important information they can use

to promote their records. The integrity of our audio channel playback is why these PUBLIC VERSION

as services monitor us when they do not monitor any streaming services (such

Pandora, Spotify, and. Tidal). includes a suKcient number (3) Ensuring the programming of each channel with the of hit records to provide familiarity to our listeners and keep them engaged overall experience; designed. mix of (4) Managing artist and music discovery through a carefully

those established. artists and. hits as well as newer and lesser-known artists; and. which can create (5) Minimizing excessive artist, title and sound duplication, the listening fatigue experiences prevalent on many of the algorithm-driven music

services.

Music Choice also considers four dynamics when programming its audio sound. sonic channels: (1) target audience (who we are trying to appeal to); (2) (the

design of a channel), (3) philosophical approach (for example, hit-based channels versus channels that play catalog music &om speci6c genres or decades); and (4)

playback (music mix and. the decision to emphasize new music or established. songs).

Using these factors, we then take the time to hand-curate each channel to ensure that our channels all follow a structured system of programming rules implemented.

using programming clocks and. music categories. This drives the Music Choice

experience. Materials describing our programming philosophy are submitted as

Exhibit MC 16.

Successful implementation of this curation system and. programming of philosophy requires Music Choice to use music programmers with a high level PUBLIC VERSION talent in identifying the hottest and. best new recordings before they have been Because Music discovered by the public or, in many instances, other broadcasters. are experts Choice invests substantial time and. money on programming staff who for in their genres and sub-genres and who are able to provide unique experiences

each channel, we are able to achieve competitive programming advantages.

However, our ability to invest in a robust programming staf'f has been

adversely impacted by the unfortunate fact that our revenues are sharply declining.

At the time of SOBS II, Music Choice's

Because many other music service providers rely on data technology instead

of the knowledge and. expertise of human curation, they are slow to react to the

dynamics of the music industry and pop culture. This over reliance on computer-

generated playlists and. "shu6le" play services limits consumer exposure to new

music and new artists. For example, some services will only give a new song "likes" significant exposure once it has received. enough "data points" such as or

ratings. On these data-driven services, there is no strategic placement of new music of or new artists, whereas new music and new artist placement is the key driver

10 PUBLIC VERSION

Music Choice's artist promotion programs.

These uncurated music services tend to play the same repetitive playlist sessions over and over again until a data point forces a change in the algorithm. session. This is why users hear the same song titles used over and over to begin a number of The reliance on these data points is also the reason why only a limited artists appear on these services, giving new artists and independent labels fewer opportunities. Music Choice's investment in human curation is an important one on because it helps to ensure that the consumer experience is reviewed. and re6ned a daily basis in real time, as the dynamics change.

B. Music Choice Creates Original Content for its Residential Music Channels

Music Choice has expended signi6cant resources to improve its residential

service in ways that are designed to sell records. For example, we have constantly

improved and upgraded. our on-screen display, which features key marketing

information such as the artist's name and. album title, as well as artist facts, album

artwork, artist images and more, all of which we curate. The on-screen display is

designed to draw customers'yes to the screen, where they can quickly learn what

music is available for purchase in the general marketplace. Once a consumer hears

a song on Music Choice, they have all the information they need to make a

purchasing decision or search for more information on the artist. While users are

listening to an audio channel, the artist, song, and album information is displayed is on the screen, along with images and. interesting facts related to the artist. This music part of Music Choice's ongoing commitment to giving consumers an engaging

11 PUELIC VERSION experience while simultaneously giving record. labels a powerful platform to promote their artists. Examples of what our on-screen display looks like while a song is playing are submitted as Exhibit MC 17.

Labels, artists and. managers love these promotional on-screen features because they realize that Music Choice's national audience rivals — and. in some

cases exceed.s — that of terrestrial radio. For example, our "Today's Country"

channel reaches people a month, which is a greater reach than the

very popular Country stations KPLX Dallas (1.2 million listeners monthly) and

WVSN Chicago (1.2 million listeners monthly) combined.. Our ability to reach a

national audience and. deliver an on-screen experience that helps identify an artist

and song title is especially meaningful because many consumers overlook or do not

have immediate access to this information while driving a car — the most popular

place to listen to radio. A comparison of Today's Country audio channel listenership

numbers and those of terrestrial radio stations is submitted as Exhibit MC 18. Because Music Choice has viewers a month who listen ~

airplay on Music Choice provides a tremendous opportunity for the record labels to present their artists and records to a highly engaged and targeted. base of music

fans who tune in to discover music. While Music Choice has made substantial investments in improving screen quality and functionality to engage all of these

customers, artists and record. labels continue to bene6t from such Music Choice

initiatives at no cost.

12 PUBLIC VERSION

0 Because we always tie our investments back to our core residential

audio service in some way, in 2018 we started. investing more heavily in

social media marketing to compliment the promotion of airplay on the audio

service. To do this, we hired. a team of social media experts to execute various

programs to help drive awareness of our artists and to help boost record sales

and. streams. Our strategy was to create a natural extension of our audio

service that would drive social media engagement between our audience and 0 artists and help sell records. Music Choice's investment in social media 0 marketing to enhance the promotional opportunities on the audio service, is a 0 great example of how we have adapted our business to stay relevant to the

needs of labels and artists looking for promotional opportunities.

0 For example, in September of 2015 Music Choice premiered the new 0 0 album "Got Your Six" by a popular metal band. Five Finger Death Punch on our Rock audio channel, with on-channel and. oK-channel social media

marketing support. This initiative demonstrates the interplay between the

' audio channels and our social media marketing initiatives. The record. label,

Prospect Park, provided Music Choice with a grant of right at no charge to air

0 the entire album at 8pm on September 8, 2015 because they knew that the 0 four million Rock fans who listen to that channel would tune in. i~ As part of the social media portion of this promotion, which was tied

back to the audio channels, Music Choice gave viewers a chance to win a

guitar signed. by every member of the band. Users entered for a chance to win PUBLIC VERSION

during by taking a selfie of themselves in &ont of the Music Choice TV screen the album premiere on Music Choice's Rock channel, and posted it to social media platforms such as Twitter using the hashtag Pgotyoursixselfie.

To support the promotion, Music Choice researched and selected

speci6c facts about the artist. The facts, which appeared on-screen on the

Music Choice residential service, let viewers know about the availability of

the album at retail and also encouraged the band's fans to promote the album

premiere on their own social media by standing in front of a television

playing the Music Choice audio service, taking a se16e, and posting it to their

social media accounts. A presentation summarizing the results of the Five

Finger Death Punch album premiere event is submitted as Exhibit MC 19.

Music Choice further expanded its efforts to tie airplay into our social media

marketing with "Brand New This Week," an initiative that highlights brand new

songs as they play on their respective channels and cross-promotes on our social

media accounts and sometimes the artists'ocial media accounts. This program

provides an extremely valuable promotional opportunity because it highlights the

record labels'riority releases each week, i.e., the releases that they are trying to

get onto the Mediabase charts. The record labels'romotion staff aggressively lobby

our programming team by calling them weekly to work their records and

specifically to ask us to consider their artists for Brand New This Week. When

Music Choice selects a song to be on the Brand New This Week Program, we

support that effort with a marketing program that includes promotional materials, PUBLIC VERSION custom artist facts, and more. A presentation providing more information about the so Brand. New This Week program is submitted as Exhibit MC 20. Artists are often the excited and. appreciative to hear their songs on Music Choice that they support media program on their own social media in collaboration with Music Choice's social

team. This cross-promotion with the artists on social media is one of the most

impactful parts of the program. Screen shots of artist's social media accounts

promoting their Brand. New This Week programming are submitted as Exhibit MC

21. Recently, in partnership with the labels, we have also launched another series of

programs called. the Artist Social Ambassador Program or "ASAP" program, which

gives artists customized promotions on our audio service and. is co-promoted. on

social media.

Equally important to our upgrades to the on-screen product and. our social

media marketing programs is the investment we have made in customized

programs to promote the record. labels'lanned album releases. Over the years

Music Choice has funded many original performances and. the creation of

performance recordings in our studio with the record. labels'ooperation. These

performance recordings have proven to be a valuable addition to the other

promotional services we have given participating artists, and have often been

played and. highlighted on our residential audio channels as part of special

promotional events, helping to create awareness for the artists. Music Choice also

creates marketing materials in connection with the recordings to further promote

the artists.

15 PUBLIC VERSION

One in-studio performance from 2014 that especially stands out was by the artist G-Easy, who is now one of the top Pop and Hip Hop artists in music and who recently appeared on Britney Spears's new single 'Make Me." In 2014, when G-Easy released. his first major label debut album, "These Things Happen," Music Choice A." paid to produce a live performance, which was featured. on our show "You 5 A video of that performance can be seen at httos://vimeo.corn/105391211. Music

Choice's promotional efforts helped G-Easy gain visibility by giving him exposure on both our audio channels and video-on-demand services at a critical point in his early career. Other promotional programs we have invested in over the years include:

"Take Back Your Music" (a program on our audio and. video channels featuring the up-and-coming artists — in the past, Robin Thicke, Florida Georgia Line, Luke

Bryan and. Fifth Harmony), "The F-Word" (referring to "fame" and. discussing the challenges that now famous artists like Justin Bieber and Kendrick Lamar faced on the road to stardom), "MC Icons" (which celebrates the success of artists, like Taylor

Swift, who were connected to Music Choice early in their careers, prior to becoming stars), and. "Primed" (our artist discovery program, which has featured. artists such as Sophia Carson). Materials Rom each of these programs are submitted. as Exhibit

MC 22.

C. Music Choice Creates Original Content for its Resid.ential Vid.eo Offerings In ad.dition to its audio programming, Music Choice has been able to make its

VOD service the 41 destination for music videos on TV. Our VOD service offers both

16 PUBLIC VERSION current hit videos that appeal to the "everyday" music consumer, niche music styles, and a deep catalog that appeals to a more diverse audience and includes original shows and. specialty feature programming like playlists and era- or mood-based

destinations, all of which help to round out the customer experience.

The artist promotions for which we produce original content feature elements

across both the audio and video channels, reach consumers on both, and. leverage

the promotional eKect of every platform. By combining the power of our audio and

video services, we give the record. labels more wide-reaching promotional

opportunities that have tangible bene6ts for the artists: First, artists reach a wider

audience when they are promoted. on both our audio service and our video services,

and the two services oKer two very different types of promotion — promo ad. panels

on the audio channel and targeted marketing in the form of pre-roll ads for the

artist's video and for similar artists. Examples of the PR we have done for artist

programs are submitted as Exhibit MC 28.

Second, artists who are featured on these programs get increased airplay on

the audio channels (which impacts Mediabase tracking, as described in more detail

below), and. are given feature placement on VOD (which leads to more views of the

artist's content). For example, for "Primed" programming, artists receive over

twenty spins a week — a huge number for a new artist. Or when an artist does an

"artist takeover" on a channel, we will feature that artist's songs every hour on the

hour on that channel. And when "MC Icons" recently named a pop music star an

"MC Icon," we produced related original content, such as special artist facts that

17 PUBLIC VERSION highlighted the artist's songs that had been played on our different audio channels,

and displayed those facts on our video products. II. Music Choice's Residential Music Service Promotes the Discovery and Paid Consumption of Sound Recordings

A. Music Choice Airplay Creates Awareness that Drives Record Sales and On-Demand Streaming, Which Generate the Vast Majority of Record Company Revenue

From my experience in the broadcasting industry, both at Music Choice and. at various terrestrial radio stations, I know that playing music on the Music Choice residential service drives record sales and the on-demand streaming of recorded Since music by creating an awareness of new (or newly discovered) music or artists.

I arrived at Music Choice in 1998, the promotional value of the Music Choice service

has significantly increased, and so has the record labels'nd artists'ecognition of

that value. In early 1998, Music Choice had approximately 5 million customers.

Today, that number has grown to roughly 50 million customers. This tremendous

increase in customers means that the nationwide promotion and exposure offered by

Music Choice reaches far more potential record.-buyers. At the same time Music

Choice's promotional effect is more targeted towards listeners who are inclined to

engage with the particular type of music played on any given channel.

1. Record Companies Acknowledge Music Choice's Promotional Impact One way that we know that Music Choice provides a valuable

promotional benefit to the record labels is that they constantly tell us. Today,

the record labels clearly recognize Music Choice as a key component of their

marketing and promotional strategy. This is because, as the Music Choice

18 PUBLIC VERSION service and the music industry both changed over time, the value of Music

Choice's promotional programs to the record labels and the artist community

has only increased, beginning in 1999 when I first changed the programming

philosophy at Music Choice to emphasize new music and new artists as a key

programming strategy — a change that had a high appeal to the record labels.

As traditional radio airplay has become more and more difficult to

obtain, Music Choice has been that open and welcoming platform for labels

and artist who want to reach a national audience through airplay. This

promotional effect is now widely acknowledged by the record labels, artists,

managers and the music industry as a whole. Record labels and artists e6'ect frequently give us verbal and written testimonials praising the positive that the Music Choice service has on their promotional efforts. Some

examples of testimonials we have received include the following:

A sampling of other emails thanking us — and often asking how else we

can support the label's artists — is submitted as Exhibit MC 24.

19 PUBLIC VERSION

The record labels'bjectives with these acknowledgements are important to note. First they want to thank Music Choice for the airplay. But they also want to send us the message that they need us to continue or even increase the airplay so that they can move the song in question up the

Mediabase charts, or more importantly, increase their sales or streams of recorded music. Record labels also recognize us in industry trade publications in the hopes that others in the music industry ecosystem will realize that

Music Choice is playing the labels'ecords, so other outlets should also play

them because their audience will respond favorably. This is an important

com&mation of our role in the ecosystem and of the way in which we help

drive the consumption of music.

The record labels have also sent us commemorative plaques. The

plaques are given to various key music programmers and stations around the

country as a "thank you" for helping the labels achieve milestone record sales

and. airplay. We receive these plaques because we play and. support artists

heavily throughout the lifespan of each of their projects. Not every station or

channel in the broadcast industry receives these plaques, only those outlets that the labels feel greatly contributed to the success of their projects and

increased sales. Receiving these plaques is a true acknowledgement of our

value to the labels. Examples of the inscriptions on these plaques include

recognition of Music Choice and its employees individually for:

20 PUBLIC VERSION

sales of the ~ Helping artist Armin Van Buren reach 500,000 digital

single "This is What it Feels Like". digital ~ Helping Steve Aoki, Chris Like and Tujamo reach 500,000

downloads of the single "Delirious (Boneless)" Featuring Kid Ink. on ~ Helping Def Jam Recordings reach a total of eleven 41 singles

the Dance charts in 2015.

Representative photographs and a listing of these and other plaques, which are proudly displayed at Music Choice, are submitted as Exhibit MC

25.

2. Record Companies Service Music Choice Both major and independent record labels service Music Choice to increase exposure for, and promotion of, their artists on the Music Choice channels. "Service" is a term used in the music industry for the various ways

that record labels try to get radio stations to play records, including by

free copies of the recordings to stations. Record labels and providing artists'anagement

companies service Music Choice because they know that Music

Choice provides a national platform to break new artists and increase their ability to sell records and drive streams of their music. Due the fact that the labels highly value the promotion on our channels, Music Choice also receives invitations from the labels to attend their exclusive events, where they introduce their artists'ew recordings.

This activity, however, was not always the norm. When I joined Music

Choice in 1998, the record labels and artists were just beginning to recognize 21 PUBLIC VERSION

Music Choice and digital broadcasting as a viable and. consistent outlet for record. and new artist promotion. Consequently, at that time, the labels, especially the major labels, did not actively lobby us to put their records on "service" Music Choice, and the promotional divisions of the labels did not

Music Choice nearly to the extent they do tod.ay.

One of the means by which labels service us is by providing free promotional

copies of every new single and or album, usually delivered via a password protected.

website. The labels do this because they know that Music Choice provides a

national platform to break new artists and increase their opportunity to sell more

records or drive more streams of their music. In addition to sending us new releases in the hopes that we will play

them, the top record label promotion executives (for both major and.

independent labels) proactively call us weekly to promote their artists, solicit

feedback on them and discuss promotional strategies. As an investment in

our relationship with the labels, several times a year Music Choice conducts

in-person presentations in &ont of all of the major labels'op promotional

executives to tell them about improvements in our products and. the

promotional opportunities on our network. In these meetings, we work with

the labels to target speci6c artists and programs to collaborate on.

For example, Music Choice met with executives &om to discuss including their new summarizing the artist, , in "Primed." A presentation

22 PUBLIC VERSION resulting "Primed." programming and. its cross- promotional effect across both our audio and video platforms is submitted as

Exhibit MC 26. In August of this year we gave a presentation in Nashville to music artists in , which represents some of the biggest Country

the world. The presentation, submitted. as Exhibit MC 27, walks the label

through the value of our audio channels when compared to terrestrial radio,

and explains our product improvements and the promotional opportunities

available through our ASAP program. executives then decided to collaborate with

Music Choice to promote one of the label's biggest and. most established artists, visited Music

Choice's studios in New York City to appear on our "Music Choice Now"

program, where we promoted the availability of his music throughout our

network. A screenshot of the Music Choice blog discussing the event is

submitted as Exhibit MC 28.

Music Choice is an important part of the music industry ecosystem — equal to terrestrial radio. Record companies spend the time and resources servicing us

because they realize, based on data &om proven case studies, that Music Choice

helps create marketplace awareness of their artists'rojects, and this awareness

ultimately leads to music sales or increased streaming (the new currency of the

music industry). If our platform were not promotional or valuable to the music PUBLIC VERSION

costs over ecosystem, the industry — which has continually consolidated and. reduced. overall the last five years — would. have cut Music Choice out the picture. If the e6'ect of the Music Choice service were not overwhelmingly promotional, the labels

and artists simply would not expend their resources to seek more airplay and.

promotions on our residential service.

8. BDS and Mediabase Tracking Demonstrates (and Increases) The Promotional Impact of Music Choice

Third. party tracking of Music Choice playlists also demonstrates — and

increases — the promotional impact of Music Choice. In 2001, Music Choice

secured a deal with Nielsen's Broadcast Data Systems ("BDS") to track our

playlists. BDS is one of two key organizations that monitor the playlists of terrestrial radio stations and other media outlets. In order for Music Choice

to secure BDS tracking, the record labels had to speak up on our behalf.

Mark Tindall, then a Senior Vice President and. General Manager of West

Coast Operations at BDS, reached. out to the labels and asked them whether

they wanted. Music Choice included in the BDS reports. The labels

emphatically said. yes.

After BDS began tracking Music Choice in 2001, the labels were able

to more easily see the huge promotional impact our residential audio service

provides, and. they began more aggressively seeking airplay on Music Choice

including by servicing Music Choice in the same manner they would with terrestrial radio.

24 0 0 PUBLIC VERSION 0 In 2014, Silvio Pietroluongo, Director of Charts for Billboard. Magazine 0 and who oversees Nielson BDS, solicited. Music Choice to expand its relationship with BDS by adding our Dance/EDM audio channel to the 0 channels it tracks because of the channel's importance to the Dance music industry. In 2016, BDS ad.ded two of our Latin channels - Pop Latino and 0 Tropicales - to their reporting panel, with the support of the Latin chart 0 0 manager Amaya Mendizabal. milestone in the music industry's acknowledgment 0 Another important of Music Choice's promotional impact came in 2018, when we were ad.ded. to the selection of broadcasters tracked. by the second, and. largest, airplay monitoring service, Mediabase. The primary difference between BDS and

Mediabase is that the latter is more widely used by major record. labels to

determine where to focus their promotional activities. Mediabase monitors in the U.S., Canada, and. 0 most commercial radio stations in over 150 markets Puerto Rico, 24 hours per day, 7 d.ays a week, and uses the most up-to-date

audio 6ngerprinting technology to recognize songs played. on over 1,900

stations and. channels. '~ Mediabase also publishes a panel chart, which is a highly coveted panel of the top media outlets in the countxy including terrestrial radio i~ stations and satellite radios stations like Music Choice. The airplay from the media outlets that are part of the published. panel charts is how songs are 0 how many times a song gets 0 charted and ranked and ultimately impact 0 25 0 PUBLIC VERSION played. If Music Choice plays a particular song seventy times a week on its Hit List channel, those seventy spins are added to the aggregate total spins for that song for the week. As record. labels and artist compete to have a number one song or top ten song, each and every airplay spin that Music

Choice provides starts to make a difference. Mediabase users can access

airplay charts by market, station call letter, market size, rank at station, spin

count, increase/decrease in spins (i.e., trending), dayparts, and audience

reach. Mediabase is the gold standard of airplay monitoring and stations that

are part of its published chart panels impact the labels'romotional efforts

more than stations that are just monitored.

In 2018, Mediabase began tracking seven Music Choice channels: Hit

List, Pop Hits, Today's Country, Hip Hop & R&B, R&B Soul, Rock and.

Alternative. With the vocal support of key record label executives, five of

those seven channels (Pop Hits, Hip Hop & R&B, R&B Soul, Rock and

Alternative) were immediately placed on Mediabase's published chart panels.

Recognizing the proven promotional impact of those channels, Mediabase

soon ad.ded. more Music Choice channels to the published chart panels,

including Hit List, Pop Rhythmic, Dance/EDM, Adult Alternative, Pop

Latino, Mexicana and Tropical.

Music Choice was able to secure the support of labels for this

Mediabase panel reporting because our early airplay provides support for

many hit records, ahead of terrestrial radio. Mediabase-monitored and. panel-

26 PUBLIC VERSION

factor reported airplay on Music Choice's residential audio service is a driving behind the development of an artist's story. Music Choice is typically the first

media outlet to play an artist's music nationally, which helps generate buzz

and which labels can leverage into further exposure and promotion. Through

this early airplay, Music Choice often starts the public "story" about an artist of or song that is trying to break into radio. Labels have been using evidence

airplay on Music Choice to convince terrestrial radio stations to start playing

a song from new and established artists for years. I have heard numerous terrestrial radio programmers admit that they played a song on their station

only after hearing it on Music Choice.

As data driven technology, such as airplay tracking and audio

fingerprinting, continues to evolve and influence the way record labels

promote artists, Music Choice's airplay continues to be part of the newer tool

sets being developed. Now that Music Choice has become an established part

of Mediabase's published chart panels, record label executives across the

board have made it clear that Music Choice's audio service is a very

important breakout platform for new music and new artists, and the fact that

that our weekly airplay spins make Mediabase's published panel charts

makes a huge difference in the success of the tracks we play.

Music Choice helps break new artists and promote their releases by

first giving them Mediabase and BDS monitored airplay (and later by

offering labels and artists promotional programs). When Music Choice starts

27 PUBLIC VERSION

and playing these new artists a "detection" of airplay appears on Mediabase labels. BDS, helping fuel an emerging success story that is very useful to the

The labels use this d.ata to do two things: either lobby a new media outlet for

airplay based. on Music Choice's initial airplay or lobby Music Choice to

labels noted. their increase the airplay. A sampling of emails in which artists'ediabase

reported play on Music Choice is submitted as Exhibit MC 29.

A recent example of record. label recognition of the impact of these

early plays is an email sent to Music Choice &om a

promotional executive at , about the song

. This email is submitted as Exhibit MC 80. Another recent

example shows how important Mediabase spins are even to established.

artists: In this case, Music Choice took the lead on playing the song

emailed Music Choice recognizing how the label

was able to add. to the song's airplay story because of the airplay Music

Choice started. for . This email is submitted as Exhibit MC 81.

Even in cases as such as this where the artists involved are mega-superstars,

Music Choice's airplay on Mediabase makes a difference as it influences other terrestrial radio stations to play the song.

Another recent experience demonstrates just how important our airplay and.

promotional support is to labels, as well as the signi6cance of our reporting to

Mediabase. In October of 2014, Music Choice mad.e the programming decision to

28 PUBLIC VERSION reduce the airplay of recently-released. song on our Rock channel. This had a signi6cant negative impact on Mediabase spins. After a one day decrease in the spins on our Rock channel, two Senior Vice Presidents of Rock Promotions at called Music Choice to expressed concern that the drop in spins would mean they would not be able to achieve their

objective of pushing new release to the number one spot on the

Mediabase published panel Rock charts, and that they would lose valuable

promotion through airplay on Music Choice, all of which in turn would jeopardize

the success of one of their biggest artists. After various discussions, we decided to

increase our plays of the song back to the previous levels of airplay, but this episode

clearly demonstrated the record. companies'cknowledgment of the importance of

plays on Music Choice's audio channels as a key promotional tool.

4. Music Choice's On-Screen Display Increases the Promotional Impact of Airplay on our Audio Channels

As noted. above, the current on-screen display of contextual information for

our channels is designed. to draw customers'yes to the screen where they can

quickly learn what music is available for purchase in the general marketplace. This

has a huge promotional impact due to the screen-viewing habits of our customers.

The top reason viewers look at the screen is to get information on a song or

artist. According to the July 2016 study by Ipsos OTX MediaCT, Exhibit MC 15,

of the viewers pay complete attention or mostly watch the screen during their listening session. Id. at slide 84. When asked how often they glance at the

television, of listeners said they do so rarely or never. Id. of PUBLIC VERSION customers look at the screen either once a song, when they want information on a don't know is song or artist, when a song they like is playing, or when a song they believe that Music playing. Id T.his study also shows that ~ oi our customers Choice audio channels are a great place to discover new artists and music. Id. at

slide 27. This study also shows that listener satisfaction is high. Id. at slide 25.

When customers do look at the screen, they immediately see important

contextual and promotional information in our on-screen display. With Music Choice dedicating ~ of its available advertising airtime and screen space to promoting label content like music videos and new release, this is valuable &ee

promotion for the labels.

5. Artists Visit Music Choice as Part of their Marketing Plans

Music Choice's studio is a home for independent artist and labels, who reap

the benefit of our investment in the promoting their music projects. This year, ~ of the artists that visited Music Choice's studios to promote their releases were signed to independent labels. One of the biggest independent label success

stories of 2015 was Think It's A Game Record's artist Rich Homie Quan, who

became the first independent artist since 2006 to have the most played record. at

Urban radio, as reported by Mediabase. Rich Homie Quan recognized Music Choice

as one of the first national media outlets to ever play his music, and personally

thanked us for our support.

80 PUBLIC VERSION

Examples of the recognition we received from Rich Homie Quan and of promotional and other materials related. to his achievement are submitted as

Exhibit MC 82.

As part of his appreciation for Music Choice, Rich Homie Quan participated in a promotion on our audio channels where consumers could hear exclusive new songs from his unreleased album before any terrestrial radio station in the country.

For the promotion, which was called the "Quan-a-Thon," Music Choice was granted the right to play his songs every hour on the hour. To support this promotion, Music

Choice invested in the production of original content and marketing materials at no cost to the label. A presentation summarizing the Rich Homie Quan event on Music

Choice is submitted as Exhibit MC 88.

The labels'ncreased requests to bring their artists to Music Choice studio and to do more promotions, such as album pre-sale campaigns, are a prime example of record labels'ecognition of Music Choice's value. On a weekly basis, artists and 0 0 VERSION 0 PUBLIC 0 meet with us to present representatives from the record. labels visit our offices or 0 their music projects to our music channel programmers. Artists and labels typically 0 time their engagement with Music Choice in conjunction with record releases, 0 award shows or tour dates. Music Choice sets aside resources to make these 0 engagements with artist and labels more productive and. more promotionally 0 valuable. We have increasingly become a regular promotional stop for artists when 0 they are introduce a new release or a live show. In 2002, the first year we had any artists in the studio, we had ~ artist visits. The number of artists who visit our music quickly and drastically increased after 2002, as studios to promote their 0 artists and labels recognized the huge promotional value of the Music Choice service 0 combined with the promotional visits. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 PUBLIC VERSION

0 0 0 0 0 A list of artists and label representatives who have visited Music Choice or engaged with Music Choice on location for promotional

- 2016 is submitted as Exhibit MC opportunities from January 2018 August Choice's Custom Artist Promotions 6. Examples of Music Music Choice is very committed. to aiding the discovery of new artists and promotional impact comes from all of the new music new music, and our biggest that is added. to our audio channels on a weekly basis. Music Choice adds hundreds audio service for airplay that helps start the story for an of new songs a week to its artist, which then leads radio and other media outlets to begin playing new tracks featured by Music Choice.

From the weekly pool of songs added to Music Choice's playlist, we select 0 artists to highlight in "Brand New This Week" or "Primed." In connection with our "Primed" programming in particular, Music Choice commits to weekly airplay and and their record.. also creates original content to amplify the promotion of the artist 0 Music Choice has featured both major label and independent label artists on 0 "Primed." This year, "Primed" featured eight artists across various genres, 88 ,0 0 PUBLIC VERSION representing a curated list of the best of emerging artists on Music Choice. For

example, earlier in 2016 Music Choice agreed to feature Def Jam Records'rtist,

Bibi Bourelly, on "Primed." The label was facing challenges deciding how to promote

this artist because she did not fit into any one particular genre. This promotion is

another good example of how unique Music Choice is in the music industry

ecosystem. Because of the diversity of our audio channels, we were able to play Bibi

Bourelly's music across multiple genres to help the label identify the best audience

for this artist. A case study of the Bibi Bourelly promotion on Music Choice is

submitted as Exhibit MC 85.

This year we also featured Mainland, an indie-rock band signed to the

independent record label 800 Records, on Primed. The promotion was supported by

airplay and marketing that would have been nearly impossible to get from terrestrial radio without an established hit record. A presentation summarizing the

Mainland Primed program is submitted as Exhibit MC 86.

Music Choice also promotes new releases by established artists. In 2015,

Music Choice and Sony Music's RCA Records jointly promoted the pre-release sale

of Chris Brown's most recent album, "Royalty." This promotion is a prime example

of the ways in which Music Choice promotes new music: The campaign, entitled

"Chris Brown Week," called for Music Choice to totally rebrand one of its most

popular audio channels, Hip Hop 8z RRB, as demonstrated in this screen shot:

84 PUBLIC VERSION

i~ ji &i i." i ~ ~ iLi i i'l.

s~ II ~ Pa ll ~ i i'0 &I» &] lYi II p p~vi.e ai I

II &0 tl ~

~ I

I(,ll 4

HIP-HOP Sc R&B: CHRIS BROWN "PREVIEW" 17 IIivs "i&li ~ 9 i,'Iailss ~i ~ ~ i f Ig 4lf P' iT: 9 iil

As part of the promotion, RCA Records granted Music Choice the right to air one unreleased. song a day every hour on the hour from "Royalty" to help drive pre- orders of the album. RCA Records provided Music Choice with a customized i-Tunes purchase link to track pre-ord.er sales.

For this promotion, which came at a crucial time in Chris Brown's career as he had recently fallen Rom grace in the eyes of the public due to highly-publicized interpersonal and. legal scand.als, Music Choice invested in signi6cant resources in customizing the on-screen experience for the promotion, rebranding the channel with the title "Chris Brown Week" and producing original content and. customized marketing promos. Music Choice promoted the album on the Music Channels, and. the album debuted at 41 on the RSB/Hip Hop albums chart, 42 on both "The Top" album sales and digital album sales charts, and 48 on the Billboard. 200 chart. 0 0 PUBLIC VERSION

via 0 During the course of Chris Brown Week, ~ preorder albums were sold 0 the Music Choice custom i-Tunes URL. Mediabase spins of Chris Brown tracks 0 increased ~ the week of the promotion versus the prior week. Chris Brown himself was so appreciative of the program that he allowed Music Choice to film 0 Choice's content for the promotion at his private residence in Los Angeles (at Music additionally supported this ASAP program with a complete social 0 expense). We media marketing campaign, which included the artist talking to his fans about how valuable Music Choice's promotion is. Materials related to the Chris Brown Week 37. promotion are submitted as Exhibit MC 7. Music Choice's audio channels provide unique promotional opportunities on a television platform, which has a different and greater impact than other platforms. Music Choice offers a very different promotional opportunity than offered by

because our service is primarily delivered through TV. For other media outlets newer artists, many of whom don't have a big enough hit 0 many artists, especially record to be on a television award show or TV programs like Jimmy Fallon that feature music, Music Choice provides their 6rst exposure on television. When consumers hear an artist's music on Music Choice and see the supporting content

like the artists images and artist facts, the effect is like a billboard or a marquee 0 that puts the artist's name in lights. Music Choice is helping to build that artist' 0 is broadcast into fifty million cable homes across the entire 0 brand as their image country. Unlike terrestrial radio or satellite radio, which is primarily used in the 0 an artist valuable TV impressions that translate 0 car, our on-screen product gives into awareness, record sales, and on-demand streams of that artist's recorded 0 36 0 0 0 PUBLIC VERSION

music.

Music Choice recently executed a promotion with a new independent

recording artists who happened. to be a YouTube star, Bugatti Beez. Bugatti Beez has generated over 101,861,870 views on YouTube and. has an established. fan base

of over 1,186,900 subscribers to his YouTube channel. Bugatti Beez's record. label

saw Music Choice as the perfect platform to launch his career as a recording artist.

The case study we made in connection with our promotion of Bugatti Beez's music 0 documented a trend that many of the record labels already recognize: his established. fan base viewed. the promotion on TV from Music Choice as evidence

0 that he is fLnally making it." This case study is submitted as Exhibit MC 88. III. Music Choice's Promotional Impact Leads to the Creation of More Recordings of Greater Diversity 0 All of the promotional effects discussed. above impact both records sales and

, ~ on-demand. streams on platforms like iTunes and. Spotify, giving the record labels

more revenue to invest in their artist and in future recordings. When Music Choice , ~ ' plays a song with the contextual content such as artist, title, artist images and artist facts, it ed.ucates the viewer about the artist and. gives them the information

they need. in order to search for more content on that artist. Typically when

someone searches the web for an artist they are directed to a webpage that directly

links to the artist's content on revenue-driving platforms such as iTunes and

Spotify.

Our audio channels especially benefit new artists and artists in genres that

are not served by terrestrial radio. As previously discussed, many of our audio 0 PUBLIC VERSION channels play music in genres and sub-genres that are not featured by terrestrial radio. Our promotional effect is targeted towards listeners who are inclined to The fact that engage with the particular type of music played on a given channel.

we have so many channels across so many genres allows us to broadcast many

different formats and sub-formats, including many formats that are often ignored

on terrestrial radio such as heavy metal, jazz, gospel, roots, dance music, easy

listening, blues, classical and certain kinds of hip-hop. Although these formats may

get some airplay on local college radio, Music Choice provides a far more powerful

promotional push because it is a national platform available to many millions of

listeners. This means that, by specifically targeting each genre's listener-base,

Music Choice can help sell records and drive on-demand streaming traffic for many

artists who are not actively promoted by the record labels or played on terrestrial

radio. Without the national promotions and resulting sales Music Choice drives,

some artists in these under-served genres would likely lose their recording contracts

and cease releasing records.

Beyond playing formats and less popular music that are otherwise not

heard, Music Choice also plays deeper cuts from an album. This has always

been part of our programming philosophy. Unlike terrestrial radio, Music

Choice plays more than just the top songs from an album. This is even more important today as the record labels have moved to a digital "singles" (e.g.,

iTunes) purchasing model versus the "album" model of the past. PUBLIC VERSION

Whereas, in the past, a consumer needed to purchase the entire album regardless of whether they wanted the deeper cut singles, in a singles market, these songs would never be individually purchased if outlets like

Music Choice did not play them and thereby give them exposure. These deep cut plays add sales value to the record labels.

The single biggest driver of record sales is exposure through airplay.

Consumers do not usually buy records that they have not heard before. It is widely

recognized throughout the music industry that terrestrial radio remains the single media biggest method for promoting music and artists. Neilsen consistently notes in

and music industry specific studies that radio is the number one source of music

discovery for consumers. In 2015, Neilsen found that 61% of consumers use radio as

their primary source of music discovery. See the Neilsen Music 860 Report 2015

Highlights, submitted as Exhibit MC 89, at p. 4. According to the Infinite Dial 2014 75% of study by Edison Research and Triton Digital, submitted as Exhibit MC 40,

consumers who believe it is very important or somewhat important to keep up to

date with music say radio is a good source for that, whereas 88% say music

channels on TV are a good source. Id. at p. 58. What helps radio is that all stations

across all formats essential play the same new music at the same time which

impacts consumers perception of what's available in the marketplace.

Although terrestrial radio remains the primary way that consumers discover

new music, over the past several decades consolidation in the industry has made it harder and harder for the record labels to promote new artists through terrestrial PUBLIC VERSION

was radio. The latest phase of this consolidation began in the late 1990s, while I still working in terrestrial radio. In the early- to mid- 1990s, you would typically formats. have two or three independent stations in larger markets playing similar

As a result, more formats were economically viable. Each station had a clear personality and. freedom to play whichever records it wanted.. Beginning in 1996, when Congress relaxed cross-ownership rules for media companies, radio consolidators like Clear Channel began to buy radio properties around the nation. went to a "cluster" By the late 1990s, the few companies owning most of the stations strategy, where each company focused. on a particular format or musical genre to

"own" in a given market. The company would. then focus its resources on that

format and. eliminate the competition for the format in that market. Unpro6table

formats were eliminated and. radio station personalities disappeared. Today, only a

few formats are economically viable.

After consolidation, the companies set up a system where programming was

taken out of the hands of the individual stations. Instead, one program director

would dictate the playlists for many of the stations in a given format. The

consolidators were able to improve the cost structure of radio stations this way.

However, it severely restricted. the number of new records that could be added to

the playlist each week and. consequently lowered the number of new records that

could be pushed by the labels.

This contrasts starkly with Music Choice's approach, which is to invest in

increasing programming staff who are tasked with creating individualized

40 PUBLIC VERSION

number programming for each channel. Where a typical market has only a limited 50 channels of formats featured on terrestrial radio, Music Choice presently offers cable on the television, with an additional 25 channels available to authenticated

subscribers through the ancillary app and website platforms. If you listen to terrestrial radio anywhere in the country for a particular format, you will basically

hear the same several songs played over and over again. This tightening of the

playlists also seriously limited the number of new artists that the record labels can

break on terrestrial radio.

Music Choice, by giving under-served artists airplay, also helps support and

promote artists whose careers might not be sustainable otherwise: Record labels

sign a large number of new artists every year to recording contracts. Getting signed.,

however, does not mean that an artist will be successful. Especially given the

consolidation of the radio industry, the record labels can only push a small fraction

of those bands on terrestrial radio. It is not surprising, then, that the vast majority to of records produced by the labels are not profitable. Instead, the labels have to try

make up for those losses with sales from the few records that they do push. Artists in under-served. genres are most at risk for losing record contracts,

but promotion through Music Choice helps keep these artists viable. For the artists

that are not actively pushed by the labels and do not get airplay on terrestrial radio,

the outlook is not good. If an artist's first record does not sell enough copies, the

artist will often be dropped by the record label and will not be able to make any

more records. Music Choice provided, and still provides, a solution to this serious

41 PUBLIC VERSION problem for artists and record labels. Without Music Choice, many artists and genres of music would not get the promotion they are getting today. We have recently seen a number of established artists lose their recording contracts and turn to independent labels where promotional dollars are limited.. In addition, many

artists in non-pop music formats would simply have no outlet to reach a national

audience at one time without Music Choice. Emails &om labels and. artists asking

for Music Choice's support in the form of airplay, studio visits or promotional

programs are submitted as Exhibit MC 41.

Because, as discussed. above, Music Choice promotes record sales in many

formats that are not covered by terrestrial radio, those sales allow the labels to

recoup their investments, help keep those artists from losing their recording

contracts and allow those artists to keep making records. Moreover, even if those

sales came at the expense of sales of recordings by other, superstar, artists pushed

by the labels on terrestrial radio (and I have not seen any evidence of such sales substitution), the net result would still be more new music being created., released

and enjoyed because the superstars would still sell enough to keep their recording

deals.

Music Choice not only supports major labels, but also extensively supports

independent labels like Think It's A Game Records and. independent distributers

like E-One, InGrooves and Empire Distribution. These independent companies have

become more important to many artists that have been displaced by the major

labels due to the evolving music business. These independents turn to Music Choice

42 PUBLIC VERSION for national airplay, promotional support, and a chance to have their songs listed on

Mediabase and BDS charts through Music Choice's channel playlist. Because terrestrial radio is diKcult to penetrate for those artists not on a major label, many of these artist would not have a national platform to promote their music without

Music Choice. Music Choice also invests in the independent artist community by

giving them access to our programming team and the opportunity to compete with

the major labels for airplay and. promotion. Regardless of the label, major or indie,

Music Choice invests the same resources into promoting artists on our network. Before the UNITED STATES COPYRIGHT ROYALTY JUDGES Library of Congress Washington, D.C.

In the Matter of: Docket No. 16-CRB-0001-SR/PSSR {201S-2022) Determination of Royalty Rates and Terms for Transmission of Sound Recordings by Satellite Radio and "Preexisting" Subseri tion Services SDARS II DECLARATION OF DAMON WILLIAMS contained in my I, Damon Williams, declare under penalty of perjury that the statements

Written Direct Testimony in the above-captioned matter are true and correct to the best of my New knowledge, information, and belief. Executed this 19th day of October 2016 in New York,

York.

Damon Williams %'ritten Direct Testimony of Cregory S Crawford PhD

00000000000000000000000000000000000000000000 PUBLIC VERSION

UNITED STATES COPYRIGHT ROYALTY JUDGES THE LIBRARY OF CONGRESS

) In re ) ) Determination of Royalty Rates and Terms for Transmission ) Docket No. 16-CRB-0001 SR/PSSR of Sound Recordings by Satellite Radio and "Preexisting" ) (2018-2022) Subscription Services ) ) (SDARS III) ) )

WRITTEN DIRECT TESTIMONY OF GREGORY S. CRAWFORD, PhD (Ou behalf of Music Choice)

October 19, 2016 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

Table of contents

.1 I. Introduction .1 I.A. Summary of qualifications and experience .. 2 I.B. Scope of charge and summary conclusion. .6 II. Overview of Music Choice's residential services business.. .6 II.A. Industry overview. 9 II.B. Music Choice's residential audio music service. .11 III. The hypothetical market and the lack of suitable benchmark markets. .11 III.A. The hypothetical market ill.A.1. Baseline Principles. 11 14 III.B. The ideal benchmark market. III.B.1. The use of benchmarks. 14 III.C. The lack of suitable benchmark markets for sound recording performance rights for PSSs ...... 16 16 III.C.1. Introduction .

IV. An economic model of the royalties that would arise in the hypothetical market for sound recording performance rights ...... 19 IV.A. Overview and summary results. 19 IV.B. Model Assumptions .. 21 IV.B.1. The hypothetical market. 21 IV.B.2. A non-cooperative bargaining model captures well markets with bilateral market power ...... 21 IV.B.3. A model with a single record label and PSS provider captures the essential insights that would arise in a model with multiple record labels and/or PSS providers 23 IV.C. A Non-cooperative bargaining model of sound recording performance rights for PSSs ...... 24 IV.C.1. Nash bargains: Surplus, threat points and bargaining power. 24 IV.C.2. Applying the bargaining model to sound recording performance rights for PSSs ...... 27

V. Quantifying the royalty that would arise in the hypothetical market for sound recording performance rights for PSSs.. 31 V.A. Summary and Conclusions ...... 31 V.B. Predicting Music Choice's future performance.. 33 V.C. Estimating Costs and Revenues for Music Choice's Residential Audio Service 35 V.C.1. Music Choice's Consolidated Financial Statements necessitate using Cost and Revenue Allocation Methods to determine the financial performance of its Residential Audio Service ...... 35 36 V.C.2. Residential Audio Music Service Costs . V.C.3. Residential Audio Music Service Revenue. .38 V.D. Quantifying the royalty that would arise in the hypothetical market for PSS sound recording performance rights. 42 V.D.1. Calculating Joint Agreement Profits. 42 V.E. Calculating Threat Points and Bargaining Power 49 V.F. Estimating a hypothetical marketplace royalty. 50 52 Vl. Evaluation of the statutory factors VI.A. Overview ...... 52 VI.B. Maximizing the availability of creative works to the public...... 53 VI.C. To afford the copyright owner a fair return for his creative work and the copyright user a fair income under existing economic conditions .. 58 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

VI.D. To reflect the relative roles of the copyright owner and the copyright user .60 Vl.E. To minimize any disruptive impact on the structure of the industries...... 61

.61 Vll. Conclusion.

Appendix A. A non-cooperative bargaining framework .63

.71 Appendix B. Exhibits ......

Appendix C. Materials relied upon. .76

Appendix D. Curriculum Vitae .80 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

I ~ Introduction

I.A. Summary of qualifications and experience

of Zurich in (1) I am Gregory S. Crawford, Professor ofApplied Microeconomics at the University Switzerland. I received a PhD in economics from Stanford University in 1998. I was an assistant professor at Duke University, an assistant and later associate professor at the University of Arizona, and Professor of Economics at the University of Warwick in the United Kingdom. In 2007-08, I served as Chief Economist at the Federal Communications Commission (FCC), an independent federal regulatory agency charged with regulating a number of media and communications industries, including the broadcast and cable television industries. I reported directly to the Chairman ofthe FCC and advised him and his staff on a number oftopics in these industries, including mergers, spectrum auction design, media ownership, network neutrality, and bundling. After my service at the FCC, I joined the Department ofEconomics at the University of Warwick as a full professor and, in 2013, moved to the University ofZurich as a (chaired) Professor ofApplied Microeconomics. I am Director of Graduate Studies for the economics department. In 2011, I was invited to be a research fellow at the Centre for Economic Policy Research ("CEPR"), one of the leading European research networks in economics. In 2014, I was asked to be one ofthe (two) co-Program Directors for the CEPR's Industrial Organization Programme.

(2) I conduct research on topics in both industrial organization and law and economics. Much of my research has analyzed the cable and satellite television industries. I have published extensively at the intersection of these fields, evaluating conditions of demand and supply within the cable television industry and the consequences of regulation on economic outcomes in cable markets.'hen the National Bureau of Economic Research (NBER) commissioned a volume analyzing the consequences of economic regulation across a number of American industries, I was asked to write the chapter on cable television.' was also asked to write a chapter for the Handbook ofMedia Economics on the economics of television and online video markets.' have published numerous academic articles in

Gregory S. Crawford, "The Impact of the 1992 Cable Act on Household Demand and Welfare," RAND Journal of Economics 31, no. 3 (2000): 422-49; Gregory S. Crawford and Matthew Shum, "Monopoly Quality Degradation and Regulation in Cable Television," Journal ofLaw and Economics 50, no. 1 (Feb. 2007): 181-209; Gregory S. Crawford and Ioseph Cullen, "Bundling, Product Choice, and Efficiency: Should Cable Television Networks Be Offered A La Carte7," Information Economics and Policy 19, no. 3-4 (Oct. 2007): 379—404; Gregory S. Crawford and Ali Yurukoglu, 2012. "The Welfare Effects ofBundling in Multichannel Television Markets," American Economic Review, 102(2): 643- 85. The NBER is a private, nonprofit research organization dedicated to studying the science and empirics of economics. It is the largest economics research organization in the United States. The chapter is titled, "Cable Regulation in the Satellite Era," Chapter 5 in Rose, N., ed., "Economic Regulation and Its Reform: What Have We Learned?" forthcoming, University of Chicago Press. Media Gregory S. Crawford, "The Economics of Television and Online Video Markets", Chapter 7 in Handboolc of Economics, Volume 1, 2015, Pages 267—339. PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

such outlets as the American Economic Review, Econometrica, the RAND Journal ofEconomics, and the Journal ofLaw and Economics.

("CRB"), as a rebuttal witness (3) I have testified twice previously before the Copyright Royalty Board for the Commercial Television Claimants in the matter ofthe distribution of copyright royalties for the distant importation of broadcast television signals in 2004 and 2005 and as a direct and rebuttal witness in the predecessor to this proceeding." My curriculum vitae is submitted as Appendix D.

hour. (4) I am being compensated for my time on this matter at a rate of $700 per

I.B. Scope of charge and summary conclusion

Choice's residential music (5) Counsel for Music Choice has asked me to evaluate the market for Music service and to recommend a range ofroyalty rates for the license of digital performance rights in sound recordings for Music Choice's residential music services pursuant to 17 U.S.C. 114(f) ofthe Copyright Act (the "Act"). It is my understanding that Music Choice is considered a pre-existing subscription service ("PSS") under the Act.

for pre- (6) For convenience, I will refer to the licensing of digital performance rights in sound recordings existing subscription services simply as the "licensing ofPSS sound recording performance rights". If ever I refer to the licensing of other rights (e.g. musical works rights) and/or the licensing of any rights, including sound recording rights, in other markets (e.g. the market for Interactive Webcasting), I will be explicit about the rights and market(s) at issue.

the (7) Because Music Choice is a PSS, the range ofroyalty rates I recommend are intended to consider following statutory objectives set forth in Section 801(b)(1) ofthe Act:

1. To maximize the availability of creative works to the public;

2. To afford the copyright owner a fair return for his creative work and the copyright user a fair income under existing economic conditions;

3. To reflect the relative roles ofthe copyright owner and the copyright user in the product made available to the public with respect to relative creative contribution, technological contribution, capital investment, cost, risk, and contribution to the opening ofnew markets for creative expression and media for their communication; and

4 See In the Matter ofDetermination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio Services, Docket No. 2011-1 CRB PSS/Satellite II. PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

4. To minimize any disruptive impact on the structure ofthe industries involved and on generally prevailing industry practices.'8)

Accordingly, the methodology I present in my testimony not only approximates the royalties that I believe would arise in a hypothetical marketplace transaction between individual record labels and a PSS provider like Music Choice, it also considers adjustments to that royalty that could be made to reflect these additional statutory factors.

I describe (9) I evaluate these questions in a four-step process. First, in Section II ofmy report, institutional features ofthe music industry, the market for cable radio services in general, and for Music Choice in particular.

(10) Second, in Section III ofmy report, I introduce the hypothetical market for PSS sound recording performance rights, discuss the use of possible benchmarks for this market, and analyze what would constitute an ideal benchmark. I next consider alternative benchmark markets and flnd that there are no suitable benchmarks among markets for other uses of sound recording performance rights. I close this step with a finding that, while the market for PSS musical works performance rights would be the best benchmark for a hypothetical market for PSS sound recording performance rights, ifthe Judges conclude, as they did in the last proceeding, that the market for PSS musical works performance rights is inappropriate, then I must conclude that there are no suitable benchmark markets.

17 U.S.C. 801(b)(1) Copyright Right Judges, functions htto://www.couvriaht.aov/title17/92chao8.htmL PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

a model of what (11) Due to the lack of a suitable benchmark, in my third step I build and quantify royalties would arise in such a hypothetical market. This model has two parts: an economic f'ramework (described in Section IV of my report) and an empirical quantification ofthe necessary inputs in this economic f'ramework (described in Section V of my report). The economic framework is one of negotiated bargains between individual record labels and Music Choice. It is a clarified version ofthe economic framework I presented in my testimony on behalf of Music Choice in the last such proceeding before the Copyright Royalty Board. The empirical quantification relies on Music Choice's financial statements adapted to measure the costs and revenues of its residential audio service only, and is an improved version of the quantification methodology I presented in my last testimony.

between (12) The economic framework introduced in Section IV is based on non-cooperative bargaining firms. This frainework maps well to the economic environment governing negotiations between copyright owners and copyright users both in general and with regard to negotiations over perfonnance rights for sound recordings. There are three key factors in this framework that determine the royalties that would likely arise as a consequence of arm'-length negotiations between Music Choice and the owners of sound recording performance rights (i.e. record labels). These three factors are the "Joint Agreement Profits", "Threat Points", and "Bargaining Power." I describe these factors in general and how they would apply in the market for sound recording performance rights for PSSs.

(13) In Section V, I describe in detail the quantification ofthese three factors in the case ofthe hypothetical market for PSS sound recording performance rights for the period 2018—2022. The primary input into this quantification were realized and forecast financial results from Music Choice's ongoing operations.

arise in a (14) I faced two complications using this information to quantify the royalties that would hypothetical market for PSS sound recording performance rights. The first related to predicting Music Choice's future financial performance. In an ideal world, I would have had perfect foresight about Music Choice's future financial results and could have used them as the primary input into my analysis. But there is uncertainty regarding any estimates of future performance. I have therefore used a blend of current financial results (historical parts of 2016) and forecast results for 2017 and 2018 about which Music Choice has reason to be very confident as a proxy for Music Choice's financial performance in the period 2018—2022.

— a (15) I faced a second complication as well. Because Music Choice offers three separate products residential audio service, a residential video service, and a commercial audio service — and the hypothetical market in question is for the residential audio service only, one of my primary tasks was to ensure the revenue and cost data I used as inputs in my calculations reflected only a residential audio service. This posed challenges on both the revenue and cost sides. On the revenue side, Music Choice often charges a single per-subscriber fee for both residential audio and (some) residential 0 0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD 0

video services, so this revenue had to be decomposed into that accruing to its component products. I 0 did so using well-established principles from the Financial Accounting literature. On the cost side, 0 Music Choice relies on many ofthe same facilities and staffto provide both residential audio, 0 residential video, and commercial audio services, so these costs had to also be decomposed into that accruing to their components. I used best-practice techniques from the Financial Accounting literature describe what information I needed from Music Choice. Senior financial officers from Music Choice (including their CEO) went through their cost line-items and, for each one, determined what they would be ifthey were to offer a residential audio service on a stand-alone basis. I looked closely at 0 each of their answers and, after several iterations, was confident that the costs they reported were indeed only those necessary to provide their residential audio service on a stand-alone basis.

0 elements in the (16) I then used this "residential-only" cost and revenue information to quantify the three noncooperative bargaining framework: the Joint Agreement Profits, Threat Points, and Bargaining Powers. Given these elements, the bargaining framework provided a recommended royalty equal to 0 Music Choice's Threat Point plus its Bargaining Power times the Joint Agreement Profits. I 0 converted this royalty into an estimate of the royalty rate that would arise in the hypothetical market for PSS sound recording performance rights by dividing by Music Choice's (unadjusted) audio 0 service revenue. market for PSS (17) Based on this analysis, I conclude that the royalty that would arise in the hypothetical 0 sound recording performance rights is 3.5% of residential audio service revenues and under no circumstances would be higher than 5.6% of residential audio service revenues. This is less than the 0 current PSS royalty rate and significantly less than what I understand is SoundExchange's proposed 0 royalty in this proceeding. is that I believe (18) Furthermore, I believe these conclusions to be very conservative. The primary reason 0 that there is an enormity of qualitative evidence that there is a promotional benefit to record labels Rom airplay on Music Choice. Quantifying these benefits is a difficult empirical challen e and I was 0 unable to do so given the data at my disposal, but I do show that it takes only [[ 0 j Music Choice listeners to download a single trackper month for the promotional benefit to record labels to be equaL to the total royalties it currently earnsfrom the PSS market. 0 Furthermore, if Music Choice and record labels had equal bargaining power and the rate of Music 0 Choice listener downloads was twice that, it would be sufficient to predict a zero royalty in the 0 hypothetical market for PSS sound recording performance rights. made several other 0 (19) But this is only one reason I feel my conclusions are conservative. I have also 0 assumptions in the course of my analysis that favor the record labels when calculating this 0 hypothetical marketplace royaltl rate. First, I assume that Music Choice's residential audio revenue 0 should only be reduced by [[Q]] to account for their Video-on-Demand service (which has always been bundled with their residential audio service), rather than the overall [[~]] adjustment to its

0 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

residential audio license fees that an internal Music Choice analysis concluded was the appropriate percentage. Second, when forecasting the future economic performance of their residential audio service, I include 2016, a year that reflected revenues both from accounts that have since been lost and from accounts for whom their per-subscriber affiliate fees have since fallen, rather than rely only on their forecast financial returns in 2017 and 2018 to construct my best estimate oftheir 2018 revenues. Finally, I have assumed that their residential audio revenues will increase at the current rate of inflation, rather than allow for the possibility of further account losses or affiliate fee declines. Accounting for any of these would imply a lower (possibly negative) royalty rate for a hypothetical PSS sound recording perfonnance royalty.

considered each (20) The 3.5% rate I conclude in this report is a hypothetical marketplace rate. I further what impact ofthe statutory objectives given by Section 801(b) ofthe Copyright Act and considered each might have on determining a "reasonable" royalty. I find that there are many reasons under each to also lower the rate (relative to this hypothetical marketplace rate). I therefore conclude that the 5.6% rate is an absolute upper bound on what would constitute a reasonable royalty rate for PSS sound recording performance rights.

II. Overview of Music Choice's residential services business

II.A. Industry overview

(21) In this section, I briefly describe how music is produced and sold to highlight elements ofthe process that are important for my subsequent analysis of digital performance rights for sound recordings.

(22) "Making music" involves a series of steps, including the creation, promotion, licensing, and distribution ofthe music. Music creation involves composing, recording, mixing, engineering, and manufacturing recorded music products. The cost of recording music of a high technical quality has decreased substantially and digital distribution has eliminated upfront manufacturing and inventory costs as a "home recording studio can be created for the cost of one day's recording at a high-end studio.""

(23) Driven by changes in technology and how consumers access music, music licensing and distribution has changed substantially in the last decade. Historically, the music industry received the bulk of its

Joel Waldfogel, 2012. "Copyright Protection, Technological Change, and the Quality ofNew Products: Evidence fiom Recorded Music since Napster," Journal ofLaw and Economics, University of Chicago Press, vol. 55(4), pages 715—740 at 735. Ex-TuneCore CEO Jeff Price: The End of the New Music Industry Transformation. htt://www.h ehot.corn/h ebot/2012/09/the-end-of-the-new-music-indus -nansformation-how-technolo destro ed-the-traditional-music-indu html PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

revenue via the distribution ofphysical media (records, cassettes, CDs, etc.). As recently as 2005, 91 percent of record industry revenues were obtained via sales ofphysical media. In the recent past, the industry has shifted &om physical sales to distributing digital recordings via online platforms, both for permanent ownership (digital downloads) and access on a real-time basis (on-demand music streaming). By 2015, digital revenues were almost 70% ofrecord industry revenue. Indeed 2015 music industry revenue showed three important pillars: music streaming at 34.3% oftotal industry revenue, digital downloads at 34.0%, and physical sales at 28.8%." The trend continues to point towards increased revenues &om streaming, which comprised 47% ofU.S. recorded music revenues in the first halfof 2016." Similarly abroad, where the global music industry is expected to increase

revenue in 2016 (and the next three years) driven by paid streaming."

it on terrestrial

(24) Traditionally, record companies made consumers aware ofnew music by promoting radio." However, the record labels produced more music than radio stations could air, encouraging was them to pay radio stations to promote their music (so-called "payola"). Even though this practice outlawed in 1960s, record companies have paid fines as recently as in 2005 and 2006 to settle chargesair.'oday, that record-company executives paid radio stations and their employees to get their music on the music consumers discover new music via a variety ofmedia, including AM/FM radio, YouTube, Pandora, Facebook, music television channels, and cable radio services like Music Choice, among others." For instance, a July 2016 study by Ipsos OTX MediaCT found that 16]]

Furthermore, exposure is a necessary ini rut into music discovery and Music Choice's Video on Demand service is very popular. For [[ [], Music Choice VOD has had [I l]." Indeed, in in the

RIAA 2007 Year-end statistics. httn://www.musicaememoria.corn/docs/riaa vear end 2007.ndf. News and Notes on 2015 RIAA Shipment and Revenue Statistics. htto://www.riaa.corn/wn- content/unloads/2016/03/RIAA-2015-Year-End-shinments-memo.ndf iu Ibid. RIAA, News and Notes on 2016 Mid-Year RIAA Music Shipments and Revenue Statistics. Figure 5: U.S. Recorded Music 1H 2016. httn://www.riaa.corn/wn-content/unloads/2016/09/RIAA Midvear 2016Final.ndf Credit Suisse, "Global Music," Apr. 4, 2016 at 1—2. SoundX 000034949-950. Joel Waldfogel, 2012. "Copyright Protection, Technological Change, and the Quality ofNew Products: Evidence fiom Recorded Music since Napster," Journal ofLaw and Economics, University of Chicago Press, vol. 55(4), pages 715— 740. '4 In fact, as late as 2005 and 2006, Sony BMG, Warner Music Group and Universal Music Group settled with the New York Attorney General by paying fines ($10 million, $5 million and $12 million, respectively) to settle charges that they bribed radio station employees to get their music on the air. See Dean Starkman, "Sony BMG Settles Radio Payola Probe, July 26, 2005. httn://www.washinatonnost.corn/wo-dvn/content/article/2005/07/25/AR2005072501624.html; Jeff Leeds, Universal Music Settles Big Payola case, The New York Times, May 12, 2016. httn://www.nvtimes.corn/2006/05/12/business/12navola.html. The Infinite Dial 2016, Edison Research, Triton Digital, "Sources used for keeping-up-to-date with music," at 34. httn://www.edisonresearch.corn/wn-content/unloads/2016/03/The-in6nite-Dial-2016.ndf. Exhibit MC 15, Ipsos OTX MediaCT, Music Choice Viewership Study July 2015, at 27. Testimony ofDavid J. Del Beccaro at 28—29.

PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

music last 12 months, Music Choice VOD had [[ ]] views representing [Wl] ofmonthly VOD snd [Q]] of total VOD views [including non-music VOD views)." the relative benefits ofthe (25) The method by which users consume music has important implications for various participants in the music industry supply chain. In the 1996 rate proceeding for the PSS sound recording performance rights, the record labels presented evidence showing that for every dollar spent — on a CD, 5 cents went to copyright owners ofmusical works, 7 10 cents went to the recording artist, and 56-88 cents went to the record companies.'imilarly for digital downloads: ofthe average digital download price of $1.20, a 2011 Rolling Stones article reported that record labels earn 70% of that and pay 9.1 cents in mechanical royalty, keeping keep 75 cents (62%). I am aware ofno evidence that the shares of different rights holders of CD sales or digital downloads are qualitatively dissimilar today. (26) What is different today is that streaming and digital downloads now form the bulk ofmusic industry'y revenue. Because it is typically perceived to be confidential business information, detailed information about sound recording and musical composition performance rights royalties and payments among rights holders and record companies for such products are hard to find. It appears, however, that major record companies keep the substantial bulk ofthe payments made by music streaming services like Spotify. A study by French Musio trade body SNEP with Ernst k Young found that major record labels keep 73% ofthe payouts made by streaming services, or about 46% of subscription payments (with artists having a 7% share and writers/publishers having a 10% share). comparison, the royalties that SoundExchange collects from PSS services is only a small fmction of overall record industry revenues. In 2015, Soundgxchsnge received [~] million in PSS royalties, which was only [l l] oftotal US music industry wholesale revenues of $4.95 billion

that year.'g

Ibid. Library of Congress, [Docket No. 96-5 CARP DSTRA], Determination ofReasonable Rates and Terms for the Digital Performance ofSound Recordings. 63 FR at 25405 (May 8, 1998). htto://www.convriaht.uov/fedrea/1998/63fr25394.odf The average price of a single download of $1.20 ($1,226.9MM/ 1,021) comes f'rom News and Notes on 2015 RIAA htto://www.riaa.corn/wo-content/unloads/2016/03/RIAA-2015-Year-End-shinments- Shipment and Revenue Statistics. " memo.ndf. The 70% share comes from Steve Knonoer "The New Economics ofthe Music Industrv The Rollina Stone Oct. 25 2011. htto://www.rollinastone.corn/music/news/the-new-economics-of-the-music-industrv-20111025. The 62%0 labels keen is calculated as ([$1.20 x 70%1 - 9.1 centsl / $1.20. Tim Ingham, "Major labels keep 73% of Spotify premium payouts — reports," Feb. 3, 2015. htto://www.musicbusinessworldwide.corn/artists-set-7-of-streamina-cash-labels-take-46/ SoundX 000028618 RESTRICTED; News and Notes on 2015 RIAA Shipment and Revenue Statistics. httn://wm.riaa.corn/wn-content/unloads/2016/03/RIAA-2015-Year-End-shinments-memo.ndf PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

II.B. Music Choice s residential audio music service

Corporation (27) I understand that the idea for Music Choice began in 1987 when General Instrument developed the technology for digital music service. In 1991, Music Choice (at the time known as Digital Cable Radio Associates), was launched as a stand-alone entity. 'usic Choice launched its residential music service at inception in 1991 and expanded their business to offer video services in 2004

offers 50 (28) From these beginnings, I understand that Music Choice's residential music service now diverse audio music channels and an additional 25 channels to authenticated television subscribers throu h Music Choice's website and mobile app. Music Choice's programming reaches nearly [[I Choice's music channels ]] residential cable subscribers across the United States. 'usic

cover a broad range of tastes, including R&B, hip hop, jazz, classical, dance, theatrical, and several

subgenres of rock. (29) The company's commercial-free music broadcasts include both an audio and visual component.'usic

Choice provides music channels curated by programming experts who create cohesive playlists

across all genres of music. They also invest in audience research to understand their customers and their programmers use that information to help optimize programming.'30)

As songs are being broadcast, Music Choice displays key marketing information including the artist' name and album title as well as interesting facts about the artist and/or song. Viewers can easily use this information to buy music through other distribution channels (e.g. CDs, digital downloads, or paid-streaming services).

(31) At its own expense, Music Choice also produces original content about individual artists and has 30 invested in social media marketing to promote user engagement and increase awareness of artists. All of these features make Music Choice music offerings attractive not only to consumers (and the

Testimony of David J. Del Beccaro at 1-2. Music Choice 2015 Annual Report, 7. MC0003221. ("Pursuant to the General Partnership Agreement dated March 1, 1991, Music Choice a Pennsylvania general partnership, was formed under the Pennsylvania Uniform Partnership Act for the purposes of further developing and maintaining a programming service consisting of digitized music and audio progrannning. Such purpose was expanded to include the provision of video services pursuant to the Second Restated Agreement [effective January 1, 2004j.") Music Choice 2016 Media Kit at 24. ("50 channels in nearly 50 million homes.") htt://cor orate.musicchoice.com/files/2214/6228/4671/2016 Music Choice Media Kit. df. h://cor orate.musicchoice.com/about-us/ roducts/. Music Choices offers three residential products: Music Choice "Music Choice play, Music Choice Video on Demand and Music Choice Music Channels. Music Choice also provides for Business" for businesses. Testimony of Damon Williams at 6. Testimony of Damon Williams at 7. Testimony ofDamon Williams at 11-12, 14, 29-30. Testimony of David J. Del Beccaro at 33-34. Testimony ofDamon Williams at 13. e 0

PUBLIC VERSION Written Direct Testirriony of Gregory S. Crawford, PhD 0 cable operators that FCC calls Multichannel Video Programming Distributors or MVPDs) that 0 provide service to them), but also to record labels and artists."

(32) I understand that Music Choice strictly complies with the sound recording performance complement Music Choice does not pre-announce 0 as required by the statutory license. In accordance with the Act, songs, play more than three different selections from any one album within a three hour period, or 0 play consecutively more than four selections by the same artist or from a compilation set of albums 0 within a three-hour period on any of its channels.' understand that these restrictions are intended to 0 limit the use of Music Choice's service as a substitute for other paid forms of music consumption 0 (e.g. CD sales or digital downloads or paid streaming services).

0 (33) Music Choice currently has three lines of businesses: a residential audio service, a residential video service, and a commercial audio service. Its primary service, both historically and today, is its residential audio service. Music Choice was formed as a General Partnership in 1991 for the purpose 0 of offering a residential audio service and it has been its flagship service (and primary source of 0 revenue) ever since." (34) Music Choice also offers a commercial audio service, but the two services are financially distinct and 0 serve distinct markets. They have separate revenues and pay separate fees for musical works and sound recording performance rights and are marketed and sold by cable operators to businesses, with 0 a share of operators'evenue passed back to Music Choice. Music Choice's commercial service is also different f'rom the residential services with respect to its content: it has no VOD, very few commercial subscribers have video channels, and it offers three audio channels beyond those offered on its residential service. (35) In the 2000s, Music Choice sought to differentiate itself f'rom other cable radio operators by offering first with a Video-on-Demand (VOD) service in 2004 and dedicated music video services. It began 0 video channels in 2010. The former is called Music Choice On Demand and the latter is called MC 0 Play (forinerly SWRV TV). 0 0 "Seventeenth Annual Report on the Status of Competition in the Market for the Delivery of Video Programming," FCC (May 6, 2016) MB Docket 15—158. - recordings, Testimony ofDavid J. Del Beccaro at 5. See also 17 U.S. Code Il 114 Scope of exclusive rights in sound h s://ss~.law.cornell edu/uscode/text/17/114 Music Choice 2015 Annual Report, 7. MC0003221. Music Choice officially changed the name of its music channels from SWRV TV to Music Choice Play (MC Play) on orate.musicchoice.com/about-us/ ress-room/ ress-article/music-choice-rebrands-swrv- October 15, 2013. See htt://co 0 tv-music-choice- la -inte rates-all-networks-one-seamless-ex erience-october-15th/. See also Music Choice 2015 Annual Report, 7. MC0003221. ("Pursuant to the General Partnership Agreement dated March 1, 1991, Music Choice a 0 Pennsylvania general partnership, was formed under the Pennsylvania Uniform Partnership Act for the purposes of further developing and maintaining a programming service consisting of digitized music and audio programming. Such 0 purpose was expanded to include the provision ofvideo services pursuant to the Second Restated Agreement [effective January 1, 2004]."1

0 10 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

its audio music (36) Music Choice does not sell its services directly to households. Instead, it provides channels along with its video offerings (VOD and MC Play) to MVPDs. MVPDs, in turn, bundle these channels with other, predominantly video, television channels and offer them for sale to households, typically in a digital cable package. Thus, households consume Music Choice's audio and video channels as part oftheir residential TV package from MVPDs. This has important implications for how outcomes would be determined in a hypothetical market for sound recording perfonnance rights.

costs for (37) While Music Choice's video services are also distinct from its audio service, revenues and the residential audio and residential video services are sometimes commingled. Music Choice's VOD service has always been sold as a bundle with its residential audio service to MVPDs. By contrast, when it was introduced in 2010, MC Play (then called SWRV) was priced separately to cable operators." Starting around 2012-13, Music Choice moved away from this strategy and began bundling its video channels with its audio channels and VOD service. Today, the vast majority of households with access to its services subscribe to a MVPD that pays a single fee for all of the residential audio and residential video (VOD and MC Play) services. This also has important implications for predicting outcomes in a hypothetical market for sound recording performance rights.

(38) While it pays separate royalties for each of its three lines of business, Music Choice maintains its Financial Statements only at the consolidated (company-wide) level and does not have a separate Profit and Loss (P&L) Statement for each business. This impacts the analysis I conduct in this report in ways I describe further in Section V below. Before doing so, I discuss the hypothetical market for sound recording performance rights, possible benchmark markets for that hypothetical market, and an economic model of how royalties would arise in such a market.

III. The hypothetical market and the lack of suitable benchmark markets

III.A. The hypothetical market

III.A.1. Baseline Principles

sound recording (39) The purpose of this proceeding is to determine reasonable rates for the license of perfonnance rights for Pre-existing Subscription Services (PSSs) pursuant to 17 U.S.C. 114(f) of the Copyright Act. In my opinion, there are two steps to this process. The first step is to determine what royalties would arise in the absence of regulatory intervention, i.e. what royalties would arise in a

Testimony ofDavid J. Del Beccaro at 28.

11 0 0

0 of S. Crawford, PhD PUBLIC VERSION 0 Written Direct Testimony Gregory

0 "hypothetical market" for PSS sound recording performance rights.'he second step is to consider how the royalties that arose in such a hypothetical market should be adjusted, if at all, to account for the four statutory objectives outlined in Section 801(b) of the Act. In this subsection, I consider what should be is an appropriate hypothetical market; In Section VI below, I discuss how such a market rate 0 adjusted for the 801(b) factors. such market 0 (40) In analyzing what might be hypothetical market outcomes, one first must decide how outcomes are determined. I understand that rates for PSSs may be set without regard to marketplace 0 rates, but that the law sets standards for royalty rates for other products and services using various descriptions of hypothetical marketplace rates. For example, the Copyright Act requires royalty rates 0 for noninteractive Webcasting that "most clearly represent the rates that would be negotiated in the marketplace between a willing buyer and a willing seller."'imilarly, royalty rates for music licensed from Performing Rights Organizations under consent decrees agreed with the Department of Justice must reflect the "fair market value" ofthose rights," market rates should be (41) In my opinion, all such language points to a similar outcome: hypothetical an economic model matched most closely to the institutional features of that detenTnned using market, including that those features are consistent with competitive markets (about which more below), For example, a hypothetical market with a small number of large firms on one side and a large number of small firms (or a large number of consumers) on the other might profitably be analyzed using economic models of oligopoly interaction. Or, as I introduce in later in this report, a hypothetical market with a small number of firms on both sides ofthe market might profitably be bargaining. While not necessarily exhaustive, I believe that analyzed using economics models of economic theory, in general, provides a wealth of models to describe almost all possible hypothetical 0 rights. marketplaces for sound recording performance (42) For the hypothetical market for PSS sound recording performance rights, there are a several possible market. One could, for example, imagine record labels scenarios for the structure of the hypothetical jointly negotiating with individual buyers of sound recording performance rights, as they currently do via SoundExchange under the guise of the compulsory license. However, I believe this would be inappropriate. The joint negotiation through SoundExchange is a reasonable solution to govern negotiations in the presence ofa compulsory license. In the absence ofthe compulsory license, 0 individual record labels would negotiate individually with copyright users like PSS providers. Indeed, individual record labels do negotiate individually with copyright users in a host of markets 0

0 While not strictly necessary given the legal standard in this proceeding (i.e. I understand that rates for PSSs can be set without regard to marketplace rates), I think that understanding what would be competitive marketplace rates is a useful starting point from which one can then evaluate the statutory factors. 114 - Scope of exclusive rights in sound recordings. h s://www.law.cornell.edu/uscode/text/17/114. 17 U.S. Code II 0 United States of America v. Broadcast Music, Inc., 426 F.3d 91, footnote 5 (2d Cir 2005). 0

0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

not governed by compulsory licenses, including rights for interactive webcasting, rights for music included in video performances, and rights for digital downloads.

IV (43) There remains the question of how competitive would be such a market. In the recent Webcasting decision, reflecting a willing buyer/willing seller rate standard, the Copyright Royalty Judges concluded that "they are required by law to set a rate [in that proceeding] that reflects a market that is effectively competitive" (emphasis added). Later in the same decision, the Judges equate "effective" with "workable" competition, with the latter defined by Dr. Carl Shapiro, representing Pandora, as "[a market in which there] is regular, significant competition among suppliers for the patronage of buyers...." Dr. Shapiro further concludes, "A market can also be workably competitive even if it is quite concentrated, so long as the suppliers compete regularly and energetically to win business from each other." Indeed, the Judges go further later in the decision, concluding "Both Dr. Shapiro and Dr. Katz opined, and the Judges agree, that effective or workable competition arises when licensees have the reasonable (albeit still constrained) ability to select sound recording inputs based upon price.""' agree with these conclusions and think a similar standard should apply in a hypothetical market for PSS sound recording performance rights (before the 801(b) factors are accounted for).

(44) As such, I think the hypothetical market is best characterized as sellers ofsound recording performance rights (i.e., individual record labels) negotiating with individual buyers ofsound recording perforinance rights for PSSs (i.e., Music Choice or any other PSS) in an environment where both buyers and sellers are competing with each other for business.

(45) While the outcomes in a hypothetical marketplace between record labels and PSS are of primary interest, there are three other related markets that also play a role in determining such outcomes. The first such group of markets is the set of markets for other inputs required by PSS providers in order to offer a music service. One important group of inputs are the digital performance rights for musical compositions, the licenses for which must be negotiated with the performing-rights organizations include the ("PROs") that control them, notably ASCAP, BMI, and SESAC. Other important inputs facilities, staff, and capital required to offer such a service.

United States Copyright Royalty Judges, The Library of Congress, In re Determination ofRoyalty Rates and Termsfor Ephemeral Recording and Webcasting Digital Performance ofSound Recordings PVeb IV), Docket No. 14-CRB-001- WR (2016-2020) at 38. 4" Ibid at 59. 4'bid at 121. The copyright owners ofmusical works are the songwriters who author them or, more typically, the music publishing companies who acquire the rights f'rom the songwriters. These music publishers and songwriters typically join PROs and appoint the PRO as their licensing agent for the public performance right. The PRO then handles the voluminous task of identifying users, issuing licenses, collecting royalties, and distributing those royalties to the appropriate music publishers and songwriters. See Marie Connolly and Alan B. Krueger, "Rockonomics: The Economics ofPopular Music" in Ha»dbooIc on the Economics ofArt and Culture (Amsterdam: Elsevier, 2006) at Chapter 20.

13 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

market is the (46) The second group ofmarkets important for understanding outcomes in the hypothetical set ofmarkets for PSS outputs. PSS providers themselves combine sound recording performance rights with other inputs to produce a set of audio "channels" that they in turn sell to cable, telecommunications, and satellite television providers (MVPDs).

market is the set (47) The final group ofmarkets important for understanding outcomes in the hypothetical ofmarkets served by the same buyers and sellers. This turns out not to be relevant for the buyers in this market, PSSs, as each PSS provider only offers residential audio services in the PSS market. It is very relevant, however, for the copyright owners of sound recording performance rights (record labels). It is widely understood that record labels license rights to perform sound recordings to a wide variety ofbuyers, including PSS providers, SDARS providers, and Webcasters. They also physically produce and sell CDs directly to consumers and license the sale of digital downloads. As I show below, understanding the relationship between the rights licensed in the PSS market and the profitability ofthese other performance and distribution channels is critical to determining a marketplace royalty for sound recording performance rights for PSSs.

III.B. The ideal benchmark market

III.B.1. The Use of benchmarks

(48) Agreeing on a hypothetical market is only the first step in the determination ofa reasonable rate for sound recording performance rights. The next question is how to use such a market. In my opinion, there are two reasonable approaches. The first is to identify a benchmark market that is close to the hypothetical market (in ways that I describe below) and rely on outcomes in that market, suitably adjusted for all differences, to inform what outcomes would arise in the hypothetical market." The second is to identify those factors that would influence outcomes in the hypothetical market and directly estimate those outcomes.

Before I (49) In this report, I will take the second approach, and describe it in detail in Sections IV and V. do so, however, I describe why, given the decisions made in the previous proceedings, I don't think that using a benchmark approach is the best option for determining reasonable rates for sound recording performance rights in the PSS market.

4'ndeed, in their decision in the previous proceeding, the Copyright Royalty Judges concluded "Where the determination standard is reasonable rates calculated to achieve the $801(b)(1) factors, the Judges have found market benchmarks, if and any, to be a useful starting point." See In the Matter ofDetermination of and Terms for Preexisting Subscription Satellite Digital Audio Radio Services, Docket No. 2011-1 CRB PSS/Satellite II. 78 FR 23056.

14 0 0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD 0 0 0 Ill.B.1.a. The ideal Benchmark Market and forces that would arise (50) The ideal benchmark market would reflect the same market environment 0 are five in the hypothetical market for PSS sound recording performance rights. In my opinion, there characteristics that a market would have that would make it an ideal benchmark for the hypothetical market. They are these: negotiation of benchmark market, the 0 1. Market lace outcomes with similar stakes: In an ideal royalty rates between the buyers and sellers of sound recording performance rights would take

place in an open market and not in a venue such as a statutory proceeding. Royalties negotiated in 0 an open market would ensure that any such buyers and sellers were not influenced by non-market 0 factors such as regulatory statutes. That being said, many licensing markets are hybrid markets. Negotiations occur in a marketplace

can to a judicial or 0 setting, but, in case ofimpasse, either party to the negotiation appeal regulatory body for a rate determination.4" These negotiations are sometimes described as occurring 'in the shadow'f a rate proceeding. Because such appeals are often time-consuming, costly, and uncertain, the ideal hybrid benchmark market would have similar stakes to each party marketplace in the benchtnark market'. as in the hypothetical market, This would avoid relying on outcomes in which one or another party could gain from a negotiation less than the cost ofthe regulatory process. sellers as those that Same sellers: In an ideal benchmark market, the sellers would be the same rights for sound recordings, i.e., individual record labels. license perfonnance

3. ~Sane bu ers: in m ideal benchmark market, the buyers would be the same as those that wish to license performance rights for sound recordings, i.e., PSS providers like Music Choice. 4. Satne ri hts sold to the same downstream bu ers: In an ideal benchmark market, the buyers would buy from the sellers the same rights, i.e., PSS sound recording performance rights, and 0 would themselves sell the same products in markets further downstream (i.e. to cable and satellite operators). In particular, buyers in the ideal benchmark market would package the sound recording performance rights along with other inputs (such as musical composition performing 0 44 For example, negotiations between the major US Performing Rights Organizations (PROs) and users of musical 0 compositions have this feature. 4's in such settings, the party whose gains are less than the cost has little bargaining power with the other party and must David either take whatever offer the other party chooses to grant or completely forgo an agreement. Music Choice CEO between J. Del Beccaro described how Music Choice felt forced to twice settle with SoundExchange (for rates paid David J. Del 0 2002—2007 and 2008-2012) due to the prohibitive costs ofthe rate litigation process. (Testimony of 0 Beccaro at 10— 11) 0 15 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

0 rights and the PSS's own original creative content) to create audio channels. They would then 0 enter into multi-year long-term contracts to sell the rights to these channels to multi-channel video distributors (MVPDs) who, in turn, would bundle them with other content (sports channels, 0 news channels and other content) and sell them in the retail market to households as part of a digital television service. 0 0 5. Satne method of end-user purchase and consumption: In an ideal benchmark market, the ultimate consumers ofthe products in which sound recording performance rights are embodied would purchase and consume them in the same manner as in the PSS market. Households (or other end- 0 users) would purchase access to MVPD television bundles for a monthly fee and then consume 0 the PSS provider's audio channels over the course of each month. Critically, ifthey chose to 0 consume, this consumption would be passive: as the PSS provider selects the particular tracks to 0 play on each of its channels, end-consumers in the ideal benchmark market could either listen or 0 not to those channels, but could not directly select to which music they would wish to listen. 0 III.C. The lack of suitable benchmark markets for sound recording 0 performance rights for PSSs 0 III.C.1. Introduction

0 (51) In an ideal world, there would be an ideal benchmark for the hypothetical market for sound recording 0 performance rights for PSS. Unfortunately, there is no such ideal. In this subsection, I briefly 0 articulate a range of possible markets and discuss their suitability as a benchmark for the hypothetical 0 market for sound recording performance rights for PSSs. 0 III.C.1.a. Sound Recording rights for a range of digital services are inappropriate as a 0 benchmark 0 (52) In the last proceeding, Dr. George Ford, an expert for Soundaxchange, proposed using the royalties paid by a range ofusers of sound recording performance rights in a broad spectrum of other markets 0 in which the record companies directly license sound recording rights as his proposed benchmarks. 0 These included royalties for sound recording performance rights for portable and non-portable 0 interactive webcasting services, cellular ringtones/ringbacks, and digital downloads.4 rights 0 (53) In my opinion, none ofthese markets are suitable benchmarks for sound recording performance 0 for PSSs. There are just too many differences: among buyers, the services offered by those buyers,

0 46 See In the Matter ofDetermination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio 0 Services, Docket No. 2011-1 CRB PSS/Satellite II.FR 7S 23057. 0

, ~ i~ PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

and how consumers use those services. These many differences make comparing outcomes in other markets for sound recording performance rights unusable as benchmarks.

rights from other (54) I am not alone in concluding that agreements for sound recording performance digital markets map poorly to the hypothetical market for PSS sound recording performance rights. Even Dr. Ford agreed, concluding "In my own review of marketplace agreements in which performance rights are sold by copyright owners to copyright users for digital transmissions, I was unable to find any contract that matched closely the peculiar nature ofthe PSS's business." Despite this, he continued to rely on these as benchmarks. The Judges were also unconvinced, in their — decision concluding "... the key characteristic of a good benchmark—comparability is not present."

III.C.1.b. Musical Works rights for PSSs are the best available benchmark, but were rejected by the Judges in the previous proceeding

the determination (55) I also presented my testimony on behalf of Music Choice in the last proceeding on

of reasonable rates for PSS sound recording performance rights. In my report in that proceeding, I concluded that the best and only usable benchmark for the hypothetical market for sound recording perfonnance rights for PSSs was the royalty paid by Music Choice to the three performance rights societies (ASCAP, BMI, and SESAC), and that the rate in the hypothetical market for PSS sound recording perfonnance rights would be no higher than this rate. '57) PSSs. has been (56) This royalty is often called the royalty for "musical works" performance rights for It approximately [[~]] of Music Choice's residential audio revenues.

I base the conclusion that a PSS musical works royalty is the best benchmark for the PSS sound recording royalty on the similarity between the markets for musical works rights and sound recording rights for PSS. Both involve the same buyers selling on the same rights downstream to households that use the rights in the same way.

rights are (58) The only differences between the markets for musical works rights and sound recording two. First, musical works royalties are marketplace agreements between buyers (PSS providers like Music Choice) and sellers (the PROs), but the two biggest sellers (ASCAP and BMI) are required by virtue of antitrust consent decrees to enter into agreements on a "reasonable fee" basis." This is only a modest difference, however, as the rate courts enforcing the consent decree set rates pursuant to a

See In the Matter of Determination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio Services, Docket No. 2011-1 CRB PSS/Satellite II. FR 78 23058. Testimony ofDavid J. Del Beccaro at 9— 10. 4p United States of America v. Broadcast Music, Inc., 426 F.3d 91, footnote 5 (2d Cir 2005). ("The BMI consent decree fee stipulates that "[i]n any [rate] proceeding, [BMI] shall have the burden of proof to establish the reasonableness of the Court shall requested by it. Should [BMI] not establish that the fee requested by it is a reasonable one, then the determine a reasonable fee based upon all the evidence." BMI Consent Decree, art. XIV(A).")

17 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD

Fair Market Value standard designed to approximate marketplace outcomes." As long as stakes for participants in the PSS musical works marketplace are comparable to those in the PSS sound recording marketplace (and I believe they are), I find that this difference isn't sufficient to rule the PSS musical works royalty as a potential benchmark.

works performances (59) The second difference is that the sellers are different: owners ofrights for music (PROs) versus owners ofrights for sound recording performances (record labels). I conducted a detailed analysis ofthe consequences ofthis difference in Section IV.C.3 in my report in the last

' proceeding and concluded that the difference in sellers would only lead to royalties for sound recording performance rights strictly less than those for musical works performance rights." Thus using the existing Music Works rate as a benchmark would be conservative: it would be an upper bound on the rate that would arise in the hypothetical marketplace for sound recording performance rights for PSSs.

for PSSs a (60) Unfortunately, the Judges did not find royalties for the musical works performance rights suitable benchmark for the hypothetical market for sound recording performance rights for PSSs, concluding "the musical works market involves different sellers (performing rights societies versus record companies) selling different rights."

(61) As the market for PSS musical works performance rights are the best possible benchmark in my ' opinion, ifthe Judges continue to conclude that they are inappropriate, then I must conclude that there

' is no appropriate benchmark.

that (62) In the absence of an appropriate benchmark, I conclude that the only way to estimate the royalties 0 would arise in a hypothetical market for PSS sound recording performance rights is to do so directly. In the next two sections, I do just this. In Section IV, I present an economic model ofthe royalties that would arise in the hypothetical market for PSS sound recording performance rights. In Section Choice's V, I quantify the elements ofthis model using the predicted financial performance ofMusic residential audio service.

i ~

United States of America v. Broadcast Music, Inc., 426 F.3d 91, at 2. (2d Cir 2005). ("A rate court's determination of — the fair market value ofthe music is often facilitated by the use ofbenchmarks agreements reached sitar arms'ength negotiation between other similar parties in the industry. See Music Choice II at 194.") See Testimony of Gregory S. Crawford, PhD, In the Matter ofDetermination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio Services, Docket No. 2011-1 CRB PSS/Satellite II. See In the Matter ofDetermination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio 0 Services, Docket No. 2011-1 CRB PSS/Satellite II. FR 78 23082. 18 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD 0

IV. An economic model of the royalties that would arise in the hypothetical market for sound recording performance rights

IV.A. Overview and summary results

0 the that would arise (63) In this section I introduce a model to describe the key factors influencing royalty in a hypothetical market for PSS sound recording performance rights. In the next section, I use financial information from Music Choice to measure the key components of the theoretical model and 0 thereby estimate the royalty that I would expect to arise in this hypothetical market. 0 between firms." (64) The economic framework that I propose is a non-cooperative model of bargaining The simplest non-cooperative bargaining models analyze negotiations in a bilateral monopoly 0 between a, single buyer and a single seller. Nobel-prize-winning economist John Nash established with 0 long ago that there is a unique solution to bargaining problems between bilateral monopolists 0 equal bargaining power that satisfies certain reasonable properties of rational behavior," Such problems are called Nash Bargaining Problems and his solution to them is called the Nash Bargaining 0 Solution. First, 0 (65) Two subsequent developments increased the relevance and applicability ofthe Nash approach. 0 an extension was developed to allow for asymmetric bargaining power, implying an associated 0 Asymmetric Nash Bargaining Solution. Second, it was shown that this Asymmetric Nash Solution 0 can be rationalized as the unique outcome of a bargaining problem in which each party to the bargain alternates making offers until an agreement is reached," This is a realistic depiction ofbargaining in 0 What I present here is a slightly clarified version ofthe same framework I first presented in my filing on behalf ofMusic Choice in the last proceeding. See Testimony of Gregory Crawford, PhD, In the Matter ofDetermination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio Services, Docket No. 2011-1 CRB PSS/Satellite Il. In this report, I propose a non-cooperative bargaining approach instead of a cooperative bargaining approach (whose most common solution concept is the Shapley value). While the two approaches are complementary, I choose the noncooperative approach for two reasons. First, I believe cooperative approaches work best where all parties benefit from the modeled activity (e.g. making contributions to a public project) and not when a gain to one party in a 0 negotiation implies a loss to another (as for royalties for sound recording performance rights). In such cases, I believe that a non-cooperative approach better captures the negotiating environment. Second, I find that the cooperative value of approach can lose its insightfulness when multiple parties to a bargain are all essential (i.e. without whom the value the the project is zero). In such cases, the Shapley value indicates that each party earns 1/N times the overall of 0 the share of the project, where N is the number of participants. When applied to negotiations in a vertical supply chain, total surplus accruing upstream or downstream depends only on the number of firms on each side of the market, and not on other features of the market being studied. As described further below, a non-cooperative approach allows generally for "threat points". In my opinion, these threat points can more flexibly capture the profit consequences of including (or not) one or more parties to a bargain than can the cooperative approach. s4 The four properties are Invariance to Equivalent Utility Representations, Pareto Efficiency, Symmetry, and Independence ofIrrelevant Alternatives. See John F. Nash, Jr. "The Bargaining Problem," Econometrica 18, no. 2 (Apr. Press, 1950): 155-62, and Abhinay Muthoo, Bargaining Theory with Applications (Cambridge: Cambridge University 1999) at Chapter 2. 0 Abhinay Muthoo, Bargaining Theory with Applications (Cambridge: Cambridge University Press, 1999), Chapter 3. 19 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

many market environments, including those for digital performance rights. The Asymmetric Nash

Bargaining Framework is among the most widely applied tools within economics to analyze

between negotiations 6rms.'66)

In what follows, I use the solutions from a bilateral monopoly Asymmetric Nash Bargaining Problem to derive certain essential results, and then demonstrate how these results can be used to understand the determinants ofthe PSS sound recording performance royalties that would arise in a hypothetical market for such rights where there was effective competition (and not monopoly) on each side ofthe market.'7

"upstream" Qrm (67) In the bilateral monopoly model that follows, I consider a single record label as the in a negotiation for sound recording performance rights and a single PSS as the "downstream" firm in the negotiation. I then show the following:

1. The factors that determine bargaining outcomes for sound recording performance rights for PSSs are three:

o The Joint Agreement Profits: The Joint Agreement Profits are the total profits (i.e. the profits to be shared between the two firms) that could be generated ifthe upstream record label and downstream PSS were able to reach an agreement.

o Bach Firm's Threat Point: Bach firm's Threat Point is the profit that would accrue to each ofthe record label and PSS ifthey were unable to reach an agreement.

~ The difference between the Joint Agreement Profits and the sum of each firm's Threat Point are called the Incrementalprofits.'hey are the "size of the pie to be split" in the negotiations.

o Bach Firm's Barpaining. Power: Bach Qrm's Bargaining Power is a number between 0 and 1 representing the strength of each ofthe record label and PSS in the

Recent empirical applications ofthe Nash Bargaining approach include determining the consequences to consumer welfare of Ii-la-carte cable television offerings (Gregory S. Crawford and Ali Yurukoglu, 2012. "The Welfare Effects of Bundling in Multichannel Television Markets," American Economic Review, 102(2): 643-85), analyzing the impact of price discrimination (i.e. charging different prices to different buyers) in negotiations between hospitals and medical device manufacturers (Grennan, Matthew. 2013. "Price Discrimination and Bargaining: Empirical Evidence fiom Medical Devices." American Economic Review, 103(1): 145-77.), asking whether grocery stores offering private-label "Do brands can negotiate more favorable wholesale prices f'rom branded manufacturers (Meza, S. dt, Sudhir, K. private labels increase retailer bargaining power7" Quantitative Marketing and Economics (2010) 8: 333.), and measuring the impact ofhospital mergers on negotiations with managed care organizations (Gowrisankaran, Gautam, Aviv Nevo, and Robert Town. 2015. "Mergers When Prices Are Negotiated: Evidence Rom the Hospital Industry." American Economic Review, 105(1): 172-203.). Appendix A provides the mathematical foundations both for this bilateral monopoly model as well as a richer model that allows for two types of inputs (sound recording performance rights and musical works performance rights), multiple firms providing each input, and multiple PSS providers. Mathematically, the Incremental Profits = The Joint Agreement Profits — (Threat Point Label+ Threat Point PSS)

20 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

hypothetical negotiation. In this modeling framework, the sum ofthe two Bargaining Powers is necessarily 1, so if one firm has a bargaining power of 0.2, the other necessarily has a bargaining power of 0.8.

2. I then describe qualitatively what each ofthese Nash Factors would be in the hypothetical market for PSS sound recording performance rights. In the next section, I estimate each of the Nash Factors based on information on the financial performance ofMusic Choice's residential audio service.

(68) Before discussing how each ofthese results was derived, I describe the relevant assumptions that underlie the model and how they map to the hypothetical market for PSS sound recording performance rights.

IV.B. Model Assumptions

IV.B.1. The hypothetical market

(69) The purpose ofthe bargaining model I introduce here is to estimate the royalty rate that would exist between individual record labels and PSSs in general, and Music Choice in particular, in the absence of a compulsory license. As described in greater detail in Section III.A above, I conclude that this market is best characterized as individual sellers of sound recording performance rights (i.e., the record labels) negotiating with individual buyers of sound recording performance rights for PSSs (i.e., Music Choice and any other potential buyer).

(70) As also described in Section III.A above, there are three other related markets that play a role in determining outcomes in the hypothetical market. These are the markets for other inputs required by PSS providers, the market for PSS outputs (i.e. the market in which PSS providers sell their services to MVPDs), and markets served by the same buyers and sellers. The importance of each ofthese additional markets will become clear once I introduce the non-cooperative bargaining model.

IV.B.2. A non-cooperative bargaining model captures well markets with bilateral market power

(71) The use of a non-cooperative bargaining model is predicated on the assumption that record companies and PSS providers each have a certain degree of market power. In the context ofthe market for PSS sound recording performance rights, this assumption is reasonable. In a typical economics textbook, marketplace outcomes between buyers and sellers are represented by the intersection of a downward- sloping marginal revenue (or demand) curve and an upward-sloping marginal cost (or supply) curve. This framework for evaluating market outcomes works well when markets are perfectly competitive

21 0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD 0 '0 are small or when there is some market power on only one side of a market.'n markets where there numbers ofboth buyers and sellers and each has some degree ofmarket power, however, this kind of "marginal analysis" isn't appropriate. Economists have found that modeling marketplace outcomes with bilateral market power as bargains is more realistic. , ~ 0 believe that both individual (72) In the hypothetical market for PSS sound recording performance rights, I 0 record companies and PSS providers would have a certain degree of market power. For the record companies, this market power would be evident in the form oftheir ownership of sound recording performance rights, without which it would not be legal to digitally broadcast most songs in the U.S. Record labels are therefore able to exercise market power over those songs for which they own the 0 sound recording performance rights. current product 0 (73) PSS providers like Music Choice also have a degree ofmarket power due to their 0 offerings and established relationship with MVPDs. For example, Music Choice possesses a bundle 0 oftechnology, programming, and staffthat would also be costly and time consuming for any other 0 Qrms to duplicate. Furthermore, they have negotiated long-term contracts with individual MVPDs that prevents another firm from easily entering the market with a rival service and obtaining a large 0 customer base quickly. PSS providers are therefore able to exercise some degree ofmarket power as '0 a buyer of sound recording performance rights in their negotiations with rights holders due to their 0 existing assets and ongoing business relationships with MVPDs. some (74) I note briefly that the assumption that individual record labels and PSS providers each has 0 market power does not mean that the hypothetical market isn't effectively or workably competitive. 0 Markets with small numbers of 6rms or firms with market power can be workably competitive as 0 long as firms are consistently trying to compete with their rivals for business opportunities. As will 0 be shown below in the analysis ofThreat Points, the model ofhypothetical market outcomes I present 0 incorporates both ofthese features: market power, but also competition among rivals both upstream (i.e. among record labels) and downstream (among PSS providers).

conclude that a non- (75) Given that both PSS providers and record companies have some market power, I cooperative bargaining model is an appropriate framework for modeling market outcomes for the PSS sound recording performance rights in the absence of a compulsory license.

0

0 In such settings, firms optimally equate marginal revenues with marginal costs to determine their profit-maximizing prices or quantities. 0 Alvin E. Roth, "Bargaining (Economic theories ofbargaining),"in Social Science Encyclopedia (London: Routledge, 1996) at 46-7. 0 22 0 0 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

0 IV.B.3. A model with a single record label and PSS provider captures the essential insights that would arise in a model with multiple record labels and/or PSS providers

0 outcomes to depend (76) Recent authors have extended Nash's bilateral concept to allow for marketplace on the actions of multiple upstream and downstream firms. 'hile these extensions are indeed relevant in the market for sound recording performance rights in PSS, the essential insights are present in Nash's solution for bilateral monopoly.

and multiple PSS (77) To see this, briefly consider a bargaining problem between multiple record labels 0 providers, In this more complex setting, one could still use a non-cooperative bargaining framework to characterize royalties for each PSS in their negotiations with each record company. One could then "total add up the predicted royalties paid by each PSS across these record companies to determine a sound recording royalty" that each PSS would pay for licenses covering all recorded music, One could then average this total sound recording royalty across PSSs and argue this average royalty is a a statutory blanket good approximation of a hypothetical marketplace rate that could be used for license for sound recording performance rights for PSSs. (78) In the case of symmetric record labels (i,e. all of a roughly comparable size) and symmetric PSS providers (similarly), an average total sound recording royalty would likely be identical to that arising PSS provider.62 Even from bilateral monopoly negotiations between a single record label and a single the absence of symmetry, for example if some record labels were "majors" and other labels were in "independents", the royalty predicted in a bilateral &amework would be conservative as the bilateral structure is modeled with a major record label in mind (and outcomes for independents would generally yield lower predicted PSS royalty rates)." what follows, I therefore assmne a bilateral monopoly fiamework and describe, where relevant, (79) In how deviations in the predicted outcomes from this framework would be affected by multiple firms up- or downstream. 0

Henriclc Horn and Asher Wolinslq, "Bilateral Monopolies and Incentives for Merger," RAND Journal ofEconomics 19, 0 no. 3 (1988): 408-19. The equivalence should be true as long as a PSS provider could not offer a viable PSS service in the absence of an agreement with all (major) record companies. This is consistent with my understanding of the business environment facing Music Choice: they would not be able to operate a viable PSS without sound recording performance licenses covering the vast majority ofthe licensable music library (Conversation with Music Choice, CEO, David J. Del 0 Beccaro, Oct. 10, 2016). with While jumping ahead, this is because the Threat Point for a PSS provider will generally be stronger in negotiations an independent than with a major record label and a stronger Threat Point for a PSS provider will lower the predicted royalty they must pay.

23 i ~

i ~ PhD PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, 0 IV.C. A Non-cooperative bargaining model of sound recording performance rights for PSSs

IV.C.1. Nash bargains: Surplus, threat points and bargaining power 0 IV.C.1.a. Overview of general factors it yields intuitive (80) One attractive feature ofthe Asymmetric Nash Bargaining Framework is that predictions about how outcomes are determined when firms negotiate or bargain. In what follows, I consider the negotiation between an upstream firm, a record label, and a downstream firm, a PSS 0 provider. that determine how (81) The Asymmetric Nash Bargaining Solution identifies three fundamental factors 0 two firms "split a pie". I call these the "Three Nash Factors." They are: 0 ~ The Joint Agreement Profits: The Joint Agreement Profits are the combined profits to both the 0 upstream and downstream firms in the market under study Rom reaching an agreement. For PSS 0 sound recording performance rights, the Joint Agreement Profits are simply the revenue earned 0 by the PSS provider in the market for PSS less all the non-sound recording royalty costs that they face.s4

~ Bach Firm's Threat Point: Bach firm in a negotiation must allow for the possibility that no

0 agreement will be reached. The pro6t each receives when no agreement is reached is called their disagreement point or "Threat Point". 0 firms'hreat o I call the difference between the Joint Agreement Profits and the sum ofthe Points the Incremental Profits. 'hey are the "size ofthe pie to be split", i.e. the profits the firms could earn &om an agreement above and beyond what they could earn 0 in the absence of an agreement. 0 ~ Bach Firm's Bareainina Power: Each firm's bargaining power is a number between 0 and 1 measuring the strength ofthat 6rm in the negotiations. The sum ofthe two Bargaining Powers

amount The "Profits" in the Joint Agreement Profits are the payments a good or service can command beyond the needed to cover its cost ofproduction. They are sometimes also called "Economic Profits". They are related to but distinct from conventional notions of profit The differences are two. First, profit in the conventional sense is typically meant to measure the return on the capital invested in a firm. Such returns to capital are not part of economic profits as provider's costs. they are part ofthe costs ofproviding a service and should therefore be counted as part of a PSS As that is Second, the profit ofthe PSS would normally count sound recording performance royalties as one ofits costs. retained in the Joint the thing being determined in the bargaining model, however, it is excluded from their costs and Agreement Profits. I incorporate both ofthese differences in the next section when I estimate the Joint Agreement Profits using Music Choice's financial information. 0 Mathematically, the Incremental Profits = The Joint Agreement Profits — (Threat Point Label+ Threat Point PSS) The bargaining power ofthe PSS provider is just one minus that ofthe record company.

0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD

of is 1, so if one firm has a bargaining power of 0.2, the other necessarily has a bargaining power 0 have historically been related to each party' 0.8. Economic estimates ofbargaining power ' 0 patience in a negotiation, with greater patience yielding greater bargaining power. understand that this is consistent with the nature ofbargaining between Music Choice and the copyright 0 owners with whom they negotiate.ss "Nash Factors" determine 0 (82) The Asymmetric Nash Bargaining Solution demonstrates that these three 0 royalties in a simple and intuitive way: The profit received by each ftrm in a bargain equals its Threat Point plus its 0 Bargaining Power times the Incremental Proftts. (83) Muthoo (1999) describes the Nash Bargaining Solution as "splitting the difference". Two simple examples based on the features ofthe hypothetical market for PSS sound recording performance 0 rights nicely demonstrate this result and provide some intuition on how each ofthe Nash Factors 0 influences bargaining outcomes. (84) Examole 1: A Bargain with no Threat Points: Consider two firms bargaining in a market: an upstream record label and a downstream PSS provider. Suppose for simplicity that the Joint Agreement Profits ifthey reach an agreement in the PSS market are 20 and that each firm's Threat Point is zero. In this case, the "size ofthe pie to be split", or Incremental Profits, are just the Joint Agreement Profits of 20. Further suppose that each firm is equally patient in the negotiations and therefore has equal Bargaining Power (equal to 0.5). The Asymmetric Nash Bargaining Solution says that each firm gets their Threat Point (zero) and one-halfthe Joint Agreement Profits. Thus, the 0 record label and PSS provider each get a profit of 10. Since the revenues in the PSS market accrue to the PSS provider, they would therefore pay the record label a royalty of 10. Nash Bargaining "splits the pie".

, ~ (85) Examnle 2: A Bargain with Threat Points: A second example demonstrates the influence Threat Points have on bargaining outcomes. As above, let the Joint Agreement Profits in the PSS market be 20. Suppose, however, that the record label earns an additional profit of 4 in another market if it is able to reach an agreement with the PSS provider in the PSS market (e.g. fiom digital downloads of the songs played by the PSS). Its Threat Point in the bargaining problem is its profit in the case of 0 disagreement. In case of disagreement, the record label loses the profit Rom digital downloads i ~

Abhinay Muthoo, Bargaining 27reory with Apphaations (Cambridge: Cambridge University Press, 1999) at 51. Conversation with Music Choice, CEO, David J. Del Beccaro, Oct. 10, 2016. * Mathematically, we would write this Royalty = Threat Point+ Bargaining Power Incremental Surplus This example assumes copyright owners benefit from a promotional effect of an agreement in the PSS market. 0 25

, ~ PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

prompted by PSS listening and its threat point is therefore -4 (i.e. less than 0). 'uppose the Threat Point for the PSS provider continues to be zero.

The Incremental (86) The profit to each party in this bargaining problem is slightly more complicated. Profits, or "size ofthe pie to be split", are now the Joint Agreement Profits of 20 from the PSS market = minus the sum of the Threat Points of the two firm, -4, for a total of 20 - (-4) 24. This is more than the profit available in the PSS market alone because ofthe additional benefit the record label gets from digital downloads stimulated by the promotional effect ofthe PSS market. As above, suppose each firm is equally patient and has equal bargaining power (0.5). The Asymmetric Nash Bargaining Solution says the profit to the PSS provider is equal to its threat point (zero) plus one halfthe Incremental Profits of 24, or 12. The profits to the record companyfrom the I'SS market are equal to its Threat Point (-4) plus one halfthe Incremental Profits (12), or 8. As all PSS revenues accrue to the PSS provider, they pay the record label a royalty of 8. Ofthe Joint Agreement Profits in the PSS market of 20, the record label gets 8 and the PSS provider 12.

4 from (87) Lest this seem unfair, recall that in the case of agreement the record label also gets a profit of outside the PSS market, e.g. due to digital downloads. It is for this reason that markets served by the saine buyers and sellers are important if one wishes to understand outcomes in a market with bargains. If one includes this into its overall profit &om an agreement with the PSS provider, then it is clear that each firm gets 12, i.e. each firm shares equally the total profit across all markets of 24.

IV.C.1.b. Royalty rates

(88) The royalty calculated above is measured in profit (dollars), not as a share of revenue (as is typical in many digital audio markets). It can easily be converted into a share of revenue, however, for a vertical market like that for PSS sound recording performance rights where the downstream PSS provider sells a product and passes on a portion of its revenue &om those sales to the upstream record company. In this case, the appropriate royalty rate is simply the royalty (measured in dollars) divided = by the PSS provider's revenue. In the first example, the royalty rate would be 10/20 50%. In the second example, the royalty rate would be 8/20 = 40%.

(89) While straightforward within a single market, care must be taken comparing royalty rates across markets. Two markets may have the same royalty (measured in dollars) but very different royalty rates. A simple example shows why.

Threat Points, and the record (90) In Example 1 above, the Joint Agreement Profits were 20, there were no label and PSS provider had the same Bargaining Power, yielding a royalty to each of 10. Ifthe Joint Agreement Profits of 20 arose from a downstream market where the PSS provider had revenue of 20 and no costs, then the appropriate revenue royalty rate would be 50 percent (record label royalty of 10

The model allows for the possibility ofpositive or negative threat points.

26 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

market divided by PSS revenue of20). Ifthe Joint Agreement Profits of 20 arose lrom a downstream with revenue of 40 and costs of 20, however, then the appropriate revenue royalty rate would be 25 percent (record label royalty of 10 divided by PSS revenue of 40). For a given amount of Joint Agreement Profits, higher costs generally imply lower (revenue) royalty rates. This can be very important if comparing possible benchmark markets with different cost structures and highlights why it is most useful to select a benchmark market with the same buyers.

IV.C.2. Applying the bargaining model to sound recording performance rights for PSSs

IV.C.2.a. Overview

(91) The Nash Bargaining Framework is extremely useful for determining the sound recording performance royalty that would arise in a hypothetical market between record labels and PSS providers. In this subsection, I use my understanding ofthe hypothetical market for sound recording performance rights for PSSs to describe, in general terms, each ofthe three Nash Factors in this hypothetical market." In Section V below, I use data on Music Choice's financial performance to formally quantify each ofthese Nash Factors in the hypothetical market for PSS sound recording performance rights.

IV.C.2.b. Joint Agreement Profits

(92) Determining what is the Joint Agreement Profits in the hypothetical market for sound recording performance rights is straightforward. It is simply the joint economic profit to be shared between the record label and PSS provider in the PSS market if an agreement is reached. This is just the total economic profits earned by the PSS provider before payment of a sound recording performance royalty.

iV.C.2.c. Each Firm's Threat Point

(93) The Threat Point for the PSS provider is similarly straightforward. In the absence of an agreement with any ofthe major record labels, the PSS provider cannot produce any music channels and therefore cannot earn any surplus. Its Threat Point, therefore, is zero."

In this subsection, I describe the mapping between the bargaining model and the hypothetical market for PSS sound recording performance rights in general terms. In the next section, I quantify these general terms using financial performance information from Music Choice. As discussed in the text in paragraph (77), it is my understanding that Music Choice could not offer a viable cable radio service without the catalog ofeven a single major record label. I note here that this assumption is conservative: if a PSS provider could offer a viable service in the absence of an agreement with a single record company, it would increase its Threat Point, increasing its portion ofthe Joint Agreement Profits, and reducing any royalty payment (or rate) to the record company.

27 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

in the

(94) The Threat Point for the record label, however, is more nuanced. Like the PSS provider, absence of an agreement, the record label receives no revenuePorrt the PSS market. As a result, that portion of its Threat Point is likewise zero.'95)

However, failure to reach an agreement with a PSS provider has additional implications for the record label. In addition to licensing sound recording performance rights in the PSS market, the record label also licenses the performance rights to its sound recordings in other, non-PSS markets. For example, they also earn profit from the licensing ofmusic for digital downloads and streaming services. Failure to reach an agreement with a PSS provider could have an adverse impact on these additional revenue streams."

on other (96) Indeed, a large body of evidence suggests there are significant promotional benefits of PSS sources ofrecord label profits, for example CD sales, digital downloads, and plays on music streaming services. The following paragraphs summarize this evidence.

noted that (97) In its 1998 ruling for PSS sound recording performance rights, the Librarian of Congress the Copyright Arbitration Royalty Panel reasonably found that the record companies did not incur a risk from lost sales due to the PSSs'ctivities.7s Indeed, the Librarian noted that the Panel believed that PSSs decreased the risk to the recording companies because digital audio services have

74 Here is one place where the simplification provided by the bilateral monopoly &amework could influence the predictions ofthe model. Allowing for multiple PSS providers would simply mean that the record label would earn no surplus in the market covered by that PSS provider. It could earn surplus Rom markets covered by other PSS providers, but that would be true with or without an agreement with a given PSS. The more important consequence is the PSS's assumption that a record label could not earn profits from the subscribers to MVPDs that currently take a single service. I strongly believe this is the case. Only ifMVPDs were able and willing to switch PSS providers in the case the given PSS provider and record label couldn't reach an agreement or households were willing to switch MVPDs in such a case could the record label earn incremental profits in the case of disagreement in an environment with multiple PSS providers. On the first point, in the case ofMusic Choice, the market leader among PSSs, I understand that the David J. Del weighted (by subscribers) average term of its current affiliate agreements is [I )] (See Testimony of f'rom Becarro at 4.) and conclude that record labels are unlikely to be willing to forgo profits the PSS market for so long. On the second point, the Music Choice audio service is but one element ofthe typical digital cable bundle. Industry "Must-have" observers call content that might induce households to switch providers "Must-have" content. Such content is typically limited to high-value sport content (e.g. Regional Sports Networks), broadcast networks like ABC, households CBS, or NBC, and select high-value cable television channels (e.g. ESPN). It is extraordinarily unlikely that would switch cable providers in order to have access to a particular PSS provider. Because these effects are likely to be negligible, I conclude that assuming that a record label's Threat Point is zero in the PSS market is reasonable. This promotional effect oflicensing PSS sound recording performance rights is a kind of opportunity cost (albeit a negative one): licensing sound recording performance rights to a PSS provider increases the record labels profits. This is analogous to lowering their costs (indeed, lowering their costs to a value less than zero, as their direct costs are zero). In a recent paper, Judge David R. Strickler raised the importance of opportunity costs in the context ofEfficient Component Pricing Rules (ECPRs) used for setting regulated rates for access to a monopolist's essential inputs. The threat points used in the noncooperative bargaining model used in this report incorporates an identical economic effect to the opportunity cost concept included in the ECPR. See Strickler, David R., Royalty Rate Setting for Sound Recordings by the United States Copyright Royalty Board: The Judicial Need for Independent Scholarly Economic Analysis (December 31, 2015). Review ofEconomic Research on Copyright Issues, 2015, 12(1/2), 1-15. Library of Congress, P3ocket No. 96-5 CARP DSTRA], Determination ofReasonable Rates and Terms for the Digital Performance of Sound Recordings. 63 FR 25407 (May 8, 1998). ("Because the record companies introduced no evidence showing decreased overall sales ofrecords and CDs, the Panel reasonably found that the record companies did not incur additional risk fiom lost sales due to the Services'ctivities.")

28 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

substantial promotional value." According to the Panel, the promotional value came from the constant airplay ofnew types ofmusic not readily accessible in the marketplace, which in turn stimulated record sales. In fact, the record companies'wn expert acknowledged, "there (are)

promotional benefits to recording companies from record companies having their music played on radio stations or digital music services.'"'98)

Testimony from Music Choice's CEO, David J. Del Beccaro, and Music Choice Senior Vice President ofProgramming Strategy and Partnerships, Damon Williams, provide many examples of promotional activities that Music Choice engages in on behalfofthe record labels and performing artists. I will not describe every one ofthese numerous activities here, but instead focus on just one ofthem: the efforts record labels devote to promoting their artists to Music Choice (or co-promoting their artists with Music Choice).

(99) For example, record labels provide Music Choice with "promotional copies" oftheir every new single or album to persuade Music Choice to play these recordings on its service. It is my understanding that the record executives "proactively" call Music Choice to both "solicit feedback" on the artists and discuss promotional strategies.'s Music Choice also conducts in-person presentations with record labels executives to discuss promotional opportunities on Music Choice and collaborate on promoting specific artists and programs. " As discussed in detail in Mr. Williams's testimony, Music Choice provides exposure to lesser-known artists that might not otherwise be heard on commercial radio.

(100) I also understand that the record labels provide commemorative plaques to Music Choice acknowledging their contributions to the record companies'bility to reach certain milestones in record sales and airplay. ' further understand that the record labels and artists regularly provide verbal and written testimonials thanking Music Choice for promoting their records (see Exhibit MC 20).

(101) Perhaps the starkest example ofthe beliefthat music creators feel that Music Choice offers them promotional benefits is the experience they had with the record label re &resenting the artist Chris Brown. As described in Dave Del Beccaro's testimony, [I

.]]" Why would they do this unless they felt that promotion on the

Library of Congress, P)ocket No. 96-5 CARP DSTRA], Determination ofReasonable Rates and Terms for the Digital Performance of Sound Recordings. 63 FR 25407 (May 8, 1998). 63 FR 25408 (May 8, 1998). 78 ibid Testimony ofDamon Williams at 22—23. Testimony ofDamon Williams at 22-23. Testimony ofDamon Williams at 30. Testimony ofDavid J. Del Beccaro at 41&2.

29 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

Music Choice platform would ultimately benefit them (and him) via increased revenue from increased licensing and/or sales of his music?

Music Choice (102) One sometimes hears arguments, if only in these proceedings, that services like cannibalize other revenue sources from record companies, for example from CD sales, digital downloads, or music streaming subscriptions and/or plays. As an economist, I find it completely implausible that record companies would devote (or forgo) substantial resources to promote and/or co-promote artists on Music Choice as I'e summarized above if this were true. Why devote effort and expense to cost yourself money? Instead, I find it far more plausible that Music Choice does provide a net promotional benefit (i.e., net of cannibalization, if any), that artists and record labels recognize that, and that they devote significant resources to maximize that benefit.

both (103) Furthertnore, there is support for the fact that Music Choice provides a promotional benefit to sound recording and musical composition performance rights holders. It is my understanding that Music Choice's on-air displays provide important information about artists that listeners otherwise might not have easy access to." For example, I understand that listeners learn what music is available for purchase using Music Choice's on-screen displays. I further understand that Music Choice has also invested heavily in social media to complement its airplay promotion, increase artist awareness, and help boost record sales and streams.' study by Ipsos OTX MediaCT in July 2016 shows that [[ ]] or Music Choice Music Channel viewers are interested in "being exposed to new artists and music."" This same 2016 study also shows that listener satisfaction is highly favorable for both Music Choice Music Channels and Music Choice On Demand. Music Choice also adds value

because it is a Nielsen-rated network and is popular with millennial audiences, particularly younger audiences.'104)

Taken together, these facts provide an enormity of evidence that Music Choice provides a promotional benefit to sound recording copyright holders. As described in Example 2 above, this promotional benefit, in turn, can have an important impact on the outcomes of negotiations in a hypothetical marketplace between PSS providers and record companies. I return to this point both when I estimate the royalty that would arise in a hypothetical marketplace for PSS sound recording perforinance rights in Section V below as well as when I evaluate the 801(b) condition regarding a fair return for copyright owners and a fair income for copyright users in Section VI below.

Testimony ofDamon Williams at 11-12, 14, 29—30. g4 Testimony of Damon Williams at 13. Exhibit MC 15, Ipsos OTX MediaCT, Music Choice Viewership Study July 2015, at 27. Music Choice Advertising, "Popular with Millennials." h://co orate.musicchoice.com/advertisin /nielsen-audience- rankers/; See also Music Choice 2016 Media Kit, at 19. htt://co orate.musicchoice.com/files/5814/5564/2367/2016 Music Choice Media Kit. df

30 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

IV.C.2.d. Each Firm's Bargaining Power

(105) The last Nash Factor, the Bargaining Power for the record label in its negotiations with a PSS over the surplus, is difficult to determine. Theoretical research in this area has linked bargaining power to each finn's patience in a negotiation." Both record labels and Music Choice have a history of successful negotiations, so there is nothing a priori to suggest that in the hypothetical marketplace, one would be more or less patient than the other. Furthermore, estimating Bargaining Parameters of firms in marketplace settings is a challenging undertaking at the frontier of economic research. In what follows, I will therefore assume that a range of Bargaining Powers is possible. As I think it unreasonable to believe that either a record label or a PSS provider could extract all the profits from a bargain, I choose a range of bargaining powers for each party between 0.2 and 0.8.

IV.C.2.e. Combining the Nash Factors

(106) Combining these Nash Factors yields a simple representation ofthe royalty to each of the record label and the PSS in the hypothetical market for PSS sound recording performance rights. This formula is qualitatively similar to that in the second example above: The PSS provider earns its threat point

(zero) plus half of the Incremental Profits given by the Joint Agreement Profits in the PSS market

plus the promotional benefit to the record labels from the PSS service." The record label does less well from the PSS market: it earns the satne share of the Incremental Profits, but loses the value of the promotional benefit f'rom PSSs embodied in its negative threat position.'107)

In the next section, I describe how I use financial information from Music Choice to estimate each of the Nash Factors and predict a range of royalties that would arise in a hypothetical market for sound recording perfonnance rights in the PSS market.

V. Quantifying the royalty that would arise in the hypothetical market for sound recording performance rights for PSSs

V.A. Summary and Conclusions

(108) The goal ofthis section is to estimate the royalty that would arise in the hypothetical market for sound recording performance rights for PSSs over the period covered by this proceeding, 2018—2022. To do

Abhinay Muthoo, Bargaining Theory with Applications (Cambridge: Cambridge University Press, 1999) at 51. The incremental profits are the difference between the combined agreement surplus and the record label's threat position, the lost promotional benefits (a negative number). As in the second example in the section above, subtracting a negative number yields a positive number and thus a larger pie: the combined agreement surplus plus the record label's promotional benefit in case of agreement. As in example 2, it earns it back on its surplus in those other markets. Thus the overall split of surplus (i.e. including the promotional benefit in the calculation) is equaL

31 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

so, I quantify the three Nash Factors underlying bargaining outcomes in this hypothetical market based on Music Choice's consolidated financial results for the recent past as well as predictions of their near-future financial performance. Where possible, I use the same financial information Music Choice itself uses for its operational and strategic decision-making.

Choice's (109) There are, however, two challenges that I had to overcome in order to use Music financial information to estimate a royalty in the hypothetical market. The first challenge was determining how best to estimate Music Choice's revenues and costs over the 2018—2022 period in light ofrecent changes in their financial performance. As I describe more fully below, I conclude that Music Choice's financial results for the period 2016-2018 represent the best estimate oftheir likely future performance over the years 2018-2022.

Choice's (110) The second challenge was that this proceeding determines reasonable rates only for Music residential audio business. As described in greater detail in Section II above, Music Choice has three lines of business: a residential audio service, a residential video service, and a commercial audio service. Some of Music Choice's subscription fee revenue bundles residential audio and video services and many ofMusic Choice's costs are common: they are used in the production ofboth the residential audio and video businesses. In my report in the last proceeding, some ofthe residential video financial data was included in my calculations, to the displeasure ofthe Copyright Royalty Judges. I have been extremely careful in this presentation to ensure that the financial results I present reflect only the performance ofMusic Choice's residential audio business.

(111) To overcome the problem ofbundled revenues and common costs, I used economic principles and drew on the Financial Accounting literature and replies to detailed questions put to Music Choice's senior financial officers to allocate common costs and bundled revenues to their component lines of business, thereby isolating the revenue and cost (and thus profit) ofMusic Choice's residential audio business. This is the key input into the first Nash Factor, the Joint Agreement Profits. I discuss each ofthese challenges in greater detail and my solutions to them in the subsections to follow.

(112) I similarly estimate each of a record label's and Music Choice's Threat Points and Bargaining Power and combine the three Nash Factors via the Asymmetric Nash Bargaining Solution into an estimate of the royalty, and royalty rate, that would arise in the hypothetical market negotiations between a major record label and Music Choice.

(113) I estimate the Joint Agreement Profits in the hypothetical market for PSS sound recording performance rights ranges from [[ ]] in 2018 to [[ I] in 2022. Despite my strong beliefthat there is a promotional impact that benefits record labels &om Music Choice's service, I am unable to quantify it and therefore conservatively assume that it is zero. In this case, I therefore estimate that each firm's Threat Point is zero. Finally, I assume a range ofBargaining Powers for each of a record label and PSS provider ofbetween 0.20 and 0.80.

32 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

the (114) Given these estimates ofthe Joint Agreement Profits, Threat Points, and Bargaining Powers, Asynuuetric Nash Bargaining Solution provides me with an estimate ofthe royalty, and royalty rate, that would arise in the hypothetical market for PSS sound recording performance rights. Assuming equal bargaining power between record labels and PSS providers, I estimate the PSS royalty would rate of pay between 2018 and 2022 varies between [I ]] to [l J], a royalty 3.5% oftheir (unadjusted) residential audio service revenues. Even in the extreme case ofrecord labels having 80% ofthe bargaining power, I conclude that the royalty rate should be no higher than 5.6% of (unadjusted) residential audio service revenues.

V.B. Predicting Music Choice's future performance

(115) As discussed above, I faced two challenges using Music Choice's financial performance to estimate the royalty that would arise in a hypothetical market for sound recording performance rates for PSSs. This subsection discusses the 6rst ofthose challenges: determining Music Choice's future financial performance.

(116) The royalty rate determined in this rate proceeding will be applicable in the 2018—2022 period. In an ideal world, I would know in advance what will be Music Choice's financial performance over this period. Absent this information, I must make a forecast ofthis performance.

(117) This is made more challenging by recent changes in the performance ofMusic Choice's residential audio business. In previous proceedings, Music Choice has [J J]. In the recent five-

year period, however, [[

]]. Music Choice's decreasing number of subscribers tracks the decline in the cable MVPD subscribers from 55.1 million households in 2013 to 53.7 million in

2014. As shown'118) in the table below, Music Choice's total residential audio revenue, as measured in its internal financial documents, [[

90 Jl See Testimony ofDavid J. Del Beccaro at 23. "Seventeenth Annual Report on the Status of Competition in the Market for the Delivery ofVideo Programming," FCC ("MVPDs. (May 6, 2016) MB Docket 15-158. httos://auns.fcc.uov/edocs uublic/attachmatch/DA-16-510A1 Rcd.pdf. In the last report, we noted that 2013 marked the first-ever decline in total MVPD video subscribers and that the decline was due to a decrease in cable MVPD subscribers. These trends continued in 2014. From year-end 2013 to year-end 2014, the total number ofMVPD video subscribers dropped from 101.7 million to 101.6 million households. While telephone MVPD subscriptions rose from 11.8 million to 13.2 million households and Direct Broadcast Satellite (DBS) MVPD video subscriptions held steady at 34.4 million households, cable MVPD subscription dropped from 55.1 million to 53.7 million households.") (emphasis added)

33 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

W WW R W W M W R

in the (119) These current and projected financials illustrate that Music Choice has had a downward step financial performance of its residential audio business. For example, as shown above actual through Au ust 2016 and ro'ected residential audio revenues for 2016 of

]] On the cost side, Music

Choice is [[ " ]]. Relying on forecasts beyond 2018 requires careful consideration, as revenue rojections are uncertain and depend on both the

(120) Based on these facts, in my opinion, the financial performance forecast by Music Choice for the period 2016-2018 represents the most reliable predictor of Music Choice's likely financial perfonnance in the 2018—2022 period at issue in this proceeding. In the calculations to follow, I

I note here that, due to the inclusion of Music Choice's Video-on-Demand service, in its residential audio revenue that Music Choice uses for internal decision-making, this measure of residential audio revenue is different from the measure that? use in my residual royalty calculation. This issue is discussed in greater detail in Section V.F below. See tMusic Choice, Financial Results as ofAug. 2016]. See Testimony ofDavid J. Del Beccaro at 21-22. l]

'4 Conversations Sept. 6, 2016, Sept. 30, 2016, and Oct. 10, 2016. Testimony ofDavid J. Del Beccaro at 27—28.

See Testimony of David J. Del Beccaro at 27—28. [] R.i] Conversation with Music Choice, CEO, David J. Del Beccaro on Oct. 10, 2016.

34 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

average their 2016-2018 values as my preferred estimates of revenues and costs to predict the royalty that would arise in the hypothetical market for sound recording performance rights for PSS for 2018. 'sing this average predicted perforinance for 2018 as a base, I follow the practice Music Choice uses in its affiliate contracts and inflate their revenues and costs by [[~]] per year to project Music Choice's financial performance for the years 2019-2022.

V.C. Estimating Costs and Revenues for Music Choice's Residential Audio Service

(121) This subsection discusses the second challenge I faced using Music Choice's financial performance to estimate the royalty that would arise in a hypothetical market for PSS sound recording performance rates. This challenge was determining the costs and revenues of Music Choice's residential audio service apart from their residential video and commercial audio businesses.

V.C.1. Music Choice's Consolidated Financial Statements necessitate using Cost and Revenue Allocation Methods to determine the financial performance of its Residential Audio Service

(122) Music Choice's year-end financial statements only include revenues and expenses (costs) at the consolidated (i.e. company-wide) level.'hile Music Choice is a private enterprise, financial reporting at the consolidated level is consistent with accounting guidelines that require that public enterprises must report separate financial inforination for operating segments that are "evaluated regularly by the chief operating decision-maker in deciding how to allocate resources and in assessing perfonnance.' A review of Music Choice's detailed financial statements supports the view that

Given that Music Choice's financial performance in 2016 was superior to that in 2017 and 2018, this is conservative relative to using Music Choice's 2017-2018 financial performance, despite the fact that performance in this (shorter) period in my opinion reflects most accurately Music Choice's predicted financial performance on a going-forward basis. predictions about 1 use the additional year as 1 think it important to rely on as many years as possible when making financial performance. This inflation, is line with Music Choice price escalation, and is lower than the forecast during this period from Congressional Budget Office's Economic Projections for Calendar Years 2016 to 2026 at 34. hti s://www.cbo. ov/sites/default/files/114th-con ress-2015-2016/re orts/51908/51908-breakout-cha ter2-2. df '"" See Music Choice Annual Reports 2011—2015. '"'ummary of Statement 131, FASB, h://www.fasb.or /summ /stsum131.shtml. '"'ASB, Statement ofFinancial Accounting Standards No. 131, 7. (FASB defines operating segment as a component of a business enterprise: ~ "That engages in business activities fi.om which it may earn revenues and incur expenses (including revenues and expenses relating to transactions with other components of the same enterprise) ~ Whose operating results are regularly reviewed by the enterprise's chief operating decision maker to make decisions about resources to be allocated to the segment and assess its performance, and ~ For which discrete financial information is available."

35 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD 0 0 Music Choice does not evaluate the financial performance of its residential audio music service 0 business separately from its other product segments. (123) As described in greater detail in Section II above, Music Choice has three lines ofbusiness: a 0 residential audio service, a residential video service, and a commercial audio service. In order to calculate the Joint Agreement Profits that would arise in a hypothetical market for PSS sound 0 recording performance rights, I must use as inputs just those costs and revenues that accrue to the residential audio business. As some of Music Choice's revenue bundles residential audio and video 0 services and many of Music Choice's costs are so-called "common costs", this is a non-trivial task. To overcome this problem, I use principles from the modern literature on Financial Accounting and replies to detailed questions put to Music Choice's senior financial officers to allocate common costs and bundled revenues to their component parts, thereby isolating the costs and revenues (and thus profit) of Music Choice's residential audio business. The balance ofthis section describes how I did 0 So.

0 V.C.2. Residential Audio Music Service Costs 0 0 V.C.2.a. Cost Allocation Methods (124) As mentioned above, Music Choice's financial statements do not include costs organized by its major 0 revenue-generating product categories — residential audio, residential video, and commercial audio- but by cost categories. In particular, these financial statements include costs under the following 0 major cost categories: (I) rights costs, (2) operations and programming costs, (3) sales and marketing costs, (4) general and administrative costs, (5) depreciation and amortization, and (6) other

0 expenses.'125) In order to calculate the Joint Agreement Profits for a residential audio service alone, the costs in each ofthese categories must be allocated into those necessary for the running ofthe residential audio business and those that aren'. For some costs this is easy: each of Music Choice's three product lines incur certain direct costs. Example of direct costs include the royalty costs for audio and video '0 services that are payable to performing rights organizations and SoundExchange. The identification 0 and allocation of these direct costs to each ofthe services, including the residential audio music ,0 service, is straightforward.

' (126) Any costs that are not directly attributable to a particular service but shared across product lines are 0 grouped together into what are called common costs.'" For example, a major component ofthe 0 '"'ee Privileged and Confidential - Audio Only Model BW 101416B. 0 '04 ("A common cost is a cost of operating a facility or activity or like cost object that is shared by two or more users. Common costs arise because each user obtains a lower cost by sharing than the separate cost that would result if each user operated independently.") Charles T Horngren, Srikant M. Datar, and Madhav V. Rajan, Cost Accounting (New Jersey: Pearson Education Inc., 2015), 611. 0 36 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

Music Choice's common costs are its labor (staffing) costs, which are costs largely incurred at the company level and cannot easily be allocated to any one of its component businesses.

(127) I have reviewed the literature for alternative cost allocation methods and determined that the most reasonable method to use for Music Choice is what is called the Incremental CostAllocation Method. In this method, there is a Primary Product that is "most responsible for the common cost.""'n this method, the analyst determines what are the costs of offering the primary product on a stand-alone basis. I call these costs the Primary Product Costs. The method then identifies Incremental Products, the first ofwhich is the most responsible for the remaining common costs. I call these costs the First Incremental Product Costs. This process continues for as many products there are that share common costs.

(128) Allocating Music Choice's costs to its various product lines proceeded as follows. First, based on answers to detailed questions put to Music Choice's senior financial officers, my team (under my supervision) and I asked Music Choice to allocate all direct costs to the product lines responsible for them.

(129) Then the Incremental Cost Allocation Method was applied to the remaining (common) costs. This required defining which ofMusic Choice's products was the Primary Product and determining what would have been the costs for that product. For Music Choice, determining the Primary Product was straightforward. Music Choice was founded to provide a residential audio service, and the residential audio service remains the most important in terms ofrevenues and company strategy.'he provision of any Incremental Products are only commercially viable due to the existence ofthe residential audio services business. Ifthe residential audio services business were to cease, then Music Choice would cease providing any services (it would close altogether), and common costs would be nil. Music Choice's Primary Product is therefore clearly its residential audio service.

(130) Given the Incremental Cost Allocation Methods, the next step was to determine what would be the costs ofproviding a residential audio service on a stand-alone basis. The next subsection describes this methodology.

V.C.2.b. Music Choice's Costs for a Residential Audio Service

(131) In order to determine what would be Music Choice's costs for a stand-alone residential audio service, I reviewed Music Choice's financial statements and asked the company to prepare a detailed cost

' Homgren, Dater, and Rajan (2015), "Allocation of Support-Department Costs, Common Costs, and Revenues, Chapter 15 in Cost Accounting: A Managerial Emphasis, 15'dition, Pearson Education. '~ Music Choice 2015 Annual Report, 7. MC0003221 ("Pursuant to the General Partnership Agreement dated March 1, 1991, Music Choice a Pennsylvania general partnership, was formed under the Pennsylvania Uniform Partnership Act for the purposes of further developing and maintaining a programming service consisting of digitized music and audio programming. Such purpose was expanded to include the provision ofvideo services pursuant to the Second Restated Agreement [effective Jan. 1, 2004].")

37 0

PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD 0

0 allocation exercise that was consistent with the Incremental Cost Allocation Method described above.'132) In the Incremental Cost Allocation Method, the goal is to identify the costs necessary to provide a residential audio music service on a stand-alone basis. As such, I asked Music Choice to go through its annual consolidated Profit and Loss statements and, for each line item in each year &om 2016- 2022, estimate what would be the costs of that line item in that year ifMusic Choice only offered a residential audio business. 108

(133) Music Choice conducted this exercise in late September and early October of 2016. I understand that senior financial management (including the CEO, David J. Del Beccario), went through each line item at a granular level to determine the stand-alone costs for its residential audio business.'" In the following weeks, my team (under my supervision) and I looked closely at each of their answers and, when we needed clarification about any particular cost allocation, went back to Music Choice with follow-up questions. At the end ofthis process, I was confident that the costs they reported were the costs necessary consistent with the Incremental Cost Allocation Method, i.e, that they represented a residential audio service on a stand-alone basis. to provide (134) Music Choice's cost allocation implied that, on an~aregate basis, a. stand-alone residential audio service would account for between [[ ]] and [[~]] of its total consolidated costs between 2016- 2018." These then were the costs I used in the calculation ofthe Joint Agreement Profits, itself an , market for PSS sound recording input into the royalty that would arise in the hypothetical perfonnance rights.

[ V.C.3. Residential Audio Music Service Revenue

(135) Having determined what would be the costs for Music Choice's residential audio service if it were offered on a stand-alone basis, I turned to the analogous exercise on the revenue side. This subsection describes the methods underlying this exercise as well as how I executed it.

V.C.3.a. Revenue Allocation Methods

(136) As described in Section II above, Music Choice has always bundled its video-on-demand (VOD) '0 service with its residential audio service, but initially at launch in 2010 sold MC Play, its video

'"' cost allocation question asl&s what would be Music Choice's costs in a world different from the one in which it currently operates. The world of interest in this proceeding is one in which Music Choice only offered its residential audio service, and not its residential video and commercial audio services. "ts In this method, the overhead costs are generally allocated to the Primary Product because absent this main product the company may not operate to only offer the Incremental Products. 0 '"9 Testimony ofDavid J. Del Beccaro at 29. "" The range is [t~tt during 2010-2022, when less of the costs re attrib table to resid tial audio. 0 38 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

channel, separately from its residential audio channels to most of its MVPD clients."'n 2012-2013, however, Music Choice began migrating to a pure bundling strategy and now licenses all of its residential audio channels, its VOD service, and MC Play in a single bundle to the vast majority of subscribers that take the MC Play service."

into its (137) Music Choice's internal financial statements breaks down some but not all ofthis revenue various product lines. For example, Music Choice's advertising revenue is always broken down by product line. Advertising revenues are easily allocated to the channels on which the advertising airs, and so can easily be broken down across the residential audio, residential video, and VOD.

revenue it (138) Music Choice does not, however, comprehensively allocate to its product lines all ofthe receives from affiliate fees paid by MVPDs. While it can quite easily separate out commercial audio revenue (as these are billed separately by the MVPDs that license this service), it cannot as easily separate out residential audio f'rom residential video revenue. Thus when an MVPD serving residential customers pays a single rate for audio channels, their video-on-demand product (VOD), and their music video channels (MC Play), I must determine how to allocate this revenue between the residential audio and video services. I do so using methods from the Financial Accounting literature analogous to those described in the previous section regarding cost allocation methods.

(139) I have reviewed the Financial Accounting literature for alternative revenue allocation methods and determined that the most reasonable method to use for Music Choice is what is called the Incremental Revenue Allocation Method. It is conceptually identical to the Incremental Cost Allocation Method described in the previous sub-section, but applied to revenues rather than costs.

(140) In the Incremental Revenue Allocation Method, the analyst first defines a Primary Product. According to the Financial Accounting literature, there are many alternative methods for defining the primary product. For example, it could be "the product in the bundle with the most sales" or it could be determined based on interviews with customers or internal managerial knowledge." The analyst then determines what would be the revenue from offering the primary product on a stand-alone basis. I call these revenues the Primary Product Revenues. The method then identifies Incremental Products, the first of which is the most responsible for any remaining revenue. I call these revenues the First Incremental Product Revenues. This process continues for as many products are bundled and sold at a single price.

(141) Allocating Music Choice's subscription revenues to its various product lines proceeds as follows. As described more fully in the subsection above, both for historical reasons as well as the nature of

"'estimony ofDavid J. Del Beccaro at 28. Conversation with Music Choice, CEO, David J. Del Beccaro, Oct. 10, 2016. '" Horngren, Datar, and Rajan (2015), "Allocation of Support-Department Costs, Common Costs, and Revenues, Chapter 15 in Cost Accounting: A Managerial Emphasis, 15'" edition, Pearson Education.

39 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

Music Choice's current sales and strategic considerations, it is clear that Music Choice's Primary Product is its residential audio service. This means that its residential video service, the only remaining product line whose revenue is bundled with residential audio revenue, is its First Incremental Product.

(142) Having made these decisions, all that remained was to determine the revenue that would arise &om providing a residential audio service on a stand-alone basis. The next subsection describes how I determined this.

V.C.3.b. INusic Choice's Revenues for a Residential Audio only Service

(143) In order to determine what would be Music Choice's revenues for a stand-alone residential audio service, I reviewed Music Choice's financial statements and asked the company to estimate current and future revenues that would be consistent with the Incremental Revenue Accounting Method described above.

(144) In the Incremental Revenue Accounting Method, the goal is to identify the revenues that would arise from offering a residential audio music service on a stand-alone basis. As such, I asked Music Choice to go through its annual consolidated Profit and Loss statements and, for each year &om 2016-2022, estimate what would be the revenues that would arise in that year ifMusic Choice only overed a residential audio business.

(145) Music Choice conducted this exercise in late September and early October of2016. It required separating revenues that accrued &om the bundled licensing ofresidential audio channels, its VOD service, and its music video channels, MC Play. For its music video channels, MC Play, Music Choice could rely on the same methods it has long used internally for allocating this revenue for the purpose of video rights fee payments. While I understand that the exact calculation varies by customer, Music Choice broadly allocates a value to the video channels equal to the value it was able to receive in the market for those channels when they were sold on a stand-aLone basis, adjusted for changes in the marketplace over time." On an aggregate basis, these estimated video channel license fee revenues account for (are estimated to account for) between [I~1] and [I~lj of its combined unadjusted residential audio and video (i.e. bundled) license fee revenues between 2016- 2022.

(146) Having pulled out its music video channel revenue, all that remained was to separate residential audio revenue &om video-on-demand (VOD) revenue. As noted earlier, Music Choice has never charged a separate price for its VOD service, instead bundling it with its audio channels in an effort to slow year-on-year decreases in the per-subscriber fee it could negotiate with its MVPD clients. This

As described more fully above, Music Choice offered its music video channels on a stand-alone basis in new contracts from 2010-2012, but migrated to bundled pricing Irom 2012—2013. Legacy contracts have maintained separate pricing for Music Choice play into 2016, but at an ever-decreasing incidence. See also Testimony ofDavid J. Del Beccaro at 30.

40 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

complicates efforts to determine a reasonable proportion ofthe bundled revenue that would accrue only to a residential audio service.

(147) That being said, Music Choice has [[

(148) I confirmed this answer (and the reasons for it) with senior Music Choice personnel in early October 2016. Indeed, an estimate of average adjustment of [[ 1'] (using the [l ]] of residential audio allocated to VOD for its customers with VOD services) is more conservative compared with the relative usage ofMusic Choice's video services and the share ofrevenue ([I ]]) allocated to their Music Channels in Music Choice's existing revenue allocation methods. In )articular, Music Choice's VOD service is available in nearly [I 1] households and has [[ 1] plays per month, whereas its Music Video channel (MC Play) is only available in [[ ]] households and has tens ofthousands ofplays per month." Despite this, I elected to be conservative and only reduced their bundled residential audio snd VOD revenue by [~]] when calculadng the revenue that their residential audio service would earn on a stand-alone basis. In the balance ofthis report, particularly the Appendices, I call the adjustment Music Choice made on residential audio revenues their "VOD-adjusted residential audio revenues", I call my more conservative, [~], reduction of their residential audio revenues their "[Wl] VOD-adjusted residential audio revenues", and I call the residential audio revenues they use in their internal financial statements their "unadjusted residential audio revenues".

(149) At the end ofthis process, I was confident that the revenues they reported were consistent with that required by the Incremental Revenue Allocation Method, i.e. that they represented the revenues that would arise from a residential audio service offered on a stand-alone basis. After making my adjustment to be even more conservative, I used the [~]] VOD-adjusted residential audio

Conversation with Music Choice CEO, David J. Del Beccaro, Sept. 30, 2016. See also Testimony ofDavid J. Del Beccaro at 30-31. ne Testimony ofDavid J. Del Beccaro at 31. "7 Conversation with Music Choice CEO, David J. Del Beccaro, Sept. 6, 2016. See also Testimony ofDamon Williams at 12.

41 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

revenues in the calculation ofthe royalty that would arise in the hypothetical market for PSS sound recording performance rights. I describe next how I calculate this royalty.

V.D. Quantifying the royalty that would arise in the hypothetical market for PSS sound recording performance rights

between (150) Given my method of calculating the costs and revenues that would arise for Music Choice 2018—2022 for a residential audio service offered on a stand-alone basis, I had all the information I needed to estimate the royalty that would arise in the hypothetical market for PSS sound recording perfonnance rights. To do so, I calculated the Joint Agreement Profits, Threat Points, and Bargaining Powers that would arise in a hypothetical negotiation between an individual major record label and a PSS. This subsection describes the details underlying these calculations.

V.D.'i. Calculating Joint Agreement Profits

V.D.1.a. Economic Principles

(151) As described above, the Joint Agreement Profits are the combined profits to both the upstreain and downstream firms from reaching an agreement. The profits" that should be used when constructing the Joint Agreement Profits are economic profits, a concept distinct from accounting profits. Accounting profits "stet with revenues and then deduct costs, such as wages, raw material costs, overhead, and taxes." '" These costs from the Income Statement are often referred to as operating costs. Operating costs are costs that are needed to operate the (here, residential audio music service) business and can be measured using the costs that would arise were Music Choice to offer a stand- alone residential audio service as described in Section V.C.2 above.

(152) As noted in the Corporate Finance literature, "there is one cost that [accounting profitsj do not commonly deduct: the cost of capital."'" The cost of capital is the "return required by investors" to invest in a business.'conomic profits are accounting profits less capital expense, the distinction between which is important: "[A] business that breaks even in terms of accounting profits is really ' making a loss; it is failing to cover the cost of capital."

'" Richard A. Brealey, Stewart C. Myers, and Franklin Allen, Principles ofCorporate Finance (New York: Music ChoiceGraw-Hill, 2006) at 310. Ibid. ' Ibid at 311. ' Ibid at 310.

42 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

or residual (153) In the literature, economic profit is often called either economic value added (EVA) income. 't is calculated as follows: EVA = residual income = income earned — capital expenses = income earned — (invested capital x cost of capital)

(154) Residual income is thus simply income earned (or accounting profit) minus capital expenses." Income earned is the total after-tsx operating income earned by the company's invested capital base.' 'he (after-tax) operating income for Music Choice's residential audio business is based on the revenue it would earn minus its direct and allocated total costs.'he capital expenses for Music Choice's residential audio business are just Music Choice's invested capital base times their cost of capital. The next subsection describes how I estimated these for the relevant period covered by this proceeding.

V.D.1.b. Estimating Music Choice's capital expenses

(155) Estimating capital expenses requires determining both the amount of invested capital and the cost of capital. I estimate Music Choice's capital expenses for its residential audio music business &om — 2018—2022 in a two-step procedure. First, I estimate their average capital expenses in the 2013 2015 period using standard techniques f'rom Corporate Finance. I then use this information to estimate what would be their capital expenses in the 2018—2022 period. I describe each ofthese steps in what follows.

V.D.1.b.i. Estimating Music Choice's Audio Only Invested Capital from 2013-2015

(156) The amount of invested capital, or invested capital base, for a company is the sum of its operating working capital (which I will just call working capital), fixed assets, intangible assets, and net other long-term operating assets.'he invested capital is the equity base on which the investors expect to

'~ Richard A. Brealey, Stewart C. Myers, and Franklin Allen, Principles ofCorporate Finance (New York: Music ChoiceGraw-Hill, 2006) at 311. See also Tim Koller, Mare Goedhart, and David Wessels, Valuattonr Measuring and Managing the Value ofCompanies, = — (New Jersey: John Wiley and Sons, Inc.: 2005), 116. (Economic profit NOPLAT (invested Capital x WACC), where NOPLAT stands for net operating profit less adjusted taxes, and WACC stands for weighted average cost of capital.) "cost '~ For clarity, I refer to the total cost to the firm ofits capital as "capital expenses". This distinguishes it Rom the of capital", which is the percentage return required by investors to invest in a business. '~ ("NOPLAT [net operating profit less adjusted taxes] represents the total afier-tax operating income (generated by the company's invested capital) that is available to all financial investors.") Tim Koller, Mare Goedhart, and David Wessels, Valuation: Measuring and Managing the Value ofCompanies, (New Jersey: John Wiley and Sons, Inc., 2005), 160. ' Music Choice has minimal taxes and for simplicity, I have excluded this in my calculation, which is also conservative. nn For discussion of direct cost and cost allocation to the residential audio music service see Section V.C.2.b. ("Most financial analysts, however, separate invested capital into operating working capital (current operating assets less current operating liabilities), fixed assets (e.g., net property, plant, and equipment), intangible assets (e.g., goodwill) and

43 0

0 S. Crawford, PhD PUBLIC VERSION Written Direct Testimony of Gregory

earn a return. For all of these components, I only used the portion that is required to support the residential audio business in my calculations.

fixed assets, 0 (157) As described above, a finn's invested capital is a combination oftheir working capital, intangible assets, and net other long-term operating assets. For Music Choice, the two most important 0 components are their working capital (consisting of their current assets less current liabilities) and 0 their long-tenn assets (i.e., fixed assets, measured by their property, plant, and equipment (PP&E)).

— (158) I estimated the working capital for Music Choice's residential audio service's during 2013 2015 by consolidated (company-wide) level and then allocating a measuring their working capital at the portion to their residential audio service based on the ratio ofthis segment's revenue to total 0 revenue.'usic Choice's consolidated working capital is the sum of their current operating assets 0 less their current operating liabilities obtained &om their balance sheet in each year.' The top panel of Appendix B.2 shows these calculations in each year from 2013— 15.

To estimate the long-term assets that support Music Choice's residential audio business, I asked (159) Music Choice, again, what portion of its fixed assets, intangible assets, and net other long-term 0 operating assets from the consolidated level are required for the audio business on a stand-alone basis. 0 The middle panel of Appendix B.2 shows the values ofthese assets for Music Choice's residential 0 audio business between 2013—2015. [160] Combining all the sources of invested capital yields values &om 2013 to 2016 ranging iiom ~ 0 ~]] to [[~]], with an average invested capital of [~]] over the 2013— 2015 period. 0 V.D.1.b.ii. Estimating Music Choice's Cost of Capital 0 (161) I estimated Music Choice's cost of capital, or the expected return to its investors, for its residential audio business using the capital asset pricing model (CAPM).' The CAPM is one ofthe most widely 0 used method for estimating a firm's cost of capital and is based on the principle that the "expected net other long-term operating assets (net of long-term operating liabilities)." Tim Koller, Mare Goedhart, and David Wessels, Valuation: Measuring and Managing the Value ofCompanies, (New Jersey: John Wiley and Sons, Inc.: 2005), 166-167. 0 I used the revenue share to allocate working capital as an important component of their working capital are their accounts receivables which are driven by revenue share, 0 The operating cash for each year is based on the lowest cash to total revenue ratio (i.e., [[~j]) during the last three years (2013 —2015). "'rofessor William Sharpe, one of the three economists who contributed to the development of the asset pricing model in 0 the 1960s, received a Nobel Prize "for his contributions to the theory of price formation for financial assets, the so- called, Capital Asset Pricing Model (CAPM)." See NobelPrize.org Press Release, "This Year's Laureates Are Pioneers in The Theory ofFinancial Economics and Corporate Finance," C]ct. 16, 1990. 0 htt://www.nobel rize.or /nobel rizes/economic-sciences/laureates/1990/ ress.html 0 44 0 0 VERSION Written Direct Testimony of Gregory S. Crawford, PhD P UB LI C

return equals the risk-free rate plus a risk premium."" I assume that the cost of capital for Music Choice's residential audio music service is equal to its overall average cost of capital.

Constant (162) The risk-free rate I use in the model is the current risk-&ee rate on long-term (i.e., 20-year) Maturity Treasury, which was 2.2% as of October 13, 2016.'he long-run historical equity market- risk premium of 6.9% is obtained f'rom the 2016 Valuation Handbook, an industry standard publication."

"beta", (163) Another piece of information needed to measure Music Choice's cost of capital is its which

measures the riskiness of investing in Music Choice relative to the overall market. If it was a publicly

traded company, Music Choice's beta could be observed based on their stock market returns."'ince Music Choice is not a public company, its beta can instead be estimated based on other publicly traded companies that are deemed comparable for the purposes of estimating beta."'164)

The beta of 0,8% used in the CAPM model is a simple average ofthe unlevered-betas of seven public companies that offer music channels and other cable channels. I estimated the beta by first identifying the companies and then using Bloomberg input data to perform beta the calculation. I identified two ofthe seven comparable entities, for the purposes of estimating beta, based on SNL Kagan data: Viacom and Scripps Networks."'o obtain a larger set of comparable companies, I selected five additional companies that offer basic cable channels: Walt Disney, Twenty-first Century Fox, Discovery Communications, Time Warner, and CBS Corp." I then obtained each company's equity betas from Bloomberg, which I adjusted for leverage." It was necessary to adjust for leverage

"2 Richard A. Brealey, Stewart C. Myers, and Fratiklin Allen, Principles ofCorporate Finance (New York: Music ChoiceGraw-Hill, 2006) at 189, 215, 219. Cost of capital = Risk free rate + risk premium = Risk free rate + (beta x market risk-premium] '" See U.S, Deparhuent of Treasury, Daily Treasury Yield Curve Rates. h s://www.tress . ov/resource-center/data- chart-center/interest-rates/Pa es/TextView.as x?data- ield. See Duff and Phelps, 2016 Valuation Handbook, Guide to Cost of Capital, (New Jersey: John Wiley & Sons, Inc., 2016), Appenix 3s. "-'eta "tells us on how much on average the stock price changed for each additional 1 percent change in the market index." Richard A. Brealey, Stewart C. Myers, and Franldin Allen, Principles ofCorporate Finance (New York: Music ChoiceGraw-Hill, 2006) at 219. 's'he following publicly traded companies were used as comparable companies for the purpose ofestimating asset beta: Viacom, Scrips Networks, Walt Disney, Twenty-First Century Fox, and Discovery Communications. "'NL Kagan data shows music cable channels offerings in the US and I mapped these music channels to three owners (Viacom, Scripps Network and MSG Networks). I ultimately excluded MSG Networks because its leverage (debt/equity ratio) is substantially higher than all the other companies in the set, and therefore an outlier. For instance, this list is similar to the peer group companies listed by Viacom, the company with the most number of music channels, in its 2015 10-K. See Viacom 2015 10-K at 30. ("... peer group of companies comprised of The Walt Disney Company, Twenty-First Century Fox Inc., Time Warner Inc., CBS Corporation, Discovery Communications, Inc. and Scripps Network Interactive Inc.") The unlevered-beta = Equity beta / [I+ (1 — tax rate) / (Debt/Equity)]. See also Aswath Damodaran, Investment Valuation, (New York: John Wiley Sc Sons, Inc., 2002), 194.

45 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

because, unlike the comparable companies, Music Choice does not have any debt. The value ofbeta for the comparable public companies is 0.8%.' '"'165)

Combining a risk-f'ree rate of 2.15%, a beta of 0.8, and a market risk premium of 6.9%, one could estimate that Music Choice's cost of capital is 7.7%.'his calculation would be more conservative than an alternative approach from Duff and Phelps that is often also used in such settings.'"'166)

The market risk premium from this is based on the average firm on the market. This is inappropriate as Music Choice is small relative to the average firm and therefore investing in it would require Music Choice to provide a higher return to investors.'"" To account for Music Choice's smaller size, I add a risk premium on top of the market-based calculation in the last paragraph. The historical premium over CAPM for smaller companies comparable in size to Music Choice is [[~]].tss Thus, the cost of capital for Music Choice after including the size premium is [[~]].'44 My estimate of cost of capital is conservative as it is lower than the average for companies in the television broadcasting industry.'"

V.D.1.b.iii. Estimating Music Choice's Capital Expenses 2016-2022

(167) Based on the analysis in the two precedin subsections, I combine my estimate of Music Choice's average invested capital base of [[ t] with my estimate of Music Choice's cost of capital

[4'loomberg. i u The unlevered bctas published by Prof Aswath Damodaran for the entertainment and broadcasting industries of 0.94 and 0.74, respectively provides additional support for my estimate of 0.80. (Aswath Damodaran, "Betas by Sector (US)," data as ofJan. 2016. As accessed at htt:// a es.stern.n u.edu/-adamodar/New Home Pa e/datafile/Betas.html on Oct. 10, 2016.) [7 7% = 2,2% + (0.8 x 6.9%)] '" normalized risk-free rate of 4.0% and recommended equity risk premium of 5.5%. In this setting, Duffremium-fi& Phelps p-5-0-to-5-5use a Music Choice's estimated cost of capital would be 8.4% (i.e. 4%+ [0.8 x 5.5%]). See htt://ww.duffand hei s.com/insi hts/ ublications/cost-of-ca ital/duff- hei s-increases-recommended-us-e ui -risl- "4 The dnsnciat statements ofMusic Choice show BBITDA of [~[] and [~tt for 2015 and 2014 respectively. Duff & Phelps publishes estimates ofthe Smooth Premium over CAPM (size premium) for companies grouped by average EBITDA level (the published EBITDA closest to Music Choice's is il). It published a premium over CAPM of [[~]l for companies of similar size with an average EBITDA of ]] for the 5- years ending December 31, 2015. [Music Choice and Subsidiary, Consolidated Financial Statements, (Dec. 31, 2015 and 2014) at 4. (MC0003221)] and [Duff & Phelps, "2016 Valuation Handbook: Guide to Cost of Capital," (Hoboken, NJ: John Wiley & Sons, 2016), Exhibit B-6, Companies Ranked by EBITDA, Premia over CAPM). ]ds Ibid. 14() [[ ll [4'y cost of capital estimate is also supported by Duff & Phelps Valuation Handbook's calculated Cost ofEquity Capital (CAPM+ size premium) for the SIC 4833, Television Broadcasting Stations, of 18.5% (median) and 14.5% (SIC Composite), using data through March 31, 2016. The industry description for SIC 4833 is "establishments primarily engaged in broadcasting visual programs by television to the public, except cable and other pay television pay stations." (Duff & Phelps, "2016 Valuation Handbook: Industry Cost of Capital," (Hoboken, NJ: John Wiley & Sons, 2016), n.p.) My estimate is also conservative because I have not included any additional risk premium for lack of marketability or company specific risk-factors that may be applicable.

46 1 ~ 0

0 Crawford, PhD PUBLIC VERSION 0 Written Direct Testimony of Gregory S. 0 between 2013 and of [[ [] to calculate an average annual capital expense of [i ]] 0 2015. be 0 (168) IfI wanted to calculate Music Choice's capital expenses only between 2013 and 2015, I would finished. But I need instead to calculate Music Choice's capital expenses between 2016 and 2022.

2016- 0 (169) Rather than ask Music Choice to predict what would be their balance sheet in each ofthe years 0 2022 and use same approach as I did above for 2013-2015, I followed a simpler approach. I 0 averaged the value oftheir capital expenses across these years and calculated the ratio ofthis average capital expense to the average oftheir unadjusted residential audio service revenues across 2013— 2015. I call this Music Choice's average capital expense ratio. Music Choice's capital expense ratio 0 varied between [[Rl] and [[R]] across these years and was [[%[i] for the average year in this 0 period. 0 (170) As the average capital expense ratio is reasonably stable across years, I assumed that Music Choice's capital expenses as a share of its unadjusted residential audio service revenues would also be stable in future years. In what follows, I therefore calculate Music Choice's capital expense for each year in the period 2016-2022 as [[Q]] of its predicted unadjusted residential audio service revenue for that year.

V.D.1.c. Joint Agreement Profits (171) I now have all the inputs needed to calculate the first ofthe Nash Factors for the hypothetical market for PSS sound recording performance rights: the Joint Agreement Profits. To estimate the Joint Agreement Profits, I performed the following'steps: hypothetical

a. As described in detail in Section V.C.2 above, I asked Music Choice's to provide income statements that would reflect the revenues and costs that would have arisen over the period stand-alone business. This 0 2016-2018 had they overed their residential audio service as a information is included in the first three columns ofAppendix B.l. In constructing this spreadsheet, I made two adjustments to the raw information given to me by Music Choice:

i. I amortized several ofMusic Choice's legal expenses over time to better reflect expected expenses on an annual basis.'

excluded legal costs related to PSS-III proceeding in 2016 and 2017 because it relates to rate proceeding period 2018- 2022. Instead, I took costs incurred by Music Choice during 2011-2014 for PSS-II rate period (2013—2017) of and used the annual average costs for 2016 and 2017 and the total forecast costs incurred in 2016-2017 , and used the annual average costs for 2018. for the (current) PSS-III proceeding [[ i] 0 Furthermore, Music Choice incurs patent litigation on a recurring basis as a regular course ofbusiness. Music Choice's 2016 and 2017 P8Q. include such costs for an ongoing patent litigation with the Canadian cable radio provider, Stingray. 0 Based on conversations with Music Choice senior management, I take these costs and average them over [~] years going forward as Music Choice has historically incurred such patent costs every [[~ years. (Conversation with

47 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

ii. As discussed above, rather than rely on the full Video-on-Demand adjustment to their residential audio revenue proposed by Music Choice, I chose to be conservative and used a smaller, [[~,]] adjustment.

b. As discussed further in Section V.B above, I believe the financial performance forecast by Music Choice in the period 2016-2018 represents the most reliable predictor of Music Choice's likely financial performance in the 2018—2022 period at issue in this proceeding. Balancing the desire to rely on multiple years of financial performance against relying too much on forecast performance, I average the values from 2016-2018 and use these as my preferred estimates ofMusic Choice's revenues and costs to predict the royalty that would arise in the hypothetical market for PSS sound recording performance rights in 2018. Using this average predicted performance for 2018 as a base, I then allow for an increase of 1.5'/o per year in both Music Choice's predicted revenues and operating expenses for the period 2019-2022. This information is included in the remaining columns ofAppendix B.l.

c. This provided me with estimates of Music Choice's operating income in the period 2018— 2022. As discussed in Section V.D.l.b above, I augmented this with an estimate of Music Choice's capital expenses over the same period. To do so, I relied on the consolidated Balance Sheet Music Choice provided to me for the period 2013 —2015 to construct an estimate ofthe average capital expense ratio (capital expenses as a share of unadjusted residential audio service revenue) of [[~]]. This information is included in Appendix B.2. I use this average capital expense ratio to calculate the capital expenses for each year, 2018— 2022, covered in this proceeding.

d. I then combined all this information to construct the Joint Agreement Profits as illustrated in Appendix B.3.

i. Appendix B.3 provide the same information as in Appendix B. 1, but with greater detail necessitated by the calculations that will be described in what follows. Reported in bold halfway through the table are Music Choice's predicted operating profits from its residential audio service if offered on a stand-alone basis (using a [[~]] VOD-adjusted estimate oftheir residential audio service revenue).

ii. As discussed in footnote 79 in Section IV.C.1, the Joint Agreement Profits require two adjustments to these operating profits. I first remove &om Music Choice's estimated 2018—2022 operating profit the royalty that would be paid by Music Choice for digital performance rights for sound recordings. The reason is that this is the object of interest in the bargaining model and should therefore not be included when calculating the Joint

Music Choice CEO, David J. Del Beccaro, Oct. 10, 2016) Music Choice had another litigation with MTV/Viacom that occurred approximately [t~t]. 48 0 0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD 0

Agreement Profits. The result is equal to Music Choice's royalty-adjusted operating 0 their operating income. 0 income reported in bold two rows below 0 iii. As described in the previous subsection and summarized above, I then multiply Music 0 Choice's unadjusted residential audio revenue by the average capital expense ratio of [[~]] to calculate in each year an estimate of Music Choice's capital expenses. In each year, I subtract Music Choice's capital expenses from its royalty adjusted operating profits 0 for its residential music services business to derive its residual profits each year. This is 0 my estimate of the Joint Agreement Profits in the hypothetical market for PSS sound 0 recording performance rights in each year. It ranges ]rom [[~]] in 2018 to [[~]] in 2022. 0 V.E. Calculating Threat Points and Bargaining Power

(172) The Joint Agreement Profits is the first Nash Factor needed to calculate an overall royalty, and 0 royalty rate, for the hypothetical market for PSS sound recording performance rights. The two others 0 are each party's Threat Point and each party's Bargaining Power. I describe how I estimate these in this subsection.

0 V.E.1.a. Estimating Each Party's Threat Point

(173) As described in Section IV.C.2 above, I conclude that Music Choice's threat point is zero. That is to say that in the absence of an agreement between Music Choice and a record label, Music Choice 0 would not be able to offer a viable residential audio service and would therefore have economic 0 profits of zero.'" 0 (174) In the same section, I conclude that a record label's threat point is likely to be less than zero. That is '0 because there are two elements to the record label's threat point. The first is their profit in the market for residential audio services in the absence of an agreement with Music Choice. I conclude that this portion oftheir threat point is zero. As for Music Choice, if a record label couldn't reach an 0 agreement, it wouldn't earn any royalties from subscribers to digital cable television packages that include Music Choice and therefore wouldn't earn any profits from the residential audio market in the absence of an agreement.

, ~ (175) But there is a second tenn in the record label's threat point: profits from other markets. To the extent that subscribing to a digital cable television package that includes Music Choice causes some of those 0 households that listen to Music Choice's audio channels to purchase music from a record label's other 0 '40 As described in detail at the beginning of Section IV.C.2.c, this assumption is conservative: if Music Choice could offer a viable service in the absence of any record label's catalog (whether major or independent), it's Threat Point would be 0 higher and it would pay a lower overall royalty. 0 49 0 0 0

0 S. Crawford, PhD PUBLIC VERSION 0 Written Direct Testimony of Gregory 0

distribution channels (e.g. they buy a CD or pay for a digital download or listen to a song on an Interactive Webcasting service), failure to reach an agreement with Music Choice will mean also the 0 loss of this additional revenue. I conclude that this portion of a record label's threat point is less than 0 zero. 0 (176) Unfortunately, I do not have an estimate ofthe economic profits associated with the promotional benefits to record companies of the Music Choice service. As a result, I exclude this promotional 0 benefit from my analysis. This lowers the Joint Agreement Profits to be divided and effectively allocates the entirety of any promotional benefit to the record labels. This has two effects: it makes 0 my estimate ofthe Joint Agreement Profits a lower bound on its true value and it overestimates the 0 royalty that should be paid by Music Choice to the record label. As a result, this analysis of Threat Points disadvantages Music Choice and is therefore likely to be conservative, V.E.1.a. Estimating Each Party's Bargaining Power (177) In Section IV,C.2 above, I survey the existing academic research on estimating firms'argaining power from both a theoretical and empirical perspective, Unfortunately, none ofthis research applies

directly to the market for residential audio services, As such, there is no direct evidence on the relative bargaining power of either a record label or Music Choice in a hypothetical market for sound

recording perfonnance rights for PSSs. (178) In the absence of evidence, I assume that there could be a range of Bargaining Powers for each patty. extract all the As I think it unreasonable to believe that either a record label or a PSS provider could 0.8."'.F. profits from a bargain, I choose a range of Bargaining Powers for each party between 0.2 and Estimating a hypothetical marketplace royalty needed to estimate the royalty, and the royalty rate, (179) I now have estimates of all of the Nash Factors that would arise in the hypothetical marketplace for sound recording performance rights for PSSs for the period 2018—2022.

(180) As described in Section IV.C.1, the Asymmetric Nash Bargaining Solution demonstrates that the 0 royalty (in dollar terins) that would arise in the hypothetical market is equal to: 0 The profit received by each firm in a bargain equals its Threat Point plus its 0 Bargaining Power times the Incremental Profits.' 0 0 "" As the Bargaining Power of one side is just I minus the Bargaining Power of the other side, an estimate of Bargaining Power of 0.8 for Music Choice implies an estimate of 0.20 for a record company. * nu Mathematically, we would write this as (Profit = Threat Point+ Bargaining Power Incremental Profits) 0 50 0

( ~ 0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD 0

(181) As also described there, the Incremental Profits, or "size ofthe pie to be split" in a bargaining environment are just the Joint Agreement Profits less the sum of each party's Threat Point. As each party's threat point in the hypothetical market is (conservatively) assumed to be zero, the Incremental Profits are the same as the Joint Agreement Profits. In this case, the royalty that would arise in the hypothetical market is equal to:

The profit received by each firm in the hypothetical market for PSS sound 0 recording performance rights equals its Bargaining Power times the Joint

Agreement Profits.'iven

(182) my estimate of Joint Agreement Profits of [[~]] in 2018 and a range of Bargaining Powers between 0.2 and 0.8, my estimate ofthe royalty that would arise in the hypothetical market 0 for sound recording performance rights for PSSs in 2018 is []~]] and certainly no higher than [['$~[]. My estimates ofthe range of royalties that would arise in the other years covered in the proceeding are included in in Appendix B.3.

0 (183) As the statutory rate is typically calculated as a percent of revenue, I next convert these estimated royalties into an estimated royalty rate. I do so with care, however, as the revenue in the first row of Appendix B 3, and used to calculate the Joint Agreement Profits is the [[~]] VODudjhsred residential audio service revenue and differs from the unadjusted residential audio revenue on which 0 Music Choice has historically paid royalties. 0 (184) I assume that the Copyright Royalty Judges will assign a royalty rate based on the long-standing Choice has used for reporting its residential audio royalties, i.e. the reporting , ~ method that Music method that does not adjust for its VOD offering. As such, to construct a royalty rate, I divide my estimates of the range ofroyalties in the hypothetical market by Music Choice's unadjusted residential audio service revenue. For the 2018—2022 period, my estimate of the royalty rate that , ~ would arise in the hypothetical market for PSS sound recording performance rights is 3.5% (assuming ,0 an equal split of Joint Agreement Profits) and certainly no higher than 5.6% (assuming SoundExchange has a bargaining power of 0.8). My estimates of the royalty rate that would arise in 0 the other years covered in the proceeding are included in the last three rows in Appendix B.3. 0 0

Mathematically, we would write this as (Profit = Joint Agreement Profits * Bargaining Power)

0 51 Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC VERSION

0 Vl. Evaluation of the statutory factors

VI.A. Overview

(185) In the last section, I concluded that the royalty that would arise in a hypothetical marketplace for PSS sound recording performance rights would be 3.5% ofPSS revenue, and under no circumstances greater than 5.6% ofPSS revenue. The PSS license standard, however, provides for a "reasonable 0 royalty" that is set without reference to a market rate other than having a market rate as its absolute

upper , ~ boundary."'186) I further understand that the reasonable royalty rate is set based upon evaluation ofthe following

i ~ policy objectives: i ~ ~ To maximize the availability of creative works to the public;

~ To afford the copyright owner a fair return for his creative work and the copyright user a fair income under existing economic conditions;

~ To reflect the relative roles ofthe copyright owner and the copyright user in the product made available to the public with respect to relative creative contribution, technological contribution, capital investment, cost, risk, and contribution to the opening ofnew markets for creative expression and media for their communication; and

~ To minimize any disruptive impact on the structure ofthe industries involved and on 0 generally prevailing industry practices."

(187) I understand that a full discussion of each ofthese factors and the impact they should have on a reasonable rate for PSS sound recording performance rights is included in the statement ofDavid J. Del Beccaro.'" My goal in this section is to augment that analysis by providing my perspective on each ofthese factors.

Testimony ofDavid J. Del Beccaro at 7. See also Recording Industry Association ofAmerica v. Librarian of Congress, 176 F.3d 528 (1999) ("The Court ofAppeals, Harry T. Edwards, ChiefJudge, held that: (1) reasonableness ofrate established for compulsory license could be determined based on four statutory objectives rather than market rates..."); 0 See also Recording Industry Association ofAmerica v. Copyright Royalty Tribunal, 662 F.2d 1 (1981) at ("The statute requires the Tribunal to establish a "reasonable" royalty rate calculated to achieve the statutory objectives. We adopt the view of RIAA that: A rate that is deliberately fixed above the level that the market can bear so that a lower rate can be negotiated in the marketplace cannot be 'reasonable.'uch a rate would yield more than the 'fair return'o copyright owners mandated by the statute.") 0 '~4 17 U.S.C. 801(b)(1) htto://www.couvriaht.sov/title17/92chapg.html. 'ss Testimony ofDavid J. Del Beccaro at 19-46. See also Recording Industry Association ofAmerica v. Librarian of Congress, 176 F.3d 528 (1999) ("The Court ofAppeals, Harry T. Edwards, ChiefJudge, held that: (1) reasonableness of 0 rate established for compulsory license could be determined based on four statutory objectives rather than market 52 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

(188) In what follows, I describe how Music Choice increases consumer welfare not only by maximizing the availability of creative works, but also providing consumer benefits beyond the production and distribution of more new artists and new songs. I also describe how Music Choice did not receive a fair income as a copyright user under the existing rate of 8.5'/o because its planned increase to 300 channels never occurred. I also describe how even under modest assumption of promotional effect of Music Choice increases record labels return, while decreasing Music Choice's fair income. I also briefly discuss the relative contributions and investments made by record labels and Music Choice in the PSS market and the potential for disruptive impact based on the sound recording performance royalty.

VI.B. Maximizing the availability of creative works to the public

(189) As discussed above, one ofthe statutory factors determining a "reasonable rate" for PSS sound recording perfonnance rights is to maximize the availability of creative works to the public.

VI.B.1.a. Products must cover their fixed costs in order to be produced

(190) For any product, it can be difficult to be sure that markets succeed in providing consumers with the goods that they desire. This is particularly true for media goods like music, where it is relatively costly to produce the first copy of a recorded song, but essentially f'ree to produce additional (digital) copies." Even more costly is to ensure even modest distribution of a new song: this requires promotion, both in the form of employee time and complementary materials, in order to make the many and varied distribution channels for music aware of the song and to encourage them to play it.

(191) Economics has long established that firms will only produce goods ifthe variable profits they can earn on those goods exceed the fixed costs required to produce and distribute them. Joel Waldfogel, one of the nation's preeminent media economists, analyzed the implications of this fact for the products that are brought to market across a variety of industries, including media industries, in a book titled "The Tyranny ofthe Market"."

rates... "); See also Recording Industry Association of America v. Copyright Royalty Tribunal, 662 F.2d 1 (1981) at 12. ("The statute requires the Tribunal to establish a "reasonable" royalty rate calculated to achieve the statutory objectives. We adopt the view ofRIAA that: A rate that is deliberately fixed above the level that the market can bear so that a lower rate can be negotiated in the marketplace cannot be 'reasonable.'uch a rate would yield more than the 'fair return'o copyright owners mandated by the statute.") "'or the first copy, the artist needs a recording studio with high-end equipment to ensure a high-quality recording. Additional digital copies are trivial to make and distribute. Joel Waldfogel, one of the nation's preeminent media economists, analyzed this question across a variety of industries, including media industries, in a book titled "The 't Tyranny of the Market" (Joel Waldfogel, The Tyranny ofthe Market: 0rhy you can always get what you want, (President and Fellows ofHarvard College, 2007). Joel Waldfogel, The Tyranny ofthe Market: Fhy you can't always get what you want, (President and Fellows of Harvard College, 2007). The book's subtitle borrows from the title ofthe famous Rolling Stones song, asking "Why you can't always get what you want." This (correctly) suggests that the music industry is one of the industries for which

53 0 ~ ( PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD i ~

, ~ record label, (192) Consider the application ofthis idea to the typical relationship between a new artist, a and the vast number ofways by which labels can distribute music to consumers (e.g. terrestrial radio, streaming services, digital downloads, and cable and satelHte radio services like Music Choice). Assume for the moment that the record label couldperfectlyprerjict what would be the demand for one oftheir new artists. Economics concludes that the label would only invest the time and money to record and promote the artist ifthe revenues that they would receive from CD sales, downloads, and royalties from streaming services exceeded the costs that they would have to pay on recording and promotion.

(193) As has long been established in the music industry and summarized succinctly by Damon Williams, Music Choice's Senior Vice President for Strategy and Partnerships at Music Choice, "The single biggest driver ofrecord sales is exposure through airplay."" One cannot help but read his testimony, or be familiar with the music industry, and not agree that airplay is a crucial factor for the profitability 0 of a recording artist, particularly a new recording artist. And key determinants of airplay are the decisions ofthe terrestrial radio stations, streaming services, and cable and satellite radio services like Music Choice. Indeed, the decisions ofMusic Choice can be particularly important as it is a national music distributor which offers service in 50 million homes.

VI.B.1.b. Music Choice increases demand for new music and new artists in ways other music services would not

(194) As described repeatedly and at length in Mr. Williams'estimony, Music Choice differentiates itself Rom other music services in its focus on identifying and promoting new artists and new music." As

' further described by both Mr. David J. Del Beccaro and Mr. Williams in their testimony, Music Choice's recent loss of a major MVPD affiliate and reductions in affiliate fees paid by three other major affiliates have reduced the programming staffMusic Choice can employ to curate their playlists." It has also reduced their ability to offer artists visits to Music Choice to promote their music.'

(195) The royalty paid by Music Choice to Sound Exchange is an important cost factor in their residential audio music service and I believe, based on Mr. Del Beccaro's testimony and an analysis oftheir financial statements, that changes in this royalty can have a major impact on their ability to avoid further staffreductions or even go out ofbusiness.'fthe royalty were to increase, for example, an

this economic relationship is particularly relevant. The "tyranny" to which the title refers arises when a market has only '0 those with enough demand for a single product: in such cases, the product will appeal to majority tastes, leaving minority tastes unable to find and consume products that interest them.

, ~ Testimony ofDamon Williams at 39. Testimony ofDamon Williams atl0-11, 17, 19, 21-22, 28, 33, 35, 38, 39-40, 42&3. Testimony ofDamon Williams at 10. nu Testimony ofDavid J. Del Beccaro at 34; Testimony ofDamon Williams at 32—33. 0 Testimony ofDavid J. Del Beccaro at 36. "The cuts Music Choice has had to make, and will have to continue making in the absence of any rate relief, will ultimately have a dangerous effect on Music Choice's ability to continue operations.")

54 Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC VERSION

important source of airplay for new artists is very likely to decline or even disappear. In the latter case, whatever replaced Music Choice would be unlikely to be as open to new artists and new music

as is Music Choice.'196)

Translating this eventuality into the language ofeconomics, in a world with perfect information, Music Choice's airplay ofnew artists and new songs is an important factor causing demand for these artists and songs to be realized, without wbich they would not likely be produced. This would reduce the availability of creative works. 't (197) This argument is even stronger when one considers that record labels don know about the demand for their artists'usic. Indeed, music is an example ofwhat economics call an "experience good": consumers don't know the value ofthe product until they consume it. Experience goods pose difficulties to suppliers, as not only must they produce and distribute such goods, but they also must inform consumers about their likely benefits.

(198) In such markets, information about new products is essential to their long-run viability. As the screenwriter William Goldman quipped about movies, "Nobody knows anything... Not one person in the entire motion picture field knows for a certainty what's going to work." While there is a burgeoning "science" of identifying the characteristics of and producing popular music hits, I believe this is generally also true for music, particularly for new artist.

(199) In this uncertain environment, early airplay by companies like Music Choice can be absolutely critical to the eventual viability of an artist.'[ ofMusic Choice's customers believe that Music Choice is a great place to discover new artists and music.'s'articularly since joining the Mediabase measurement system in 2013, Music Choice plays can raise artists'isibility, enhancing their likelihood ofbeing picked up by other distribution channels like terrestrial radio and streaming services. A reduced Music Choice further reduces these opportunities and would absolutely reduce the availability of creative works.

VI.B.'l.c. Music Choice provides additional consumer benefits beyond the increased availability of creative works

(200) As discussed above, increased availability of creative works provides benefits to consumers in the form of an increased production ofnew music. It is no doubt with these consumer benefits in mind that the Copyright Act introduced this first 801(b) statutory factor. As an economist, however, I think

' Damon Williams describes in detail how terrestrial broadcast radio is becoming ever less open to new artists (Testimony ofDamon Williams at 24, 38-40) and how data-driven, algorithmic programming methods often require sufEcient plays via other distribution channels in order to "break into" their playlists (Testimony ofDamon Williams at 7, 10). '64 See Testimony ofDamon Williams at 27. '~ Testimony ofDamon Williams at 30 citing Exhibit MC 15, Ipsos OTX MediaCT, Music Choice Viewership Study July 2015, at 27.

55 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

it important to highlight additional consumer benefits fiom PSS providers like Music Choice beyond their role in increasing the availability of creative works.

preference". (201) One principle economists use when analyzing market outcomes is that of "revealed Revealed preference merely says that when you see market actors making choices, you can assume that the choices they make are the best they could do among the options available to them. Thus cable operators that choose to license Music Choice's residential audio channels do so because they think the benefits to the households that purchase their digital cable packages would be higher than if they offered another digital music service (or none at all). Similarly for households: those households that choose to listen to Music Choice do so because they think that the benefits of doing so are greater than the benefits of alternative uses oftheir time.

(202) If, due to increased pressure on the profitability of its residential audio service, Music Choice had to reduce investments in the quality of its service or, worst case, go out ofbusiness, not only would this reduce consumer benefits due to reduced availability of creative works (and thus reduced discovery of new music), but it would also reduce these other consumer benefits.

(203) For example, a reduced or shuttered Music Choice would cause some households to substitute to other music distribution channels (e.g. radio). As they could have chosen to listen to radio anyway, by revealed preference they would necessarily be worse off. Worse from a music industry perspective, a reduced or shuttered Music Choice would cause some households to substitute to other, non-music, alternatives (e.g. other television channels on their digital cable package). As these alternatives are already present, not only would such consumers be worse off, but so too would the record industry as a whole.

VI.B.1.a. PSS providers are inconsequential for helping record labels cover the costs of music production, promotion, and distribution

(204) My analysis to this point has focused exclusively on the consequences ofthe PSS sound recording performance royalty for the leading PSS, Music Choice. Here I broaden that analysis to consider the consequences ofthis royalty for record labels.

(205) As discussed in Section II.A above, PSS revenues represent less than [I ofUS music industry wholesale revenues. This is a miniscule fiaction and it strains credulity to believe that any decision by an artist or record label could be impacted by such a small consideration.

(206) This conclusion is further supported by academic research on the impact of copyright protection on the quantity and quality ofrecorded music. In a recent working paper titled, "Copyright protection, technological change, and the quality ofnew products: evidence from the recorded music industry since Napster," Joel Waldfogel constructs three measures ofmusic quality and analyzes their PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

evolution from 1960 to the present time." He finds that all three indices are consistent with each other and, despite the sharp decline in industry revenue since 2000, all point to the conclusion that there is "no evidence of a reduction in the quality of music released since Napster. Indeed, the1999."'n two usage-based indices suggest that the quality of music has increased fairly substantially since a companion paper, Professor Waldfogel further concludes that there is "no evidence that changes since Napster have affected the quantity of new recorded music or artists coming to market."'

(207) If the huge decline in revenue since 2000 hasn't changed the quantity and quality ofmusic coming to market, it's even less likely that the miniscule changes in record label revenue caused by changes in the PSS sound recording performance royalty would have any impact. I therefore conclude that the PSS royalty impacts the decision-making of artists or record labels at all, and that any royalty set in this proceeding would have an inconsequential impact on their production of creative works.

VI.B.1.b. My analysis of this statutory factors suggest a reasonable royalty would be, if anything, lower than the hypothetical marketplace royalty

(208) Based on my analysis in this subsection, I conclude that PSS providers like Music Choice not only increase consumer benefits by increasing the availability of creative works, but also by increasing consumers'njoyment of existing music distributed via such providers.

(209) For this reason, the marketplace royalty I estimate as that which would arise in the hypothetical market for PSS sound recording performance rights is an upper bound on the royalty that should be set accounting for this first statutory factor. Music Choice serves an important role in the music industry for music discovery, music promotion, and consumer enjoyment. A lower royalty would help it continue to fill this role in the face of declining financial conditions of its core residential audio business.

'66 Joel Waldfogel, 2012. "Copyright Protection, Technological Change, and the Quality ofNew Products: Evidence from Recorded Music since Napster," Journal of Law and Economics, University of Chicago Press, vol. 55(4), pages 715—740 at 717. The tlrree indices are (1) critics'etrospective lists, (2) sales data, and (3) airplay data. N7 Ibid, 4. He further concludes that this result can perhaps be explained by technological changes that not only increasing piracy (thus reducing revenue), but also reducing the costs of music production, promotion, and distribution, emphasizing in particular the rise of independent labels at identifying and promoting high-quality music (at 21-24). "'oe Waldfogel, (2011), "Bye, Bye, Miss American Pie? The Supply ofNew Recorded Music Since Napster", at 1. NEER Worldng Paper No. 16882. h://www.nber.or / a ers/w16882. df.

57 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

VI.C. To afford the copyright owner a fair return for his creative work and the copyright user a fair income under existing economic conditions

(210) As discussed above, another ofthe statutory factors determining a "reasonable rate" for PSS sound recording performance rights is to afford a copyright owner a fair return for his creative work and the copyright user a fair income under existing economic conditions.

(211) In my opinion, there are two features ofmy analysis to date that speak to the marketplace royalty I estimated above as being higher than one that would provide Music Choice with a "fair income".

VI.C.1.a. The 1% increase in the royalty rate in the last proceeding reduced Music Choice's fair income

(212) The first was described in detail by Mr. David Del J. Beccaro in his testimony regarding the rate

decision in the last, PSS II, proceeding. There he describes how the Judges determined that a

reasonable royalty which applied to Music Choice should be 8.5% (phased in during the early part of the rate period with 8.0% in 2013 and 8.5% in 2014-2017) due to their planned increase in channels from 46 to 300 and a consequent predicted increase in consumer 'usage'fmusic.'213)

As noted by Del Beccaro, in the end Music Choice did not increase its number of audio music channels. Furthermore, Music Choice has found essentially [[

]] Last, even ifthere had been [[ ]], Del Beccaro correctly notes that this would have increased the value ofthe service, ultimately increasing the revenues Music Choice could earn and directly increasing royalties without necessitating a higher royalty rate.

(214) In my opinion, this increase in the royalty rate from the previous base of7.5% &om which the Judges based their analysis to the 8.5% that applies now and in the latter years ofthe previous proceeding's period reduced Music Choice's fair income in the period 2013—2017. Even if only temporarily, the

Judges should correct this error and select a royalty rate that would compensate Music Choice for this

overcharge in the previous period.'s

Testimony ofDavid J. Del Beccaro at 15. Mr. Del Beccaro concludes that a reduction in the rate from 7.5% to 5.7% ofrevenue would be suf5cient to do so. (Testimony ofDavid J. Del Beccaro at 18)

58 0 Crawford, PhD PUBLIC VERSION Written Direct Testimony of Gpegory S.

VI.C.1.b. Any unmeasured promotional benefit unfairly increases record labels returns and 0 unfairly reduces Nlusic Choice's income Choice has a (215) As described at length in Section IV.C.2.c above, it is my opinion that Music promotional benefit on CD sales, digital downloads, and music performances on paid streaming services. As discussed in Section V.B.l.a above, as I am unable to quantify these promotional benefits, I left them out ofmy calculation ofthe royalty that would arise in a hypothetical market for PSS sound recording performance rights and concluded that my predicted marketplace royalty rate is therefore conservative.

would (216) I expand on that idea here and argue that any promotional benefits &om Music Choice services increase record labels return and therefore decrease Music Choice's fair income. I base this conclusion on the simple fact that, while I am unable to estimate these promotional benefits, record are likely to have better information than I and would include these benefits labels and a PSS provider 0 in their negotiations.' (217) Furthermore, it takes a very littlepromotional benefit to rationalize a significant impact on 0 bargaining outcomes iri the PSS market. In 2016, Music Choice is estimated to pay royalties to SoundExchange in the amount of When spread over its 40 million monthly listeners, this amounts to a payment of [iI ii cents per subscriber per year. If, as described in Section iI.A above, a typical digital download sells for $1.20 and record labels keep 62% ofthat ($0.75), it takes track month only [[ ] j Music Choice listeners to download a single per 0 for the promotional benefit to record labels to be equal to the total royaliies it currently earnsPom

the PSS rptarket.'~ Put another way, if even a tiny &action ([I~i]) of Music Choice listeners

record labels would earn revenue download songs due to exposure &om its audio music channels, equivalent to that they will receive &om Music Choice in 2016."'218) Pushing this point further, ifthe current (regulated) royalty rate were the outcome that would arise in a hypothetical market with equal bargaining power and no promotional effect, a promotional benefit

'n For example, Damon Williams demonstrated the im &act reducing the airplay of a song on one oftheir Mediabase ianel charts had on record executives'ehavior. [I reaction Gom !]] (Testimony ofDamon Williams at 29.) I would expect a correspondingly greater labels to the possibility oflosing a/l Music Choice airplay in the case ofimpasse in a hypothetical marketplace negotiation and that such promotional benefits would therefore be factored into the ability to conclude the negotiation at a rate that reflected such benefits. 172 companies also lose when we have to pull back on [[ s]] Music Choice's CEO agrees: "[R]ecord [investment in curation, original content, technology, and developing new markets and media]. Almost all ofthese investments enhance the promotional benefits Music Choice provides to the record companies and artists. These benefits are far more valuable to record companies and artists, because they affect their primary revenue streams of sales, downloads, and streaming, than any potential increase or decrease in the PSS rate." The assumption ofone track/month is reasonable. It implies a Music Choice subscriber buys $ 14.42 in songs in ayear "The Price (12 x $1.20), which compares with $48 estimated to be spent by an iTunes user per year. See David Pakman, ofMusic," Mar. 18, 2014. htto://www.nakman.corn/2014/03/18/the-nrice-of-music/ 173 [I "]]

1~ j~ PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

track month would be ofjust [[ ]] Music Choice listeners downloading a single per sufficient to establish a zero royalty. Given the relatively small benefit to record labels ofPSS royalty payments, almost any promotional effect would have a significant impact on bargaining outcomes in the hypothetical market for PSS sound recording performance rights.

(219) While I am not able to adequately measure this promotional effect, assuming it is zero is very likely to reduce Music Choice's fair income from the use of sound recording performance royalties. While it is difficult to quantify what would be a reasonable downward adjustment from a marketplace rate would allow such a fair income, it is clear to me that any such adjustment would be downward.

VI.D. To reflect the relative roles of the copyright owner and the copyright user

(220) As discussed above, another of the statutory factors determining a "reasonable rate" for PSS sound recording performance rights is to reflect the relative roles ofthe copyright owner and the copyright user in the product made available to the public with respect to relative creative contribution, technological contribution, capital investment, cost, risk, and contribution to the opening of new markets for creative expression and media for their communication.

(221) I have reviewed Music Choice's CEO Mr. David J. Del Beccaro's testimony on this topic and find it convincing. While record labels are of course responsible for essential creative contributions in the production of recorded music, Music Choice also makes creative content to promote individual artists. Furthermore, with respect to each other category in the PSS market, record labels make no significant contributions or investments, while Music Choice makes considerable technological contributions, capital investments, and contributions towards the opening of new markets and media. Many of these contributions are designed to directly benefit record labels via their promotional effects, for example technology to identify which song is currently being played, as well as it's artist and album, investments in on-screen display technology and design, and production studios to make creative content. Music Choice also bears all the costs and risks ofthese contributions and investments.

(222) While one should not diminish the contributions record labels make in identifying new creative talent, as described above, I conclude that these decisions are not made with the PSS market in mind. Thus, when considering the PSS market as a whole, the scales of relative contribution are tilted almost completely in favor of Music Choice and a reasonable royalty should reasonably take this into account.

60 0 0 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

VI.E. To minimize any disruptive impact on the structure of the 0 industries.

rate" sound 0 (223) As discussed above, another ofthe statutory factors determining a "reasonable for PSS 0 recording performance rights is to minimize any disruptive impact on the structure ofthe industries involved and on generally prevailing industry practices.

to (224) As forcefully argued in Del Beccaro's testimony, even the original PSS rate of 6.5% was damaging Music Choice's financial health. The current, 8.5% rate, as well as the expense ofrecurring rate proceedings like this one, has been more damaging, particularly in light ofthe downward turn in their 0 top-line revenue. Setting a rate above that which would arise in a hypothetical marketplace (which I'e estimated at 3.5% oftheir residential audio service revenue), much less one as high as the current rate or at the rate being asked by SoundExchange, risks forcing them out ofbusiness.

(225) The closure ofMusic Choice would have a very disruptive impact on the PSS market, and the market for cable audio generally. Music Choice is clearly the market leader for cable audio for many reasons, including high quality curation, innovative on-screen technology, and providing an important music discovery channel. Any replacement would cause a loss ofthese services, and leave both consumers and record labels worse off. I conclude that a reasonable royalty should therefore also 0 take this into account. 0 VII. Conclusion

(226) In this report, I evaluate the market for Music Choice's residential music service and to recommend a range ofroyalty rates for the license ofPSS sound recording performance rights. To do so, I describe 0 what the hypothetical market for PSS sound recording performance rights would look like and 0 evaluate potential benchmarks for that market. I conclude that the market for PSS musical works performance rights is the only suitable benchmark. That being said, ifthe Judges conclude, as they did in the last proceeding, that this is an unsuitable benchmark, then I must conclude there are no 0 suitable benchmarks. (227) In the absence of a suitable benchmark, I model and estimate directly the royalty rate that would arise 0 in the hypothetical market for PSS sound recording performance rights. I do so by introducing a widely-used non-cooperative bargaining &amework, the Asymmetric Nash Bargaining Model, which predicts market outcomes when small numbers ofparties, each with some market power, bargain over a royalty. I explain why this model is suitable to analyze the marketplace determination of sound recording performance rights in general and as applied to the PSS market. 0 61

'

0 0 Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC VERSION

, ~ (228) This framework identifies three fundamental "Nash Factors" that determines how royalties would be paid in the hypothetical market. Simply put, the profit to be received by each firm in the bargain is its Threat Point (the profit it receives ifno agreement is reached) plus its Bargaining Power (ranging from 0 to 1) times the Incremental Profits (Joint Agreement Profits minus the sum ofThreat Points). I use the predicted financial performance ofMusic Choice's residential audio service (alone) to quantify these Nash Factors and estimate the royalty that would arise in the hypothetical market.

(229) Assuming equal Bargaining Power, I estimate that the hypothetical PSS sound recording performance 0 royalty rate should be 3.5% ofresidential audio revenues. Even ifthe Judges were to grant record labels 80% ofthe bargaining power in the hypothetical market (and there is no reason to do so), the 0 hypothetical royalty should be no higher than 5.6% ofMusic Choice's unadjusted residential audio revenues. In addition, each ofthe statutory factors would support a rate below this hypothetical marketplace rate. If any adjustments to a marketplace rate are made, they should clearly be 0 downward.

, ~ 0

0 ,0

0 0 0 0

62 0 0 0 PUBLIC VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD ,0 Appendix A. A non-cooperative bargaining framework

In Section IV of my report, I introduce a non-cooperative bargaining &amework to understand how royalties would be determined in a hypothetical market for the digital performance right for sound recordings on pre-existing subscription services (PSSs). In this Appendix, I present the details of this 0 framework that are summarized in the body ofmy report. 0 In the report, I analyze negotiations between a single upstream seller (a copyright owner) and a single 0 downstream buyer (a copyright user). This is a bilateral monopoly &amework. In many of the actual 0 markets where royalties for the use of copyrights are negotiated, however, there may be multiple 0 sellers and multiple buyers. In what follows, I first present the bilateral monopoly framework and derive the predictions from that &amework. I then extend it to the case of multiple buyers and sellers 0 and demonstrate that the qualitative features from the bilateral framework remain. 0 For the bilateral monopoly framework, I introduce the model in terms of the negotiations between a 0 single PSS provider and a single record label given a pre-existing agreement with a single perforining rights organization (PRO).'his is for expositional convenience. The model applies equally well to any negotiation between an upstream copyright owner and a downstream copyright user in the presence of a pre-existing agreement with another copyright owner. For example, in what follows, I use the same framework to analyze negotiations between a PSS and a PRO in the presence of a pre- existing agreement with a record label and consider the equilibrium ofthese two negotiations. I also 0 use the saine naming convention for the framework with multiple copyright owners and users. '

1. Bargaining for PSS sound recording performance rights under bilateral monopoly

In this section I describe a model ofnon-cooperative bargaining between a single PSS provider, indexed 0 the 0 by d, and a single record label, indexed by r, in the presence of a pre-existing agreement between PSS and a performing rights organization, indexed by w.

0 The bargaining problem is specified as follows. I assume that the upstream record company and downstreain PSS provider meet to negotiate terms ofthe royalty paid by the PSS provider to the record company. I further assume that they bargain over whether to form an agreement and, if so, at what royalty (measured in profits paid from a downstream firm to an upstream firm)."'he ultimate profits 0 to each player depends on whether an agreement was reached, what royalty was paid, and the nature of 0 musical works royalties) in the upstream and downstream markets. 0 demand and cost (including 0 "4 I choose to include a PRO for convenience. From an expositional perspective, it is useful to include in the model 0 another input provider, and PROs provide an important input for PSS services. The analysis here would apply equally to any other input provider to PSS services. 0 ' This is for analytical convenience. It is straightforward to translate a profit-based royalty into a royalty rate as a share of 0 revenue by simply dividing the profit-based royalty for the upstream firm by the downstream firm's revenues. 63 0 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

Let r, be the royalty (measured in profits) paid by a PSS provider to a record label as a result of their r"r = negotiation, let rd, be the portion ofprofits the PSS provider retains, and let (rd„r,) denote the pair t"'"' of royalties arising from negotiations between the PSS provider and the record label. Likewise, (rd, r„) denote a similar pair ofroyalties arising from separate negotiations between the PSS provider and a PRO.

I assume that these bilateral meetings result in the Asymmetric Nash Bargaining Solution. This solution concept can be rationalized as the unique outcome of a bargaining problem in which each party to the bargain alternates making offers until an agreement is reached.'his is a realistic depiction of bargaining in many market envirornnents, including those for digital performance rights.

Under the Asymmetric Nash Bargaining Solution, the royalty resulting &om a bilateral meeting

between each PSS, d, and each record company, r, maximizes the bilateral Nash Product:

— — rr = arg max NP (rr, r„,) = (rr', (r,, r„) trr(-, r,„)) '&„(rr, r,„) rr„(-, r,„))' (I)

where rr is royalty paid from d to r, 7t„(rr, r ) and z'd (r„, r,) are the economic profits (also called

are the "surplus") to r and d, respectively, in the case of agreement, z'„(-, r,) and 7t„(-,r ) economic profits to r and d, respectively, in case of disagreement, and bz, is the bargaining power of the record label with respect to the PSS provider.'

There is a straightforward intuition for the Nash Product. It says that the optimal royalty that yields the Asymmetric Nash Bargaining Solution is obtained as the value that maximizes a (geometrically) weighted average oftwo incremental profits: the incremental profit the record label gets.'he from an — incremental profit agreement at royalty r, compared to no agreement, z;, (I"r, r„,) Ter (-, r ), and the the PSS provider gets at royalty r, compared to no agreement, W„(r„, r ) — W„(-, r„)

weighting is done according to the bargaining power of the record label. Economic estimates of bargaining power have historically been related to each party's patience in a negotiation, with greater patience yielding greater bargaining power." As bd, approaches 1, it is only the record label's incremental profits that matter (it has all the bargaining power), and the PSS distributor gets no more than it would in the absence of an agreement. As 4r approaches 0, it is only the PSS distributor's incremental profits that matter, and the record label gets no more than it would in the absence of an

Abhinay Muthoo, Bargaining Theory with Applications (Cambridge: Cambridge University Press, 1999) at Chapter 2. '" I only need to solve for the royalty to the record label, r,, as the profit retained by the PSS provider will just be the portion of the joint agreement profits that remains atter paying the record label's royalty, r,. '7'ote that the same royalty enters both weighted averages as each additional dollar paid from the PSS distributor, d, to the record label, r, in royalty, rn reduces his own retained profits by a dollar. '~~ Abhinay Muthoo, Bargaining Theory with Applications (Cambridge: Cambridge University Press, 1999) at 51.

64 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

agreement. As we will see below, if bz, = 0.5, the PSS provider and record label split equally the incremental profit created by an agreement.

Other Markets Matter

To use this fratnework to understand outcomes in the market for PSS sound recording performance rights, it is useful to dig deeper into the components of each ofthe profit measures above. In particular, it is important to understand whether the profit either the PSS provider or record label earns from an agreement is influenced by outcomes in markets other than the PSS market.

PSS providers only produce PSS. They do not produce any other service that could be influenced by the outcomes of its bargaining with record labels. The profits to the downstream PSS, 7r~ (r„, r ) and/or 7t~ (-, rr), therefore depend only on the profits in its own (PSS) market.

In contrast, record labels license the performance rights to their sound recordings across a wide variety of markets. They also sell physical CDs and license rights for digital downloads and streaming music services. The profits to the upstream record label, W„(r„, r ) and/or 7t„(-, r„,), therefore depends on the profits in each ofthese markets (PSS and non-PSS).

As such, I separate the profits for the record labels into two parts: that part arising from PSS, which I denote "PSS", and that part arising f'rom all markets excluding PSS, which I denote "-PSS". This is true both when an agreement is reached and when it has not:

As noted above, downstream PSS providers earn profits only in their own (PSS) markets. For consistency, I therefore restate profits to downstream firms as PSS profits:

Nash Bar ainin and Threat Points

Once we allow for profits to record labels from other markets, it is convenient to re-write the Nash Product above to separate out the profits coming from the PSS market versus other markets: r, =argmax NP(r,.,r,„) =(z',. (r,.,r ) — TP.(-,r )) '(m~ (r,.,r ) — TP~(-,r ))'"'5 0 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

0 and PSS provider, where TP„(-, r„,) and TP~ (-, r„,) are the 'T'hreat 'P'oints for the record label respectively. I show below how I have incorporated profits in other markets into these threat points. i 0 0 Threat points are very important elements in any bargaining framework. They capture the profit to 0 each firm in case of disagreement. If a party has a stronger threat point, they are better offif an agreement can't be reached and, as I'l show below, earn a larger royalty. , ~ '00 The threat points above are defined as follows: 0 TP (-, r,„) = n'-, r„,) — = 2'p,, (-, r„,) = z„(-, r„) + (z„(-, r„) n'„(r„, r„,)) sc„(-, r ) + her'„(r„, r ) 0 For the PSS provider, d, the threat point is no different than in the original Nash Product. It is its profit in the PSS market in case of disagreement with the record label. For the record label, r, 0 however, there are two terms in its threat point. The first is analogous to the PSS provider's: it is the economic profit the record label earns in the PSS market in case of disagreement with the PSS 0 provider. The second is different, however. It measures the extra profit it earns in other (non-PSS) ss(-, - . For markets if it can' reach an agreement in the PSS market, x,. r ) n'„(r„, r ) convenience, I label this extra profit in non-PSS markets in case ofthe inability to reach an agreement e in the PSS market as her„(r„, r„) . 0 The Musical Works Market and Euuilibrium 0 Everything to this point has taken the musical works royalty, r„, paid by the PSS provider, d, to the PRO, w, as given. In practice, both the sound recording and musical works royalties, r„and r„, will be determined as the result of an economic equilibrium. In other words, the sound recording royalty, r,; that is determined as a function ofthe musical works royalty, r, must be consistent with the value 0 ofthe musical works royalty that is determined as a function ofthe sound recording royalty.

To achieve this equilibrium, I assume that each set ofnegotiations are simultaneous and separate, so other royalty agreements are not known but conjectured. The equilibrium ofthis game, interpreted as 0 a Nash equilibrium between Nash Bargains (or Nash-in-Nash) by Horn and Wolinsky (1988),' 0 yields royalties given by the Asymmetric Nash Bargaining Solution:

0 = — TP» (-, — TP,. r„,)) 0 r,. TP„(r„, r„) + b„„(rr'r„) r„) (r„, "~ 0 where z (r„) = ms~ (r„, r„) + gr„~ (r„, r„) is the combined economic profit (surplus) to both the PSS provider and the record label &om reaching an agreement in the PSS market. In my report, I call

Henrick Horn and Asher Wolinsky, "Bilateral Monopolies and Incentives for Merger," RAND Journal ofEconomics 19, 0 no. 3 (1988): 408-19. 0 66 0 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

this the "Joint Agreement Profits". It depends only on the musical works royalty, r„, as the sound should to recording royalty, r„, simply determines what portion of the combined agreement surplus go the record label versus the PSS provider (and therefore doesn't affect the Joint Agreement Profits itself). I call the difference between the Joint Agreement Profits and the sum ofthe Threat Points, — — Profit". (r„, ) TP~ (-, r„, ) TP,, (r,, r,„), the "Incremental

that aren't paid as royalties to Let r„,, be the profits retained by PSS provider, d, (i.e. those profits record company r). These are given by a symmetric formula

— — — r„,. = TP~(-, r„,)+(1 b~,,)(z (r„,) TP~(-, r,„) TP,,(r,,r ))

In the body of the text, I describe these formulas as follows:

The royalty received by each firm in a bargain equals its Threat Point plus its Bargaining Power times the Incremental Profit.

2. Bargaining for PSS sound recording performance rights with multiple upstream and downstream firms

I now extend my non-cooperative bargaining framework to allow for multiple upstream record labels, multiple upstream PROs, and multiple downstream PSS providers.

The bargaining problem is specified as follows. Let r = 1, ..., R index (upstream) record companies = W'ndex negotiating digital performance rights for sound 'r'ecordings, let w 1, ..., upstream PROs = negotiating digital performance rights for musical 'w'orks, and let d 1, ..., D index ('d'ownstream) PS S providers.

I assume that each upstream record company and downstream PSS provider meets bilaterally in simultaneous negotiations. I further assume that they bargain over whether to form an agreement and, if so, at what royalty (measured in profits paid from the downstream PSS provider to the upstream record company). The ultimate profits to each player depend on whether an agreement was reached, what royalty was paid, and the nature of demand, cost (including musical works royalties), and competition in each upstream and downstream market.

Let r = (rq,.}, d = 1, ..., D and r = 1, ..., R be the full set of royalties paid by downstream firm d to record company r and divide these royalties into those paid from d to r, rz,, and r', all other royalties paid fromthis and other PSS to other record companies. Letw= (wz }, d= 1, ...,D and w= 1, ..., W be the full set of royalties paid by downstream firm d to PRO w. Similarly, partition w into wz and w', all other royalties paid from this and other PSS to other PROs.

67

PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

I assume that these bilateral meetings result in the Asymmetric Nash Bargaining Solution. In this case, the royalty resulting f'rom a bilateral meeting between each PSS, d, and each record company,'"'2) r, maximizes the bilateral Nash Product:

= r', — B', r', r', — Ted (-, r', w)) r„,. = arg max NP (rd,, r', w) (7t',, (rd,,, w) (-, w))" (7td (r~,, w)

where r~, is royalty paid from d to r, 7c„(rdr, r', w) and e„(r„„,r',w) are the profits to d and r, respectively, in the case of agreement, Ted (-, r, w) and Ter(-, r, w) are the profits to d and r, respectively, in case of disagreement, and bd; is the bargaining power of record label r with respect to PSS provider d. The only difference in the Nash Product under bilateral oligopoly relative to that under bilateral monopoly is the dependence of profits to each firm on the royalties paid by this and other PSS providers to other record labels, r'.

a. PSS versus Other markets

As in the bilateral monopoly case, to use this framework to understand bargaining outcomes, it is useful to dig deeper into the components of each ofthe profit measures above. In particular, it is important to understand whether the profit either the PSS provider or record label earns &om an agreement is influenced by outcomes (I) in the PSS market from other providers of PSS services and/or other record labels and (2) in markets other than the PSS market. I describe the influence profits &om these other markets have on bargaining outcomes in what follows.

A record label, r, earns profits in the PSS market both from PSS operator d as well as other PSS operators, which I will label "-d". As above, they also license the performance rights to their sound recordings in other markets (e.g. for digital downloads and music streaming services) and sell physical CDs in other, non-PSS, markets, which I will label "-PSS".

As such, I separate the profits to record company, r, into three parts: that part arising from sales of PSS provider d, which I label "PSS,d", that part arising from sales of other PSS providers, which I label "PSS,-d", that part arising from all markets excluding PSS, which I denote "-PSS". This is true both when an agreement is reached and when it has not:

7t,(- r')=7K. (- r')+Pl, (- r't)=7K. '- r't)+7K. '- r')+7K. (- f"'t)

Unlike for record labels, downstream firms only earn profits in their own PSS market. They have no profits from other markets, "-PSS". Furthermore, these profits include profits from both record label r

68 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

d's and other record labels, "-r". For consistency, I restate profits to PSS provider d as PSS profits in market, which I denote "PSS,d":

n'„(-, r',w) = m', '"(-,r',w)

We can use these terms to reorganize the Nash Product in Equation (2) above, this time focusing on profits within the PSS market serviced by PSS provider, d:

r« = arg max NP(r„„,r',w) = (rr„'r~„,r',w) — TP„(-, r',w))'rr, 'r~„,r',w) — TP~(-, r',w))'"

where, as above, TP„(-, r', w) and TP„(-, r', w) are the 'T'hreat 'P'oints for record label r and PSS provider, d, respectively.

The threat points are defined as follows:

TP»(-,r,w) =rr~ '-,r,w)

TP (r~,,r',r,) =rr '-,r',r )+(x„'-,r',r„)-rr'„'(r~„,r',r,„))+(rr'„(-,r',r„) — m'„(r„,r',r ))

For the PSS provider, d, the threat point is no different than in the original bilateral monopoly Nash Product. It is its profit in the PSS market in case of disagreement with the record label. For the record label, r, however, there are now three terms in its threat point. The first is analogous to the PSS provider's: it is the economic profit the record label earns in PSS provider d's PSS market in case of disagreement with the PSS provider. The third measures the extra profit it earns in other (non-PSS) markets if it can't reach an agreement in the PSS market, non-PSS (-, r', r„) — x„(r~,, r', r„) . For convenience, as above, I label this extra profit in markets h,rr,. (r~„, r', r,„) .

It is the second term that is due to the presence ofupstream and downstream firms. It measures the extra profit record label r earns from sales in other PSS markets in case of disagreement with PSS provider d. I label this extra profit in non-d PSS markets b,rr„~'r„„, r', r ) .

The Musical Works Market and Eauilibrium

Everything to this point has taken the musical works royalty, r~, paid by each PSS provider, d, to each PRO, w, as given. In practice, all the sound recording and musical works royalties, rz, and ra~, will be determined as the result of an economic equilibrium. In other words, the sound recording

69 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

royalties, rz;, that are determined as a function ofthe musical works royalties, rz, must be consistent with the value ofthe musical works royalties that is determined as a function ofthe sound recording royalties. Shnilarly each sound recording royalty must be consistent with each other sound recording royalty and all the musical works royalties.

To achieve this equilibrium, I assume that each set ofnegotiations are simultaneous and separate, so other royalty agreements are not known but conjectured. The equilibrium ofthis game, interpreted as a Nash equilibrium between Nash Bargains (i.e. a Nash-in-Nash Equilibrium), yields royalties given by the Asymmetric Nash Bargaining Solution:

r~„= TP„(r„„, r', r„) + b~„(n'" (r', r,„) — TP, (-, r', w) — TP„(r~, r', w))

where rt' (r',r„,) = tf."~ 'r„r', r„.)+ e„'r„r', r,„)is the combined economic profit (surplus) to both the PSS provider and the record label Rom reaching an agreement in the PSS market served by PSS provider d. This is again the "Joint Agreement Profits" in the PSS market served by PSS provider d. I call the difference between the Joint Agreement Profits and the sum ofthe threat points, '" (r', r,„) — TP, (-, r', w) — TP„(r„, r', w), the "Incremental Surplus".

Let r~'„be the profits retained by PSS d, i.e. those profits that aren't paid as royalties to record company r. These are given by a symmetric formula

r~ = TP„(-, r', r,„) + (1 — b~„)(x 'r, w) — TP4 (-, r', r„) — TP„(r~, r', w))

These formulas show that the qualitative features ofroyalties this more complicated model are effectively the same as in the bilateral monopoly model:

The royalty received by eachfirm in a bargain equals its Threat Pointplus its Bargaining Power times the Incremental Profit.

70 PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

Appendix B. Exhibits

verage, $ in thousands 2018-202+ R W R RW W H W W W R~ R~ R~ ~ R R~ R~ M~ ~ R VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

72 ~ 00 ~ ~ 00 ~ ~ 00 ~ ~

PUBLIC VERSION Written Direct Testimony of Gregory S. Crawford, PhD

~Average Period f19 I 8- I $ in thousands ~2016 2017 2018 2016-2018 I: ~» ~i ~i

~ ~ ~ H % % H W

L W Q R R R R 0 L R W W W

I I I I I I

73 ~ 0 OO ~ ~

VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

74 ~ ~

VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

I- ~%~ =, Descriptiong ~ISource~

75 0

0 Crawford, PhD PUBLIC VERSION 0 Written Direct Testimony of Gregory S. 0

Appendix C. Materials relied upon 0 0 Music Choice and SoundExchange discovery documents 0 ~ J. Del Beccaro and exhibits therein. 0 Testimony of David 0 ~ Testimony of Damon Williams and exhibits therein. ~ Music Choice Consolidated Financial Statements, 2013—2015. (MC0003221, MC0003130) 0 ~ Music Choice Unaudited Income Statements, 2013—2015. (MC0014595, MC0014587, MC0014590) - ~ Music Choice Audio Only and Residential PAL, 2016-2022. [Privileged and Confidential Audio 0 Only Model BW 101416B] ~ Music Choice P&L, 2016-2022. [Privileged and Confidential - Consolidated BW 101316] 0 ~ Exhibit MC 15, Ipsos OTX MediaCT, Music Choice Viewership Study July 2015, at 27. 0 ~ SoundX 000028618 RESTRICTED. 0 ~ Suisse, "Global Music," Apr. 4, 2016 at 1-2. SoundX 000034949-950. 0 Credit Copyright Royalty Board documents, statutes, and legal documents

- ~ 17 U.S. Code $ 114 Scope of exclusive rights in sound recordings. 0 htt s://www.law.cornell.edu/uscode/text/17/114.

~ 37 CFR 383.3 - Royalty fees for public performances of sound recordings and the making of 0 ephemeral recordings, htt s://www.law.cornell.edu/cfr/text/37/383.3 ~ Copyright Law ofthe United States of America and Related Laws Contained in Title 17 of the 0 ri ht. ov/title17/92cha 8.html. 0 United States Code, 17 U.S.C. 801(b)(1). h://www.co ~ Library of Congress, [Docket No. 96—5 CARP DSTRA], Determination of Reasonable Rates and 0 Terms for the Digital Performance of Sound Recordings. 63 FR at 25405, 407, 408. h://www.co ri ht. ov/fedre /1998/63fr25394. df

0 ~ Library of Congress, Copyright Royalty Board, In the Matter of Determination of and Terms for Preexisting Subscription and Satellite Digital Audio Radio Services, Docket No. 2011-1 CRB PSS/Satellite II. Determination.

~ Recording Industry Association of America v. Copyright Royalty Tribunal, 662 F.2d 1 (1981). 0 ~ Recording Industry Association of America v. Librarian of Congress, 176 F.3d 528 (1999). VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

~ United States Copyright Royalty Judges, The Library of Congress, In re Determination of Royalty Rates and Terms for Ephemeral Recording and Webcasting Digital Performance of Sound Recordings (Web IV), Docket No. 14-CRB-001-WR (2016-2020).

~ United States ofAmerica v. Broadcast Music, Inc., 426 F.3d 91, footnote 5 (2d Cir 2005).

Books and papers

~ Abhinay Muthoo, Bargaining Theorywith Applications (Cambridge: Cambridge University Press, 1999) at Chapter 2, 3.

~ Alvin E. Roth, "Bargaining (Economic theories ofbargaining),"in Social Science Encyclopedia (London: Routledge, 1996) at 46—7.

~ Aswath Damodaran, Investment Valuation, (New York: John Wiley 0 Sons, Inc., 2002, 194.

~ Charles T Horngren, Srikant M. Datar, and Madhav V. Rajan, "Allocation of Support-Department Costs, Common Costs, and Revenues, Chapter 15 in Cost Accounting: A Managerial Emphasis, (New Jersey: Pearson Education Inc., 2015).

~ David Pakman, "The Price ofMusic," Mar. 18, 2014. htto://www.oakman.corn/2014/03/I 8/the- nrice-of-music/.

~ David R. Strickler, Royalty Rate Setting for Sound Recordings by the United States Copyright Royalty Board: The Judicial Need for Independent Scholarly Economic Analysis (December 31, 2015). Review ofEconomic Research on Copyright Issues, 2015, 12(1/2), 1-15.

~ Gautam Gowrisankaran, Aviv Nevo, and Robert Town. 2015. Mergers When Prices Are Negotiated: Evidence from the Hospital Industry." American Economic Review, 105(1): 172- 203.)

~ Gregory S. Crawford and Ali Yurukoglu, 2012. "The Welfare Effects ofBundling in Multichannel Television Markets," American Economic Review, 102(2): 643-85.

~ Henrick Horn and Asher Wolinsky, "Bilateral Monopolies and Incentives for Merger," RAND Journal ofEconomics 19, no. 3 (1988): 408-19. ~ Joel Waldfogel, 2012. "Copyright Protection, Technological Change, and the Quality ofNew Products: Evidence Rom Recorded Music since Napster," Journal ofLaw and Economics, University of Chicago Press, vol. 55(4), pages 715 — 740.

~ Joel Waldfogel, The Tyranny ofthe Market: JFhy you can 't always get whatyou want, (President and Fellows ofHarvard College, 2007).

~ John F. Nash, Jr. "The Bargaining Problem," Econometrica 18, no. 2 (Apr. 1950): 155-62

77 Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC VERSION

~ Marie Connolly and Alan B. Krueger, "Rockonomics: The Economics ofPopular Music" in Handbook on the Economics ofA&"t and Culture (Amsterdam: Elsevier, 2006) at Chapter 20.

~ Matthew Grennan, "Price Discrimination and Bargaining: Empirical Evidence from Medical Devices." American Economic Review, (2013) 103(1): 145-77.

~ Meza, S. & Sudhir, K. "Do private labels increase retailer bargaining power?" Quantitative Marketing and Economics (2010} 8: 333.)

~ Richard A. Brealey, Stewart C. Myers, and Franklin Allen, Principles ofCorporate Finance (New York: McGraw-Hill, 2006} at 189, 215, 219, 311.

~ Roman L. Weil, "Chapter 16: Allocations of cost and revenue," in Roman L Weil, and Michael W. Maher, Handbook ofCost Management, (New Jersey: John Wiley & Sons, Inc., 2005), 465— 482.

~ Tim Koller, Mare Goedhart, and David Wessels, Valuatio&t& Measu&"ing and Managing the Value ofCompanies, (New Jersey: John Wiley and Sons, Inc.: 2005), 116, 160, 166— 167.

Other

~ "Seventeenth Annual Report on the Status of Competition in the Market for the Delivery of Video Programming," FCC (May 6, 2016) MB Docket 15-158. htt s://a s.fcc. ov/edocs ublic/attachmatch/DA-16-510A1 Rcd. df.

~ 37 CFR 383 ath://www.soundexchan e.com/service- rovider/other-service- roviders/new- subscri tion-service/.

~ Congressional Budget Office's Economic Projections for Calendar Years 2016 to 2026. htt s://www.cbo. ov/sites/default/files/114th-con ress-2015-2016/re orts/51908/51908- breakout-cha ter2-2. df..

~ Dean Starkman, "Sony BMG Settles Radio Payola Probe, July 26, 2005. htt://www.washin ton ost.com/w -d n/content/article/2005/07/25/AR2005072501624.html. a Duff & Phelps, "2016 Valuation Handbook: Guide to Cost of Capital," (Hoboken, NJ: John Wiley & Sons, 2016), Appendix 3, Exhibit B-6, Companies Ranked by EBITDA, Premia over CAPM. s- a Duff &, Phelps. h://www.duffand hei s.com/insi hts/ ublications/cost-of-ca ital/duff- hei increases-recommended-us-e ui -risk- remium-from-5-0-to-5-5.

~ Ex-TuneCore CEO Jeff Price: The End Of The New Music Industry Transformation. htt://www.h ebot.com/h ebot/2012/09/the-end-of-the-new-music-indust -transformation- how-technolo -destro ed-the-traditional-music-indu.html.

78 VERSION Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

~ Jeff Leeds, Universal Music Settles Big Payola case, The New York Times, May 12, 2016. h://www.n times.com/2006/05/12/business/12 a ola.html.

~ Music Choice 2016 Media Kit. h://cor orate.musicchoice.com/files/2214/6228/4671/2016 Music Choice Media Kit. df.

~ Music Choice Advertising, "Popular with Millennials." h://co orate.musicchoice.com/advertisin /nielsen-audience-rankers/.

~ Music Choice Press Release, Music Choice Rebrands SWRV TV Into Music Choice Play Integrates All Networks Into One Seamless Experience October 15, (Oct. 14, 2013). htt://co orate.musicchoice.com/about-us/ ress-room/ ress-article/music-choice-rebrands-swrv- tv-music-choice- la -inte rates-all-networks-one-seamless-ex erience-october-15th/.

~ Music Choice Products page, h://co orate.musicchoice.com/about-us/ roducts/.

~ News and Notes on 2015 RIAA Shipment and Revenue Statistics. htt://www.riaa.com/w- content/u loads/2016/03/RIAA-2015-Year-End-shi ments-memo. df.

~ NobelPrize.org Press Release, "This Year's Laureates Are Pioneers in The Theory ofFinancial Economics and Corporate Finance," Oct. 16, 1990. h://www.nobel rize.or /nobel rizes/economic-sciences/laureates/1 990/ ress.html.

~ RIAA 2007 Year-end statistics. htt://www.musicaememoria.com/docs/riaa ear end 2007. df.

~ RIAA, News and Notes on 2016 Mid-Year RIAA Music Shipments and Revenue Statistics. Figure 5: U.S. Recorded Music 1H 2016. h://www.riaa.com/w- content/u loads/2016/09/RIAA Mid ear 2016Final. df.

~ Steve Knopper, "The New Economics ofthe Music Industry," The Rolling Stone, Oct. 25, 2011. htt://www.rollin stone.com/music/news/the-new-economics-of-the-music-indust -20111025.

~ Surmnary of Statement 131, FASB, h://www.fasb.or /stnnm /stsum131.shtml.

~ The Infinite Dial 2016, Edison Research, Triton Digital, "Sources used for keeping-up-to-date with music," at 34. h://www.edisonresearch.com/w -content/u loads/2016/03/The-Infinite- ~Dial-2016. dr.

~ Tim Inghatn, "Major labels keep 73% of Spotify premium payouts — reports," Feb. 3, 2015. htt://www.musicbusinessworldwide.com/artists- et-7-of-streamin -cash-labels-take-46/.

~ U.S. Department of Treasury, Daily Treasury Yield Curve Rates.

htt s://www.treasur . ov/resource-center/data-chart-center/interest- rates/Pa es/TextView.as x?data= ield.

79 0 0 VERSION 0 Written Direct Testimony of Gregory S. Crawford, PhD PUBLIC

D. Curriculum vitae Appendix

0 0 0 0 0 0 0 '0,0 0 0 0 0 0 0 0 0 0

0 80 0 0 0 Gregory S. Crawford

Business Address Home Address Department of Economics Burgrain 37 University of Zurich 8706 Meilen Schonberggasse 1 Switzerland CH-8007 Zurich Mobile: +41 (0)79 194 6116 Switzerland Email: gregory.crawfordOecon.uzh.ch Phone: +41 (0)44 634 3799

Education

Ph.D. in Economics, Stanford University, Stanford, CA, 1998 B.A., Economics (with Honors), University of Pennsylvania, Philadelphia, PA, 1991

Professional Experience University of Zurich, Department of Economics

Professor of Applied Microeconomics, May 2013-current

Courses taught: Graduate: Structural Estimation in. Applied Microeconomics (PhD), Empirical Industrial Organization (PhD), Cross-Section and Panel Data Econometrics (MSc) Centre for Economic Policy Research (CEPR) Co-Director, Industrial Organisation Programme, September 2014-present Research Fellow, Industrial Organisation Programme, February 2011-present Institute for Fiscal Studies (IFS)

International Research Fellow, August 2014-present Centre for Competitive Advantage in the Global Economy (CAGE)

Research Fellow, April 2011-present

Association of Competition Economists (ACE)

Steering Committee, January 2016-present University of Warwick, Department of Economics

Professor of Economics, September 2008-July 2013

Gregory S. Crawford cv, October 2016 Director of Research Impact, August 2012-July 2013 Director of Research, September 2009-July 2012 Courses taught: Graduate: Empirical Industrial Organization (MSc/PbD), Empirical Methods. Undergraduate: Introductory Econometrics (time series, limited dependent variables, panel data), Undergraduate Business Strategy. Federal Communications Commission (FCC)

Chief Economist, September 2007- August 2008

Reported to the then-FCC Chairman, Kevin Martin. Primary responsibilities were to advise the Chairman and his stafF regarding the economic issues facing the Commission, to formulate and implement desired policies, to communicate and discuss these policies with senior Commission staff, and to assist as needed the 40+ stafF economists. Main workstreams focused on the cable and satellite industries, including bundling and tying in wholesale and retail cable and satellite television markets and the economic analysis of XM/Sirius satellite radio merger. Also consulted on spectrum auction design, net neutrality, access pricing, ownership rules, and various international policy issues. Previous to joining the Commission, wrote a sponsored study analyzing media ownership and its impact in television markets. University of Arizona, Department of Economics

Associate Professor of Economics, September 2008-August 2009 (on leave) Assistant Professor of Economics, September 2002-August 2008 (on leave, 2007-08)

Courses taught: Graduate: Empirical Industrial Organization (2nd-year PhD), Business Strategy (MBA) Undergraduate: Introductory Econometrics (cross-section). Duke University, Department of Economics

Assistant Professor of Economics, September 1997-August 2002

Courses taught: Graduate: Empirical Industrial Organization (2nd-year PhD), Graduate Econometrics (1st-year PbD), Undergraduate: Introductory Econometrics (cross-section), Introductory Microeconomics, The Economics and Statistics of Sports. Other Academic Appointments

Visiting Professor, European School of Management and Technology, Berlin, Summer 2007.

Visiting Professor, Fuqua School of Business, Duke University, 2000-2001

Consulting Experience (Country)

A la carte offerings on pay television system (South Africa), 2016-present, consulting expert — Advising pay-television operator regarding regulatory submission to require them to provide television channels on an a la carte basis.

Gregory S. Crawford cv, October 2016 Rules governing sale of football rights (EU), 2015-2016, consulting expert — Advised major pay-television distributor on regulatory filing challenging how rights are sold for a major European football league.

Geographic restrictions on sport TV broadcasts and Internet distribution (US), 2014-15, consulting expert — Advised on class-action lawsuit challenging geographic restrictions placed on member teams and regional sports networks regarding television broadcasts and Internet distribution by US sports leagues Major League Baseball (MLB) and the National Hockey League (NHL). Cases settled.

Royalties for sound recording performance rights by non-interactive webcasters (US), 2014-15, testifying expert — Prepared testimony for copyright royalty judges regarding reasonable rates for sound recording performance rights by a non-interactive webcaster. Client decided not to File a report.

Royalties for sound recording performance rights on cable television systems (US), 2011-12, testifying expert — Submitted direct and rebuttal testimony to copyright royalty judges on behalf of Music Choice regarding reasonable rates for sound recording performance rights on U.S. cable television systems. Testified before judge panel.

Evaluating "neighborhooding" of news channels on Comcast cable systems (US), 2011, lead expert — Designed and executed expert reports for complaint to FCC by Bloomberg (Television) L.P. (BTV) that Comcast was not fulfilling the neighborhooding conditions imposed during Comcast-NBCU merger. Defined news neighborhoods and investigated incidence of carriage of BTV on such neighborhoods. Compared patterns to neighborhooding of sports channels on Comcast and news channels on other operators and analyzed Comcast channel changes over time. Complaint largely granted by the FCC.

Evaluating switching costs in fixed voice telephony markets (UK), 2010-11, lead expert— Designed and executed reports for Office of Communication (Ofcom) evaluating the impact of automatically renewable ('rollover') contracts (ARCs) introduced by British Telecommunications (BT) in the UK fixed voice telephony market. Based on this analysis, Ofcom prohibited rollover contracts in all residential and small business fixed voice and broadband markets.

Evaluating competitive harms (US), 2010, consulting expert — Helped design and execute economic and econometric analyses in support of client opposed to major media merger. Analysis included market, definition and quantifying the potential harms of the merger, including refusal to carry (foreclosure).

Analysis of advertising market regulations (UK), 2009-10, consulting expert — Advised project team on. analysis of demand for advertising for the purpose of evaluating changes in regulation of advertising minutes on public-service broadcasters in the United Kingdom. Designed econometric model and supervised implementation and description. of results. Report submitted to media regulator (Ofcom).

Distribution of cable copyright royalties (US), 2009-10, testifying expert — Submitted rebuttal testimony to copyright royalty judges regarding relative market value of programming provided on the distant broadcast signals carried by U.S. cable systems. Testified before judge panel.

Gregory S. Crawford cv, October 2016 Video chain merger (US), 2005, consulting expert — Supported lead expert in a challenge of a proposed merger of video chains. Merger denied.

Echostar/DirecTV (US), 2002-03, consulting expert — Supported analysis of liability for proposed merger. Helped design econometric model of pay-television demand and participated in conference calls with opposing lawyers and experts. Plurimus / Foveon (US), 1999-00, consultant and advisory board member — Conducted market research and helped design business plan for Internet start-up seeking to enter the Internet audience measurement business. Projects included conducting a survey and strategic analysis of the early (June 1999) E-commerce market, presenting a framework for analyzing household choice (demand) on the Internet, conducting a strategic analysis of tbe company's business model, and advising on the design of the company's academic program. Company initially named Foveon; later renamed Plurimus.

Advisory roles: Cartel case in. the computer industry (US), 2009; German media market (Germany), 2007; Major price-fixing litigation (US), 1999-2001

Bates White LLC, Academic Afliliate, 2005-present

Publications

"The Economics of Television and Online Video Markets," Chapter 7 in Anderson, S., Waldfogel, J., and D. Stromberg, Handbook of Media Economics, volume 1A, 2015 Elsevier Press.

"Cable Regulation in the Internet Era," Chapter 3 in. Rose, N., ed, "Economic Regulation and Its Reform: What Have We Learned?", 2014, University of Chicago Press.

"Accommodating Endogenous Product Choices: A Progress Report," International Journal of Industrial Organization, v30 (2012), 315-320.

"The Welfare Effects of Bundling in Multichannel Television Markets," (with Ali Yurukoglu), American Economic Reuters, vl02n2 (April 2012), 643-685 (lead article).

"Price Discrimination in Service Industries," (with A. Lambrecht, K. Seim, N. Vilcassim, A. Cheema, Y. Chen, K. Hosanger, R. Iyengar, O. Koenigsberg, R. Lee, E. Miravete, and and O. Sabin), Marketing Letters, v23 (2012), 423-438.

"Economics at the FCC: 2007-2008," (with Evan Kwerel and Jonathan Levy), Iteuiem of Industrial Organization, v33n3 (November 2008), 187-210.

"The Discriminatory Incentives to Bundle: The Case of Cable Television," Quantitative Marketing and Economics, v6nl (March 2008).,41-78. — Winner, 2009 Dick Wittink Prize for the best paper published in the QME

"Bidding Asymmetries in Multi-Unit Auctions: Implications of Bid Function Equilibria in the British Spot Market for Electricity, (with Joseph Crespo and Helen Tauchen),

Gregory S. Crawford cv, October 2016 0 0 0 0 International Journal of Industrial Organization, v25n6 (December 2007), 1233-1268. 0 "Bundling, Product Choice, and Efficiency: Should Cable Television Networks Be Offered A La Carte'," (with Joseph Cullen), Information Economics and Policy, v19n3-4 (October 2007), 379-404. 0 "Monopoly Quality Degradation and Regulation in Cable Television," (with Matthew Shum), 0 Journal of Law and Economics, v50nl (February 2007), 181-209. "Uncertainty and Learning in Pharmaceutical Demand," (with Matthew Shum), 0 Econometrica, v73n4 (July 2005), 1137-1174. 0 "Recent Advances in Structural Econometric Modeling: Dynamics, Product Positioning, and Entry," (with J.-P. Dube, K. Sudhir, A. Ching, M. Draganska, J. Fox, 0 W. Hartmann, G. Hitsch, B. Viard, M. Villas-Boas, and N. Vilcassim), ,0 Marketing Letters, v16n2 (July 2005).

Demand Welfare," "The Impact of the 1992 Cable Act on Household and RAND Journal of Economics, v31n3 (Autumn 2000), 422-449. Reports "Empirical analysis of BT's automatically renewable contracts," (with ESMT Competition 0 Analysis, Commissioned Research Study for the OfFice of Communications), August 2010. Also Supplementary Report, February 2011. "Television Station Ownership Structure and the Quantity and Quality of TV Programming," (Commissioned Research Study for the Federal Communications Commission), July 2007. 0 Work in Progress 0 Articles Under Review 0 "The Welfare Effects of Vertical Integration in Multichannel Television Markets," (with Robin Lee, Michael Whinston, and Ali Yurukoglu), mimeo, University of Zurich, December 2015, 0 revise resubmit at Econometrica. 0 and "Asymmetric Information and Imperfect Competition in Lending Markets," (with Nicola Pavanini and Fabiano Schivardi), working paper, University of Zurich, April 2015, revise and resubmit at American Economic Review.

0 "The Welfare Effects of Monopoly Quality Choice: Evidence from Cable Television Markets," 0 (with Matthew Shum and Alex Shcherbakov), mimeo, University of Zurich, August 2015, revise and resubmit at American Economic Reuters.

"The impact of 'rollover'ontracts on switching in the UK voice market: Evidence from 0 disaggregate customer billing data," (with Nicola Tosini and Keith Waehrer), Working paper, University of Warwick, June 2011, revise and resubmit at Economic Journal. 0 Gregory S. Crawford cv, October 2016 0 %forking Papers

"Demand estimation. with unobserved choice set heterogeneity," (with Rachel Griflith and Alessandro Iaria), University of Zurich, April 2016.

"The (inverse) demand for advertising in the UK: Should there be more advertising on television?," (with Jeremy Smith and Paul Sturgeon), working paper, University of Warwick, October 2011.

"The Empirical Consequences of Advertising Content in the Hungarian Mobile Phone Market," (with Jozsef Molnar), University of Arizon.a, March 2008.

Vilork In Progress

"Accommodating choice set heterogeneity in demand: Evidence from retail scanner data," (with Rachel Griffith and Alessandro Iaria), University of Warwick, October 2011.

"Orthogonal Instruments: Estimating Price Elasticities in the Presence of Endogenous Product Characteristics," (with Dan Ackerberg and Jin Hahn), mimeo, University of Warwick, June 2011.

"Channel 5 or 500? Vertical Integration, Favoritism, and Discrimination in Multichannel Television," (with Robin Lee, Breno Viera, Michael Whinston, and Ali Yurukoglu), mimeo, University of Zurich, October 2013.

"The Welfare Effects of Vertical Integration in Multichannel Television Markets," (with Robin Lee, Michael Whinston, and Ali Yurukoglu), mimeo, University of Zurich, March 2014.

Grants

"Endogenous Product Characteristics in Empirical Industrial Organization," Economic and Social Research Council, X,140,000 ( $220,000), 2010-2012.

"The Empirical Consequences of Advertising Content" (with Jozsef Molnar), Hungarian Competition Commission, 10,000,000 Hungarian Forint ( $50,000), 2007-2008

Other Professional Activities Editing/Refereeing

Associate Editor, International Journal of Industrial Organization, October 2005 - present.

Editorial Board, Information Economics and Policy, December 2007 - present.

Excellence in Refereeing Award, American Economic Review, 2009.

Gregory S. Crawford cv, October 2016 Referee for Econometrica, American Economic Review, Review of Economics Studies, RAND Journal of Economics, Review of Economics and Statistics, Quantitative Marketing and Economics, National Science Foundation, International Journal of Industrial Organuation, Journal of Industrial Economics, Journal of Applied Econometrics, Information Economics and Policy, Management Science, Southern Economic Journal

Keynote Lectures (previous and planned)

"Vertical Integration in Media and Communications Markets": 5th Workshop on the Economics of ICTs (Oporto, Portugal, 3/14), FSR/EUI Annual Seminar on the Economics and Policy of Communications and Media 2014 (Florence, 3/14)

"How much is too much? A closer look at choice in the entertainment industry," The Future of Broadcasting Conference (London, 6/12)

Academic Presentations (previous and planned)

2016 Presentations: Winter Marketing-Economics Summit (Denver, 1/16), University of Bern (2/16), ESMT (Berlin, 6/16), Pompeu Fabra (Barcelona, 11,16) 2015 Presentations: NYC Media Seminar (2/15), Empirical Models of Differentiated Products (IFS, London, 6/15), Advances in the Economics of Antitrust and Consumer Protection (Paris, 9/15), University of Pennsylvania (Wharton, 9/15), 15th Media Economics Workshop (Cape Town, 11/15), Bocconi (12/15), ECARES (Brussels, 12/15) 2014 Presentations: Winter Marketing-Economics Summit (Wengen, Switzerland, 1/14), Industrieokonomischer Ausschuss (Hamburg, 2/14), Network of Industrial Economists (Manchester, UK, 10/14) 2013 Presentations: Tilburg University (11/13) 2012 Presentations: University of East Anglia / Centre for Competition Policy (5/12), PEDL Inaugural Conference (5/12) 2011 Presentations: University of Cyprus (3/11), CREST (Paris, 6/11), EARIE (Stockholm, 9/11), University of Zurich (9/11), University of Mannheim (10/11). 2010 Presentations: LBS (1/10), UCL (4/10), Oxford (5/10), Invitational Choice Conference (5/10), Manchester University (9/10), EIEF (Rome, 10/10), University of Venice (10/10), University College Dublin (11/10). 2009 Presentations: ESMT, Berlin (5/09), CEPR IO, Mannheim (5/09), University of Leuven (9/09), University of Toulouse (Econometrics Workshop and Competition Policy Workshop), (11/09)

Conference Organization: CEPR Applied IO Workshop: Jerusalem (Hebrew University, 2017), London (IFS, 2016) Zurich (UZH, EARIE 2010-2016: Scientific Committee Economics of Media Markets 2010: Scienti6c Committee, Triangle Applied Economics of Media Markets 2010: Scientific Committee, Triangle Applied Micro Conference 2000: Organizer, Triangle Applied Micro Conference 1999: Co-organizer

Non-Academic Presentations

Gregory S. Crawford cv, October 2016 0

"Damages Litigation: Issues and Challenges in Complex Antitrust Cases," CRESSE 2016 (Panel, 0 Rhodes, 7/16) "Multichannel Distribution: Experimentation, Innovation and Enforcement," CRA Conference on Economic Developments in European. Competition Policy (Panel, Brussels, 12/15)

0 "Understanding 'New Media'nd its lessons for non-media in.dustries," University of Zurich 0 Dept. of Economics, Advisory Board Meeting (Zurich, 11/13)

"New Media: Economic Perspectives," University of Warwick, Window on Research 0 (Coventry, UK, 6/11)

"Doing Good with (Good) Econometrics," Warwick Economics Summit, University of Warwick, (Coventry, UK, 2/11)

0 0 0 0 0

Gregory S. Crawford cv, October 2016 Before the UMTED STATES COPYRIGHT ROYALTY JUDGES Library of Congress Washington, D.C.

In the Matter of: Docket No. 16-CRB-0001-4RIPSSR (2018-2022) Determination of Royalty Rates and Terms for Transmission of Sound Recordings by Satellite Radio and "Preexisting" Subscription Services (SDARS IHj

DECLARATION OF DR. GREGORY CRAWFORD. PH.D..

I, Dr. Gregory S. Crawford, Phd., declare under penalty ofperjury under the laws ofthe

United States ofAmerica that the statements contained in my Written Direct Testimony in the above-captioned matter are true and correct to the best ofmy lmowledge, information, and belief.

Executed this 19th day ofOctober 2016 in Zurich, Switzerland.

Dr. Grbgor+8. Craw8fid, Ph.D.,