Determination of Royalty Rates and Terms for Transmission of Sound Recordings by Satellite Radio and “Preexisting” Subscription Services (SDARS III)

Federal Register, Volume 83 Issue 243 (Wednesday, December 19, 2018)
[Federal Register Volume 83, Number 243 (Wednesday, December 19, 2018)]
[Rules and Regulations]
[Pages 65210-65270]
From the Federal Register Online via the Government Publishing Office []
[FR Doc No: 2018-26922]
[[Page 65209]]
Vol. 83
No. 243
December 19, 2018
Part II
Library of Congress
Copyright Royalty Board
37 CFR Part 382
Determination of Royalty Rates and Terms for Transmission of Sound
Recordings by Satellite Radio and ``Preexisting'' Subscription Services
(SDARS III); Final Rule
Federal Register / Vol. 83 , No. 243 / Wednesday, December 19, 2018 /
Rules and Regulations
[[Page 65210]]
Copyright Royalty Board
37 CFR Part 382
[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 III)
AGENCY: Copyright Royalty Board, Library of Congress.
ACTION: Final rule and order.
SUMMARY: The Copyright Royalty Judges announce their final
determination of the rates and terms for the digital transmission of
sound recordings and the reproduction of ephemeral recordings by
preexisting subscription services and preexisting satellite digital
audio radio services for the period beginning January 1, 2018, and
ending on December 31, 2027.
    Effective Date: December 19, 2018.
    Applicability Date: The regulations apply to the license period
beginning January 1, 2018, and ending December 31, 2027.
ADDRESSES: The final determination is posted in eCRB at For access to the docket to read the final determination
and submitted background documents, go to eCRB and search for docket
number 16-CRB-0001 SR/PSSR (2018-2022).
by telephone at (202) 707-7658 or by email at
I. Introduction
    The purpose of the Copyright Royalty Judges (Judges) in the present
proceeding is to determine the royalty rates and terms applicable to
Preexisting Subscription Services (PSS) and Satellite Digital Audio
Radio Services (SDARS) for licenses established by the Copyright Act
(Act) to utilize copyrighted sound recordings. See 17 U.S.C. 112, 114.
The Act requires the Judges to determine applicable rates and terms
every five years. See 17 U.S.C. 801(b)(1), 804(b)(3)(B).
    In determining the PSS rates, the Judges considered proposals from
both Music Choice and SoundExchange as guideposts rather than as
benchmarks and determined a rate based upon the current statutory rate
as adjusted to meet statutory requirements. In determining the SDARS
rates, the Judges relied most heavily on the opportunity cost approach
proffered by SoundExchange, but the Judges utilized opportunity cost
survey data that they found more appropriate than the data relied on by
    After the Judges issued the Initial Determination in this
proceeding on December 14, 2017, both Sirius XM Radio, Inc., (Sirius
XM), the lone SDARS, and Music Choice filed timely motions for
rehearing. SoundExchange filed responses opposing each rehearing
motion, and Sirius XM and Music Choice filed replies. On April 17,
2018, the Judges ruled on the rehearing motions. See Order Granting In
Part and Denying In Part . . . Motion[s] for Rehearing (Apr. 17, 2018).
By this order, the Judges denied the Music Choice motion and asked for
additional briefing on the primary issue Sirius XM raised, viz.,
whether the Judges should reduce the royalty rate for SDARS set in the
Initial Determination to a rate not lower than 14.7% of Gross Revenues.
Id. at 9. The parties filed briefs and responses and the Judges took
the issue under advisement.
    On October 11, 2018, the President signed into law the Orrin G.
Hatch-Bob Goodlatte Music Modernization Act, Public Law 115-264, 132
Stat. 3676 (Oct. 11, 2018) (MMA). That law includes a provision
amending section 804(b)(3)(B) of the Copyright Act (Act) to state that
``with respect to pre-existing satellite digital audio radio services,
the terms and rates set forth by the Copyright Royalty Judges on
December 14, 2017, in their initial determination for the rate period
ending on December 31, 2022, shall be in effect through December 31,
2027, without any change based on a rehearing under section 803(c)(2) .
. . .'' Id. sec. 103. As a consequence of this statutory provision, the
Judges dismissed the pending rehearing as moot. See Order Dismissing
Rehearing Proceeding (Oct. 11, 2018).
    Based upon the totality of the record, and in accordance with the
following reasoning and analysis, the Judges determine that the
applicable rates and terms for the period beginning January 1, 2018,\1\
shall be:
    \1\ In the Judges' Initial Determination in this proceeding,
they established rates for the period January 1, 2018 through
December 31, 2022. Under the MMA, these rates shall remain in effect
until December 31, 2027. See 17 U.S.C. 804(b)(3)(B) (as amended by
the MMA). Note that all redactions in this publication were made by
the Copyright Royalty Judges and not by the Federal Register.
    For PSS: 7.5% of Gross Revenues, as that term is defined for PSS.
    For SDARS: 15.5% of Gross Revenues, as that term is defined for
II. Background
A. Statutory Licenses
    In 1995, Congress granted to sound recording copyright owners the
exclusive right ``to perform the copyrighted [sound recording] publicly
by means of a digital audio transmission.'' \2\ 17 U.S.C. 106(6).
Concurrently, Congress limited that exclusive right by creating two
statutory licenses that would enable certain users, including SDARS and
PSS, to transmit digitally sound recordings without obtaining a
voluntary license from each copyright owner. See 17 U.S.C. 112(e),
114(d). The section 112 license (ephemeral license) allows an entity
that transmits a sound recording digitally to make ephemeral
phonorecords of the sound recording to facilitate the transmission.
Section 112(e) describes conditions under which an entity may license
the ephemeral sound recording.\3\ Section 114 describes limits that
apply to the digital transmission license.\4\
    \2\ See Digital Performance Right in Sound Recording Act of
1995, Public Law 104-39, 109 Stat. 336 (1995).
    \3\ Section 112 provides that a sound recording transmitter may
make no more than one ephemeral phonorecord, ``unless the terms and
conditions of the statutory license allow for more.'' 17 U.S.C.
    \4\ Specifically, section 114 excludes from the statutory
license transmissions by interactive services. See 17 U.S.C.
B. The Standards for Determining Royalty Rates
    Section 801(b)(1) of the Act provides that the Judges shall ``make
determinations and adjustments of reasonable terms and rates of royalty
payments'' for the statutory licenses set forth in, inter alia, section
114(f)(1) (``digital performance license'').\5\ The digital performance
license requires that the Judges set rates and terms that are
``reasonable.'' Id. In addition, section 801(b)(1) provides that these
``reasonable'' rates shall be calculated to achieve four specific
    \5\ Sirius XM and SoundExchange agree in substance that the
Judges should conform the SDARS regulations regarding ephemeral
licenses to the language adopted by the Judges in Web IV. See SEPFF
] 2371; SXMPFF ] 492. The Judges approve this agreement and adopt it
in the regulations for the forthcoming rate period. See infra,
section III.
    (A) To maximize the availability of creative works to the
    (B) 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 reflect the relative roles of the copyright owner and the
copyright user in the product made available to the public with
[[Page 65211]]
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).
    In SDARS 1, the Judges detailed the historical treatment of these
section 801(b)(1) standards. See Determination of Rates and Terms . . .
73 FR 4080, 4082-84 (Jan. 24, 2008) (SDARS I). There, the Judges noted
that the section 801(b)(1) factors originated in the protracted
legislative process that ultimately produced the Copyright Act of 1976.
The SDARS I Judges examined the legislative history of the 1976 Act and
noted that the motivation for adopting the four itemized 801(b)(1)
factors arose from an exchange between two law professors, Professor
Ernest Gellhorn, on behalf of certain copyright users, and Professor
Louis H. Pollack, on behalf of certain copyright owners. The issue
between the professors was the constitutionality of the Copyright
Royalty Tribunal (CRT), a predecessor of the Copyright Royalty Board.
As recounted in SDARS I: ``Professor Gellhorn had recommended that, in
order to bolster the constitutionality of the Tribunal, the Congress
should, inter alia, adopt statutory standards beyond the vague
criterion of `reasonableness.' '' SDARS I, 73 FR at 4082 (citing
Hearings on H.R. 2223 before the Subcomm. on Courts, Civil Liberties,
and the Administration of Justice of the House Comm. on the Judiciary,
94th Cong., 1922 (1975).\6\ After consideration of alternative
potential statutory language, Congress adopted the four-part itemized
factors included in section 801(b)(1) to supplement the ``reasonable''
rate requirement. Id.
    \6\ The SDARS I Judges also noted that, in like fashion, the
Register of Copyrights concluded that it would be ``wise to
establish, in the statute, certain criteria beyond `reasonableness'
that each Panel is to apply to its decision-making.'' Id. (citing
Second Supplementary Report of the Register of Copyrights on the
General Revision of the U.S. Copyright Law, Chapter XV, at 31
    There is additional legislative history regarding the itemized four
factors in section 801(b)(1) that aids in understanding how those
factors should be applied and informs economic analysis under these
statutory provisions. This legislative history is highlighted by
dueling positions taken in Congressional testimony in 1967 by the
licensors, through the National Music Publishers Association (NMPA) and
its economic witness, Robert R. Nathan, and by the licensees, the
Recording Industry Association of America (RIAA), through their
counsel, Thurman Arnold, Esq., a well-known advocate of strong
antitrust enforcement. See Hearing on S. 597, Subcomm. on Patents,
Trademarks and Copyrights of the S. Committee on the Judiciary, (Mar.
20-21, 1967) (Senate Hearing).
    Mr. Nathan criticized any proposed legislation that would subject
the songwriting industry to a statutory mechanical licensing scheme.
Id. at 382. He did not agree that licenses in the music industry should
be treated differently than how ``we generally function under
competitive marketplace bargaining arrangements whereby most entities
in our economy bargain for that which goes into the creation of goods
and services and also bargain the price for which those goods and
services are sold.'' Id. He further noted that the statutory mechanical
royalty rate was in part a reaction to an early 20th century concern
regarding a Supreme Court decision allowing a player-piano manufacturer
to play songs through the use of perforated paper rolls fed into the
new devices (player pianos), without a license and without a duty to
pay royalties to the songwriters and publishers. White-Smith Music
Publishing Company v. Apollo Company, 209 U.S. 1 (1908). As Mr. Nathan
explained: ``[T]he Aeolian Co.[,] had gained control of some 80 percent
of the musical compositions and Congress . . . fear[ed] the threat of
monopoly in the mechanical reproduction of music.'' Senate Hearing at
382-83. The Copyright Act of 1909 superseded the effect of White-Smith
by creating a statutory license and imposing a fixed statutory rate for
mechanical reproduction of musical compositions.
    In his 1967 testimony, Mr. Nathan advocated that Congress eliminate
the compulsory license and the statutory rate, and he specifically
urged Congress to resist replacing the fixed statutory fee with a
regulatory standard to be implemented by a quasi-adjudicatory body. As
to the latter point he explained to Congress: ``[O]ne might ask . . .
whether the music publishing industry has any characteristics of a
public utility? I submit . . . that there is nothing in the music
publishing industry which gives [it] the characteristics or the
elements of a public utility . . . .'' Id. at 383. Mr. Nathan noted
what he felt was a key distinction: Unlike traditional public utilities
such as ``railroad systems'' or ``streetcar lines,'' the songwriting
and publishing industry is ``a creative and nonstandardized area,'' and
``[m]onopoly and public utility aspects are just not prevalent in this
industry.'' Id.
    The licensees' opposing position, expressed by Mr. Arnold on behalf
of the RIAA, contained the seeds of the standard ultimately adopted in
section 801(b)(1). As Mr. Arnold testified, the statute should include,
inter alia, ``accepted standards of statutory ratemaking,'' including a
rate ``that insures the party against whom it is imposed a reasonable
return on . . . investment'' and ``that divides the rewards for the
respective creative contributions of the record producers [the
licensees] and the copyright owners . . . equitably between them.'' Id.
at 469.
    Mr. Nathan criticized this approach on two fronts. First, he argued
that the ``personal service'' nature of the songwriting and publishing
industry precluded application of a ``reasonable rate of return''
requirement for establishing the compulsory royalty rate. Second, with
regard to the division of the ``rewards'' proposal, Mr. Nathan stated
that ``I have never in all my experience encountered this novel concept
of dividing rewards for creative contributions as a meaningful and
relevant standard of ratemaking.'' Id. at 1093-94.\7\
    \7\ As the present record (and the record in Phonorecords III)
demonstrates, subsequent to Mr. Nathan's 1967 testimony, the
economic concept of ``dividing rewards for creative contributions as
a meaningful and relevant standard of ratemaking'' has blossomed,
with the application of Opportunity Cost/Efficient Component Pricing
approaches, Nash Bargaining Solutions, and Shapley Value analyses.
    Resolution of this 1967 dispute languished until 1976, when
Professor Gellhorn successfully convinced Congress to adopt an itemized
standard in the final statute. See F. Greenman & A. Deutsch, The
Copyright Royalty Tribunal and the Statutory Mechanical Royalty:
History and Prospect, 1 Cardozo Arts & Ent. L.J. 1, 53, 59 (1982). In
so doing, Congress did not explicitly address the economic dispute
between Mr. Arnold and Mr. Nathan regarding the relative merits of a
market-based rate versus a rate established in some other manner.
    Under the itemized section 801(b)(1) standard, the Judges have the
discretion to choose a market rate, a market-based rate, or a rate
unrelated to market evidence. Music Choice v. Copyright Royalty Bd.,
774 F.3d 1000, 1010 (D.C. Cir. 2014) (and citations therein). Any such
rate would be legally appropriate provided it was not ``arbitrary,
capricious, an abuse of discretion, or otherwise not in accordance with
law, or if the facts relied upon by the [Judges]
[[Page 65212]]
have no basis in the record.'' Id. at 1007. Indeed, in Music Choice,
the D.C. Circuit reaffirmed that ``the Copyright Act gives the Judges
of the Copyright Royalty Board broad discretion to set rates and terms
for compulsory licenses of the digital performance of sound
recordings.'' Id. at 1016 (emphasis added).
C. Prior Proceedings
    This proceeding is not the first in which the Judges or their
predecessors have applied the section 801(b) factors to determine
royalty rates.\8\ In SDARS I, the Judges detailed the historical
treatment of these factors by their predecessors, the Copyright Royalty
Tribunal and the Librarian in his administration of the Copyright
Arbitration Royalty Panel (CARP) system. See Determination of Rates and
Terms . . . , 73 FR 4080, 4082-84 (Jan. 24, 2008) (SDARS I). In SDARS
I, the Judges chose to ``begin with a consideration and analysis of the
[market] benchmarks and testimony submitted by the parties, and then
measure the rate or rates yielded by that process against the [section
801(b)] statutory objectives'' to reach a decision. Id. at 4084.
    \8\ The Copyright Royalty Tribunal (CRT) applied the 801(b)
factors in a section 116 (Jukebox) rate adjustment and a section 115
(Phonorecords) rate adjustment. The Librarian of Congress, as
administrator of a Copyright Arbitration Royalty Panel (CARP) issued
a determination for the section 114 satellite radio license (SDARS
I). In 2017, the Judges presided over a contested Phonorecords rate
hearing, the determination of which will issue after the present
determination and will involve application of the 801(b) policy
factors to the Phonorecords license.
    The precedent guiding the present panel of Judges signals an
analysis in which the Judges may weigh the evidence presented to
support the rate proposals, including marketplace benchmarks, apply the
section 801(b) policy factors to assure the final rates are consonant
with those factors and, if the evidence permits, also establish a zone
of reasonableness within which the rate shall be set.\9\
    \9\ The U.S. Court of Appeals for the D.C. Circuit has also
concluded that the Judges may apply the ``[section 801(b)] . . .
objectives [to] determine a range of reasonable royalty rates that
would serve all these objectives adequately but to differing
degrees, [and] the [Judges are] free to choose among those rates,
and courts are without authority to set aside the particular rate
chosen . . . if it lies within a ``zone of reasonableness.'' See
Recording Indus. Ass'n of America v. Copyright Royalty Tribunal, 662
F.2d 1, 9 (D.C. Cir. 1981) (footnotes omitted). Thus, the Judges may
establish such a zone of reasonableness, but are not required to do
D. The Present Proceeding
    The Judges commenced the present proceeding with publication of
notice seeking petitions to participate. See 81 FR 255 (Jan. 5, 2016).
Seven entities filed petitions to participate.\10\ The Judges dismissed
the petitions of Music Reports, Inc. and David Powell. Muzak LLC
withdrew its petition to participate. The parties participating in the
hearing were George Johnson d/b/a GEO Music Group (GEO), Music Choice,
Sirius XM Radio, Inc. (Sirius XM), and SoundExchange, Inc.
    \10\ Original petitioners included George Johnson d/b/a GEO
Music Group; Music Choice; Music Reports, Inc.; Muzak LLC; Sirius XM
Radio, Inc.; SoundExchange, Inc. (SoundExchange); and David Powell.
SoundExchange appeared on behalf of itself and its members, the
American Association of Independent Music; the American Federation
of Musicians of the United States and Canada; the Recording Industry
Association of America; the Screen Actors Guild and the American
Federation of Television and Radio Artists; Sony Music
Entertainment; Universal Music Group; and Warner Music Group.
    The Judges presided over an evidentiary hearing that commenced on
April 12, 2017, and ended on May 18, 2017. Parties to the hearing
presented oral closing argument on July 18. The parties called 35
witnesses,\11\ including 15 experts.\12\ Of the 856 exhibits marked for
identification for the hearing (not including illustrative
presentations by various witnesses) the Judges admitted 511 (including
those admitted for limited purpose) into evidence during the
hearing.\13\ On June 14, the parties filed their respective Proposed
Findings of Fact (PFF) and Proposed Conclusions of Law (PCL). Parties
filed Reply PFF and PCL on June 29.
    \11\ In addition to live witnesses, participants also designated
prior testimony of witnesses in prior proceedings. See 37 CFR
    \12\ GEO Music Group (GEO) presented the testimony of George
Johnson. Mr. Johnson asked to be qualified as an expert in the music
sound recording business. There being no objection, the Judges
acknowledged his experience as a songwriter, singer, and independent
record producer for approximately 30 years and qualified him for
purposes of the present proceeding as an expert in the music
    \13\ Immediately prior to and during the hearing in this
proceeding, participants filed motions seeking to limit or exclude
opposing parties' evidence. The Judges' conclusions on those motions
are issued by separate order or orders. References to evidence in
this Determination are to evidence admitted to the record.
III. The Section 112 Ephemeral License
    The ephemeral license rates that the Judges are to determine in
this proceeding shall ``most clearly represent the fees that would have
been negotiated in the marketplace between a willing buyer and a
willing seller.'' 17 U.S.C. 112(e)(4). All parties to the present
proceeding agree that the value of the section 112 ephemeral license is
linked to the value of the section 114 performance license.\14\ Music
Choice asked that the Judges include the section 112 rate in the
overall rate. Sirius and SoundExchange asked the Judges to determine
that the value of the licenses be allocated 5% to the ephemeral license
and 95% to the performance license, consistent with the current
regulations applicable to SDARS, webcasters, and new subscription
(CABSAT) services. See, e.g., Sirius XM . . . Proposed Findings . . .
and Conclusions at 234 (SXM PFFCL); Proposed Findings . . . and
Conclusions of SoundExchange . . . at 938 (SX PFFCL); see 37 CFR
382.3(c), 382.12(b) (2016).
    \14\ See Music Choice Written Direct Statement at 6;
Introductory Memorandum to the Written Statement of Sirius Radio
Inc. at 1; Proposed Rates and Terms of SoundExchange, Inc. and
Copyright Owner and Artist Participants at 5.
    The parties' positions and the Judges' decisions concerning the
ephemeral license regulations are detailed in section XI.C of this
Determination; the regulatory language adopted by the Judges is
attached as Appendix A.
IV. PSS Performance License
A. Background
    The Act defines a PSS as ``a service that performs sound recordings
by means of noninteractive audio-only subscription digital audio
transmissions, which was in existence and was making such transmissions
to the public for a fee on or before July 31, 1998 . . . .'' 17 U.S.C.
114(j)(11). When Congress enacted that definition, there were three PSS
entities in existence. See H.R. Rep. No. 105-796, at 81, 85, 89 (Oct.
8, 1998). Only two remain, and Music Choice was the only PSS that
participated in this proceeding.\15\ SoundExchange represented
Copyright Owners in the PSS portion of the proceeding. George Johnson,
an individual licensor, also proposed a PSS rate.
    \15\ The other remaining PSS entity, Muzak LLC, filed a Petition
to Participate, but withdrew it before the deadline for filing
Written Direct Statements.
    Music Choice operates a residential audio service that consists of
50 channels of audio programming delivered to subscribers' televisions.
Written Direct Testimony of David J. Del Beccaro, Trial Ex. 55, at 4
(Del Beccaro WDT). Music Choice's services are delivered to customers
by cable operators and other multichannel video programming
distributors (MVPDs) as part of customers' digital basic cable service.
    In addition to its cable TV-based service, Music Choice makes its
50 cable channels, plus an additional 25 channels of audio programming,
available to authenticated television subscribers through its website
and a
[[Page 65213]]
mobile app. Id. Music Choice describes these internet transmissions as
``an ancillary part of its residential music business . . . .'' Written
Rebuttal Testimony of David J. Del Beccaro, Trial Ex. 57, at 25 (Del
Beccaro WRT).
1. PSS Rates From SDARS II
    The parties in the prior proceeding (SDARS II) reached agreement on
the rates and terms of the section 112 license prior to the hearing.
See 78 FR at 23054-56.\16\ Therefore, the Judges' focus in that
proceeding was limited to determining the appropriate rates and terms
for the section 114 license. The Judges began with a consideration and
analysis of the market benchmarks and testimony submitted by the
parties and then measured the rate or rates yielded by that process
against the Section 801(b) statutory objectives to reach a decision. 78
FR at 23055. The Judges repeat that approach in the current proceeding.
    \16\ In the SDARS II proceeding, SoundExchange and Music Choice
submitted a joint stipulation with respect to the Section 112(e)
ephemeral license, and the Judges adopted the proposal based on the
stipulation. 78 FR at 23055-56. The provision addressing the Section
112(e) license appears in current CRB Rule 382.3(c). It states that
``[t]he royalty payable under 17 U.S.C. 112(e) for the making of
phonorecords used by the Licensee solely to facilitate transmissions
for which it pays royalties as and when provided in this subpart
shall be included within, and constitute 5% of, the total royalties
payable under 17 U.S.C. 112(e) and 114.''
    In SDARS II, Music Choice advocated adoption of the annual
royalties it pays to performing rights societies (PROs) (i.e., ASCAP,
BMI, and SESAC) for the right to perform musical works to subscribers
of its residential audio service as a precedential benchmark. Indeed,
Music Choice asserted that the Judges were required to rely on that
musical works rate. The Judges rejected that contention but analyzed
whether the rates that Music Choice paid the PROs were a useful
benchmark. 78 FR at 23056. Music Choice contended that two pieces of
evidence corroborated use of the musical works rates as a benchmark:
(1) Decisions from Canada and the United Kingdom concluding that
royalty rates for sound recordings and musical compositions have
equivalent value \17\ and (2) results of an economic model called the
Asymmetric Nash Bargaining Framework (Nash Framework) \18\ offered by
Music Choice's expert, Professor Gregory Crawford.\19\ Based on his
analysis, Professor Crawford concluded the PRO rates were an
appropriate benchmark for the sound recording license at issue.
    \17\ The Judges dismissed Music Choice's reliance on foreign
jurisdictions because of a lack of proof of comparability between
foreign markets and U.S. markets. Further, Music Choice failed to
convince the Judges that the governing laws were sufficiently
similar to U.S. law to offer even analogous reasoning. See 78 FR at
    \18\ The Nash Framework, as presented in the instant proceeding,
is discussed in greater detail infra, section IV.C.1.a.
    \19\ Professor Crawford's Nash Framework from SDARS II (as well
as the Judges' reasons for rejecting it) is described at length in
the determination and need not be repeated here. See SDARS II, 78 FR
at 23056-57, 23058. As discussed below, in the current proceeding
Music Choice does not premise its Nash-based model (or any other
model) on an asserted equivalency between the value of sound
recordings and musical works, in light of the Judges' rejection of
that argument on the record presented in SDARS II. Nonetheless,
Professor Crawford's Nash Framework in the instant proceeding is
strikingly similar to his Nash Framework in SDARS II.
    The Judges disagreed and found that the musical works benchmark
lacked comparability to the hypothetical PSS market. Id. at 23058. The
Judges found that the musical works market involved different sellers
(PROs versus record companies) selling different rights (musical works
performance rights versus sound recording performance rights) than
those at issue in this proceeding.\20\
    \20\ The Judges acknowledged that musical works performance
rights and sound recording performance rights are likely perfect
complements, but concluded that, based on the record, such
complementarity had not been shown to inform the decision regarding
relative value of the rights.
    With regard to the Nash Framework, the Judges noted:
    The Nash Framework is a theoretical concept whose goal is to
evaluate how the surplus from a hypothetical transaction might be
divided between negotiating parties. Even assuming that the Nash
Framework has predictive value in some real-world contexts, Music
Choice provided no data to support the theoretical approximations in
the market for any intellectual property rights, much less those
that the Judges are charged with evaluating. Therefore, the Judges
find that the Nash Framework is not useful corroborating evidence.
78 FR at 23058.\21\
    \21\ The Judge who dissented from the majority decision offered
what the majority characterized as a ``more spirited rejection of
the probative value of the Nash Framework as proffered in this
context.'' The majority concurred with this assessment but concluded
that ``as a threshold matter, [the] Nash Framework, without real-
world data to support its predictive capacity, is unworthy of
further consideration. 78 FR at 23058, n.17.
    For its part, SoundExchange offered certain marketplace agreements
executed by interactive music streaming services as a benchmark. The
Judges also rejected this proposed benchmark on comparability grounds.
78 FR at 23058.\22\
    \22\ The markets that the proffered agreements covered were
subscription interactive webcasting, ringtones/ringbacks, and
digital downloads. The Judges concluded that these markets involve
the licensing of products and rights separate and apart from the
right to publicly perform sound recordings in the context of the PSS
proceeding. The Judges noted that the buyers are different from the
target PSS market. Thus, the key characteristic of a good
benchmark--comparability--was not present. 78 FR at 23058. The
Judges noted that the bundling of Music Choice's services with
multiple channels of video and other non-music programming
significantly dim the possibility of market comparators. The Judges
concluded that ``in the absence of some rational, reasoned
adjustment to make the music agreements data more comparable to the
PSS market, the Judges find its probative value in this proceeding
of only marginal value.'' Id.
    The Judges concluded that the evidence presented by Music Choice
framed the lower end of a range of reasonable rates and that presented
by SoundExchange framed the upper end. 78 FR at 23059. Having rejected
the parties' respective proffered benchmarks (and proposed
corroborating evidence) for any purpose other than to frame a range of
potential rates, the Judges were left with a consideration of the then-
prevailing royalty rate of 7.5% of gross revenues, which fell within
that range. The Judges started with the then-prevailing rate and
applied the Section 801(b) factors. Consideration of the section 801(b)
factors persuaded the Judges that they should adopt that rate, but
adjust it up to 8.5% based on Music Choice's planned expansion of its
service from 46 channels to up to 300. The Judges concluded that the
planned expansion would result in a substantial increase in the number
of plays of recorded music without any corresponding increase in
compensation. 78 FR at 23059-60. Nevertheless, the Judges acknowledged
that the upward adjustment of the benchmark rate was based on projected
usage that was likely to occur during the rate period. The Judges noted
that ``[s]hould Music Choice alter its anticipated usage under the
statutory license in the future, such evidence can be taken into
account in a future rate proceeding. . . .'' Id. at 23061.
2. Standard for PSS Royalty Rates
    When the Judges determine a section 114 rate for PSS, they
generally begin with an appropriate rate (or range of rates) and adjust
it, as appropriate, in accordance with the section 801(b)(1) statutory
factors. By contrast, the section 112 ephemeral license requires the
Judges, among other things, to ``establish rates that most clearly
represent the fees that would have been negotiated between a willing
buyer and a willing seller.'' 17 U.S.C. 112(e)(4).\23\
[[Page 65214]]
The ephemeral license also requires a minimum fee for each type of
service offered by a transmitting organization.\24\
    \23\ Section 112(e)(4) also directs the Judges to base their
decision on such factors as (1) whether use of the service may
substitute for or promote the sale of phonorecords or otherwise
interferes with or enhances the copyright owner's traditional
streams of revenue and (2) the relative roles of the copyright owner
and the transmitting organization in the copyrighted work and the
service made available to the public with respect to relative
creative contribution, technological contribution, capital
investment, cost, and risk. 17 U.S.C. 112(e)(4).
    \24\ The ephemeral license for both PSS and SDARS is addressed
in section XI.C.
    Consistent with this process, in determining the appropriate rate
for the PSS market for the upcoming rate period, the Judges must first
identify a starting point for applying the Section 801(b) policy
factors. A marketplace benchmark, if available, can be a useful
starting point for applying the Section 801(b) factors. See SDARS II,
78 FR at 23056. A key component of a marketplace benchmark is that the
market it purports to represent is comparable to the hypothetical
target market in the proceeding. See SDARS I, 73 FR at 4088 (``
`comparability' is a key issue in gauging the relevance of any
proffered benchmarks.''). In determining whether a benchmark market is
comparable, the Judges consider such factors as whether it has the same
buyers and sellers as the target market and whether they are
negotiating for the same rights. 78 FR at 23058. ``Although the
applicable Section 114 statutory standard provides a broader scope for
analyzing relevant `benchmark' rates than the `willing buyer/willing
seller standard' . . . , nevertheless potential benchmarks are confined
to a zone of reasonableness that excludes clearly noncomparable
marketplace situations.'' 73 FR at 4088.
    In the hypothetical PSS market the buyers are the PSS services, and
the sellers are the copyright owners of the sound recordings that are
being transmitted (which most often means record companies). The buyers
and sellers are negotiating for the same bundle of rights as those
granted to a PSS under section 114(f)(1)(A) of the Copyright Act to
make digital subscription transmissions of the copyrighted works.
    When the parties (or the Judges) identify variances in the
comparability of the hypothetical target market and the proffered
benchmark market, the Judges will consider reasoned adjustments that
might more closely align the two markets.\25\ Even when a proffered
benchmark is not comparable to the target market, however, the Judges
may use the rates derived from the proffered benchmark as a reference
point (or guidepost) to help frame a zone of reasonableness within
which to set an appropriate rate for the upcoming rate period (as they
did in SDARS II).\26\
    \25\ When the Judges are faced with proposed benchmarks that are
not comparable and cannot be made so with reasoned adjustments, the
Judges reject the proffered benchmarks. See, e.g., SDARS II, 78 FR
at 23058; SDARS I, 73 FR at 4089-90.
    \26\ See supra, section IV.A.1.
B. The Parties' Rate Proposals
1. Music Choice's Proposal
    Since 1998, the PSS have paid a fee based on a percentage of gross
revenues, as that term is defined by regulation.\27\ See SDARS II, 78
FR 23054, 23056; 63 FR 25394, 25413 (May 8, 1998). Music Choice has
proposed continuing that rate structure but seeks at least a 34%
reduction in the current rate of 8.5% of gross revenues, to a rate no
higher than 5.6% of gross revenues. MC PFF ] 30.
    \27\ Music Choice also does not propose an alternative per-
subscriber rate should the Judges adopt such a rate structure rather
than a percent-of-revenue structure. Neither party has proposed to
combine both rate structures (e.g., in a greater-of structure).
Given that neither party has advocated a hybrid rate structure nor
provided sufficient evidence to support such a rate structure in the
current proceeding, the Judges weigh the arguments and evidence in
the record to determine the applicable rate structure from the two
structures that the parties proposed.
2. SoundExchange's Proposal
    SoundExchange requests that the Judges change the PSS rate
structure. Rather than the percentage-of-revenue formula, SoundExchange
proposes that PSS pay a per-subscriber fee that would begin at $0.0190
in 2018, the first year of the new rate period, and rise to $0.0214 in
2022, the last year of the rate period. Amended Proposed Rates and
Terms of SoundExchange, Inc. and Copyright Owner and Artist
Participants at 7. Although SoundExchange does not offer a percent-of-
revenue alternative to its proposed per-subscriber rates, it
acknowledges that converting its proposed rates to a percentage-of-
revenue rate would plausibly yield a rate of [REDACTED] % for 2018, the
first year of the upcoming rate period. SX PFFCL ] 1949; see Written
Rebuttal Testimony of Gregory Crawford, Trial Ex. 59, ] 113 (Crawford
WRT).\28\ The evidence in the record supports that this conversion
estimate is correct; thus the lowest rate that SoundExchange proposes
([REDACTED] %) exceeds the highest rate that Music Choice proposes
(5.6%) by [REDACTED] %; it exceeds the current rate by [REDACTED] %,
assuming no increase in subscribers.\29\
    \28\ Music Choice's expert, Professor Gregory Crawford,
estimates that Music Choice would pay [REDACTED] % of its unadjusted
residential service revenue in sound recording performance royalties
in 2018 under the CABSAT rates, the basis for SoundExchange's rate
proposal, compared to the 8.5% it currently pays. Crawford WRT at ]
113, Table 6. This estimate appears consistent with the effective
rate that Stingray, a Music Choice competitor, paid in 2015 under
the CABSAT rates. SX PFFCL ] 1949; Trial Ex. 1017 at SoundX
    \29\ Assuming that the number of subscribers that carried Music
Choice's service remained flat over the upcoming rate period, the
annual 3% increases SoundExchange proposes would bring the rates to
[REDACTED] % for 2019, [REDACTED] % for 2020, [REDACTED] % for 2021,
and [REDACTED] % for 2022, or [REDACTED] % over the current rate.
This estimate is consistent with SoundExchange's estimate that a
CABSAT service pays almost [REDACTED] times as much on a per-
subscriber basis as a PSS. SX PFFCL ] 1940 and evidence cited
therein. See id. ]] 1934-35 (estimating that Music Choice's PSS
statutory royalty payment amounts to [REDACTED] cents per listener
per year whereas for a CABSAT service, the annual per-subscriber
royalty for 2017 is 22.2 cents).
    SoundExchange also proposed a separate rate for internet
transmissions by a PSS, leading to a dispute between the parties over
whether a PSS's internet transmissions are included in the PSS license
and subject to the PSS rate standard. The Judges referred the question
of categorization of Music Choice's streaming service to the Register
of Copyrights (Register) for a legal opinion. Analysis of the
Register's opinion follows in Section IV.D.2.
3. GEO's Rate Proposal
    George Johnson, d/b/a GEO Music Group (GEO) proposed that PSS pay a
per-subscriber rate of $0.10 in 2018 rising to $0.20 in 2022. Johnson
WDT at 14. He also proposed a percentage-of-revenue rate of 45% of
gross revenues. It is unclear whether he proposed that PSS pay both
components or that they pay them as a greater-of or lesser-of
structure. Mr. Johnson did not proffer a benchmark or any other
evidence to support his rate proposals for PSS. He merely stated that
``[t]hese are estimates from public data and actual royalty statements.
If the Sirius XM and Music Choice would provide number of listeners per
station and on a per-play basis, that would help GEO to better
establish a more reasonable rate.'' Id. The Judges find that there is
no evidence in the record to support the PSS rates that Mr. Johnson
proposed and therefore decline to adopt them.\30\
    \30\ Mr. Johnson also proposed requiring the PSS to install a
``buy button'' on their services to promote sales of music
downloads. 5/3/17 Tr. 2232, 2238 (Johnson). Such proposal is beyond
the scope of the Section 114 and 112 licenses and therefore beyond
the Judges' authority in the current proceeding.
C. Rates for Music Choice's Cable Radio Service
1. Analysis of the Parties' Proffered Benchmarks
a. Music Choice's Proffered Nash Model
    Music Choice, through its expert, Professor Crawford, contended
that in the absence of an appropriate marketplace benchmark, the best
way to
[[Page 65215]]
estimate the royalties that would arise in a hypothetical effectively
competitive market for the PSS sound recording rights is to use an
economic model. Professor Crawford chose as that model one based upon
the Nash Bargaining Solution, developed by Nobel-prize-winning
economist John Nash. Crawford WDT ]] 62, 64. Professor Crawford offered
a variation of the Nash Framework that the Judges rejected in SDARS II
as a means of corroborating the proffered musical works benchmark.
Crawford WDT ] 65.\31\
    \31\ Music Choice acknowledged that the Judges rejected its
proposed musical works benchmark as a marketplace benchmark in SDARS
II. Rather than proffer a marketplace benchmark from another market,
however, Music Choice proffered Professor Crawford's Nash Framework,
not to corroborate the musical works benchmark rejected in SDARS II,
but as a stand-alone benchmark.
    In his Nash Framework proposal, Professor Crawford modeled a single
record label as the ``upstream'' firm in the negotiation of sound
recording performance rights to be licensed to a single PSS, the
``downstream'' firm in the negotiation. Id. ] 67. The Nash Framework is
based on the assumption that the record label and PSS provider each
have a certain degree of market power. Id. ] 71. Professor Crawford
asserted that this assumption is applicable with respect to Music
Choice given its current product offerings and established
relationships with MVPDs. Id. ] 73. According to Professor Crawford,
Music Choice has negotiated long-term contracts with the MVPDs and
possesses a unique bundle of technology that would be costly and time
consuming for other firms to duplicate. Id. ] 73. Professor Crawford
concluded that because both PSS providers and record labels have some
market power, a non-cooperative bargaining model such as the Nash
Framework is an appropriate framework for analyzing market outcomes for
the PSS sound recording performance rights in the absence of a
compulsory license. Id. ] 75.
    In the Nash Framework three fundamental factors determine how two
firms would ``split a pie'' in a hypothetical negotiation. These ``Nash
Factors'' are: (1) The Joint Agreement Profits; (2) each firm's Threat
Point; and (3) each firm's bargaining power. Id. ] 81.\32\ To determine
the royalty that would arise in the hypothetical market for sound
recording performance rights for the PSS over the 2018-2022 rate
period, Professor Crawford quantified the Nash Factors based on Music
Choice's costs and revenues of its residential audio service as a
standalone business. Id. ] 110.
    \32\ Joint Agreement Profits are the combined profits to both
the upstream and downstream firms in the market under study from
reaching an agreement. For the PSS this means the revenue the PSS
earns for the PSS less all non-PSS royalty costs that they incur.
Crawford WDT ] 81. The Threat Point for each firm is the profit it
would receive when no agreement is reached. Id. The difference
between the Joint Agreement Profits and the sum of the firms' Threat
Points is called the ``Incremental Profits'' which are the profits
the firms could earn by reaching an agreement above and beyond the
profits they could earn in the absence of an agreement. Id. The
profits each firm receives in a bargain equals its Threat Point plus
its Bargaining Power times the Incremental Profits. Id. ] 82. Dr.
Crawford communicated this formula in mathematical terms as Royalty
= Threat Point + Bargaining Power * Incremental Surplus. Id. at
i. Joint Agreement Profits
    Because Music Choice keeps its books on a consolidated basis,
Professor Crawford analyzed Music Choice's costs and revenues to
determine how they would have been allocated if Music Choice operated
its residential audio service as a standalone business. Id. ]] 122-149;
4/24/17 Tr. 733-38 (Crawford); 5/18/17 Tr. 4549-52 (Del Beccaro).\33\
This process was conducted not in the ordinary course of business but
to isolate Music Choice's residential audio business for use in the
Nash Framework and in response to the Judges' observation in SDARS II
that the residential audio service is the applicable Music Choice
business line in analyzing the section 114 license. Crawford WDT ] 110.
Professor Crawford also asserted that isolating the residential audio
service is necessary to ensure that Music Choice does not subsidize
this business line with profits from other business lines, which
Professor Crawford believes would be inconsistent with economic policy
and the statutory objectives of the PSS license as he understands them
to be. Id. ] 176; 4/24/17 Tr. 787 (Crawford).
    \33\ Music Choice has three business lines: 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. Many of Music Choice's costs
are used in the production of both the residential audio and video
business lines. Crawford WDT ] 110. According to Professor Crawford,
the residential audio service remains the most important in terms of
revenues and company strategy. Professor Crawford asserted that if
the residential audio service were to cease, Music Choice would
cease providing any services and would close altogether. Crawford
WDT ] 129.
    It would not be fruitful to detail the multistep process Professor
Crawford conducted to disaggregate costs and revenues to derive inputs
for the Nash Framework analysis. Nonetheless, it is worth noting that
many of the steps required judgment calls on Professor Crawford's part
that undoubtedly affected the inputs he later plugged into the Nash
Framework.\34\ The Judges do not suggest that Professor Crawford's
adjustments were erroneous or inappropriate under the circumstances but
only mention them to highlight the level of discretion and subjectivity
that Professor Crawford employed in developing the inputs that he fed
into the Nash Framework. Given the extreme complexity of the process
that Professor Crawford developed, it would be impracticable if not
impossible for the Judges to ``back out'' one or more of the
adjustments Professor Crawford made in developing the model if the
Judges found they were unwarranted. The discretion that Professor
Crawford exhibited in disaggregating Music Choice's costs and revenues
pales, however, in comparison to that he exercised in choosing other
Nash Factors, such as bargaining power and Threat Point. The great
degree of discretion in quantifying the inputs in the Nash Framework as
proposed by Professor Crawford underscores the inherent weakness in the
Crawford model. The Judges concerns about the model are more applicable
in the current proceeding than they were in SDARS II because Music
Choice seeks to elevate the model to benchmark status rather than as
information to corroborate a proffered rate as was the case in SDARS
    \34\ For example, Dr. Crawford chose to exclude certain legal
costs that Music Choice incurred or expected to incur related to the
PSS III proceeding in 2016 and 2017 because those costs relate to
litigating the 2018-2022 rate proceeding. Instead he substituted
costs that Music Choice purportedly incurred during the PSS II rate
period (2013-2017). He also chose to average certain patent
litigation costs over an eight-year period that Music Choice
incurred during 2016-2017 because, based on his discussions with
Music Choice executives, Music Choice historically has incurred such
patent costs every eight years. Crawford WDT ] 148. Of course, as a
practical matter, no individual company can know with any reasonable
degree of certainty when, in the future, it may be sued for patent
infringement or sue another that allegedly violates one of its
    Professor Crawford used the disaggregated costs and revenues to
begin the Nash Framework calculations. The first step in that process
is to create the first Nash Factor--Joint Agreement Profits--the joint
economic profits to be shared between a record label and PSS provider
in the PSS market if an agreement is reached. It is the total economic
profits that the PSS provider earns before payment of a sound recording
performance royalty. Crawford WDT ] 92.
    Based on his analysis of Music Choice's financial information as
discussed above, Professor Crawford estimated the Joint Agreement
Profits in the hypothetical market for PSS sound
[[Page 65216]]
recording performance rights would range from [REDACTED] in 2018 to
[REDACTED] in 2022. Crawford WDT ]] 113, 171.
ii. Threat Points
    Professor Crawford then calculated each party's Threat Point, the
second factor in the Nash Framework. A Threat Point is a theoretical
construct representing the profit that would accrue to a record label
and a PSS provider if they are unable to reach an agreement. Each firm
in a hypothetical negotiation will have a Threat Point. Crawford WDT ]
67. Under the model, threat points can be positive, negative, or zero.
Id. at 26 n.71. For a record label, a negative threat point could occur
where the record label could earn additional profit in a non-PSS market
(e.g., music downloads) if it reaches an agreement with a PSS in the
PSS market. If the record label fails to reach the agreement with the
PSS provider, it loses all prospective profits it would have earned in
the PSS market and the profits it could have earned in the non-PSS
market. Id. ] 85.
    The profit each firm earns in a bargain equals its threat point
plus its bargaining power (discussed below) times incremental profits.
Id. ] 82. Incremental profits are the difference between the joint
agreement profits and the sum of the firms' threat points. Id. ] 81.
Professor Crawford determined that Music Choice's threat point would be
zero because, in the absence of an agreement between Music Choice and a
theoretical record label, Music Choice would not be able to offer a
viable residential audio service and therefore would have economic
profits of zero. Id. ] 173. Professor Crawford asserted that assigning
a zero threat point to Music Choice is conservative because it is based
on an assumption that Music Choice could not offer a viable service in
the absence of an agreement with a single label.\35\ If Music Choice
could offer such a service in the absence of the catalog of any record
label, then Music Choice's threat point would be higher than zero,
which would suggest that Music Choice should pay a lower royalty rate
under the model. Id. at 49 n.149.
    \35\ Rather than postulate the hypothetical PSS market as a
negotiation between a single PSS and a single record label Professor
Crawford could have constructed the model as a negotiation between a
single PSS and a group of record labels. Under this scenario, the
PSS might reach agreements with some labels but not others. The
failure of an agreement with certain labels (i.e., smaller labels)
might not preclude the PSS from offering a service whereas the
failure of the PSS to reach an agreement with any of the larger
labels might preclude the PSS from offering any type of service
(i.e., PSS service or non-PSS service). Under this scenario, the
assignment to the PSS of a negative threat point might be more
appropriate than assigning a zero threat point because if Music
Choice failed to reach an agreement with one major label then it
might be precluded from offering any service.
    Outside of the threat point discussion, however, Professor Crawford
asserted that Music Choice's residential audio service remains the most
important in terms of revenues and company strategy. Indeed, Professor
Crawford asserted that if the residential audio service were to cease,
Music Choice would cease providing any services and would close
altogether. Id. ] 129. Placed in the context of the threat point
discussion, this concession strongly suggests that Music Choice
deserves a negative threat point under Professor Crawford's model, the
extent of which would be measured by the amount of profits Music Choice
would lose if it closed its non-PSS business lines. SoundExchange's
expert pointed out this inconsistency in Professor Crawford's
presentation. 5/3/17 Tr. 2461, 2343 (Wazzan) (``Dr. Crawford concedes
that Music Choice would go out of business altogether without the
residential music business. So they would lose their commercial and
video revenue streams. And if you look at the financials, we know that
Music Choice is forecasting significant profits in its non-PSS lines of
    Music Choice's responses to this disconnect between Professor
Crawford's threat point assessment and his statements about the primacy
of Music Choice's residential audio business are unavailing. For
example, Music Choice contended that the SDARS II decision is precedent
for treatment of the threat point analysis that Professor Crawford
employed. Music Choice Reply to SE PFF 2044 at 817-18. The passage from
SDARS II that Music Choice referred to pertained to an analysis of
Factor B in Section 801(b)(1), regarding the setting of a rate that
provides a fair return (for the service) and a fair income (for the
copyright owners) under existing market conditions. The Judges were
concerned in that context that Music Choice was making claims of
unprofitability of its business as a whole to support a downward
adjustment in the rates under the Section 801(b) factors. The Judges
pointed out that the subject of the section 114 license was Music
Choice's residential audio business rather than its entire business,
which included non-PSS lines. 78 FR at 23059. By that point in the
determination, the Judges had already discounted the use of the
Crawford model and the proffered musical works benchmark the results of
which the model purportedly corroborated. The Judges did not opine on
how Professor Crawford should have calculated the threat point for his
own model because the Judges dismissed the usefulness of the model. 78
FR at 23058 (``without real world data to support its predictive
capacity [Professor Crawford's application of the Nash Framework] is
unworthy of further consideration.'').
    Therefore, the Judges agree with SoundExchange's criticism that
Professor Crawford incorrectly assigned a threat point of zero to Music
Choice when, under Professor Crawford's own testimony, Music Choice
would lose profits from non-PSS business lines if Music Choice could
not reach an agreement with one or more record labels. Based on that
fact alone, the results of Professor Crawford's model in the current
proceeding are suspect, but the flaws in Professor Crawford's
presentation do not end there.
    With respect to the threat point for a hypothetical record label,
Professor Crawford asserted that it would be zero in the PSS market. As
for the label's threat point in the non-PSS market (e.g., sales of CDs
and downloads), Professor Crawford asserted that the analysis was more
``nuanced.'' Crawford WDT ]] 94-95, 174-175. Due to an alleged
promotional effect that the PSS has on the label in the non-PSS market,
Professor Crawford concluded that the record label's threat point could
be negative. Professor Crawford has no way of estimating the purported
promotional effect of Music Choice's services in the non-PSS market so
he assigned a zero threat point to the hypothetical record label. Id.
]] 175-176. We concur with Professor Crawford's decision not to attempt
to assign any promotional value to Music Choice's service in the non-
PSS market. The evidence he cited to support such an effect is either
dated (i.e., from a 1998 CARP decision) or anecdotal (i.e., record
labels provide Music Choice with ``promotional copies'' of new singles
or albums). Id. ]] 97-104. The Judges do not doubt that record labels
seek exposure for the artists they promote, and digital platforms like
Music Choice may provide meaningful exposure to the artists that appear
on its PSS service. The Judges find no evidence in the record in this
proceeding that they can use to quantify what impact, if any,
promotional activities on Music Choice's platform would have on artists
(and the labels that sign them) in non-PSS markets.
    The Judges are less sanguine, however, about Professor Crawford's
assignment of a zero threat point to the
[[Page 65217]]
first portion of a record label's threat point (i.e., that dealing with
the PSS market). It is not at all clear that a record label's failure
to reach an agreement with Music Choice would mean a loss of all record
company profits in the PSS market if that market includes all providers
of residential audio services. There is evidence in the record that at
least one Music Choice competitor, Stingray Music, provides a service
that is comparable to the residential audio service that Music Choice
provides, but pays a much higher royalty rate than Music Choice
pays.\36\ Although that competitor, which is a recent entrant to the
U.S. market, has not sought a royalty rate closer to that which Music
Choice pays, it certainly could in the future, perhaps using the lower
rate paid by Music Choice as a comparable to support its own rate
reduction. In other words, the lower rate that Music Choice pays as a
PSS could put downward pressure on the rates that competing services
pay to record labels.
    \36\ Stingray Music is a Canadian digital pay television audio
service owned and operated by Stingray Digital. It has about 50
music channels that are available to television service subscribers
of several cable and IPTV providers in the U.S. Like Music Choice,
Stingray also has a business service and streams to individuals who
subscribe to television services that provide Stingray Music. Wazzan
WDT ] 62. The PSS and services such as Stingray, which SoundExchange
refers to as CABSAT (cable/satellite) services compete for the same
MVPD wholesale buyers. Stingray bought Music Choice's European
affiliate, which it operates as Music Choice International. In the
U.S., Music Choice and Stingray are direct competitors. Id. ] 62(g),
    By contrast, if Music Choice and the theoretical record label were
unable to reach an agreement, the rate that Music Choice pays could no
longer be used by providers of comparable services to justify lower
royalty rates. Under that scenario, a record label could actually
benefit from the loss of Music Choice to the extent that the rate it
pays could be shown to be below a market rate, which would result in a
positive threat point for the record label.\37\ As with the asserted
promotional effect, however, such an effect is impossible to estimate
with any accuracy. The Judges do not conclude from this discussion that
zero is the correct threat point for the hypothetical record label but
rather confirm the lack of usefulness of the Crawford model because
critical components of the model, at least as presented by Dr. Crawford
in the current proceeding, allow a broad level of discretion and
subjectivity, which undermines the credibility of the results.
    \37\ See Wazzan WRT ] 57 (``there is considerable reason to
believe that the existence of Music Choice imposes significant
opportunity costs on record companies in today's market [in that]
record labels receive substantially higher revenues from interactive
and non-interactive music services than from the PSS'').
SoundExchange's expert, Dr. Wazzan, attempted to correct this error
and others in Professor Crawford's model and derived a range of
rates that are several times greater than those Professor Crawford's
estimated. Wazzan WRT ] 48. SX PFFCL ] 2046 (comparing Crawford's
range of 1.4% to 5.6% to Wazzan's ``Corrected'' range of 9.0% to
36%). If the Judges were to rely to some extent on the Crawford
model, the evidence in the record does not support a rate outside of
this wide range of 1.4% to 36% of gross revenues. After reviewing
each party's evidence regarding the Crawford model, however, the
Judges do not have a high level of confidence regarding where within
that broad range a reasonable rate might lie. Nevertheless, the many
flaws in Professor Crawford's model suggest that the lower end of
the range of rates that the Crawford model yields is likely outside
the zone of reasonableness.
iii. Relative Bargaining Power
    Professor Crawford's assignment of the parties' respective
bargaining powers (the last element of the Nash Framework) was also
based on faulty reasoning. Under the Nash Framework, each firm's
bargaining power is a number between 0 and 1, which measures the
strength of that firm in the negotiation. Crawford WDT ] 81. The sum of
the two parties' bargaining powers equals 1. Id. Professor Crawford
related each firm's bargaining power to each party's patience in a
negotiation. The party with greater patience also has greater
bargaining power. Professor Crawford contended that that comparison is
consistent with the nature of bargaining between Music Choice and the
copyright owners. According to Professor Crawford,
[b]oth 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. . . . 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.
Crawford WDT ] 105. The Judges interpret Professor Crawford's statement
regarding relative bargaining power as saying he has no way to quantify
what the relative bargaining powers are between Music Choice and the
record labels. Ultimately, the Judges believe that this is an accurate
statement that further undermines the usefulness of the Nash Framework
in the proceeding. That being said, what evidence there is in the
record regarding the relative bargaining power of Music Choice and the
record labels suggests that the record labels have much greater
bargaining power than Music Choice (or a similarly situated PSS in the
hypothetical market).
    Mr. Del Beccaro, Music Choice's President and CEO testified about a
history of ``inequality in bargaining power'' between Music Choice and
the record labels that forced Music Choice to accept rates that were
higher than it would have otherwise. See, e.g., Del Beccaro WDT at 10
(``Music Choice had no choice but to accept a rate increase to 7
percent for 2002 to 2003 and 7.5 percent for 2004 through 2007''); id.
at 11 (``[d]espite repeated efforts by Music Choice to engage in
settlement negotiations, when the royalty rate came up for adjustment
for the next rate period, SoundExchange did not negotiate a settlement
until directed to by the Judges during the direct trial opening
statements of the SDARS I proceeding in June 2007''); id. at 12 (``[In
SDARS III] 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.'').
    Professor Crawford contended that ``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 performance rights for
PSSs.'' Crawford WDT ] 177. But he needed look no further than Mr. Del
Beccaro's statements about Music Choice's efforts to negotiate
settlements with SoundExchange. These statements strongly suggest that
Music Choice has very little if any bargaining power in its
negotiations with the labels. The greater the bargaining power by the
record labels, the higher the rates that Music Choice would be required
to pay. Crawford WDT at 73, Ex. B.3. Therefore, the Judges find no
support in the record to suggest that Music Choice or a similarly
situated PSS would enjoy anything but minimal bargaining power in
negotiations with the labels, particularly any of the major labels. As
a result, even under the fundamentally flawed Crawford model, nothing
but the highest projected rate of 5.6% would even be considered to fall
within a zone of reasonableness. Given the inherent subjectivity of the
model, however, the Judges continue to conclude that it provides no
useful information regarding the royalty rates that a PSS should pay,
other than perhaps to eliminate from a potential zone of reasonableness
all rates at or below 5.6%. Therefore, the Judges reject, for the
second time in two consecutive PSS
[[Page 65218]]
proceedings, the usefulness of Professor Crawford's presentation of the
Nash Framework as a model for determining reasonable royalty rates for
the PSS.
b. SoundExchange's Proffered CABSAT Rate
i. The CABSAT Benchmark
    SoundExchange asserted that there is no applicable marketplace
benchmark suitable for the PSS market, even with a comparability
adjustment. See Wazzan CWDT ] 12. According to SoundExchange ``nobody
has identified any agreements relating exclusively to a PSS, or even
relating in material part to a PSS.'' Id. ]] 45, 47.\38\ SoundExchange
observed that even if such agreements existed, one would expect the
rates under those agreements to be influenced by the statutory license.
Id. ] 44.
    \38\ SoundExchange acknowledged that the record includes
evidence of two Muzak agreements that address Muzak's PSS service,
but SoundExchange asserted that these agreements are concerned
primarily with Muzak's business establishment service. Trial Exs.
401, 402. In any case, SoundExchange asserted that there are a
number of reasons why these agreements would not make suitable
benchmarks. See Wazzan CWDT ] 45.
    Rather, SoundExchange proffered as its benchmark a royalty rate
developed in a settlement under section 114 of the Act and applicable
to certain ``new subscription services'' that offer digital music
transmissions to cable or satellite television subscribers.\39\
SoundExchange referred to these new subscription services' rates as
``CABSAT'' rates. The Judges adopted the ``CABSAT'' rates in a separate
proceeding under a statutory provision that prescribes a rate-setting
standard different from the one at issue in the present proceeding. See
Written Direct Testimony of Paul Wazzan, Trial Ex. 27, ] 11 (Wazzan
    \39\ See 37 CFR part 383. Three services currently offer
residential audio services through cable and satellite television
providers and pay royalties under part 383 regulations as New
Subscription Services: Stingray, Sirius XM, and Muzak's legacy DMX.
    SoundExchange asserted that the CABSAT rates are set in a
``hybrid'' market in which negotiations occur in a marketplace setting
but, in the case of an impasse, either party can appeal to a judicial
or regulatory body for a rate determination. SoundExchange contended
this ``hybrid'' environment makes CABSAT rates an appropriate benchmark
if the parties have similar stakes in the benchmark and target markets.
See Crawford WDT ] 50; Wazzan CWRT ] 20. SoundExchange concluded that
while no party has identified a suitable marketplace benchmark for the
PSS that is not constrained by regulation, the statutory CABSAT rates
are ``a market-like rate.'' See Crawford WDT ] 58.
    SoundExchange argued that the two services that use the statutory
PSS license (i.e., Music Choice and Muzak's Dish CD service) ``are in
all important respects functionally equivalent to the three services
``that use the statutory CABSAT license.'' See SX PFFCL at xxiv.
SoundExchange asserted that both services are cable radio services that
are delivered to consumers through MVPDs; both provide a similar number
of channels and similar genres of music; both would negotiate in the
hypothetical market for the same rights from the same entities; and
PSSs would meet every element of the regulatory definition of a CABSAT
service. SoundExchange argued that PSSs and CABSAT services compete
head-to-head for carriage on MVPDs. In short, according to
SoundExchange, the only material difference between the two types of
services is the date on which they commenced operation. See, e.g.,
Wazzan CWDT ]] 59, 60, 66; Crawford WDT ] 50; 5/3/17 Tr. at 2305-06
(Wazzan); 4/24/17 Tr. at 714 (Crawford); Written Direct Testimony of
Jonathan Bender, Trial Ex. 29, at 29 (Bender WDT). For that reason and
``because setting relatively lower rates for the PSS would distort the
market in their favor'' SoundExchange asserted: ``the CABSAT rates
present an appropriate benchmark in the absence of any clearly-
appropriate unregulated marketplace benchmark.'' Id.
    According to SoundExchange, Music Choice considers Stingray, one of
the CABSAT services, to be its primary competitor. 5/18/17 Tr. 4641-42
(Del Beccaro). SoundExchange acknowledged that the CABSAT rates, which
are statutory rates set in the context of a CRB rate proceeding, are
not unregulated marketplace rates. Nevertheless, SoundExchange asserted
that the CABSAT rates represent the ``best available benchmark for the
PSS rates.'' SX PFFCL at xlv. SoundExchange acknowledged that the
statutory rate standard for CABSATs is a willing buyer/willing seller
standard. Nonetheless, SoundExchange contended that no adjustment would
be required to the CABSAT rates under the Section 801(b) factors before
applying them to the PSS services. See Wazzan CWDT ] 18. The extant
CABSAT rates only apply for three of the five years of the current PSS
rate period (2018-2020), therefore SoundExchange proposed that the
Judges apply a 3% per year rate increase (the size of the CABSAT rate
increases in 2018-2020) to the 2020 CABSAT rate to derive the rates for
2021 and 2022, the last two years of the upcoming PSS rate period. See
Bender WDT at 29-31.
    Under SoundExchange's proposal, the rate for PSSs' residential
audio services would be a monthly per-subscriber rate of $0.0190 for
2018, $0.0196 for 2019, $0.0202 for 2020, $0.0208 for 2021, and $0.0214
for 2020. SX Amended Rate Proposal at 7, 10. These rates would cover
the PSSs' royalty obligations under the section 114 and 112(e)
licenses. Id.
    For PSSs' webcasting activities,\40\ SoundExchange proposed that
the PSS pay the same rates that apply to commercial webcasters
providing a subscription service under 37 CFR 380.10. Through 2020,
that rate would be a per-performance rate of $0.0022, adjusted for
inflation. For PSSs that are unable to measure performances, the rate
would be based on the average number of recordings on the service
played per hour multiplied by the Aggregate Tuning Hours. SX Amended
Rate Proposal at 8.
    \40\ This proposed rate would apply to ``all licensed
transmissions and related ephemeral recordings through an internet
streaming service qualifying as a PSS (or any similar service
capable of tracking the individual sound recordings received by any
particular consumer).'' SX Amended Rate Proposal at 8.
    In advocating for the CABSAT benchmark, SoundExchange also stressed
the importance of changing the current rate structure from a percent-
of-revenue to a per-subscriber structure, because CABSAT rates are
calculated per-subscriber. SX PFFCL ] 1949. SoundExchange acknowledged
that one could convert the proffered CABSAT-based rates to a percentage
rate. SoundExchange estimated that in 2015, Stingray paid an effective
percentage royalty rate of ``just under [REDACTED] %'' of its revenues.
SX PFFCL ] 1949. This converted CABSAT rate compares to Professor
Crawford's estimate that Music Choice would pay between [REDACTED] %
and [REDACTED] % of its unadjusted residential audio service revenue
under SoundExchange's rate proposal. See Crawford WRT ] 113. Given this
perceived equivalence, SoundExchange perceived no reason to adopt a
percent-of-revenue rate structure for PSS in the current proceeding.
    SoundExchange contended that a per-subscriber rate structure is
preferable because Music Choice is paid under such a structure by its
MVPD customers. Id. ] 1950. SoundExchange also argued that a per-
subscriber rate is easier to apply and more transparent than a
percentage-of-revenue rate. See Orszag AWDT ] 27; Crawford WDT ]] 147-
[[Page 65219]]
148; Bender WRT at 13.\41\ Of particular concern to SoundExchange was
the perception that a revenue-based structure gives Music Choice the
flexibility to reduce the amount of royalties it pays by charging its
affiliated owners discounted prices. According to SoundExchange, Music
Choice is partially owned by cable companies, including Comcast, Time
Warner Cable, and Cox, and charges lower prices to its MVPD owners than
it charges to other MVPDs. See Wazzan CWDT ] 90. SoundExchange argued
that ``the Judges should be suspicious of commercial arrangements
between Music Choice and its MVPD partners.'' Id. ] 1979.
    \41\ SoundExchange acknowledged that although allocation
disputes can arise under a percent-of-revenue structure, ``such
disputes have not materialized between SoundExchange and Music
Choice in recent memory.'' SXPFFCL at ] 1952.
    SoundExchange disputed Music Choice's attestations that its MVPD
partner affiliate fees are a function of the relative size of
affiliated MVPDs vis-[agrave]-vis non-affiliates. See Del Beccaro WDT
at 22-23; but see Wazzan WDT ] 91. SoundExchange contended that
evidence in the record shows that all affiliates received discounted
rates from Music Choice, regardless of the number of subscribers they
had at the time. SX PFFCL ] 1990; Trial Ex. 410, Music Choice Partner
Affiliation Agreement, Sch. B at MC0012247-48; 5/3/17 Tr. 2333
(Wazzan). SoundExchange contended that this purported affiliate
discount, which remains in effect, represents a [REDACTED] % discount
to fees that non-affiliated MVPDs are required to pay. 5/3/17 Tr. 2333-
37 (Wazzan). SoundExchange represented that Music Choice's non-partners
with the largest number of subscribers are expected to pay $[REDACTED]
or $[REDACTED] per subscriber per month in 2018, while the partners are
expected to pay $[REDACTED] per subscriber per month, about one third
as much. See Wazzan CWRT, App. C. at 43-44.
ii. Music Choice's Opposition to the CABSAT Benchmark and Per-
Subscriber Rate Structure
    Music Choice opposed SoundExchange's proffered CABSAT benchmark and
proposed per-subscriber rate structure. As a preliminary matter, Music
Choice contended that SoundExchange's identification of the necessary
components of a comparable market for benchmarking purposes omits two
key requirements, namely that the benchmark represent a workably
competitive market and that the buyers and sellers in both the target
market and the benchmark market have similar stakes. See Crawford WDT ]
50; 5/24/17 Tr. at 695-96 (Crawford). Music Choice contended that the
proffered CABSAT benchmark fails on both accounts because the CABSAT
rates and terms were set by a settlement between SoundExchange and
Sirius XM. Crawford WDT ]] 55-56. According to Music Choice, the
settlement did not reflect any sort of competitive marketplace. Id.
    Music Choice asserted that Sirius XM is not an active participant
in the CABSAT market, providing its CABSAT service to only one
affiliate (DISH Network). Music Choice contended that Sirius XM's
CABSAT service is merely a promotional vehicle to drive subscriptions
to its primary business, the satellite radio service. See Crawford WRT
] 43. In support of this argument, Music Choice noted that Sirius XM's
CABSAT service generates only [REDACTED] % of Sirius XM's revenues.
Given that the CABSAT service generates such a miniscule percentage of
Sirius XM's revenues, Music Choice argued that Sirius XM had no real
incentive to vigorously negotiate the CABSAT settlement let alone incur
the costs of a rate proceeding.\42\ Id. ]] 55-56. By contrast, Music
Choice has a far different stake because the PSS service is its primary
business. Crawford WDT ] 129 (Music Choice's residential audio service
remains its most important business in terms of revenues and company
    \42\ Music Choice argued that the ``bargaining and market
dynamics that led to the settlement from which the current CABSAT
rates and terms are derived also make clear that those rates are not
market rates, or even market-like . . .'' MC Reply to
SoundExchange's PFFCL at 68. According to Music Choice, Sirius XM
had no rational business incentive to litigate the last CABSAT
proceeding, so it had little choice but to settle. Id. at 69-70 (and
evidence cited therein).
    In Music Choice's estimation, the proffered CABSAT benchmark lacks
a key indicator of comparability--similar stakes--which Music Choice
asserted must be present when using a benchmark from a hybrid market
(i.e., a market in which negotiations occur in a marketplace setting
but, in the case of an impasse, either party can appeal to a judicial
or regulatory body for a rate determination). See Crawford WRT ]] 55-
56. Music Choice also argued that the ``sellers'' in the proffered
CABSAT market and the hypothetical PSS market are not comparable
because in the CABSAT market SoundExchange represents the entire record
industry as opposed to individual record companies which purportedly
would reflect the sellers in the hypothetical PSS market. Id.
    Music Choice also argued that the proffered CABSAT benchmark is
flawed because the underlying CABSAT market is neither competitive nor
stable. See Del Beccaro WRT at 5-6. According to Music Choice,
``[t]here has never been a CABSAT licensee that has proven able to
operate a long-term profitable business from its CABSAT operations, nor
have the majority of participants in the CABSAT market actively or
successfully sought new affiliates or competed in the marketplace.''
Id. Music Choice asserted that Stingray is the ``only active CABSAT.''
Id.\43\ According to Music Choice, after six years in the CABSAT market
Stingray has captured only 6% of the MVPD market and, until recently,
all of its affiliates were small cable operators that pay high rates,
which have sustained Stingray. See Del Beccaro WRT at 10. Music Choice
projected that if it left the market, Stingray could not replace it
because Stingray would have to reach agreements with larger MVPDs at
lower rates while still paying the high per-subscriber CABSAT rates.
Id. Over time under this market dynamic Music Choice contended Stingray
would be forced to exit the CABSAT market. Id.
    \43\ According to Music Choice, the only companies ever to enter
the CABSAT market are MTV, DMX, Sirius XM, and Stingray. Music
Choice represented that MTV and DMX have since exited the CABSAT
market. According to Music Choice, Sirius XM has only one affiliate,
which it purportedly uses as a promotion tool, and is not competing
for new business. MC Reply to SX PFFCL at 71-72 (and evidence cited
    Music Choice also faulted SoundExchange for glossing over the
legislative history of the PSS license and the Section 801(b) policy
standard, which, Music Choice contended, reflects Congressional intent
to ``protect the unique business expectancies of the PSS, even against
later market entrants, which is inapplicable to other statutory
licensees and must inform any interpretation or application of the
801(b)(1) policy standard to the PSS.'' MC Reply to SX PFFCL at 66.
Music Choice noted that ``Congress `grandfathered' the three PSS, Music
Choice, DMX and Muzak, which were already in operation at the time
Congress passed the Digital Millennium Copyright Act (DMCA) \44\
allowing the PSS to continue operating under the 801(b)(1) policy-based
rate standard rather than be subjected to the new [willing buyer/
willing seller] marketplace standard.'' Id. at 65. Thus, Music Choice
concluded, ``the mere fact that non-comparable services pay different
rates provides no useful data for setting the PSS rates.'' Id. at 66.
    \44\ Public Law 105-304, 112 Stat. 2860 (Oct. 28, 1998).
    Music Choice agreed with SoundExchange (and Dr. Wazzan) that
[[Page 65220]]
there are no types of licensed music services comparable to the PSS.
Id. at 67. Music Choice disagreed, however, that the current PSS rate
is below market. In fact, it contended that the current PSS rate is an
above-market rate, given that it is the result of settlements that
Music Choice had little choice but to accept to avoid litigation costs.
Id. (MC Reply to SX PFFCL ] 1789).
    Music Choice contended that, despite SoundExchange's claims to the
contrary, the reason Music Choice has not sought direct licenses is not
because it would not get a better rate than the statutory rate but
because the cost of direct license negotiations would be too high. Id.
Music Choice also noted that since the current statutory rate does not
exclude revenues from direct licenses for PSS, Music Choice would still
have to pay a share of revenues attributed to the sound recordings from
the direct licenses in addition to the royalties required by those
direct licenses. Id. at 67-68. According to Music Choice, direct
licensing would only make sense if it could directly license 100% of
its music. Id. at 68 (Reply to SX PFFCL ] 1789).
    Music Choice acknowledged that PSS providers and CABSAT services
both sell cable radio to MVPDs but contended that material differences
in quality, programming, on-screen displays and other features set the
PSS (or at least Music Choice's) service apart from that of the
CABSATs. Id. at 77. Music Choice contended that its screen displays
provide significantly more promotional impact than those of any CABSAT
service. Id. at 78-79.\45\
    \45\ Music Choice cited the fact that it bundles its residential
PSS with its video offerings as ``critical and relevant, because
those bundled offerings provide a value proposition that is
appealing to MVPD providers and allows [Music Choice] to compete
effectively against the Stingray and Sirius XM's CABSAT services.''
MC Reply to SXPFFCL at 85.
    Music Choice also opposed the per-subscriber rate structure that
SoundExchange proposed. Music Choice contended that the proposal is
based on the false premise that Music Choice provides unfairly
advantageous discounts to cable providers with which Music Choice is
affiliated. MC PFF ] 279; Wazzan WDT at 37-38; 5/3/17 Tr. 2330
(Wazzan). Music Choice represented that a supermajority interest in
Music Choice is owned by non-cable companies, including some affiliated
with record companies, which would be harmed if Music Choice gave
below-market rates to its cable affiliates. Therefore, according to
Music Choice, doing so would make no economic sense. MC PFF ]] 283,
285-288; Del Beccaro WRT at 16, 19-20.
    Music Choice asserted that any preferential pricing it offers is
the result of the size of the cable company, although factors such as
long-term commitment to the Music Choice service may also play a role.
Del Beccaro WRT at 16-17. Indeed, Music Choice represented that at
times its cable affiliates have made concessions on price just to help
Music Choice survive. MC PFF ] 299; 5/18/17 Tr. 4593-94 (Del Beccaro);
4/24/17 Tr. 804-05 (Crawford).
iii. Judges' Analysis of SoundExchange's Proffered CABSAT Benchmark and
Proposed Per-Subscriber Rate Structure
    In determining whether a proffered marketplace benchmark is
comparable to the hypothetical target market the Judges have looked at
the comparability of the buyers, sellers, and rights over which the
parties negotiated.\46\ When the two markets were comparable (i.e., the
buyers, sellers, and rights are the same), the Judges have found that
the rate that the buyers and sellers have negotiated in the market can
provide useful guidance in determining the rate for the target
market.\47\ In the present proceeding, SoundExchange conceded that
``[t]he CABSAT benchmark is not a marketplace benchmark. It is instead
a regulated rate.'' SX PFFCL ] 1847 (and evidence cited therein). The
prevailing CABSAT rates were agreed to by SoundExchange and Sirius XM,
the only remaining participants, in a CRB rate-setting proceeding. See,
e.g., Crawford WRT ] 33.\48\
    \46\ See, e.g., SDARS I, 73 FR 4080, 4088 (Jan. 24, 2008).
    \47\ See id. (`` `comparability' is a key issue in gauging the
relevance of any proffered benchmarks . . . potential benchmarks are
confined to a zone of reasonableness that excludes clearly
noncomparable marketplace situations'').
    \48\ The rates that the participants agreed to and the Judges
adopted based on that agreement were monthly per subscriber payments
of: 2016: $0.0179; 2017: $0.0185; 2018: $0.0190; 2019: $0.0196; and
2020: $0.0202. 80 FR at 36928 (37 CFR 383.3(a)(1)).
    As a threshold matter, the Judges note that in setting a statutory
rate for PSS they are not required to approximate a market rate.\49\
Rather, the Judges' mandate is to set a reasonable rate consistent with
the Section 801(b) factors.\50\ In enacting the DMCA, Congress carved-
out the PSS from application of the willing buyer/willing seller
standard intended to approximate a market rate.\51\ The intent of the
carve-out was to acknowledge the pioneering status of the PSS, which
invested in a new type of digital audio service (i.e., transmission of
noninteractive audio to the television) in reliance on the existing
801(b) rate standard and to protect their prior investments.\52\ The
PSS took the risks and received the benefits, one of which was a
statutory exception from the rate-setting provisions in the DMCA that
were designed to ``move the industry to market rates.'' \53\
SoundExchange now argues, however, that the Judges should adopt the
proffered CABSAT rate benchmark as a market-like rate. The Judges
    \49\ Music Choice v. CRB, 774 F.3d. 1000, 1012 (``nothing in the
statute requires the Judges to rely on market rates or agreements
when setting Section 114 rates'').
    \50\ 78 FR at 23055; Music Choice v. CRB, 774 F.3d at 1013
(``The Copyright Act does not `clearly require[ ] the use of `market
rates'. [I]nstead, `reasonable rates' are those that are calculated
with reference to the four statutory criteria.'').
    \51\ SoundExchange v. Muzak, 854 F.3d 713, 714-15 (D.C. Cir.
    \52\ See Id. at 719.
    \53\ Id.
    Notwithstanding the similarities in PSS and CABSAT service
offerings that SoundExchange noted, the Judges do not find the
proffered CABSAT rate benchmark is a useful starting point from which
to apply the Section 801(b) factors.
    First, it is not at all clear to the Judges that the proffered
CABSAT benchmark market and the hypothetical PSS market offer the same
rights. As discussed below in reference to the Register's Memorandum
Opinion regarding the scope of the PSS market, the rights that the PSS
can exercise while maintaining the grandfathered rate-setting
methodology are limited to PSS entities' existing service offerings and
expanded service offerings, as the Register defines those terms.
Services that a PSS entity provides outside the scope of the
grandfathered categories constitute different service offerings, i.e.,
rights outside those offered in the hypothetical PSS market. Although
the types of activities that PSS and CABSAT entities perform may
overlap in certain respects, for purposes of determining comparability
of the hypothetical market to the target market, the relevant service
bundle is limited to those activities that the hypothetical PSS entity
may provide consistent with the grandfathered rate methodology.
    PSS entities, such as Music Choice, and CABSAT entities may (and
do, subject to an appropriate royalty rate) provide services outside
the scope of the PSS license (e.g., internet-based and mobile
application-based services that are consumed outside the home).\54\
[[Page 65221]]
These different services, however, are not included within the bundle
of rights that PSS entities would negotiate for in the hypothetical
market. Although it is theoretically possible to adjust the proffered
CABSAT benchmark to accommodate for the difference in the bundle of
rights that the CABSAT and PSS services negotiate for, SoundExchange
acknowledged no such difference and, consequently, offered no
adjustment in the current proceeding to account for the difference. The
Judges can find no persuasive evidence in the record that would allow
the Judges to develop such an adjustment sua sponte.\55\
    \54\ Dr. Wazzan referenced some of the differences he perceived
between the services that PSS and CABSAT entities provide. Wazzan
WDT ]] 67-72. For example, he noted that Music Choice provides
``internet simulcasts of its channels to subscribers of the MVPDs
that distribute Music Choice'' but took no position on whether such
streaming is part of its PSS. Id. ]] 67-68. He continued that ``the
CABSAT rates in Part 383 are quite clearly limited to a service
`transmitted to residential subscribers of a television service'
through an MVPD using `a technology that is incapable of tracking
the individual sound recordings received by any particular
consumer.'' Id. ] 70. According to Dr. Wazzan, ``internet streaming
is something else, because streams are typically transmitted to
devices other than televisions, over the public internet.'' Id. Dr.
Wazzan noted that Sirius XM and Stingray both provide internet
streaming services but do so under a different rate structure than
that applicable to the CABSAT service. Id. ] 72. In finding that the
rights conveyed to the CABSAT services are not comparable, for
benchmarking purposes, to those for which a theoretical PSS would
negotiate, the Judges do not take a position on whether the rights
conveyed to the theoretical PSS entities are broader or narrower
than those conveyed to the CABSAT services. They could be broader in
some senses and narrower in others, but the evidence in the record
shows that there are meaningful differences. All differences could
affect the value of the underlying license and therefore are
relevant in assessing the comparability of the proffered benchmark
market and the target market. Ultimately, a detailed analysis might
support a finding that, on balance, the differences are a wash,
which would support a finding that, notwithstanding the differences
in the rights granted, no comparability adjustment was necessary.
Based on the record in the current proceeding, however, the Judges
are not in a position to make such an assessment and therefore are
left with a record that shows a lack of comparability of rights with
no adjustment to sufficiently align the markets.
    \55\ Although the Register's Memorandum Opinion was issued after
the record was closed in the current proceeding, the D.C. Circuit's
Muzak decision, which highlighted the limitations in the rights that
a PSS could exercise consistent with the grandfathered rate
methodology, was issued during the proceeding. As a party to the
case, SoundExchange advocated for the restrictions on the PSS
license that the D.C. Circuit found. SoundExchange certainly could
reasonably anticipate the impact that the Muzak decision would have
on the rights that other PSS entities could exercise consistent with
the grandfathered rate-setting methodology. Indeed, one of
SoundExchange's witnesses referenced the decision and the
limitations it placed on the rights that a PSS entity could exercise
consistent with the grandfathering provision. SX PFFCL ] 1807; 5/10/
17 Tr. 3205 (Bender); see SX PFFCL ] 1807 (``[d]uring the hearing in
this case, the Court of Appeals for the D.C. Circuit held that
Muzak's PSS status is limited to its historic DishCD service.'')
Therefore, SoundExchange had notice that the rights that a
hypothetical PSS entity could exercise consistent with the
grandfathering provision were limited to providing the types of
services (i.e., existing and expanded service offerings) that the
Register set forth in her Memorandum Opinion addressing the scope of
the PSS license.
    SoundExchange attempted to conflate what the PSS services and
CABSAT services do (as represented by SoundExchange) with what they
have the right to do either in the hypothetical PSS market or in the
CABSAT market. SX PFFCL ] 1794 (``the same rights are conveyed, because
both create audio music channels incorporating the licensed sound
recordings and sell them to MVPDs, who in turn resell those channels to
consumers as part of subscription bundles.''); see 5/3/17 Tr. at 2305-
06 (Wazzan); see also SX PFFCL ]] 1797-1799 (``CABSAT Services And PSS
Are Functionally Equivalent Cable Radio Services And So Implicate the
Same Rights''). Similarities in service offerings do not necessarily
equate to comparability of rights that each of the service types is
authorized to exercise.
    SoundExchange's attempted direct compare-and-contrast of the
various activities in which the two types of entities engage also
ignores the fundamental, statutory difference between PSS and CABSAT:
Legislative intent that PSS and non-PSS be treated differently with
respect to the way in which their respective royalty rates are
determined. By SoundExchange's own admission, the CABSAT rates were
based on a settlement agreement negotiated in the context of a
proceeding in which the applicable rate standard was a willing buyer/
willing seller standard. In adopting the DMCA, Congress expressly
carved-out the PSS from that standard. The Judges conclude that
applying the CABSAT rate benchmark as proffered by SoundExchange in the
current proceeding would effectively subject the PSS to the willing
buyer/willing seller standard, which, in the Judges' view, would be
inconsistent with Congress's intent in adopting the PSS rate-setting
methodology in the DMCA.
    The proffered CABSAT benchmark also raises concerns because of the
enormous difference between the current PSS statutory rate of 8.5% of
gross revenues and the rates proposed under the CABSAT benchmark
(converting to approximately [REDACTED] % of revenue in the first
year). In SDARS II, the Judges characterized a difference between the
prevailing statutory rate of 8% and a proposed rate as high as 32.5%
(for SDARS services) as a ``yawning gap'' that raised concerns about
the reasonableness of the proffered benchmark that yielded such rates.
See 78 FR at 23066. The Judges have the same concerns about the rates
derived from the proffered CABSAT benchmark and find that the wide gap
strongly suggests that the buyers in the CABSAT market lack
comparability with those in the theoretical PSS market. This difference
in comparability of buyers is supported by SoundExchange's own
admission that Sirius XM, which negotiated the CABSAT rates with
SoundExchange, ``is first and foremost the provider of an SDARS'' that
``also provides a CABSAT service.'' SX PFFCL ] 1838. The PSS in the
theoretical market are buyers negotiating for rights to operate their
core business and therefore will have a greater stake in negotiating
the most favorable rate. On the other hand, a buyer negotiating for
rights for a non-core service might be more willing to settle for an
acceptable rate rather than the best possible rate. Significant
differences in the stakes of the respective buyers between the PSS and
the CABSAT services suggest a lack of comparability between the two for
benchmarking purposes.
    The Judges conclude that the CABSAT benchmark as proposed in the
current proceeding is not sufficiently comparable to the hypothetical
PSS target market and that the CABSAT rates are outside of the zone of
reasonableness for determining PSS rates for the upcoming rate period.
The only useful information that the proffered CABSAT benchmark
provides is to identify a rate ceiling that any reasonable PSS rates
must remain below. In other words, a reasonable PSS rate for the
upcoming rate period must be lower than the lowest rate proposed by
SoundExchange based on the CABSAT benchmark (i.e., $0.0190 per
subscriber or [REDACTED] % of gross revenues).
    By rejecting the proffered CABSAT benchmark, the Judges also reject
one of SoundExchange's arguments in support of abandoning the current
percent-of-revenue rate structure in favor of a per-subscriber rate
structure. See SX PFFCL ] 1949. The Judges find SoundExchange's other
reasons in support of a per-subscriber rate structure equally
unpersuasive. Even reviewing the evidence SoundExchange presents in a
light most favorable to SoundExchange, the Judges do not find that
Music Choice's arrangements with its affiliated MVPD customers support
a change in the rate structure to a per-subscriber structure. In this
regard, the Judges accept as credible the evidence that Music Choice
presented that historically it has charged and currently charges
similarly situated non-partner
[[Page 65222]]
affiliates rates that are the same as or lower than those charged to
its partners. MC Reply to SX PFFCL at 188 (Reply to SX PFFCL ] 1960)
(and evidence cited therein); see, e.g., 5/18/17 Tr. 4582, 4593-94 (Del
Beccaro); Del Beccaro WRT at 18.
    If SoundExchange and Music Choice were to agree to a per-subscriber
rate structure, that structure would not, on its face, be inconsistent
with the Copyright Act. Without a persuasive argument, supported by the
evidentiary record, however, the Judges are reluctant to change the
existing rate structure, which has thus far seemingly operated
effectively. The arguments and record in the current proceeding do not
support such a change. Therefore, the Judges reject SoundExchange's
request to change the rate structure to a per-subscriber structure.
    After reviewing and dismissing both proffered benchmarks, the
Judges are left with the broad parameters of a zone of reasonableness
that must be higher than 5.6% of gross revenues \56\ and lower than
[REDACTED] % of gross revenues (or $0.0190 per subscriber). The current
rate of 8.5% of gross revenues falls within that range, albeit toward
the lower end. In SDARS II, the Judges could endorse no proffered
benchmark as an appropriate starting point for application of the
Section 801(b)(1) factors. See 78 FR at 23059. Therefore, the Judges
looked to the prevailing statutory rate to begin the analysis of the
Section 801(b)(1) factors. Id.
    \56\ See supra, section IV.C.1.a.
    Notwithstanding that no party advocated using the statutory rate as
the starting point of the Section 801(b)(1) analysis and that the rate
was negotiated in the shadow of the statutory license, the Judges found
in SDARS II that the current rate was neither too high, too low, nor
otherwise inappropriate. Id. The Judges reach the same conclusion in
the current proceeding. As was the case in SDARS II, neither party has
proposed using the current statutory rate as the starting point for
applying the Section 801(b)(1) factors. SX PFFCL ] 1889; 4/25/17 Tr.
848 (Crawford). Music Choice contended that the current rate is too
high and SoundExchange contended that it is too low. The parties do not
contend that the previous PSS proceeding was ``necessarily'' wrongly
decided, only that the Judges now must look elsewhere to find a
reasonable rate. See 5/3/17 Tr. 2305 (Wazzan).\57\
    \57\ As discussed below, Music Choice did fault the Judges'
decision to make an upward adjustment to the prevailing statutory
rate to account for Music Choice's anticipated increase in the
number of channels it offered.
    Both parties' disdain for the current statutory rate appears to
stem primarily from the fact that in the first proceeding to set a rate
for the PSS, which occurred about twenty years ago, the CARP looked to
the musical works royalty rate to help determine what the rate should
be for the PSS. See, e.g., SX PFFCL ]] 1894-1900. Since then, the
parties have either agreed to a royalty rate or, as occurred in SDARS
II, the Judges selected a rate after fully reviewing the evidence in
the record. The Judges and their predecessors each chose a rate that
they viewed as reasonable and supported by the evidence before them at
the time. The fact that once upon a time one decision-maker relied on a
type of evidence that the Judges do not find persuasive in the current
proceeding on the current record is irrelevant in the current
proceeding. Unlike Music Choice and SoundExchange, the Judges are not
convinced that the specter of the musical works rate on the prevailing
PSS rate is so great as to preclude the Judges from using the current
PSS rate as the starting point for applying the Section 801(b)(1)
    The Judges must continue to have the flexibility to rely on the
best evidence they have available on the record before them in
selecting reasonable rates and terms for the upcoming rate period. At
this time, in this proceeding, on this record, the best available
evidence is the prevailing statutory rate, which falls within the broad
parameters of the zone of reasonableness indicated by the evidence that
the parties presented. Therefore, the Judges look to the prevailing
statutory rate of 8.5% as the starting point for the Section 801(b)
2. Application of the 801(b)(1) Factors
    The digital performance license requires that the rates (but not
the terms) be determined to achieve the statutory objectives detailed
above. See 17 U.S.C. 801(b)(1). SoundExchange asserted that if the
Judges use the prevailing statutory rate as the starting point of the
section 801(b) factor analysis then they should adjust the rates upward
to provide copyright owners a fair return (Factor 2), to reflect their
greater contributions to the product made available to the public
(Factor 3), and to avoid further disruption of the industries involved
(Factor 4). SX PFFCL ] 2112.
    Music Choice contended that the Judges should not have adjusted the
prevailing statutory rate upward in SDARS II to account for Music
Choice's projected increase in usage of sound recordings. Music Choice
argued that the PSS license is not a general ``usage'' license (in that
making more channels available does not necessarily lead to a greater
number of performances), and that listeners can only listen to one
channel at a time, regardless of how many channels are available for
them to choose from. Del Beccaro WDT at 16.\58\ The Judges find this
claim somewhat peculiar. Music Choice appears to assume that all
members of a household are transfixed to the same television set as
they might have been at the dawn of the television age. Modern viewing
habits, however, are far different. Televisions and other comparable
electronic devices abound in modern households. It is not unreasonable
to assume that each individual in a modern household could have access
to his or her own viewing or listening device, any one of which might
be capable of viewing or listening to the Music Choice service.
    \58\ Music Choice contended that had there been any increase in
revenues due to the increase in the number of channels that Music
Choice offered, that SoundExchange would have reaped the benefits
through increased royalties under as a percentage of revenues. In
SDARS II, the Judges found no evidence to support a projected
increase in revenues. 78 FR at 23060 (``Music Choice provided no
evidence, however, to suggest that the planned expansion in usage
would result in increased revenues to which the statutory royalty
rate is to be applied''.) Indeed, Music Choice represented that even
though it added 25 channels to its app and internet platforms during
the current rate period, its listenership remained flat while its
revenues actually decreased. Del Beccaro WDT at 16, 18.
    In SDARS II, the Judges found evidence of Music Choice's then
current intention to increase the number of Music Channels offered from
46 to 300. 78 FR at 23059. Music Choice does not dispute that
intention. Del Beccaro WDT at 15. A greater variety of channels could
reasonably be expected to attract its own audience.\59\ The Judges may
rely on a party's present intentions as to future actions. Of course,
present intentions of future actions do not ensure that the latter will
come to fruition. In this instance, the Judges' finding was based on
the evidence in the record before them. Music Choice represented in the
current proceeding that in actuality, the expansion of its service was
far more limited than it had anticipated in the last rate period. Del
[[Page 65223]]
Beccaro WDT at 18. Consequently, Music Choice contended that it has
been overpaying for the past rate period because the rate should have
been kept at 7.5% of gross revenues. Del Beccaro WDT at 18. Indeed,
Music Choice argued that this alleged ``overpayment justifies a rate
reduction in the next rate period'' below the previous period's 7.5%
rate. Crawford WDT ] 214.
    \59\ Mr. Del Beccaro suggested that the Judges should follow the
principle that PSS royalties should only be payable based on actual
performances, which occur when a song is actually received by a
listener as is the case with respect to webcasters. He quickly
cautioned, however, that Music Choice is not able to track the
actual number of performances to enable such a per-performance rate.
Del Beccaro WDT at 16-17 and n.2.
    While Music Choice chose not to expand its channel offerings as it
had anticipated, it had the right to do so consistent with the
statutory license, and the rate that the Judges adopted reflected Music
Choice's stated intention regarding that projected expansion. A
licensee has no general statutory or regulatory right to a rebate in a
subsequent proceeding. Nevertheless, the Judges specifically noted in
SDARS II that if Music Choice's projected increase in channels did not
materialize the Judges could take that fact into account in a future
proceeding. 78 FR at 23061. In SDARS II, the Judges found the increase
from 7.5% to 8.5% was consistent with the second section 801(b) factor
(fair return to copyright owners).\60\ In this proceeding, the Judges
examine again whether the basis for that increase continues to exist in
the present market.
    \60\ 78 FR at 23059.
a. Factor A: Maximize Creative Works to the Public
    Music Choice contends that the PSS services are favored under this
factor because the PSS (and Music Choice in particular) generate
original content (such as on-screen displays and curated channels) in
providing the PSS service. MC PFF ] 334-335. Music Choice contends that
this original creative content has great promotional impact on the
sound recordings they play on the service, which is illustrated by the
fact that record labels lobby to get their sound recordings played on
the service. Id. ]] 352-362. The Judges do not doubt that Music Choice
expends resources promoting the artists that appear on the service and
that such exposure can be promotional to the artists and their record
labels. These efforts are already incorporated into the current
statutory rate and therefore no downward adjustment is justified to the
extent Music Choice promotes artists.
    Music Choice contended that the current rate is actually hindering
it in providing the types of promotional services that help artists and
labels. Del Beccaro WDT at 17-18. By Music Choice's own admission,
however, much of the decline is due as much to Music Choice's declining
revenues as to the royalty rate it pays. Written Direct Testimony of
Damon Williams, Trial Ex. 56, at 32-33 (Williams WDT). Indeed, since
the royalty is currently based on a percent of revenue, a decrease in
revenues would actually result in a decrease in the royalties Music
Choice pays. Nevertheless, Music Choice provided no quantification of
the promotional effects, if any, its service has on the artists it
promotes. Moreover, it provided no persuasive evidence to connect the
current statutory rate with any decrease in such artist services. Given
the record before them, the Judges do not find that the evidence
supports a decrease from the current rate based on this section 801(b)
    SoundExchange limited its discussion regarding this factor to
arguments in support of adoption of the CABSAT rate and arguments
against lowering the current PSS rate. The Judges do not adopt the
CABSAT rates and find no persuasive evidence in the record to support a
lower rate based on the first section 801(b) factor.
b. Factor B: Afford Fair Return and Fair Income
    The second section 801(b) factor requires the Judges to assess
whether the rate (or rates) they have chosen to begin the section
801(b) analysis affords 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).
    As discussed above, in SDARS II the Judges found that an increase
from the then-prevailing statutory rate was warranted because Music
Choice anticipated greatly expanding the number of channels of music it
would offer without any anticipated increase in revenues that would
adequately compensate copyright owners for this increase. 78 FR at
23060. In actuality, Music Choice's expansion was far more modest than
it had anticipated. Del Beccaro WDT at 18; 5/18/17 Tr. 4521 (Del
Beccaro); Del Beccaro WDT at 4 (Music Choice currently provides 50
television-accessible music channels). Given that the basis for the
Judges' increase in the royalty rate after the SDARS II hearing was a
projected expansion of music channels that did not materialize, the
Judges find that, all things being equal, a downward adjustment to the
PSS rate from 8.5% back to 7.5% is most supported by the evidence and
by SDARS II. See 17 U.S.C. 803(a)(1) (``The . . . Judges shall act in
accordance with . . . prior determinations . . . of . . . the Judges .
. . .'').
    According to Music Choice, the current rate has not provided the
service a fair income under existing economic conditions. MC PFF ] 395.
Music Choice asserted that, due to changes in Music Choice's downstream
MVPD market, it anticipates losing money on its residential audio
business over the next two years under the current rate. Id. ]] 395-
396. Music Choice's main contention was that a hyper-competitive market
for its services is making it more difficult for it to remain
profitable and provide the same level of services to copyright owners
under current market conditions. Nevertheless, all of the conditions
that Music Choice cited to support a downward adjustment are already
incorporated into the current statutory rate. Music Choice provided no
evidence that any new threat is on the horizon that might warrant a
downward adjustment from the current statutory rate going forward.
Moreover, as SoundExchange correctly noted, no copyright user, not even
a PSS, is guaranteed any level of profitability.
    Music Choice argued that a decrease from the current rate would not
have a material effect on the copyright owners and artists. MC PFF ]
409. SoundExchange contended that the PSS pay lower royalty rates than
any other music service and that these rates have a negative effect on
copyright owners and artists who receive these low rates. See 5/18/17
Tr. at 4621-23 (Del Beccaro) (PSS pay lower rates than other music
services); Harrison WDT ] 29 (record companies would not agree to
current PSS rates). SoundExchange contended that the PSS rate is so far
below a market rate that it would be ``foolish'' for any record company
to attempt to directly license their sound recordings at rates near the
current rate. SX PFFCL ] 2131 (and evidence cited therein).
SoundExchange also asserted that a higher rate for PSS would not be
unfair because Music Choice could continue to operate; it would only
make less money doing so, and the Copyright Act does not guarantee a
copyright user a certain minimum level of profits. Id. ] 2134.
    The Judges do not mean to discount the fact that the market for
providing content to cable and satellite providers is competitive and
perhaps likely to grow more competitive in the future. Nevertheless,
nothing in section 114 of the Copyright Act would authorize the Judges
to shield PSS services from market forces and the Judges see no reason
to do so in the absence of such a mandate. Music Choice's argument that
a rate reduction would not materially affect the return that record
labels receive for the sound recordings
[[Page 65224]]
they put into the marketplace is also misplaced. The relevant market
for determining whether an adjustment is warranted is the market for
PSS services, not the sound recordings market as a whole. As a
percentage of total royalties, the amount copyright owners receive from
the PSS services may be low. Nevertheless, all revenue sources are
important for those that have earned them, and the rate charged for the
use of sound recordings by the PSS must ensure that the copyright
owners receive a fair return. Therefore, no additional downward
adjustment is warranted.
    SoundExchange claimed that the PSS pay the lowest royalty rates of
any type of music service. Even if true, those comparative rates are
already reflected in the current statutory rate. Section 114 is clear
that the PSS that qualify for the grandfathered rate methodology are
sui generis. At the time the grandfathered provision was adopted the
number of qualifying services was very limited and has become more
limited over time. Only two companies qualify for the grandfathered
rate methodology and only for portions of their respective businesses.
Therefore, consistent with the section 114 grandfathering provision,
the correct question to ask is not whether the current statutory rate
(or whatever rate the Judges choose to begin analysis of the section
801(b) factors) offers copyright owners a fair income vis-[agrave]-vis
the rate they would earn from non-PSS music services but whether the
current statutory rate offers copyright owners an unfairly low return
that warrants an upward adjustment to ensure that copyright owners
receive a fair return in the upcoming rate period. Admittedly, it is a
difficult standard to meet, but SoundExchange has not provided
sufficient persuasive evidence to support such an upward
    \61\ Having determined that a downward adjustment is justified
by the second section 801(b) factor, the Judges have reassessed the
first section 801(b) factor and determined that no further
adjustment is warranted notwithstanding the rate decrease supported
by the second factor. The Judges review the evidence with respect to
the third and fourth factors with the assumption that a rate
reduction is already supported based on the second factor. 78 FR
31842, 31843 (May 28, 2013).
    After reviewing the evidence provided by both parties, the Judges
conclude that (outside of a 1 percentage point reduction due to the
anticipated expansion of the number of music channels that did not
materialize) neither party has provided sufficient evidence to support
a change from the current rate based on the second Section 801(b)
c. Factor C: Reflect Relative Roles
    The third section 801(b) factor requires the Judges to assess
whether the rate they have chosen to begin the section 801(b) analysis
reflects 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.
    Music Choice contended that with respect to this factor it has made
a much stronger evidentiary showing than SoundExchange and therefore a
lower rate should be warranted. MC PFF ] 426. For example, Music Choice
noted that it makes significant creative contributions in terms of
original programming, curation, and promotional content that increases
subscribers' engagement with the music and increases the promotional
impact of the Music Choice service. Williams WDT at 56; 5/18/17 Tr. at
4693 (Williams). Music Choice noted that it expends substantial
resources on improving its service offerings but that declining
revenues over the past rate period have forced Music Choice to cut
staff that are used to provide these services. Williams WDT at 7. Music
Choice discounted the record labels' contributions in this regard,
arguing that they apply only to the sound recordings and not
specifically to the PSS service. See MC PFF ] 447. Music Choice also
noted that historically it has had to invent the technology necessary
to get high-quality digital music programming to subscribers, but that
the current rate has limited its ability to continue investing in
improving its technology. Id. ]] 450-52.
    Music Choice asserted that the risks it faces are increasing
relative to those faced by the record companies. Music Choice also
contended that it (and other PSSs) has fewer opportunities for
profitability. Del Beccaro WDT at 20. Music Choice noted that its
residential business has still not become profitable on a standalone
basis. Id. at 19-20. Music Choice pointed to consolidation among MVPDs
and shrinking margins in the cable industry combined with competitive
pressures that have led to a rapid deterioration of Music Choice's
subscriber fees. Id. at 21. Music Choice represented that this changing
MVPD market has fundamentally changed the financial outlook for Music
Choice's residential audio service. Id. at 24-25.
    Music Choice disputed SoundExchange's assertions that the Music
Choice service is substitutional. See 5/16/17 Tr. 4076-77 (Harrison);
5/15/17 Tr. 3882 (Walker). Finally, Music Choice argued that it
contributed more to the opening of new markets for creative expression
and new media for its communication than the record companies. For all
of these reasons Music Choice believes that this factor warrants a
downward adjustment. MC PFF ]] 500-501.
    Not surprisingly, SoundExchange argued that no downward adjustment
is warranted under this factor. SoundExchange believes that ``Music
Choice's wholesale distribution model seems to be relatively
inexpensive to operate.'' See Wazzan CWDT ] 80. By comparison, record
companies spend far more on artists, repertoire, and marketing. Id.
SoundExchange countered Music Choice's argument that the record
companies' expenditures are not PSS-centered, arguing that without the
record companies' expenditures the PSS would have no sound recordings
to use for their services. Id. ] 80. SoundExchange further disputed
Music Choice's contentions that past expenditures by investors in Music
Choice warrant a rate reduction. According to SoundExchange, these
capital costs were invested long ago and the investors have made no
investments in the last eighteen years. See SX PFFCL ] 2141; but see
Del Beccaro WDT at 20. SoundExchange contended that these investors
have realized returns on their investments and that those investments
have helped fuel Music Choice's non-statutory video service line of
business. See SX PFFCL ] 2141; but see 5/18/17 Tr. at 4630-31 (Del
    With the exception of Music Choice's assertion that market
conditions have deteriorated recently, neither party made a persuasive
argument that a further change in the current statutory rate is
warranted, in either direction. Virtually all of the evidence that the
parties present reflects conditions that have occurred under the
current statutory rate. Therefore, all of the relative contributions of
SoundExchange and Music Choice are already incorporated into that rate
and no adjustment is warranted. The small rate reduction from the
current statutory rate that the Judges found warranted under the second
section 801(b) factor does not change the Judges' assessment.
    As for the negative change in market conditions, Music Choice only
noted a decline in the resources it spends and the staff it intends to
employ to improve the service. If anything, a decrease in the resources
it spends on the service, if quantifiable, would militate against a
rate reduction. At this time, it is unclear
[[Page 65225]]
how market conditions will affect Music Choice's business in the
upcoming rate period. Conceivably, persuasive evidence of dramatically
deteriorating conditions in the market for PSS service might militate
against an upward rate adjustment if such adjustment could be deemed
disruptive but any such adjustment would be warranted under the fourth
section 801(b) factor rather than the third. At this point, on the
current record, the Judges find no persuasive evidence to support an
adjustment from the current statutory rate in either direction under
the third factor.
d. Factor D: Minimize Disruptive Impact
    The fourth and final section 801(b) factor requires the Judges to
assess whether the rate (or rates) they have chosen to begin the
Section 801(b) analysis minimizes any disruptive impact on the
structure of the industries involved and on generally prevailing
industry practices. 17 U.S.C. 801(b)(1)(D). A royalty rate may be
considered disruptive ``if it directly produces an adverse impact that
is substantial, immediate and irreversible in the short-run because
there is insufficient time for [the parties affected by the rate]
adequately to adapt to the changed circumstances produced by the rate
change and, as a consequence, such adverse impacts threaten the
viability of the music delivery service currently offered to consumers
under this license.'' SDARS I, 73 FR 4080, 4097 (Jan. 24, 2008).
    Music Choice argued that the current statutory rate has had a
disruptive effect on the PSS market. As support for this premise, Music
Choice noted the previously discussed deterioration of Music Choice's
financial condition, which it contended is due, in part, to the fact
that the rate was increased in SDARS II. MC PFF ] 503. Music Choice did
not argue that profits from Music Choice's other business lines should
be considered in determining the possible disruptive effect of the PSS
rate. Id. ] 506.
    SoundExchange contended that if Music Choice and other PSSs cannot
continue to operate then the market will adjust by allowing other
competitors to take their place. See Wazzan CWRT ]] 83, 86. From
SoundExchange's perspective, Music Choice's quest for a lower rate is
motivated by increased competition from Stingray. According to
SoundExchange, Music Choice seeks a lower rate that would serve as a
subsidy that would allow Music Choice to maintain its unfair advantage
and its market share over non-PSS competitors. See 5/18/17 Tr. at 4532-
37 (Del Beccaro). SoundExchange asserted that such a subsidy ``fosters
Music Choice's inefficient operation and risks disrupting the market
for residential audio services.'' Wazzan CWDT ] 84. From
SoundExchange's perspective, the PSS rates are already artificially low
and merely serve to insulate Music Choice from market forces at the
record companies' expense. See Wazzan CWRT ]81, n. 112; 4/25/17 Tr. at
933-34 (Crawford). SoundExchange argued that the current statutory rate
is disruptive because it provides Music Choice a significant barrier to
entry in the market for non-PSS (CABSAT) services. 5/3/17 Tr. at 2318
(Wazzan); SX PFFCL ] 2147. SoundExchange did not accept that a higher
rate (even one as high as SoundExchange proposes) would be disruptive
to the PSS market. Rather it contended that an upward adjustment would
introduce a needed element of competition. See 4/25/17 Tr. at 902-03
(Crawford); Wazzan CWRT at 76, 83.
    The Judges find that neither party provided persuasive evidence to
warrant any further adjustment of the current statutory rate (other
than that warranted by the second 801(b) factor) in either direction.
Music Choice argued that the ``significant deterioration of its
financial condition'' is due in part to the current statutory rate but
the only evidence it cited deals with the effects of market
competition. See Del Beccaro WDT at 21. The competitive pressures that
Music Choice faces were not caused by the current statutory rate. While
the rate increase that the Judges approved in SDARS II may have
negatively affected Music Choice's margins, the Judges addressed any
potential disruptive effect of that increase by phasing it in over the
first two years of the rate period. The grandfathered rate calculation
methodology was not intended to shield Music Choice from all negative
impacts arising from competitive pressures. The reversal of that
increase that the Judges find warranted under the second section 801(b)
factor only makes Music Choice's arguments on this point less
    The reality of the marketplace contradicts SoundExchange's
contention that the current rate is disruptive. As SoundExchange
pointed out, Music Choice faces stiff competition in the market. SX
PFFCL ] 1879. The modest decrease in the statutory rate that the Judges
find warranted under the section 801(b)(1)(B) factor does not change
the Judges' assessment on this point.
    On balance, the Judges find that neither party has provided
persuasive evidence to support a finding that, under current market
conditions, an adjustment to the current statutory rate (other than
that discussed with respect to the second section 801(b) factor), is
warranted under the fourth Section 801(b) factor. Therefore, the Judges
determine that the appropriate rate for PSS services in the upcoming
rate period shall be 7.5%. This rate shall apply to the gross revenues
that the PSS services earn for all ``existing service offerings'' in
addition to all ``expanded service offerings'' as those terms are
defined and used at pages 15-16 of the Register of Copyright's
(Register's) Memorandum Opinion On Novel Material Questions of Law
(Memorandum Opinion) (Nov. 20, 2017). Based on the limited evidence in
the record, the Judges find no justification for applying a different
rate methodology to these two types of services at this time.
    The Judges accept as credible Music Choice's evidence that
additional channels that might conceivably fall within the expanded
service category currently constitute a marginal portion of Music
Choice's PSS service in terms of music usage. See Del Beccaro WDT at
16. While those types of services may increase over time, at this point
the Judges do not find that the service offerings that fall within this
category are sufficiently distinct from the existing service offerings
to justify the creation of a separate rate methodology. Nevertheless,
the Judges acknowledge SoundExchange's assertion that PSS services that
might fall within the expanded service category have recently increased
and may warrant a different rate methodology in the future. See Del
Beccaro WRT at 25; 5/18/17 Tr. 4658-59, 4661 (Del Beccaro).
D. Music Choice's Internet Streaming Service
    For the first time, in the present proceeding, SoundExchange
proposed a separate rate for PSS that stream their services over the
internet. For all licensed transmissions and related ephemeral
recordings through an internet streaming service qualifying as a PSS
(or any similar service capable of tracking the individual sound
recordings received by any particular consumer), SoundExchange
requested that the per-performance royalty fee for a commercial
webcaster set forth in 37 CFR 380.10 apply.\62\ Music Choice
[[Page 65226]]
contended that its streaming activity is already included within the
PSS statutory license and the royalty rate that PSSs pay already
includes this service. See Del Beccaro WRT at 27. As a result, Music
Choice contended that no additional royalty payment should apply for
internet streaming of the PSS service. Id.
    \62\ If a PSS does not have the technological capability to
track individual performances, SoundExchange proposes that the PSS
estimate its performances by multiplying its Aggregate Tuning Hours
by the average number of recordings played per hour across its
service. SX Amended Proposed Rates and Terms at 8.
1. Referral to the Register of Copyrights
    The Judges concluded that the threshold issue of whether the
streaming activities of a PSS were included within the scope of the PSS
license was a novel material question of copyright law that the Judges
must refer to the Register of Copyrights (Register). 17 U.S.C.
802(f)(1)(B). Hence, the Judges referred the issue to the Register,
    (1) Are a preexisting subscription service's transmissions of
multiple, unique channels of music that are accessible through that
entity's website and through a mobile application ``subscription
transmissions by preexisting subscription services'' for which the
Judges are required to determine rates and terms of royalty payments
under Section 114(f)(1)(A) of the Copyright Act?
    (2) If yes, what conditions, if any, must the PSS meet with
regard to streaming channels to qualify for a license under Section
114(f)(1)(A)? For example, must the streamed stations be identical
to counterpart stations made available through cable television? Is
there a limitation on the number of channels that the PSS may
stream? Is there a limitation on the number or type of customers
that may access the website or the mobile application?
Order Referring Novel Material Question of Substantive Law and
Setting Briefing Schedule at 3-4 (Oct. 5, 2017).
2. Register's Conclusions
    The Register concluded that
transmissions by a PSS entity that are accessible to a cable or
satellite television subscriber through that entity's website and
through a mobile application can be ``subscription transmissions by
preexisting subscription services'' for which the CRJs must
determine rates and terms of royalty payments under section
114(f)(1)(A), but only if such transmissions are sufficiently
similar to the transmissions made to those subscribers via the
entity's preexisting residential cable or satellite music service.
Memorandum Opinion at 12.
    As a preliminary matter, ``the preexisting services must be limited
to the three named entities in the [DMCA] Conference Report, i.e., DMX
(operated by TCI Music), Music Choice (operated by Digital Cable Radio
Associates), and [DiSHCD] (operated by Muzak).'' Id. at 14, internal
footnotes omitted. Moreover, the Register noted that ``not every
subscription transmission made by a PSS entity is subject to section
114(f)(1).'' Id. at 13. The Register observed that the DMCA's
amendments to section 114 of the Copyright Act were designed to move
the industry to market rates. Id. at 23. Nevertheless, the Register
noted that ``Congress intended for PSS entities to be able to expand
their service offerings to some limited extent and still have those
service offerings be considered PSS offerings.'' Id. at 14.
    According to the Register, the ultimate question is ``whether a
particular program offering by a PSS entity qualifies as a PSS offering
within the meaning of section 114(j)(11), and is therefore subject to
the grandfathered rate standard under section 114(f)(1).'' Id. at 15.
    The Register distinguished among three different types of service
    (1) A service offering identified by Congress as being a PSS
offering as of July 31, 1998, that is still offered today in the
same transmission medium identified by Congress in 1998. (The
Register refers to this type of offering as an ``existing service
offering''). According to the Register, an existing service offering
would be entitled to both a rate established under the grandfathered
rate standard under section 114(f)(1) and the grandfathered license
requirements in section 114(d)(2)(B). Id.
    (2) A service offering identified by Congress as being a PSS
offering as of July 31, 1998, that is still offered today, but in a
different transmission medium than the one identified by Congress in
1998, where only transmissions similar to the existing service
offering are provided. (The Register refers to this type of offering
as an ``expanded service offering''). According to the Register, an
expanded service offering would be entitled to a rate established
under the grandfathered rate standard in section 114(f)(1), but
would not be able to take advantage of the grandfathered license
requirements in section 114(d)(2)(B). A PSS that offered this type
of service would be required to comply with the more detailed
license requirements in section 114(d)(2)(C). Memorandum Opinion at
    (3) A service offering that is not an existing service offering
or an expanded service offering. (The Register refers to this type
of offering as a ``different service offering''). A ``different
service offering'' is insufficiently similar to an ``existing
service offering'' to be considered an ``expanded service offering''
and would not be entitled to either a rate established under the
grandfathered rate standard under section 114(f)(1) or the
grandfathered license requirements in section 114(d)(2)(B). Instead,
the royalty rate for a different service offering would be set under
the willing buyer/willing seller standard in section 114(f)(2). A
PSS marketing a different service offering would be required to
comply with the license requirements in section 114(d)(2)(C).\63\
    \63\ The Register's categorizations of service types presumes
that a service offering is eligible for the section 114 license. The
categorization is meant to delineate whether the rate for a license-
eligible service is determined pursuant to section 114(f)(1) or
section 114(f)(2). If a PSS entity began offering an interactive
service, for example, that service offering would not fall into one
of the categories and would not be eligible for the statutory
license. Memorandum Opinion at 16-17.
Memorandum Opinion at 16.
    The Register noted that ``an existing service offering can grow and
expand significantly within the same transmission medium while
remaining a PSS offering.'' Id at 19. Consistent with this
understanding, the Register noted that
[t]he user interface can be updated, certain functionality can be
changed, the number of subscribers can grow, and channels can be
added, subtracted, or otherwise changed. The only restriction is
that the existing service offering as it is today must be
fundamentally the same type of offering that it was on July 31,
1998--i.e., it must be a non-interactive, residential, cable or
satellite digital audio transmission subscription service.
Id. at 19-20 (internal footnotes omitted).
    With respect to the second category of offerings (i.e., expanded
service offerings) ``a [PSS] does not lose its designation as such in
the event the service decides to utilize a new transmission medium,
provided that the subscription transmissions are similar.'' Id. at 20
    \64\ For a service offering to qualify as an expanded service
offering, the PSS entity must continue to operate its existing
service offering. According to the Register, ``[a] service offering
that is not an existing service offering can only be subject to the
grandfathering provision if it provides transmissions similar to
their existing service.'' Memorandum Opinion at 20, internal quotes
    In assessing whether a service offering is an expanded service
offering and thus qualifies as a PSS offering, the Judges must compare
the service offering in question to the existing service offering as it
exists at the time of the comparison (rather than as it existed on July
31, 1998). Id. at 21. To aid the Judges in this comparison, the
Register offers a non-exhaustive list of factors:
    (1) Whether the service offering has a similar effect on
displacing or promoting sales of phonorecords.
    (2) Whether the quantity and nature of the use of sound
recordings by the service offering is similar.
    (3) Whether the service offering provides similar content to
similar user groups.
    (4) Whether the service offering is consumed in a similar
manner, provides a similar user experience, and has similar form,
feel, and functionality.
    (5) Whether and to what degree the service offering relates to
the pre-July 31, 1998 investments Congress sought to protect.
    (6) Whether and to what degree the service offering takes
advantage of the capabilities of the medium through which it is
transmitted (i.e., whether and the extent to which differences
between the service offerings are due to limitations in the existing
[[Page 65227]]
offering's transmission medium that are not present in the other
service offering's transmission medium).
Id. at 21-22.\65\
    \65\ Even if a service offering is found to be an expanded
service offering (rather than an existing service offering)
qualifying for the section 114(f)(1) grandfathering provision for
purposes of rate calculation, it would still not be eligible for the
section 114(d)(2)(B) grandfathering provision (regarding license
requirements) because it uses a different transmission medium than
the existing service offering. Such an offering would be subject to
the license requirements in section 114(d)(2)(C). Memorandum Opinion
at 22.
    A ``different service offering'' (the third category the Register
identified) can never qualify as a PSS offering because it would not be
one of the specifically identified pre-July 31, 1998 business
operations (i.e., the three PSS offerings) Congress sought to protect
when it enacted the DMCA. This is true regardless of whether the
service offering is developed internally or acquired. Id. at 22. When a
PSS entity expands its operations and provides additional transmissions
to subscribers to a different service, this is an entirely new
investment and is not a PSS offering. Id. at 23.
    The Register offered guidance regarding applications of the above
categorization of service offerings. First,
in accordance with the principles of narrow construction afforded to
grandfathering provisions, the Register finds that, as a matter of
law, it is irrelevant whether or not Music Choice or another PSS
entity, to some limited degree, was making transmissions via a
different medium than those specified in the legislative history on
July 31, 1998, such as the internet. If such a service was in fact
doing so, it would not be as part of an existing service offering--
any such transmissions today would be considered either an expanded
service offering or a different service offering. . . .
Id. at 19.
    The Judges must determine a royalty rate for the former type of
service (i.e., expanded service offering) in the current proceeding.
The latter type of service (i.e., different service offering) is
outside the scope of the current proceeding; a royalty rate for any
different service offering by a PSS (if any) must be determined by
reference to existing rate regulations covering that type of service
offering, in a separate, future proceeding under the willing buyer/
willing seller standard, or through voluntary negotiations.
    The Register observed that
the mere fact that a service offering is transmitted to cable or
satellite television subscribers over the internet does not
automatically disqualify the service offering from being an expanded
service offering subject to the grandfathered rate standard, so long
as the service offering, as a factual matter . . . is sufficiently
similar to the PSS entity's existing cable or satellite service
Id. at 25.
    In assessing whether an internet-based service offering is
sufficiently similar to a PSS entity's existing cable or satellite
service offering, the Judges should consider ``the degree to which
making the existing service offering accessible outside the home of the
subscriber constitutes a fundamental change to the offering.'' Id.
    According to the Register:
At least in the cable television market, there appears to be a
distinction drawn between accessing content within the home and
accessing that same content outside of it. To be clear, this
distinction is one based on the location where the PSS offering is
consumed, not the type of device on which the service is accessed.
If the service offering is available through an internet-connected
smartphone or tablet, but is designed so that the service offering
will only work when accessed within the confines of the subscriber's
residence, then it would be within the home and more similar to the
PSS entity's existing cable or satellite service offering.
Id. at 26 (internal footnote omitted).
    With respect to the impact that the number and type of channels
offered by a service has in determining its categorization for rate-
setting purposes, the Register identified examples of factors the
Judges could consider, such as how many additional or fewer channels
there are, how many channels offer different programming, and how
different that programming is from that offered by the existing service
offering. Id. The Register also notes that the Judges should consider
the reasons why any such differences exist. If the service offering has
more channels because of some benefit the internet provides (e.g.,
greater bandwidth or different contractual arrangements with cable
operators), then the PSS entity could be taking advantage of the
capabilities of the internet as a transmission medium, which could tend
to disqualify that service offering from the grandfathered royalty
calculation method. Id. at 26-27. A similar analysis could be conducted
with respect to the number and type of customers. Id. at 27.\66\
    \66\ Differences in a service offering that directly and solely
result from the imposition of the section 114(d)(2)(C) requirements
that do not apply to the existing service offering (which is subject
to section 114(d)(2)(B)) should not alone disqualify the service
from the grandfathered royalty calculation methodology necessitated
by the change in medium, nor should minor differences in the user
interface or in the visual presentation. Memorandum Opinion at 27.
    The Register noted that if a service offering qualifies for the
grandfathered rate-setting methodology, the Judges still have the
authority under section 114(f)(1)(A) to distinguish among the different
types of digital audio transmission services in operation. If material
differences between an existing service offering and an expanded
service offering exist, the Judges may set separate rates based on
those difference, using the section 801(b)(1) standard. Id. at 27-28.
3. Application of Register's Conclusions to Current Proceeding
    Music Choice provides 50 channels of audio music programming
delivered to subscribers' televisions (the Cable Radio Service). It
also makes these 50 channels, plus an additional 25 internet-only
channels, available to authenticated television subscribers through its
website and a mobile app (the internet Service). Del Beccaro WDT at
    \67\ See also Wazzan CWDT at ] 62(e) (``Music Choice provides 75
audio channels through various MVPDs, . . . and streaming to
subscribers of the cable services that carry its channels, through a
family of apps and a web portal.'') (internal footnotes omitted).
    The Register has determined, as a matter of law, that Music
Choice's internet Service \68\ is not an ``existing service offering.''
Memorandum Opinion at 19. Consequently, the internet Service is either
an ``expanded service offering'' (i.e., qualifying for grandfathered
royalty determination under the Section 801(b) factors but subject to
the expanded license requirements under section 114(d)(2)(C)) or a
``different service offering'' outside the scope of the PSS license.
    \68\ Neither party asked the Judges to determine whether Music
Choice's Cable Radio Service, as it exists today, constitutes an
``existing service offering'' or and ``expanded service offering''
by a PSS. As the Judges have already determined that the PSS rate
covers both types of offerings, the question is moot and the Judges
need not address it.
    By reference to the Register's six-factor list of criteria to
differentiate an expanded service offering from a different service
offering, the Judges find that an internet-based service that allows
subscribers to access music outside their residences is a ``different
service offering'' and is not eligible for grandfathered PSS rate
structures or license requirements applicable to PSS. The regulations
in Appendix A, therefore, exclude internet-based transmissions to the
extent they are available outside a subscriber's residence.
[[Page 65228]]
V. SDARS Performance License--Rate Structure
A. Rate Structure Arguments
1. Maintaining the Current Rate Structure
    Sirius XM emphasized that the Judges have utilized a percent-of-
revenue rate structure for ten years, and that absent any new and
sufficient factual bases to deviate from that history, the Judges
should continue to adopt this rate structure. SXMRPFF ] 384 (and record
citations therein). Moreover, it noted that SoundExchange itself
proposed a percent-of-revenue rate structure, not a ``greater-of''
structure, as recently as in the SDARS II proceeding. SXMPFF ] 253 (and
record citations therein).
    SoundExchange did not take issue with the historical bona fides of
the current rate structure. However, SoundExchange noted that it urged
the Judges to adopt what it describes as a simpler percent-of-revenue
approach in SDARS II, but the Judges refused, opting instead for a more
complicated structure that led to substantial disputes. SERPFF ] 253.
    The Judges are not convinced by Sirius XM's argument that the rate
structure should be maintained merely because it has been in place over
the past two rate periods. The Judges are charged with setting rates
and terms de novo for each period. If there are sufficient valid
reasons why the rate structure should be changed, then the Judges will
adopt those changes. Accordingly, the Judges consider the issues to
determine whether to change the existing rate structure.
2. Factors Relating to a Change in Structure
a. Lack of Expert Support
    SoundExchange advocated a deviation from the percent-of-revenue
rate structure that has existed throughout the SDARS I and SDARS II
rate periods. SoundExchange asked the Judges to establish a ``greater
of'' structure, by which the royalty rate is calculated ``on a calendar
year basis,'' but payable monthly, as the greater-of a specified
percentage of revenue or a specified per subscriber dollar value. See
Amended Proposed Rates and Terms of SoundExchange, Inc. and Copyright
Owner and Artist Participants App. A at 14-15. (Jun. 14, 2017).
    Sirius XM noted that no economist appearing in this proceeding
endorsed the use of a greater-of formula. SXM RPFF ] 383. Moreover,
Sirius XM pointed out that Mr. Orszag, an economic witness appearing
for SoundExchange, expressly testified that he advocated either a
percent-of-revenue rate structure or a per subscriber structure, and
that he did not testify in support of a structure incorporating those
two approaches in a single greater-of approach. SXM PFF ] 251. In
response, SoundExchange did not identify any testimony that explicitly
or adequately endorsed the use of a greater-of formula from an economic
point of view.
    The Judges are troubled by the lack of a cogent explanation from
the licensors' economic witnesses as to the merits, on balance, of a
greater-of rate formula. The absence of such evidence could be overcome
by explanations derived from other evidence or testimony. Not having
that further evidence, the Judges find it significant that no economist
has sufficiently explained the benefits of this greater-of approach.
b. Impact on the Parties' Risks and Rewards
    SoundExchange maintained that its proposed greater-of approach is
warranted because it allows record companies to share in the growth of
Sirius XM's revenue, while offering protection to the record companies
on the downside if revenues are too low. SEPFF ] 252 (and record
citations therein). Sirius XM argued, in essence, that this approach
smacks of a heads I win, tails you lose approach, whereby record
companies share the upside of Sirius XM's success, but have protection
in the form of a default to the per subscriber rate if the upside does
not materialize. SXM PFF ] 252.
c. Benchmarks Include a Greater-Of Rate Structure
    SoundExchange emphasized that many interactive license agreements
utilize the greater-of approach that SoundExchange advocates here,
demonstrating the market's adoption of this approach. SEPFF ]] 164-165
(and record citations therein). However, Sirius XM noted that these
interactive agreements were all negotiated in a market characterized by
the lack of effective competition, and that the lack of competition
would affect the structure as well as the level of rates. SXMPFF ] 385
(and record citations therein).
    The Judges find Sirius XM's effective competition point well-taken
in this context. Given that SoundExchange's expert economic witnesses
acknowledged the need for rates that reflect an effectively competitive
market, it is no surprise that none of their economists touted the
greater-of structure as a reflection of effective competition. The
Judges find that the greater-of rate structure, advantageous to
licensors through the shifting risks, may well represent an example of
what licensors can and would obtain when they exploit their ``must
have'' status for a special competitive advantage. The Judges do not
find it persuasive that interactive streaming services and record
companies adopt the greater-of structure in their negotiated licenses.
d. Impact on Royalty Disputes
    SoundExchange argued at length that a greater-of rate structure
that contains a per-subscriber prong will eliminate disputes regarding
the definition of revenue under the percent-of-revenue approach. SEPFF
]] 1646-1650 (and record citations therein). However, Sirius XM
convincingly countered that a greater-of formula will not eliminate the
issues of revenue definition and identification, because the issue of
which prong creates the ``greater'' royalty will itself be dependent on
the definition, identification, and calculation of the revenue-based
royalty prong. SXM PFF ] 386.
    The Judges agree with Sirius XM. If SoundExchange had proposed a
per-subscriber rate only, then the issues surrounding the percent-of-
revenue rate would be eliminated. But SoundExchange did not proposed a
pure per-subscriber rate; nor did Sirius XM. Thus, the problems
regarding the revenue-based royalty would continue to be present
(albeit perhaps less often than under a pure revenue-based rate).
e. The Greater-Of Rate Structure and Trial Subscriptions
    SoundExchange argued that its greater-of proposal helps to obviate
the dispute between the parties regarding the length of free trials
offered to potential subscribers by new owners of automobiles.
SoundExchange noted that interactive services are generally required to
pay royalties for any free trial that exceeds [REDACTED]. Orszag AWDT ]
85.\69\ By contrast, Sirius XM typically offers free trials to new and
used car purchasers that last three to twelve months.\70\ Id. ] 81.
SoundExchange argued that ``there is no
[[Page 65229]]
sound economic basis for the present disparate treatment, under which
Sirius XM is permitted to offer the repertoires of rights owners for
durations greater than one month without the payment of royalties,''
id. at ] 85, and proposed to eliminate that disparate treatment by
classifying trial users as ``subscribers'' for royalty purposes, and
setting a per-subscriber rate that varies depending on how long the
user has been in the free trial period.\71\ Thus, it would be
irrelevant to the licensors if the free trial generated no revenue or
lower revenue from automobile Original Equipment Manufacturers (OEMs)
during the period offered free to the listener. SEPFF ]] 1657-1665.
    \69\ The benchmark interactive services agreements address free
trials longer than [REDACTED] by imposing a [REDACTED] royalty. See
Orszag AWDT ] 89.
    \70\ Some paid promotions (where the automobile Original
Equipment Manufacturer pays a reduced subscription fee to Sirius XM
during the free (to the consumer) trial period) may last longer than
[REDACTED] months. See Trial Ex. 322 at 14, 15 ([REDACTED]-month
free trial for purchasers of certain high-end luxury cars
([REDACTED])). Under a percentage revenue rate structure Sirius XM
pays a royalty on this discounted subscription revenue. See Orszag
AWDT ] 82.
    \71\ SoundExchange's amended rate proposal would charge no
royalties for subscribers who are in the first month of their free
trial. During the second and third months of a free trial,
SoundExchange proposes a per-subscriber royalty rate that represents
a discount of approximately 42% off SoundExchange's proposed full
per-subscriber rate. The full per-subscriber rate would apply to all
free trials after three months. See Amended Proposed Rates and Terms
of SoundExchange, Inc. and Copyright Owner and Artist Participants,
at 3 (Jun. 14, 2017).
    Sirius XM argued that trials, both paid and unpaid, provide value
to licensors to the extent they entice new subscribers whose
subscription revenue is then shared by the licensors. Sirius XM
described the trials as a ``joint effort'' by Sirius XM and the record
companies to attract more Sirius XM subscribers and produce future
subscription revenues that inure to their mutual benefit. Corrected
Written Rebuttal Testimony of Carl Shapiro, Trial Ex. 9, at 55 (Shapiro
CWRT). Sirius XM further argued that it is in the best position to
determine the most beneficial length of the trial period, and that
requiring Sirius XM to pay per-subscriber royalties without recompense
from the trial users would act as a disincentive to Sirius XM to
utilize what it otherwise understood to be the optimal trial period.
SXMRPFF ] 388.\72\
    \72\ Sirius XM also argued that the record companies have a
higher benefit/cost ratio from trial subscriptions than Sirius XM,
and would thus agree in an unregulated market to waive royalties
``for as long as Sirius XM would choose to run unpaid trials.''
Shapiro CWRT at 55-56. SoundExchange rejected this argument because
Professor Shapiro assumed, in computing his benefit/cost ratio, that
no record company is a ``must have'' for Sirius XM. SEPFF ] 1619;
see 4/24/17 Tr. 562 (Shapiro). As a result of this assumption,
Professor Shapiro's benefit/cost calculation relied on a much lower
record company opportunity cost than that adopted by the Judges. See
infra, section VI.B.3. The Judges do not rely on this Sirius XM
argument, therefore, in rejecting SoundExchange's proposal with
regard to trial subscriptions.
    The Judges agree with Sirius XM. Under a percent-of-revenue royalty
structure, Sirius XM and the record companies are aligned in their
interest to minimize the time period for unpaid trials and trials paid
by OEMs at less than the full subscription rate. Moreover, because
Sirius XM is in the business of recruiting and interacting with
potential subscribers, it would be less efficient for the licensors (or
the Judges) to second-guess Sirius XM's downstream (retail) business
model as it relates to the optimal period of trial use. Although it
would appear from a cursory analysis that artists and record companies
suffer from the use of their recordings without recompense (or
sufficient recompense) during trial periods, the fuller view, given
Sirius XM's aligned economic incentive to maximize revenues,
demonstrates that the length and terms of trial periods are likely
consonant with the interests of the licensors. This record evinces no
evidence to the contrary.\73\
    \73\ For example, there is no credible evidence that Sirius XM
is interested in growing market share irrespective of revenue
growth, in order to compete for the market (rather than merely in
the market). This is unsurprising, because Sirius XM has already
captured the satellite radio market. See infra, text following note
B. Conclusion Regarding the Rate Structure
    For the foregoing reasons, the Judges adopt a percent-of-revenue
rate structure in this proceeding for the 2018-2022 rate period.
VI. SDARS Performance License: SoundExchange Proposal
    SoundExchange proposed a royalty fee that is the greater-of a per-
subscriber rate and a percent-of-revenue rate. With regard to the
percent-of-revenue prong, SoundExchange requested a rate equal to 23%
of Sirius XM's ``Gross Revenues,'' as that quoted term shall be defined
in the forthcoming regulations. See SoundExchange's Proposed Rates and
Terms, at 2-3.\74\ The per-subscriber rate proposed by SoundExchange is
set forth in the table below:
    \74\ The definition of ``Gross Revenues'' for the forthcoming
rate period is discussed infra, section XI.A.2.
                                            Free trial
                                            subscribers      All other
                  Year                      (months two     subscribers
                                            and three)
2018....................................           $1.45           $2.48
2019....................................            1.49            2.55
2020....................................            1.54            2.63
2021....................................            1.58            2.71
2022....................................            1.63            2.79
    For affirmative economic support of its rate proposal,
SoundExchange relied principally on the expert opinions of two economic
witnesses, Mr. Jonathan Orszag and Professor Robert Willig. Mr. Orszag
used a ``ratio equivalency'' analysis, which he applied through two
separate approaches. Professor Willig considered several economic
models: (1) An ``Opportunity Cost'' analysis; \75\ (2) a ``Ramsey
Pricing'' analysis; \76\ and (3) a ``Nash Bargaining Solution''
approach.\77\ Professor Willig also discussed a fourth model--the
Efficient Component Pricing Rule (ECPR), which he noted in his oral
testimony as analytically analogous to his ``Opportunity Cost''
analysis, and yielded the same rate.\78\
    \75\ See infra, sections VI.A-VI.C.
    \76\ See infra, section VI.G.
    \77\ See infra, section VI.F.
    \78\ See infra, section 0.
A. Professor Willig's Opportunity Cost Model
1. ``Walk-Away'' Opportunity Cost
    SoundExchange called Professor Robert Willig in support of its
proposed rates. Professor Willig approached the rate determination
using an opportunity cost model. As Professor Willig testified,
opportunity costs are incurred when ``sales through one distribution
channel reduce (i.e., substitute for, or ``cannibalize'') sales through
other distribution channels (thereby reducing compensation earned by
content creators from those other channels . . . ).'' Written Rebuttal
Testimony of Robert Willig, Trial Ex. 46, ] 20 (Willig WRT); see also
Written Direct Testimony of Carl Shapiro, Trial Ex. 8, at 19 (Shapiro
WDT) (sellers incur opportunity cost when sales in one market diminish
sales in other markets). Based upon his interpretation of survey
evidence, Professor Willig established a walk-away opportunity cost of
$2.55 per subscriber, which he equates to [REDACTED]% of Sirius XM's
relevant revenue.
    SoundExchange asserted that the appropriate opportunity cost for
rate-setting purposes is the ``walk-away'' opportunity cost. SE PFF ]]
486-95. Professor Willig defined a record label's walk-away opportunity
cost as ``compensation that it would earn from other sources of
distribution,'' if a label were ``to literally walk away from a
distributor.'' 5/2/17 Tr. 2014-15 (Willig). Professor Willig referred
to the opportunity cost as ``creator compensation cannibalization,''
\79\ and observed that ``the need to cover opportunity cost is part of
what assures efficiency in the ultimate choice of the
[[Page 65230]]
balance of . . . varieties of modes of distribution.'' 5/2/17 Tr. 2019-
20 (Willig). In an unregulated market, a supplier (record label or
copyright owner) will not sell (license) to a service unless the
supplier is compensated at or above its walk-away opportunity cost. See
id. at 2019. In this regulated market, however, the creators do not
have the option to walk away; the licenses are compulsory. Id. at 2015.
Professor Willig thus perceived the role of the Judges to ``redress
that imbalance created by the statutory license.'' \80\ Id. at 2017.
    \79\ Sirius XM's rebuttal economic expert, Professor Farrell,
concurred with the substance of this definition, agreeing that walk-
away opportunity cost ``is the profit that a label would realize
elsewhere'' if it did not license to Sirius XM. 4/24/17 Tr. 607
    \80\ Opportunity costs are more than a theoretical concept. For
example, UMG recognizes that on-demand subscription services may
substitute for sales of digital downloads. See Written Direct
Testimony of Aaron Harrison, Trial Ex. 32, at ] 17 (Harrison WDT).
Accordingly, when UMG licenses fully interactive streaming services,
it [REDACTED]. Because the direct marginal costs of distributing
additional sound recordings to Sirius XM are ``zero or nearly
zero,'' the principal marginal cost to a record company of licensing
to a service is its opportunity cost. Shapiro WDT at 19; see also
SEPFF ] 460 (not disputing Professor Shapiro's point that physical
marginal cost is zero and that the only marginal cost at issue is
marginal opportunity cost).
    As a matter of economic principle, Sirius XM did not dispute the
use of an opportunity cost approach as appropriate in identifying a
market-based SDARS royalty rate. See SX RPFF ] 109. However, Sirius XM
disagreed with Professor Willig's use of ``walk-away opportunity
cost,'' as he defined that phrase. Id.
    The Judges summarize the parties' opportunity cost dispute as:
Whether, in a hypothetical market with freely negotiated rates,
opportunity cost should (1) include the value of each Major's ``must-
have'' status which gives each Major the theoretical ability to put
Sirius XM out of business by refusing to grant it a license at a
royalty less than opportunity cost; or (2) exclude this value--a
complementary oligopoly power--by which each Major hypothetically could
put Sirius XM out of business.\81\
    \81\ Professor Willig calculated walk-away opportunity cost on
the tautological assumption that, because each Major is a ``must
have,'' its refusal to provide a license to Sirius XM would cause
Sirius XM to go out of business. As discussed elsewhere in this
Determination, Professors Shapiro and Farrell proposed the use of a
different form of opportunity cost, one that does not assume that
the loss of any one Major would cause Sirius XM's demise.
    Professor Willig asserted that the walk-away opportunity cost for a
``must-have'' label is effectively the same as the label's pro rata
share of the industry-wide opportunity cost.\82\ See 5/2/17 Tr. 2137
(Willig); \83\ see also SEPFF 502 at ] 502. Professor Willig's
opportunity cost calculation thus measures what a must-have single
record label would earn elsewhere, and proposes it as an industry-wide
measure, even if that single record label is the only label that
declines to license. On this theoretical point, Professor Farrell, one
of Sirius XM's economic experts was in basic agreement. See, e.g., 4/
24/17 Tr. 665-66 (Farrell) (for label to recover pro rata walk-away
opportunity cost, industry-wide royalty rate would have to be at least
equal to industry-wide opportunity cost).
    \82\ The evidence in this proceeding strongly demonstrates the
``must have'' status of each Major. See SE PFF ]] 517-525 (and
record citations therein). Indeed, Sirius XM implicitly acknowledged
the ``must have'' status of a Major, citing a steering adjustment as
a method by which to mitigate the ``must have'' status and
complementary oligopoly power of a Major to allow for an effectively
competitive market.
    \83\ Professor Willig did not cite any authority that has
previously used the phrase ``walk-away opportunity cost.'' Sirius
XM's economic experts asserted that Professor Willig's ``walk-away
opportunity cost'' is actually the ``monopoly'' or ``cartel''
opportunity cost. For the sole purpose of referring to and
discussing Professor Willig's approach, the Judges will use his
``walk-away'' terminology; that usage does not suggest an
equivalence with, or distinction from, monopoly or cartel
opportunity cost.
    Further, Professor Willig opined that individual labels would
bargain with an understanding that a royalty unacceptable to that label
is likely also unacceptable to other labels. As a result, a label
inclined to reject a proposed royalty will expect that other labels
will do the same, with the result that each label's opportunity cost
will equate to an industry-wide opportunity cost. See 5/2/17 Tr. at
2030 (Willig).
2. Sirius XM's Criticism of Willig's Use of ``Walk-Away'' Opportunity
    Sirius XM disputed the notion that opportunity costs should be
defined and calculated on an industry-wide basis; rather, it asserted
that the appropriate calculation must be undertaken in a ``label
specific'' manner. Sirius XM asserted an essential and disqualifying
premise: The opportunity cost Professor Willig calculated is the
opportunity cost of ``either a single monopoly record label or a fully
effective cartel.'' \84\ Farrell WRT ] 27; see also id. ] 31. As
Professor Shapiro noted:
    \84\ Professor Farrell testified that if a particular label's
decision to license is based on ``the profit impact on the industry
as a whole, that's what we would normally describe as monopoly or
cartel behavior.'' 4/24/17 Tr. 614 (Farrell).
    Most fundamentally, Professor Willig is asking the wrong
question. Rather than attempting to calculate the opportunity cost
to an individual label of having its sound recordings performed on
Sirius XM, Professor Willig calculates the opportunity cost to the
entire recorded music industry, as if a single entity (or a fully
functioning cartel) controls the rights to all sound recordings.
Shapiro CWRT at 34. Moreover, Sirius XM claimed Professor Willig
acknowledged that his opportunity cost calculation was identical to the
opportunity cost that would apply ``if there were a single monopoly
seller of sound recordings. . . .'' 5/2/17 Tr. at 2140 (Willig); see
also Farrell WRT ]] 67-71 (Willig's calculation is ``extreme'' and
leads to inflated opportunity costs).
    According to Sirius XM, Professor Willig's opportunity cost
approach ignores the goal of determining a statutory rate reflective of
an effectively competitive marketplace (as tempered by the enumerated
section 801(b)(1) factors). See 4/20/17 Tr. 418 (Shapiro) (``he is
measuring the wrong thing by looking at the monopoly opportunity
cost.''). Thus, Professors Shapiro and Farrell both opined that a rate
based on this industry-wide opportunity cost would be inconsistent with
the economic concept of ``workable competition.'' \85\ See Shapiro CWRT
at 37; Farrell WRT ]] 27-29.
    \85\ ``Effective'' competition, as used in this Determination is
synonymous with the term ``workable competition'' that is more
commonly used by economists.
    Sirius XM candidly admitted that its criticism of Professor
Willig's walk-away opportunity cost analysis is premised on the
assumption that a single label ``does not have the `must-have' monopoly
power to effectively shut-down Sirius XM's music offering . . . .'' SXM
PFFCL ] 367 (and record citations therein). Having made this
assumption, Sirius XM's witnesses explained what they characterize as a
fairly simple intuition grounded on their economic modeling in the
record: ``[A] change in Sirius XM's music mix (i.e., something less
dramatic than losing access to all music) is likely to result in only
some relatively modest loss in subscribers, if any--not, as Professor
Willig models, every Sirius XM subscriber seeking music elsewhere. See
Farrell WRT ] 67.\86\
    \86\ Professor Farrell's argument ``demonstrated mathematically
that if Sirius XM's failure to obtain a license from a record label
led to the loss of some, but not all, subscribers, then the walk-
away opportunity cost for that label would be significantly less
than that label's pro-rated share of the monopoly opportunity cost
calculated by Professor Willig, the difference between the two
depending on the fraction of Sirius XM subscribers who would cancel
their subscriptions in response to the failure of Sirius XM to
secure a license from the individual label.'' Farrell WRT ]] 68, 71.
    Sirius XM lodged another fundamental objection to Professor
Willig's opportunity cost approach. As Sirius XM noted, Professor
Willig's $2.55 opportunity cost calculation was
[[Page 65231]]
derived by applying the royalties alternative services pay to record
companies. In Web IV, the Judges found these rates to be inflated by
the complementary oligopoly power of the Majors. Sirius XM criticized
``importing'' that ``supracompetitive'' rate into this statutory
setting in the absence of any adjustment or allowance for effective
competition. The royalty with the most disproportionate impact in this
regard is the $[REDACTED]/month royalty charged to subscription
interactive services.\87\ See Written Direct Testimony of Robert
Willig, Trial Ex. 28, ] 41 & Table 2 (Willig WDT). Professor Farrell
argued that Professor Willig's calculations are significantly infected
by the noncompetitive market for licenses to interactive services. See
4/24/17 Tr. 636, 640 (Farrell). Professor Farrell cautioned against
putting ``heavy weight on a rate that has been found to be
supracompetitive and driven by complementary oligopoly . . . .'' 4/24/
17 Tr. at 641 (Farrell). Even Professor Willig agreed that a lack of
steering in the interactive market could inflate the opportunity cost
calculation for Sirius XM. 5/2/17 Tr. at 2037-38 (Willig).
    \87\ Based upon Professor Dhar's survey, interactive services'
diversion ratio of 31% comprises 70% of Professor Willig's $2.55
opportunity cost. The Judges examine the survey data infra, section
    Further, Sirius XM chastised Professor Willig for a claimed
inconsistency. Professor Willig acknowledged on the one hand that
benchmarks from other distribution channels, such as the interactive
services benchmark, must be free of the effects of complementary
oligopoly. Nonetheless, he applied the rates from these same
distribution channels without a downward adjustment to offset the
upward impact of the complementary oligopoly effect when computing
opportunity cost. See 5/2/17 Tr. 2152-54 (Willig).
    Sirius XM also criticized Professor Willig for his second
alternative justification for using the industry-wide opportunity cost;
that is what Sirius XM labeled his ``unilateral alignment'' approach.
See SXM PFFCOL ]] 379-382. Sirius XM characterized this as the
``conscious parallelism'' of like-minded oligopolists, viz., a form of
anticompetitive ``tacit collusion which, even though not a violation of
any antitrust laws, would nonetheless lead to results that would be
inconsistent with the necessity that rates be consistent with the
principles of effective (workable) competition.'' Id. ] 381 (and record
citations therein).
3. The Judges' Use of the Opportunity Cost Model To Set the SDARS
Royalty Rate
    The Judges find that Professor Willig's industry-wide walk-away
opportunity cost approach is an appropriate tool, on the present
record, to apply as an interim step in crafting the statutory rate. On
the one hand, there is no dispute between the parties that the Majors
would use this industry-wide opportunity cost calculation to set
royalty rates in an unrestricted market. On the other hand, the Judges
find there is no bona fide dispute but that these rates would partially
reflect the complementary oligopoly effect of Majors.
    Standing alone, the complementary oligopoly effect within the walk-
away opportunity cost model would inflate the rate above the
``reasonable rate'' the Judges must determine. However, the Judges may
mitigate the industry-wide walk-away opportunity cost that incorporates
complementary oligopoly effects, as they do in their ``fork in the
road'' approach later in this Determination. Thus, even if one could
construe Professor Willig's ``walkaway'' approach, standing alone, as
inconsistent with the concept of effective competition, that
inconsistency can be--and is--mitigated because the because the Judges
have considered and accounted for such ``must have''/complementary
inefficiencies by also accepting Professor Willig's practical and
reasonable ``fork in the road'' approach, discussed below.
    The Judges find unhelpful SoundExchange's alternative justification
for the use of walk-away opportunity costs in the marketplace. This
alternative point simply noted that the major record labels, who are
oligopolists, would engage in some form of what is known as ``conscious
parallelism'' when negotiating royalties. See 5/2/17 Tr. 2027 (Willig)
(``decision-making is unilateral, but parallel, across the record
[l]abels''); see also SE PFF ] 526. This exposition explains why
oligopolists would move in concert without engaging in explicit
collusion, but begs the question whether that concerted price movement
would incorporate walk-away opportunity cost ab initio. It is Professor
Willig's first point--that each Major's knowledge of its ``must have''
status imbues it with individual market power to walk-away--that is
sufficient to demonstrate the market logic of the industry's collective
exploitation of walk-away opportunity cost. See 5/2/17 Tr. 2031-34
    The Judges also find unhelpful Sirius XM's argument that Professor
Willig's opportunity cost approach is the equivalent of a benchmarking
approach. To be sure, the point is correct, but its advancement as a
criticism is wrong. When properly weighted, the opportunity cost
approach is tantamount to a useful benchmark, because the weightings
are quite analogous to (and more precise than) the ``adjustments'' the
Judges consistently make to proposed benchmarks. To the extent the
opportunity cost is infected by complementary oligopoly inefficiencies
that increased the rates from which that opportunity cost is derived,
the Judges look to the entire record to ascertain whether and how to
account for that factor, as they have by applying Professor Willig's
``fork in the road'' approach.
B. Application of the Opportunity Cost Approach
    To apply the walk-away opportunity cost approach in the satellite
radio market, Professor Willig utilized the survey conducted by
Professor Ravi Dhar (Dhar Survey) to calculate his $2.55 per subscriber
per month opportunity cost of licensing sound recordings to Sirius XM.
Willig WDT ] 41. Professor Willig's analysis is built upon two
principal elements: Diversion ratios and creator compensation data.
    Professor Willig derived the first element (his ``diversion
ratios''), from substitution data which indicate the other sources and
modes of distribution of recorded music to which Sirius XM subscribers
would gravitate if Sirius XM were no longer available at acceptable
prices. 5/2/17 Tr. 2057-58 (Willig). More particularly, the Dhar Survey
examined how Sirius XM subscribers would react to a higher price for a
subscription to Sirius XM. 5/2/17 Tr. 2057-58 (Willig). The Dhar Survey
first asked respondents if they would discontinue their Sirius XM
service at various higher prices. Willig WDT ] 40. Those respondents
who answered these ``pricing questions'' by stating they would cancel
their Sirius XM subscriptions were then asked certain ``switching
questions.'' The respondents were asked how they would listen to music,
and specifically which of the alternative distribution channels
presented in the survey question they would select. Willig WDT ] 40
(summarizing relevant aspects of Dhar Survey); Corrected Written Direct
Testimony of Ravi Dhar, Trial Ex. 22, ]] 58-60 & App. D at 69-70 (Dhar
    With the foregoing information in hand, Professor Willig needed to
assign monetary values to the diversion ratios.
[[Page 65232]]
This second element, for which Professor Willig coined the phrase
``creator compensation data,'' is the amount of compensation that would
flow to sound recording licensors from the distribution platforms to
which Sirius XM subscribers would migrate. 5/02/17 Tr. 2058-59
    To link the diversion ratio and creator compensation data for each
alternative distribution mode to which Sirius XM subscribers would
migrate, Professor Willig multiplied the diversion ratio by the creator
compensation data (per subscriber). The product according to Professor
Willig equals the opportunity cost associated with consumers listening
to Sirius XM as opposed to each alternative distribution mode. 5/2/17
Tr. 2059-60 (Willig).
    Professor Willig then added each of the positive weighted levels of
monthly creator compensation for each alternative distribution mode.
Willig WDT ] 41. According to Professor Willig, this summation
represents the total opportunity cost of licensing Sirius XM across all
alternative modes of distribution. He summarized his calculations in
the following table.
                       Opportunity Cost Based on Dhar Survey Responses--Summary of Results
                                                                   Alt. mode mix   Unit creator    Wghtd creator
              Distribution across alternative modes                     (%)       comp $/Sub-Mo.  comp $/Sub-Mo.
Paid Interactive................................................              31      [REDACTED]      [REDACTED]
Paid Noninteractive.............................................              15      [REDACTED]      [REDACTED]
Purchase CDs/downloads..........................................              10      [REDACTED]      [REDACTED]
Ad-supported Noninteractive.....................................               4      [REDACTED]      [REDACTED]
Ad-supported Interactive........................................               3      [REDACTED]      [REDACTED]
Music video.....................................................               2      [REDACTED]      [REDACTED]
Cable/satellite music channels..................................               2            0.00            0.00
Other (zero creator comp).......................................              32            0.00            0.00
    Total/Weighted-Average......................................             100            2.55            2.55
Willig WDT ] 41, Table 2.
    \88\ Professor Willig detailed how he derived the creator
compensation data for each line item in his table. See Willig WDT ]]
477-485. (The calculation methods are not in dispute.)
    As his tabular data demonstrate, Professor Willig calculated the
full opportunity cost across all alternative modes of distribution as
totaling $2.55 per subscriber per month. Willig WDT at ] 41. This
opportunity cost calculation is consistent with SoundExchange's
proposed per-subscriber royalty range of $2.48 in 2018 to $2.79 in
2022. Given Sirius XM's ARPU of $[REDACTED] per month, Professor
Willig's $2.55 per subscriber rate is equivalent to [REDACTED] % of
revenue.\89\ Thus, Professor Willig's conclusion is consistent with
SoundExchange's 23%-of-revenue rate proposal covering all five years in
the forthcoming rate period.
    \89\ ARPU is the industry acronym for ``Average Revenue per
User.'' See also infra note 142 regarding the quantification of
1. Survey Data Underlying ``Opportunity Cost'' Approach
    Professor Willig's opportunity cost approach is dependent upon the
weights he placed on various distribution channels. The Judges,
therefore, test the underlying survey data on which he relied to assess
their reliability or, more specifically, their strength in supporting
Professor Willig's conclusions.
    The Dhar Survey was conducted as an online survey. The purpose was
to measure, inter alia, the preferences of Sirius XM subscribers who
would choose to cancel their Sirius XM subscriptions at a given price.
Dhar CWDT ] 10; 5/8/17 Tr. 2728 (Dhar). The survey respondents
consisted of current paid Sirius XM subscribers who stated they have
the Sirius Select package, as well as current users of a free trial
subscription to Sirius XM (typically available with certain new or used
vehicle purchases). Dhar CWDT ] 10. Accordingly, the potential
population of survey respondents excluded those who understood
(correctly or incorrectly) that they subscribed to any other Sirius XM
package, such as ``XM Select,'' ``Mostly Music,'' or ``All Access.''
    Professor Dhar directed and conducted the survey between September
14 and September 22, 2016. To ensure the reliability and validity of
his online survey results, Professor Dhar designed and administered the
survey by applying principles of survey research applicable to online
surveys. In total, 2,602 respondents completed the survey. Dhar CWDT ]]
    \90\ An online survey obtains respondents from existing panels
of individuals who have expressed a willingness to participate.
Thus, the respondents are not randomly selected from a statistical
perspective and, accordingly, no margin of error or confidence
interval can be applied to the results. However, Professor Dhar used
what is known as a ``bootstrapping procedure,'' by which a sampling
of the survey respondents is itself randomly selected and thereby
created a confidence interval around each of the reported survey
results. Dhar CWDT ] 90.
    As noted above, the Dhar Survey consisted of two broad types of
questions: ``pricing questions'' and ``switching questions.'' The
pricing questions measured the preferences of Sirius XM subscribers who
would choose to cancel their subscriptions at given prices. The Dhar
Survey results demonstrated that 76% of Sirius XM subscribers would
cancel their subscriptions to Sirius XM at various prices between
$11.49 and $20.49 per month.
    The first of the ``switching questions'' asked the 76% who said
they would cancel their Sirius XM subscription (at any of the price
levels examined) to identify the type of music distribution channel to
which they would subscribe. The results showed that 28% of Sirius XM
subscribers said they would switch to a paid on-demand (i.e.,
interactive) music streaming service and 14% said they would switch to
a paid not-on-demand (i.e., noninteractive) music streaming service. 5/
8/17 Tr. 2761-62 (Dhar).\91\ In offering survey respondents
[[Page 65233]]
alternative subscription services, the Dhar survey specified a cost of
$9.99 per month for interactive services and $4.99 per month for
noninteractive services. Respondents were prompted to choose only ``a
new subscription . . . not . . . a music service that you currently
subscribe to.'' Dhar CWDT App. D at 69.
    \91\ The percentages of respondents selecting an alternative
service are stated as a portion of the entire population of the
Sirius XM respondents in the survey, rather than as a portion of
those who would choose to cancel their Sirius XM subscription. There
were 388 respondents who stated they would cancel their Sirius XM
subscription at various price points, which is the denominator
Professor Dhar used in his trial testimony to arrive at the 28% and
14% figures. Dhar CWDT ] 92. Professor Willig's percentages were
higher because he excluded 33 respondents who answered ``Don't Know/
Unsure'' to the switching question. Professor Willig thus determined
that 31% (not 28%) of the relevant universe would switch to a paid
on-demand service and 15% (not 14%) to a paid not-on-demand service.
Willig WDT, App. B at B-2. Sirius XM's witness, Professor Farrell,
did not dispute that the relevant denominator is the number of
respondents who would choose to cancel their Sirius XM subscription.
He used the same adjustment in his rebuttal opportunity cost
analysis, as explained elsewhere in this Determination.
    The Dhar Survey also explored preferences of respondents who
indicated they would not subscribe to a paid music service. Respondents
were permitted to choose more than one alternative music source from
among: (1) Purchased physical or digital tracks or albums, (2) free
music, (3) other, (4) none of the above, and (5) ``don't know/unsure.''
The follow-on question to those respondents who chose ``free music''
asked them to identify all of the free music sources they would choose.
Dhar CWDT at 59-60. The free music options listed included, inter alia,
(1) free not-on demand (including AM/FM radio over the internet), (2)
free (ad-supported) on-demand music services, (3) borrowed recordings,
(4) recordings the respondent already owns, and (5) AM/FM or AM/FM HD
broadcast radio. Id.
    Professor Willig used the results of this Dhar Survey to identify
the ``Alternative Mode Mix'' in his Opportunity Cost analysis, and
presented his results in the previous table.
2. Professor Hauser's Criticisms of the Dhar Survey
    Sirius XM called Professor John Hauser as a rebuttal expert witness
on survey design and methodology. In his written and oral testimony,
Professor Hauser leveled a number of criticisms at the Dhar Survey. In
particular, he criticized the switching questions and accompanying
response choices in the Dhar Survey. Professor Hauser testified that
the Dhar Survey was constructed in a manner that biased its results
because it: (1) Over-emphasized paid interactive and paid
noninteractive subscriptions in a biased and artificial manner; (2)
``buried'' the choice of free music, such as terrestrial radio \92\ as
an alternative to Sirius XM; and (3) failed to give respondents the
option of replacing a Sirius XM subscription with increased listening
to an existing (as opposed to a new) paid interactive or non-
interactive subscription. Rebuttal Expert Report of John Hauser, Trial
Ex. 11, ]] 66-69 (Hauser WRT).
    \92\ In this Determination, ``terrestrial radio'' refers to
free, over-the-air AM/FM and AM/FM HD radio, but not to AM/FM radio
streamed over the internet.
    As a preliminary matter, Sirius XM and Professor Hauser asserted
that Professor Dhar's tilt toward paid subscription services was the
consequence of his understanding that the relevant inquiry was ``if
[respondents] cancelled their [Sirius XM] subscription, what would they
subscribe to.'' 5/8/17 Tr. 2886-87 (Dhar). Accordingly, Sirius XM
asserted that the Dhar Survey was tainted from the inception because it
presented respondents only with definitions for three types of
services: Satellite radio, on-demand services, and non-on-demand
services. Dhar CWDT at 66 (Question 200), 69 (Question 200 and 210).
According to Professor Hauser, putting only these three types of
services in respondents' minds immediately prior to asking the
switching questions ``emphasize[d] both on-demand and not on-demand
services.'' 5/9/17 Tr. 3034-35 (Hauser). Professor Hauser contended
that the Dhar survey ``provided no cues to aid in the recall of other
music options (e.g., terrestrial radio) to which respondents could
switch.'' Hauser WRT ] 68. As Professor Hauser explained, ``[b]y aiding
in the recall of paid music services, but relying on unaided recall for
other music options (including free music options), Professor Dhar
biase[d] his results in favor of switching to paid music services.''
    According to Professor Hauser, this phrasing and choice selection
inevitably skewed responses in a way that did not reflect real-world
behavior. Specifically, he opined that the non-subscription option that
Professor Dhar provided as a potential response (``No, I would not
subscribe to a paid music service'') was not nearly specific enough to
capture a wide range of non-paid music options that respondents might
consider, including terrestrial radio. He further testified that, if
Professor Dhar had ``provided a list of non-paid alternatives or
existing paid subscriptions to which respondents might reasonably
switch, respondents may have been more likely to select non-paid
alternatives or existing paid subscriptions and less likely to select
new paid subscriptions.'' Hauser WRT ] 69; see also 5/9/17 Tr. 3034-35
(Hauser) (discussing ``availability heuristic'' and how ``when you show
people something, it becomes available in memory and they're much more
likely . . . to choose it'').
    Accordingly, Professor Hauser concluded that the Dhar Survey
wrongly buried other switching options such as listening to terrestrial
radio and omitted altogether listening to services to which the
respondents already paid to subscribe. Hauser WRT ]] 65-70.\93\ He
described the terrestrial radio option as buried because, for a Dhar
Survey respondent to select terrestrial radio as a choice, he or she
would first need to indicate an unwillingness to subscribe to a paid
music service in place of Sirius XM. Only then would the respondent be
shown the undifferentiated choice of listening to ``free music.'' Even
then, the respondent would need to indicate that he or she would
``listen to free music,'' and still would not be offered the explicit
choices of listening to terrestrial radio or to increase listening to a
streaming service to which he or she already subscribed or listened.
Only if the survey respondent selected the ``free music'' option would
he or she be presented--for the first time--with terrestrial radio as
an optional answer. See SXM PFF ] 390 (citing Dhar CWDT at 69; 5/8/17
Tr. 2916-20 (Dhar)).
    \93\ Confirming the importance of this criticism, Professor
Willig criticized the survey by Joseph Lenski, on behalf of Sirius
XM, for the same failure to offer the alternative of more intense
listening to an existing subscription service. Willig WRT ] 48. This
is an important failure, according to Professor Willig, because a
survey that does not offer respondents the option of listening more
to an existing subscription ``cannot provide the information needed
to assess the relevant effect, namely, the impact on creator
compensation.'' Willig WRT ] 46.
    In addition to critiquing the Dhar Survey's switching questions,
Professor Hauser created and implemented a ``Modified Dhar Survey.'' In
the Modified Dhar Survey, he essentially repeated Professor Dhar's
pricing questions, but attempted to reformulate the switching questions
in order to provide respondents with the immediate and explicit choices
of replacing Sirius XM with either terrestrial radio or increased
listening to streaming services to which they already subscribed.\94\
    \94\ Professor Hauser also criticized the ``pricing'' questions
in the Dhar Survey for listing from ``low to high'' the choice of
prices at which Sirius XM subscribers would not renew their
subscriptions, rather than also randomly reversing the order to
``high to low'' for 50% of the surveys. He also found fault with the
overall Dhar Survey because it only permitted participation by
individuals who thought they were subscribers to Sirius Select. Only
about 27% of all Sirius XM subscribers subscribe to the Sirius
Select package, and it was unclear whether subscribers knew the name
of the Sirius XM product to which they subscribed. Hauser WRT ] 124
& Figure 13; see also 5/8/17 Tr. 2858-2859 (Dhar). However,
Professor Hauser essentially utilized the same predicates to the
``switching'' questions in his Modified Dhar Survey.
    In the Modified Dhar Survey, Professor Hauser first moved the
option of listening to terrestrial radio forward in the survey. 5/9/17
Tr. 3049-50 (Hauser). He also added additional alternative responses to
the options of
[[Page 65234]]
choosing ``new CDS and/or music downloads,'' the respondent's
``existing collection of CD and/or music downloads,'' and ``other free
music option(s) (e.g., free, ad-supported Pandora or Spotify, AM/FM
radio over the internet, and YouTube.)'' Hauser WRT ] 79; id. App. I at
10. Professor Hauser then added yet more response options to allow
respondents to choose explicitly to switch to existing music service
subscriptions. Hauser WRT ]] 79, 88, App. I at 10; 5/9/17 Tr. 3061
    When Professor Hauser administered his Modified Dhar Survey to a
group of on-line survey respondents, he obtained results significantly
different from those Professor Dhar reported. Specifically, Professor
Hauser's modifications led to a material drop in the percentage of
Sirius Select respondents who indicated that they would replace their
Sirius XM subscription with a new paid on-demand service: From 28% of
respondents in Professor Dhar's survey (31% as measured by Professor
Willig) to only 15% in the Modified Dhar Survey. See Hauser WRT Table 1
& ]] 101, 104; 5/9/17 Tr. 3056 (Hauser).
    In addition, when Professor Hauser provided respondents the
terrestrial radio option early and explicitly, approximately 78% of
Sirius Select respondents indicated they would switch to terrestrial
radio. Hauser WRT Figure 11-A; 5/9/17 Tr. 3059 (Hauser). This result
was in stark contrast to the results from the original Dhar Survey,
which indicated that only 29% of the total Sirius Select respondents
would replace Sirius XM with terrestrial radio. Hauser WRT Fig 10-B;
Dhar CWDT ] 52, Table 1. Sirius XM notes that Professor Dhar himself
was unsurprised by these results. He testified at the hearing that he
anticipated that, if he had explicitly offered respondents the choice
of free music or AM/FM radio from the outset, he would have expected
the number of people who chose those options to be higher. 5/8/17 Tr.
2920-22 (Dhar).
    The Judges find the original Dhar Survey to be seriously flawed.
The Dhar Survey failed to make prominent to respondents the option of
selecting terrestrial radio as an alternative source of music if they
made a price-based decision not to renew their Sirius XM subscriptions.
Equally problematic are the absences from the Dhar Survey of any choice
for a respondent to state that he or she would either increase
listening to a streaming service to which he or she already subscribed,
or to increase listening to downloads or CDs that the respondent
already owned.
    Professor Dhar testified that the purpose of the study, as
explained to him by the SoundExchange economic expert witnesses, was to
estimate the number of cancelling Sirius XM subscribers who would then
subscribe to an on-demand or a ``not-on-demand'' music streaming
service. He explained that he did not make alternative free choices
more prominent and explicit because the ``marketplace context'' that
``the [SoundExchange] economists . . . were really interested in'' was
the subscription streaming context. Tr. 5/18/17 2752 (Dhar); see also
id. at 2751, 2752, 2754, 2810, 2889, 2921 (multiple instances of
justifying the original formulation by reference to ``marketplace
context''). The Judges find this testimony to be credible, and it
suggests that Professor Dhar was not engaged to prepare a study that
would give equal prominence to the potential alternative that Sirius XM
subscribers might choose free alternatives. Thus, the Judges agree with
Sirius XM that, by his own admission, Professor Dhar did not
comprehensively measure what Sirius XM subscribers would do if they
stopped using Sirius XM. By focusing myopically on what he
(misleadingly) was told was the ``marketplace context'' of subscription
streaming, the Dhar Survey essentially assumed its conclusion. This is
a crucial defect, given that the use for which the Dhar Survey was
intended was to weight ``opportunity costs'' in a manner that expressly
included at least one free alternative, i.e., the substitution of
terrestrial radio. It is disingenuous for SoundExchange to argue,
through Professor Dhar, that its intention was not to identify the
percent of Sirius XM listeners who would choose terrestrial radio (or
any other free alternative), given that the Dhar Survey actually did
solicit such responses, albeit in a fashion that reduced the frequency
of that response, particularly in contrast with the results of the
Modified Dhar Survey.
    The switching questions in the original Dhar survey are problematic
for additional reasons. First, the power of a ``free'' alternative is
well-understood. See C. Anderson, Free: The Future of a Radical Price
4, 2 (2009) (``Free is both a familiar concept and a deeply mysterious
one. . . . `Free-to-air' radio . . . created the mass market.''); D.
Ariely, Predictably Irrational at 51-52 (2009) (when offered a Lindt
Truffle for 26 cents and a Hershey's Kiss for 1 cent, 40% opted for
each choice; when price of each decreased by one cent (making the Kiss
free), 90% opted for free chocolate).
    Second, as the Lenski Survey \95\ made clear, 62% of Sirius XM
subscribers had listened primarily to terrestrial radio before
switching to Sirius XM. Written Direct Testimony of Joe Lenski, Trial
Ex. 7, at 8 (Lenski WDT). Notwithstanding any problems in the Lenski
Survey, it is not disputed that a substantial portion of the Sirius XM
listener base migrated from listening to terrestrial radio. Sirius XM
also presented testimony that the ``vast majority'' of Sirius XM
listening, occurs in the automobile, and most listeners in automobiles
still utilize terrestrial radio as their primary music source. See
Written Direct Testimony of James Meyer, Trial Ex. 1, ] 21 (Meyer WDT).
Simply put, the marketplace is suffused with evidence of the
substantial past and present use of terrestrial radio.
    \95\ Sirius XM commissioned a listener survey to determine the
sources of Sirius XM listeners and the destinations to which they
would migrate if Sirius XM were not available. The Lenski survey is
discussed infra, section VII.D.
    These data underscore the Judges' finding that the Dhar Survey's
burying of the terrestrial radio alternative fails to depict the
marketplace reality. Indeed, it is surprising that Professor Dhar (and
anyone who directed him regarding the purpose of his survey) would
repeatedly rely on the ``marketplace context'' rationale to justify the
construction of the switching questions in the Dhar Survey and the
results those questions elicited. The failure of the Dhar Survey
explicitly to offer to a respondent, in any set of responses to any
questions, the choice of increased listening to a streaming service to
which the respondent has an existing subscription is especially
problematic. From an economic perspective increased listening by a
respondent to a service to which a respondent already subscribes is
marginally ``free,'' because there is no increase in cost to access an
existing monthly ``all-you-can-eat'' subscription to a music service in
the car. More egregiously, the Dhar Survey explicitly instructs
respondents before presenting the first switching question:
    Keeping in mind all other music services you subscribe to would
you or would you not subscribe to a paid music service in place of
Sirius? This would only include a new subscription, and would not
include a music service that you currently subscribe to.
Dhar CWDT, at 69, App. D. Thus, not only did the Dhar Survey fail to
provide respondents with an explicit choice to utilize a music
streaming service to which they had an existing subscription, it
explicitly primed them to think specifically of such services
[[Page 65235]]
and then to consciously NOT select that service as an alternative.
    The Judges' foregoing critique should not be understood as a
finding that Professor Hauser's Modified Dhar Survey is without
defects. Professor Hauser altered the composition of the survey
population by excluding respondents who had recently taken a music
survey (in an attempt, he claimed, to eliminate respondents who
participated in the original Dhar Survey). Hauser WRT ] 96. Professor
Hauser's different population renders the Modified Dhar Survey less
than perfectly analogous to the original Dhar Survey. The record does
not reflect that this alteration of the survey population biased the
results; nor is there any evidence that the change was in any way
material. Consequently, the Judges do not find this defect to render
the Modified Dhar Survey unreliable.
    In addition, 24 participants in the Modified Dhar survey said they
would listen to an on-demand service to which they already subscribe,
even though they had answered the ``pricing question'' by stating that
they were not then subscribing to such a service.\96\ See 5/8/17 Tr.
2822 (Dhar); Trial Ex. 293, at 1. In his defense, Professor Hauser
explained that he used Professor Dhar's non-switching (i.e., pricing)
questions verbatim in order to tease out any differences arising from
the switching questions, and that the non-switching questions listed
only Spotify and Apple Music as interactive services, and Pandora, then
a noninteractive service. See Dhar CWDT, App. D at 61, 63. Professor
Hauser testified that, in his opinion, the anomaly could be explained
by the fact that respondents who used other interactive streaming
services, such as those offered by Amazon or Google, might have thought
the ``pricing'' question about existing subscriptions to interactive
services was limited to Apple Music and Spotify. Thus the respondents
indicated they did not subscribe to either of them, but could respond
affirmatively that they would listen to another On-Demand service to
which they subscribed. 5/9/17 Tr. 3104-05 (Hauser). While that
explanation is plausible, it is unsupported by record evidence.\97\ As
Professor Dhar demonstrated, this anomaly materially affected the
survey results: If one were to re-categorize those 24 responses as
having stated that they would subscribe to a new on-demand service, the
percentage of respondents who would switch to a new interactive service
would increase from 15% to 19%. 5/8/17 Tr. 2822-23 (Dhar).\98\ The
Judges adopt Professor Dhar's re-categorization to correct this anomaly
in the Modified Dhar Survey.
    \96\ Professor Dhar identified a potential similar problem with
regard to respondents who indicated they would switch to an existing
noninteractive service, but had previously indicated they did not
subscribe to such a service. However, he did not make any
adjustments to correct this problem.
    \97\ Professor Dhar posited a different explanation for this
anomaly. See 5/8/17 Tr. 2814-16 (Dhar). In light of Professor
Hauser's failure adequately to explain the anomaly, the Judges need
not consider Professor Dhar's alternative explanation.
    \98\ Professor Hauser also conceded that he checked all the
numbers in Trial Ex. 293 (in which Professor Dhar tabulated
inconsistent answers in Professor Hauser's survey and listed the
sources for the data), and Professor Hauser found them to be
correct. 5/9/17 Tr. 3143-44 (Hauser).
    Finally, Professor Hauser did not identify confidence intervals
around his survey results which could have been estimated by use of the
``bootstrap'' method. Such a subsequent sub-sampling and calculation
would have bolstered Professor Hauser's weighting based on the Modified
Dhar Survey. Cf. Dhar CWDT ] 90. There is no evidentiary requirement
that an on-line survey that, by its non-random nature, fails to produce
a statistical random sample must be subjected to a bootstrapping
approach to carry evidentiary weight. Indeed, the requirements for
precise statistical reliability that exist in the academic world should
not constrain Judges from accepting and relying on evidence that is
otherwise probative when considered in the context of the entire
evidentiary record. See, e.g., Matrixx Initiatives, Inc. v. Siracusano,
563 U.S. 27, 44 (2011) (demonstration of ``statistical significance''
not required to demonstrate reliable causal relationship when
relationship demonstrated through ``content and context'' evidence).
Moreover, the standard-setting organization for survey work, the
American Association for Public Opinion Research (AAPOR), upon which
Professor Dhar relied to use a bootstrapping approach, is by its
express language a ``nonbinding document,'' and thus does not require
the use of the bootstrapping technique through which statistical
significance could be ascertained. See Dhar WDT, Ex. G, at 1(AAPOR
Guidance on Reporting Precision for Nonprobability Samples).
    On balance, the Judges find the Modified Dhar Survey (corrected by
Professor Dhar, as noted supra) to be more probative than the original
Dhar Survey. Once corrected to account for the anomalous responses
described above, the potential deficiencies in Professor Hauser's
Modified Dhar Survey appear to the Judges to be of relatively marginal
significance when compared with the defects in the original Dhar
survey. The Modified Dhar Survey came closer to the core of the issue
at hand: Distinguishing among the alternative distribution channels to
which erstwhile Sirius XM subscribers would migrate if the Sirius
subscription price became so high as to dissuade renewal.
3. Re-Weighting Opportunity Cost Calculation With Modified Dhar Survey
    Professor Farrell took Professor Hauser's data from the Modified
Dhar Survey and plugged them into Professor Willig's opportunity cost
calculations. In so doing, Professor Farrell persuasively demonstrated
that Professor Willig's opportunity cost fell significantly below the
$2.55 per subscriber per month level, and thus below the [REDACTED]%
royalty rate Professor Willig found to be implied by that $2.55
figure.\99\ See 4/24/17 Tr. 636-37 (Farrell); Farrell WRT ]] 62-66.
    \99\ To be clear, Professor Farrell did not agree with the
opportunity cost values that Professor Willig calculated, because
Professor Farrell described them as monopoly-based opportunity costs
(as noted, supra, Professor Willig called them walk-away opportunity
costs). However, Professor Farrell's re-working of Professor
Willig's opportunity cost analysis utilizes, arguendo, Professor
Willig's ``walk-away'' opportunity costs.
    Professor Farrell noted that the Modified Dhar Survey had 498
respondents who self-identified as paid Sirius XM subscribers. Among
these 498 respondents, 13 answered the survey's pricing questions by
stating that they would continue to subscribe to Sirius XM at any
price. Therefore, like Professor Willig, Professor Farrell excluded
these 13 from the pool used to weight the opportunity cost calculation.
Another 22 respondents to the Modified Dhar Survey answered ``Don't
know/unsure'' to whether they would cancel at various hypothetical
Sirius XM subscription prices. Again, consistent with Professor
Willig's treatment of respondents who answered in this manner,
Professor Farrell excluded these 22 respondents from the pool used to
weight the opportunity cost calculation. The remaining 463 respondents
were then asked what source of music they would switch to in lieu of
listening to Sirius XM. Farrell WRT, App. F at F-1.
    Professor Farrell presented in tabular form (1) the options from
which the 463 respondents in the Modified Dhar Survey could choose; (2)
the counts of respondents who chose each option; (3) the ratio by which
the respondents would divert to each option; and (4) the creator
compensation for each option. His calculations are detailed on the
following table.
[[Page 65236]]
                       Modified Dhar Survey Responses--Diversion and Creator Compensation
                                                                                     Diversion     Creator comp/
                        Respondent choice                              Count         ratio (%)    subscriber/mo.
Alternate paid interactive service (e.g., Spotify/Apple Music)..              69          15.10%     $[REDACTED]
Existing paid interactive service (e.g., Spotify/Apple Music)...              57           12.50            0.00
Alternate paid non-interactive service (e.g., Pandora One etc.).              45            9.90      [REDACTED]
Existing paid non-interactive service (e.g., Pandora One etc.)..              30            6.60      [REDACTED]
Alternate CDs or music downloads................................              97           21.30      [REDACTED]
Existing CDs or music collection................................             240           52.60            0.00
AM/FM radio.....................................................             359           78.70            0.00
Other free options..............................................             184  ..............  ..............
    Free, ad-supported non-interactive service..................             138           30.30      [REDACTED]
    Free, ad-supported interactive service......................              92           20.20      [REDACTED]
    Free, ad-supported music video sites........................              70           15.40      [REDACTED]
    Music channel included in existing cable/SAT TV subscription              59           12.90            0.00
    Peer-to-peer file sharing or free download sites............              17            3.70            0.00
    Borrow CDs, vinyl or tapes from friends or a library........              52           11.40            0.00
    Other free services.........................................              13            2.90            0.00
    Don't know/unsure...........................................               9            2.00      [REDACTED]
Other...........................................................              15            3.30            0.00
None............................................................               8            1.80            0.00
Don't know/unsure...............................................               7
    Total.......................................................             463
Farrell WRT, App. F at F-2 (Table 3).\100\
    \100\ Professor Farrell used the creator compensation figures
from Table 2 in the Willig WDT whenever available. However,
Professor Willig had not covered in his Table 2: Peer-to-peer file
sharing or free download sites, borrowed CDs, vinyl or tapes from
friends or a library, other free services, don't know/unsure
regarding free options, and ``other.'' Professor Farrell discounted
this point, noting that (with the exception of ``Don't know/unsure''
under free options), these other services not in Professor Willig's
Table 2 have zero creator compensation value.
    Professor Farrell used the above data to calculate the opportunity
cost (i.e., the walk-away opportunity cost). More particularly,
Professor Farrell engaged in a nine-step calculation to compute
opportunity costs.
    Professor Farrell first eliminated the seven respondents who chose
``Don't know/unsure,'' noting that this was equivalent to assuming that
these seven would divert to the different options in the same
proportions as the remaining 456 respondents.\101\ He calculated the
diversion ratio for each option as the number of respondents who chose
that option divided by 456. Professor Farrell then used the same values
for ``creator compensation per subscriber per month'' as set forth in
Table 2 of Professor Willig's WDT, including Professor Willig's
adjustments for intensity of use.\102\ See Farrell WRT, App. F at F-2.
    \101\ Professor Willig adopted the same approach when treating
``Don't know/unsure.'' Willig WDT at B-3.
    \102\ Professor Farrell did not opine on the appropriateness of
Professor Willig's adjustment for intensity of use. Farrell WRT at
    Professor Farrell noted that in both the Dhar Survey and the
Modified Dhar Survey, many respondents chose multiple nonsubscription
options. Professor Farrell generally matched Professor Willig's
approach, assuming equal intensity of use for the multiple options
chosen by a given respondent.\103\ Professor Farrell calculated the
overall intensity of use for a given option across all respondents who
selected that option as equal to the average intensity of use for that
option across all respondents who selected that option. See Farrell
WRT, App. F at F-3. Applying this foregoing approach for each option,
Professor Farrell calculated an ``intensity-adjusted creator
compensation.'' \104\ Professor Farrell's calculation generated an
opportunity cost of $1.44 per subscriber per month.\105\ (Professor
Farrell also applied the diversion data from the Lenski Survey
(discussed later in this Determination) and arrived at a similar
opportunity cost estimate of $1.43. Farrell WRT ] 66.\106\)
    \103\ See Willig WDT at B-3 and B-4. Unlike Professor Willig,
Professor Farrell assumed equal intensity of use percentages
whenever individuals selected combined free options and paid
services in in their multiple option choices, whereas Professor
Willig assigned 50% to alternate CD or music downloads, and 25% to
each of the free options. According to Professor Farrell, this
difference did not have a large impact on the size of the
opportunity cost.
    \104\ Professor Farrell assumed that creator compensation for
the option ``Other'' to be zero. See Farrell WRT, App. F, at F-3.
Professor Willig appeared to make the same assumption. See Willig
WDT at B-8.
    \105\ Professor Farrell recognized that the value (unweighted)
of the monthly ``unit creator compensation $ per subscriber'' could
decrease if a lower intensity of use (fewer plays) among those who
selected multiple options also reduced the overall revenue base
under a per play royalty structure as calculated under Professor
Willig's assumptions. The $1.44 opportunity cost set forth in the
accompanying text assumes (in favor of the licensors) that creator
compensation for paid services and paid non-interactive services
does not decrease for decreased intensity of use. Professor Farrell
opined that--if noninteractive services alone would pay a lower
royalty (because their royalty payments are based on a per-play/
intensity-based formula), but interactive service royalties would
not be similarly reduced because of a reduction in intensity of use
(i.e., if they more likely to pay royalties on a per-subscriber or
percent-of-revenue basis)--his opportunity cost calculation would
generate a lower opportunity cost of $1.35. See Farrell WRT, App. F,
at F-3. However, Professor Farrell does not provide in his written
or oral testimony a basis to make this ``creator contribution''
adjustment based on relative changes in intensity, and the Judges
therefore do not credit his argument that--under his reworking of
Professor Willig's opportunity cost calculations--the opportunity
cost can be reduced from $1.44 to $1.35.
    \106\ As explained elsewhere in this Determination, the Lenski
Survey did not provide pricing information to respondents, making it
a less valuable tool for estimating opportunity cost. Accordingly,
the Judges do not rely on Professor Farrell's $1.43 opportunity cost
calculation that is based on the Lenski Survey as an independent
basis to calculate opportunity cost, but rather consider it as
confirmation that Professor Willig's opportunity cost calculation
(based on the original Dhar Survey) was too high.
    Professor Farrell used the same methodology for survey respondents
who were Sirius XM free trial subscribers. See id., App F at F-3-F-4.
However, the Judges do not find the trial subscriber population to be
an appropriate universe from which to calculate opportunity cost
because trial subscribers have not demonstrated a positive WTP.
    SoundExchange failed to raise persuasive objections to Professor
Farrell's opportunity cost calculation
[[Page 65237]]
based on the Modified Dhar Survey. In its PFF, SoundExchange asserts
    Professor Farrell also revised Professor Willig's opportunity
cost calculations to show what the industry-wide opportunity cost
would be if one used diversion ratios from the Hauser and Lenski
surveys. Trial Ex. 10 at 17-21 (Farrell WRT); 4/24/17 Tr. 636:2-7
(Farrell). It is not clear what the point of this exercise was --
neither the Lenksi nor the Hauser survey can reliably be used to
calculate opportunity costs, as Sirius XM's own experts admit.
SEPFF561. Likewise, in its RPFF, SoundExchange does not attack any
aspect of Professor Farrell's application of the Modified Dhar Survey,
but rather renews its attack on the underlying work of Professor
    Professor Farrell's recasting of Professor Willig's calculations
using the Hauser survey is invalid since the Hauser survey entirely
misstated the switching question, see SE FOF ]]614-22, and since
Professor Hauser conceded unequivocally that the economists should
not rely on his survey, see SE FOF ]619 (citing Hauser testimony).
SERPFF, Response to ] 408 at 266.
    SoundExchange's objection to the use of Professor Farrell's
approach is dependent on its antecedent criticism of Professor Hauser's
analysis. As discussed, however, the Judges have found the Modified
Dhar Survey results to be more accurate and probative than the results
produced by the Dhar Survey. Accordingly, SoundExchange's criticism is
without merit.\107\
    \107\ In its RPFF, SoundExchange added to its argument:
``Professor Hauser conceded unequivocally that the economists should
not rely on his survey.'' However, Professor Hauser made this
comment because he also objected to other aspects of the Dhar
Survey, particularly with regard to its ``pricing'' questions, that
he nonetheless retained in the Modified Dhar Survey. Thus, he argued
that these antecedent deficiencies in the Modified Dhar Survey
precluded reliance on the results derived from his modified
``switching'' questions in the Modified Dhar Survey. The Judges
disagree with Professor Hauser's characterization of the
deficiencies he identified in the Dhar Survey that were unrelated to
the ``switching'' questions. Thus, the Judges can and do give
considerable weight to the Modified Dhar Survey, which they find
sufficiently credible and probative.
    Using Professor Dhar's corrected calculation indicating that 19% of
Sirius XM subscribers would switch to a new interactive subscription
service, the per Sirius XM subscriber opportunity cost increases from
$1.44 to $[REDACTED].\108\ Given Sirius XM's ARPU of $[REDACTED], the
percent-of-revenue royalty rate derived from the $[REDACTED] per
subscriber per month opportunity cost is 15.5%.\109\
    \108\ 15.1% of the ``creator contribution'' value of $[REDACTED]
equals $[REDACTED]. 19% of $[REDACTED] equals $[REDACTED]. The
difference is $[REDACTED] ($[REDACTED] - $[REDACTED] = $[REDACTED]).
When that $[REDACTED] is added to the $1.44 calculated by Professor
Farrell, the full opportunity cost based on the Modified Dhar Survey
(as adjusted for the foregoing anomaly in the Hauser survey answers)
    \109\ Professor Willig attempted to corroborate Professor Dhar's
diversion ratios with a regression analysis seeking to measure
relative cross-elasticities. The Judges do not apply that analysis
because: (1) The Dhar Survey results are without value (as discussed
previously) and therefore cannot be ``corroborated''; and (2) there
were significant disputes regarding the accuracy of Professor
Willig's regression that rendered the value of that analysis
inconclusive. See Shapiro WRT at 27-37.
C. Opportunity Cost Model and Effective Competition
    In Web IV, the Judges reconfirmed that a statutory willing-buyer,
willing-seller royalty rate is one that would emerge in a market that
is effectively competitive. See Web IV, 81 FR at 26334. Both
SoundExchange and Sirius XM acknowledged that the rate set in this
proceeding must reflect a market with such effective competition. 4/26/
17 Tr. 1103 (Orszag) (agreeing that ``the rates to be set here by the
Judges . . . must reflect the workings of effective competition'');
Shapiro CWDT at 21 (``My approach here is consistent with the one taken
by the Judges in Web IV . . . . I use the terms `workably competitive'
and `effectively competitive interchangeably.''); 4/20/17 Tr. 366
(Shapiro) (``prices . . . at a complementary oligopoly level [are] not
[at] a workably competitive level.'').
    The Judges defined an effectively competitive market In Web IV as
one that ``mitigate[s] the effect of complementary oligopoly on the
prices paid by . . . services . . . .'' Web IV, 81 FR at 26366. To
obtain the rate that is effectively competitive, the Judges considered
the services' ability to ``steer'' listeners as a sufficient
counterweight to the Majors' complementary oligopoly power. Id. at
26343. The Judges also noted in Web IV that SoundExchange had correctly
described the concept of effective competition as ``fuzzy'' and that
``no `bright line' can be drawn between effectively competitive and
noncompetitive rates.'' Id. As the Judges further noted, the
implication of this ``fuzziness'' was not that the principle of
effective competition should be discarded, but rather that this ``fuzzy
line'' needs to be drawn on a case-by-case basis, from the evidence and
testimony adduced at the hearing.'' Id. (emphasis added).
    In the present proceeding, the parties' economists proposed that
the Judges once again adjust for improper market power by applying a
steering adjustment. SoundExchange proposed that the Judges select from
one of three possible adjustments: (1) The 12% steering adjustment
revealed by the specific steering evidence in Web IV; (2) a [REDACTED]%
steering adjustment allegedly implied by the provisions of ``Mid-tier''
agreements \110\ between record companies and streaming services, see
4/25/17 Tr. at 1053 (Orszag); or (3) a [REDACTED]% steering adjustment
implied by rates in direct licenses between Sirius XM and certain
Indies. See Written Rebuttal Testimony of Jonathan Orszag, Trial Ex.
43, ] 70 (Orszag WRT). However, in this proceeding, these proposed
adjustments are unacceptable.
    \110\ ``Mid-tier'' services means internet streaming services
that offer only limited interactivity, and thus offer a tier of
service between a noninteractive service and a fully interactive
service. The limited interactive functionality of the mid-tier
service offerings includes limited caching and playbacks.
    The Judges cannot simply import the 12% steering adjustment from
Web IV into the satellite market; that 12% figure was derived from
highly specific evidence presented in Web IV. There is not an adequate
basis in the present record to support a finding that the
noninteractive market from which that steering adjustment arose is
sufficiently similar to the satellite radio market to render reasonable
an importation of the 12% steering adjustment here. In particular, the
record shows that Sirius XM does not steer in the satellite market
despite the ability of its human programmers (as opposed to algorithmic
programmers) to do so in order to potentially reduce rates in exchange
for additional plays, which is the essence of steering. See infra,
section VII.C.
    For two reasons, the Judges cannot accept the proffered [REDACTED]%
steering adjustment that SoundExchange divined from the Mid-tier
agreements. First, there is no evidence in the record to indicate
whether that proposed adjustment may reflect a premium that a Major may
impose not to prohibit a licensee from steering away from the licensor,
rather than a discount offered to encourage a licensee to steer toward
the licensor. Further, the rate differentials in those agreements on
which SoundExchange's economic expert, Mr. Orszag, relied appear to be
the product of many other differences in those agreements in addition
to the steering/no-steering distinction, as Mr. Orszag candidly
acknowledged. 4/26/17 Tr. 1155-56 (Orszag); see also SXM RPFF ]] 85-86
(and record citations therein).
    Finally, the Judges reject any steering adjustment based on the
direct licenses between Sirius XM and various Indies. As explained in
the discussion of
[[Page 65238]]
Professor Shapiro's reliance on these direct licenses as benchmarks,
the record is clear that multiple other provisions of those direct
licenses provided substantial consideration to the Indie licensors to
justify their willingness to enter into those deals. Moreover, the
Indie direct licenses contain neither legal guarantees nor economic
incentives that would compel or motivate steering by Sirius XM in favor
of direct licensors.
    Accordingly, the Judges must review the record in this proceeding
to identify a means to establish rates that are consistent with
effective competition. The Judges accept certain principles regarding
the nature of effective competition. ``Between the extremes of a market
with `metaphysically perfect competition' and a monopoly (or collusive
oligopoly) market devoid of competition there exists `[in] the real
world . . . a mindboggling array of different markets' . . . all of
which possess varying characteristics of a `competitive marketplace.'
'' Web IV at 26333 (citing Web III Remand, 79 FR at 23114, n.37).\111\
Economists have long understood that the ``fuzzy'' nature of the
concept of effective competition is inescapable, yet the concept must
be applied, lest pragmatic economic analysis be straightjacketed by
rigid textbook models such as perfect competition and simple
monopoly.\112\ The D.C. Circuit has recognized this conceptual
fuzziness, acknowledging in the rate-setting context the need for
pragmatic market analysis, establishing rates intermediate between the
pedagogical poles of perfect competition and pure monopoly. See
Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Board, 574 F.3d
748, 757 (D.C. Cir. 2009) (IBS) (statutory provisions ``do[ ] not
require that the market assumed by the Judges achieve metaphysical
perfection in competitiveness'' (emphasis added)).\113\
    \111\ See J. M. Clark, Toward a Concept of Effective
Competition, 30 a.m. Econ. Rev. 241, 243 (1940) (``The specific
character of competition in any given case depends on a surprisingly
large number of conditions . . . .'').
    \112\ See A. Kahn, Antitrust Policy, 67 Harv. L. Rev., 28, 35,
(1953) (``[T]here exists no generally accepted economic yardstick
appropriate for incorporation into law with which objectively to
measure monopoly power or determine what degree is compatible with
workable competition.''); J. Markham, An Alternative Approach to the
Concept of Workable Competition 349, 361 (1950) (The concepts of
``market performance and workable competition are essentially
pragmatic''); G. Stocking, Economic Change and the Sherman Act: Some
Reflections on ``Workable Competition,'' 44 Va. L. Rev. 537, 553
(1958) (``the economists' concept of workable competition . . . is
vague . . . .).
    \113\ The quoted language refers to section 114(f)(2)(B), which
governs the compulsory license for eligible nonsubscription services
and new subscription services. Under the license at issue in the
present case, the D.C. Circuit has not required the Judges to adopt
market rates. However, to the extent that the Judges choose to use
market rates as an input for the development of rates under section
801(b)(1) (as they do here), the quoted language from IBS is
D. Professor Willig's ``Fork in the Road'' Approach and Sirius XM's Own
Market Power
    The Judges find no basis to lock themselves into a Hobson's choice
by which they must either adopt an inapplicable steering adjustment as
a proxy for an adjustment to reflect effective competition, or accept a
rate that is higher than an effectively competitive rate.\114\
``Steering'' is not the only way the inefficient market power of
complementary oligopoly can be offset or mitigated in order to
establish an effectively competitive rate.
    \114\ A third possibility would be to utilize an otherwise
appropriate market benchmark rate that is effectively competitive.
However, the Judges cannot identify such a rate in the present
    In this regard, in his hearing testimony, Professor Willig
explained how and why his opportunity cost approach would result in a
rate that is effectively competitive. Professor Willig described a
``fork in the road'' for the Judges as follows:
    [T]he fork in the road is whether, in considering the comparison
between the opportunity cost and the royalty rate in the target
market, should you take the other markets as they are or should you
bring in hypotheticals and make adjustments to the opportunity cost
based on . . . changes in the other markets? And that to me is a
very consequential fork in the road . . . .
5/2/17 Tr. 2040 (Willig); see id. at 2047, 2153. Professor Willig
opined that attempts to adjust one rate downward, such as the
interactive rate, to account for the complementary oligopoly effect,
would be incomplete, because other distribution modes, such as
terrestrial radio, do not generate sound recording royalties and thus
do not create a positive opportunity cost. Thus, Professor Willig
described as a ``morass'' any attempt to take the ``fork-in-the-road''
by which the Judges attempt to adjust every rate that fails to reflect
market forces. See id. at 2057, 2048. Rather, he recommends that the
Judges ``should take the fork in the road that says take those markets
as they are because that's what drives honest-to-goodness opportunity
cost.'' Id. at 2057.
    This is precisely what the Judges accomplish by taking the
opportunity cost analysis that results in the 15.5% rate.\115\ The
Judges further note that Sirius XM did not challenge Professor Willig's
``fork in the road'' concept, either in cross-examination or in its
post-hearing proposed findings and replies to proposed findings.
    \115\ The Judges' rate is less than the rate proposed by
Professor Willig, because the Judges give less probative weight to
the Dhar Survey, not because they disagree with Professor Willig's
opportunity cost approach.
    Accordingly, the Judges find that the 15.5% opportunity-cost
derived rate: (1) Reflects the offsetting market forces of higher
complementary oligopoly rates and lower (zero) opportunity costs
attributable to listeners who otherwise would migrate to terrestrial
radio; and (2) is consistent with Professor Willig's opinion regarding
the need for a consistent treatment of market forces, as described in
his ``fork in the road'' analysis.
    This ``fork in the road'' approach is also consistent with a
recognition of the countervailing downstream market power that Sirius
XM, the sole SDARS licensee, possesses as a monopolist in that
downstream market, narrowly defined as the market for the sale of
subscriptions to satellite radio. To be sure, this narrow definition of
the market ignores various other forms of music distribution, such as
terrestrial radio and all other alternative distribution channels
identified in the survey analyses. However, as that survey evidence
makes clear, even terrestrial radio, which is free to the listener,
cannot attract sufficient listeners to deprive Sirius XM of the
substantial profits it realizes from its unique position as the only
supplier of satellite radio in the market. Further, Sirius XM is priced
higher than interactive (and noninteractive) streaming services. Yet,
despite their differentiated features, those services to date have been
unable to convince enough Sirius XM subscribers to convert to a new
paid subscription service to reduce the revenues and profits realized
by Sirius XM. Clearly, Sirius XM's uniquely differentiated service has
struck a chord with music listeners--particularly those who listen to
Sirius XM in the car. This point was made clearly by Professor Shapiro,
who testified:
    Sirius XM spends substantial sums of money on its infrastructure
and satellites. In doing so, it creates a unique differentiated
service. That is quite valuable to consumers. That's why they are
willing to pay for the service and, of course, most of the listening
is in the car.
5/4/17 Tr. 2550 (Shapiro).\116\
    \116\ Sirius XM is both a monopolist, in the sale of satellite
radio subscriptions, and a competitor among the various distribution
channels more broadly. This is not an inconsistency. Since 1933,
economists have recognized that a firm may be a ``monopolistic
competitor,'' with the power of a monopoly (as reflected in the
downward sloping demand curve it faces) but the restraints of
competition (making that demand curve relatively elastic compared to
the demand curve for the product of a full-fledged monopolist). See
E. Chamberlin, The Theory of Monopolistic Competition (1933).
[[Page 65239]]
    Correspondingly, Sirius XM bears all the hallmarks of a ``natural
monopoly.'' A natural monopoly develops when ``it is cheaper for [an]
entrepreneur to produce q units than it is to have those units produced
by two [or more] smaller firms . . . .'' A. Schotter, Microeconomics: A
Modern Approach 416 (2009); see also W. Baumol and R. Willig, Fixed
Costs, Sunk Costs, Entry Barriers, and Sustainability of Monopoly, 96
Q.J.Econ. 405, 409, 418 (1981) (``[A]n industry has been called a
natural monopoly if . . . industry outputs can be produced more cheaply
by a single firm than by any combination of several firms. These per
unit costs arise from relatively large sunk costs (compared to marginal
costs) and those sunk costs act as ``barriers to entry [that] . . .
impede the establishment of new firms [because] [t]he need to sink
money into a new enterprise, whether into physical capital,
advertising, or anything else imposes a difference between the
incremental cost and the incremental risk that are faced by an entrant
and an incumbent''); H. Varian, Intermediate Economics: A Modern
Approach at 453 (``When there are large fixed costs and small marginal
costs, [that] situation is referred to as a natural monopoly.''). As a
natural monopolist in the satellite radio market, Sirius XM can, and
does, realize substantial profits, as demonstrated in fine detail by
Professor Lys. The history of Sirius XM bears out this point. When
there were only two satellite firms--Sirius and XM--both were on the
brink of bankruptcy. See SDARS II, 78 FR at 23069. After they merged,
they were transformed from two pumpkins into a single coach, as it
were, realizing profits across many financial measures. See Lys WDT,
    In the hypothetical market the Judges construct in this proceeding,
they identify significant power on both the licensor side and the
licensee side. On the licensor side, that power is reflected in the
opportunity cost analysis--the ``creator contribution'' values
identified by Professor Willig. Those values embody the complementary
oligopoly features that flow from the ``must have'' nature of the
Majors' repertoires. On the licensee side, there are profits that flow
from two sources: (1) The highly differentiated nature of Sirius XM's
offerings that permits it to attract listeners who otherwise would
listen to free terrestrial radio; and (2) the entrepreneurial ability
by which Sirius XM has harnessed the natural monopoly structure of
satellite radio delivery to its financial benefit.
    The Judges find from this record that the hypothetical upstream
market negotiations between such economically powerful entities would
resemble a bilateral monopoly. Thus, as Professor Willig testified, the
record companies would be expected to recover their opportunity costs
(inclusive of any complementary oligopoly profits). Through its own
market power, Sirius XM could afford to pay those opportunity costs
because, as Professor Lys explained,\117\ it earns sufficient profits
to pay those opportunity costs and still earn a significant profit.
    \117\ Professor Lys's detailed examination of Sirius XM's
profitability is discussed later in this Determination.
    Thus, Professor Willig's ``fork in the road'' approach, and Sirius
XM's capacity to pay the market-based opportunity costs, taken together
or separately, are supportive of the 15.5% rate determined by the
E. The ``Efficient Component Pricing Rule''
    Professor Willig identified another approach to rate-setting: The
Efficient Component Pricing Rule (ECPR). As he described this approach:
    The ECPR rates would be calculated by adding on to the direct
cost of providing access the opportunity cost of the competitive
entry; i.e. the margin on the competitive business that the
copyright owners would lose if the entrant won that business away.
In short, ECPR prescribes rates for access equal to direct plus
competitive opportunity costs.
Willig WDT ]35.
    Professor Willig testified that the ECPR could be ``somewhat
relevant here since the statutory royalty at issue can be construed as
the price of access to the copyrights protecting the sound recordings,
and since the various modes of distribution of the sound recordings do
compete with each other to various extents.'' Willig WDT ]14. Moreover,
Professor Willig noted that ``by its very design, ECPR is arguably
consistent with the policy objectives (a), (b), and (c) of section
801(b)(1).'' Id. At first blush, it is puzzling that Professor Willig
did not include in his written testimony an explicit application of the
ECPR model.\118\ However, in a colloquy with the Judges, Professor
Willig acknowledged that his ``opportunity cost'' model constituted an
application of the ECPR model.\119\
    \118\ The absence of a more explicit application of the ECPR
approach by Professor Willig in his Written Direct Testimony is also
somewhat surprising because Professor Willig has been identified by
his colleagues as the economist who first developed the ECPR
approach, also known as the ``parity pricing'' principle. See W.
Baumol, J. Ordover, and R.D. Willig, Parity Pricing and Its Critics:
A Necessary Condition for Efficiency in the Provision of Bottleneck
Services to Competitors, 14 Yale J. Reg. 145, 148 n.4 (1997) (``So
far as we have been able to determine, the ECPR proposal stems from
Willig's work. Robert D. Willig, The Theory of Network Access
Pricing, in Issues in Public Utility Regulation 109 (1979).'').
    \119\ See 5/2/17 Tr. at 2107.
    Professor Willig testified that he was reluctant to rely solely on
the ECPR approach because it is intended to establish rates that
correct for the case in which an owner of an upstream essential (``must
have'') input also competes downstream in the retail market (i.e., a
vertically-integrated firm) but refuses to make the essential input
available to would-be competitors (i.e., the upstream firm engages in
what is known as ``foreclosure'').
    The Judges find the Opportunity Cost/ECPR approach to be more
applicable here than Professor Willig suggested. Although the Judges do
not constitute an ``antitrust court,'' the parties acknowledge that the
Judges must establish rates that are effectively competitive, i.e.,
that adjust or offset sufficiently for any complementary oligopoly
power in the benchmark markets or in the markets from which opportunity
costs arise. Whereas an ``antitrust court'' would seek to remedy, ex
post, pricing that was in excess of an ECPR-derived price, the Judges
here are charged with setting a rate, ex ante, that reflects an
effectively competitive rate. There is no reason why an ECPR rate could
not accommodate ex ante rate-setting as well provide an ex post
    \120\ One of Professor Willig's colleagues and frequent co-
authors, and a developer of the ECPR approach, the late Professor
William Baumol, explicitly noted the appropriateness of applying the
ECPR approach to the setting of royalties for licenses in the music
industry. W. Baumol, The Socially Desirable Size of Copyright Fees,
1 Rev. Econ. Res. on Copyright Issues 83 (2004).
    Moreover, a particular limitation of the Opportunity Cost/ECPR
approach is expressly accounted for in the present statutory and
regulatory structure. That is, some economists have questioned whether
the ECPR truly models for an efficient and competitive price, because
the opportunity cost of the upstream supplier(s) that must be covered
by the rate has embedded within it supracompetitive profits that are
not the consequence of more efficient operations. See generally C.
Decker, Modern Economic Regulation 151 (2015) (``[T]he ECPR does not
seek to
[[Page 65240]]
address concerns about monopoly pricing . . . . [T]he ECPR approach
effectively guarantees the pre-entry profits of the incumbent,
including any inefficiency associated with its historic
activities.'').\121\ In rate-setting proceedings, when presented with
sufficient evidence, the Judges can and do expressly adjust or offset
marketplace rates in order to reduce the royalty to a level that better
reflects effective competition, rather than simply allowing the rate to
incorporate (without a downward adjustment or offset) the full
complementary oligopoly effect baked into the opportunity cost.
    \121\ The inefficiently high downstream price is set when, in
the usual situation, the vertically-integrated supplier sells at a
monopoly retail price. In the present context, the Majors, as
complementary oligopolists, price their sound recordings in the
unregulated interactive market above even the monopoly level and the
retail interactive services must cover their input costs through
retail prices higher than they would be in the absence of such
inefficiently high input prices. See Web IV, 81 FR at 26343.
    On balance, the Judges find Professor Willig's discussion of the
ECPR approach to be persuasive confirmation of the Judges' finding that
his Opportunity Cost approach provides an appropriate basis for setting
a reasonable rate when the proper survey data are used as inputs.\122\
    \122\ As discussed in connection with Factor C in the itemized
801(b)(1) factors, Sirius XM's development of a differentiated
product through its satellite-based network constitutes a form of
product differentiation that creates value and profits that, under
Factor C (and under an appropriate consideration of the ECPR
approach) should continue to inure to the benefit of Sirius XM, net
of the licensors' opportunity costs.
F. Professor Willig's Nash Bargaining Solution Approach
    Professor Willig asserted that the walk-away opportunity cost he
calculated, $2.55 per subscriber per month, represented only the
minimum that each label would accept in unregulated negotiations with
Sirius XM. As he further explained, in an unregulated market, even
after receiving the full walk-away opportunity cost, the label would
still negotiate with Sirius XM for a portion of the surplus value
(revenue over costs) that remained. In order to quantify this surplus,
and to calculate and then add the label's share of the surplus to the
label's walk-away opportunity cost, Professor Willig applied what is
known in game theory and in economics as the ``Nash Bargaining
Solution,'' which he described as a type of price discovery engaged in
by an ``unregulated profit-maximizing firm.'' Willig WDT ]38. The Nash
Bargaining Solution is an analytic approach that identifies a price
agreed to in a bilateral negotiation between one buyer and one seller,
in which each party will refuse to accept a value below that which it
would receive absent an agreement (referred to as its ``threat,''
``disagreement,'' or ``fallback'' point), and each party uses its
``bargaining power'' to negotiate for itself the greatest share of any
surplus value (i.e., value in excess of the sum of both parties'
``threat/disagreement'' point values). See id. Under this model, the
surplus that can be created may be split evenly between the parties. 5/
2/17 Tr. 2116-18 (Willig). A 50:50 split of the surplus assumes the
parties have equal bargaining power and means the parties benefit
equally by executing the agreement.\123\ 5/2/17 Tr. 2110 (Willig).
    \123\ Importantly, this does not mean each party enjoys equal
profit. The parties may not profit equally ``because their fallback
values (opportunity costs) may have been different.'' 5/2/17 Tr.
2110 (Willig). Even if parties do not possess equal bargaining
power, and even if that disparity in bargaining power is
incorporated into a Nash model, neither party would be compelled by
the assumptions of the model to accept less than its fallback value,
i.e., its opportunity cost. Id. at 2110-11 (Willig).
    In this model a record label's fallback point would be its walk-
away opportunity cost, which Professor Willig calculated to be $2.55
per subscriber per month. Willig WDT ]48; 5/2/17 Tr. 2110-11 (Willig).
Sirius XM's fallback point would be its projected ARPU in the absence
of music programming, less variable costs (i.e., its earnings in a
world absent an agreement with the single seller (record company) in
this model).\124\ Professor Willig computed this amount to be
$[REDACTED] per subscriber per month. See Willig WDT ]48.
    \124\ Professor Willig based his projection on the finding in
the Boedeker Survey that 70% of Sirius subscribers would leave in
the absence of music programming. See Willig WDT ]48 & n.22. He
computed variable costs as [REDACTED]% of ARPU, based on Professor
Lys's testimony. See id. ]48 & n.21.
    Professor Willig calculated the total earnings created by Sirius
XM's compulsory license (Sirius XM's ARPU less variable costs exclusive
of royalties) as $[REDACTED] per subscriber per month. This resulted in
a surplus from the agreement of $2.78 per subscriber per month.\125\
Assuming that the parties would divide the surplus equally, Professor
Willig opined that the record labels would earn from the agreement
their opportunity cost of $2.55 plus one-half of the surplus ($1.39)
for a total of $3.94 per subscriber per month. See id. ]49. Given a
Sirius XM ARPU of $[REDACTED], this per subscriber rate is equivalent
to a percent-of-revenue rate of [REDACTED]%.
    \125\ Professor Willig computed the surplus as the total
earnings from the agreement less the sum of the parties' fallback
points. See Willig WDT ]48.
    Based on this alternative approach, SoundExchange concluded that
``Professor Willig's Nash Bargaining Solution therefore appropriately
suggests a rate above the copyright owners' opportunity costs.'' SEPFF
]725 (emphasis added). As such, SoundExchange argued that this approach
confirms the reasonableness of its even lower $2.55 per month
subscriber royalty and the equivalent 23%-of-revenue rate implied by
that per-subscriber proposal.
    Sirius XM leveled two basic criticisms at Professor Willig's Nash
Bargaining Solution model. First, it asserted that Professor Willig's
Nash Bargaining Solution posited a monopoly seller of sound recording
performance licenses, which is antithetical to the requirement that the
statutory rate must represent the product of a hypothetical market that
is effectively competitive. SXMRPFF ]196 (and record citations
therein).\126\ Second, Sirius XM noted that SoundExchange's proposal
that the Nash surplus be deemed split 50/50 (rather than in favor of a
record company) is irrelevant, because the opportunity cost figure of
$2.55 is already inflated by the complementary oligopoly effect in that
opportunity cost figure. See id. ]197 (and record citations therein).
    \126\ Sirius XM also relies on Professor Farrell's ``Nash-in-
Nash'' model, as a counterpoint to Professor Willig's Nash
Bargaining Solution. Professor Farrell injects a second record
company to the Nash approach, as contrasted with the single record
company assumed by Professor Willig. However, Sirius XM acknowledged
that Professor Farrell's ``Nash-in-Nash'' approach was not intended
to provide a separate rate proposal, but rather to demonstrate the
fact that the absence of competition would inflate the rate above an
effectively competitive rate. Id. ]]198-200 (and record citations
    As the Judges have held previously, a significant problem with a
Nashian analysis is that the bargaining power of the respective parties
is speculative and thus the outcome of the bargain is indeterminate.
See SDARS I, 74 FR at 23058; see also id. at 23083 (dissenting opinion)
(concurring on the indeterminacy of a ``surplus-splitting'' analysis).
In the present case, the Nash Bargaining Solution again was not
developed sufficiently in the record for the Judges to rely on that
approach as an independent useful tool for setting the statutory rate.
G. Professor Willig's ``Ramsey Pricing'' Approach
    In another pricing approach, Professor Willig applied the economic
concept of ``Ramsey Pricing.'' This approach is
[[Page 65241]]
designed to address the economic issue of ``[h]ow to price various
products or services whose supply draws on common assets in a fashion
that maximizes consumer welfare while also providing enough net revenue
to meet an overall financial target.'' Willig WDT ]13.\127\ In the
context of this proceeding the ``common assets'' are the sound
recordings supplied by the record labels. Professor Willig did not look
to the Ramsey Pricing approach to recommend an SDARS royalty rate;
rather, he used the Ramsey Pricing approach as ``directional'' guidance
to substantiate his conclusion that the SDARS royalty rate should be
higher than the current statutory rate. 5/2/17 Tr. 2086 (Willig).
    \127\ Ramsey pricing is frequently employed as an analytic
framework for such applications as sales taxes levied to raise
sufficient revenue to meet a government financial target, prices for
various telecommunications services that all are enabled by the same
underlying electronic network, and prices for various railroad
services that all make use of the same track infrastructure. Willig
WDT ]13 n.4.
    Ramsey pricing requires that for different modes of distribution of
sound recordings, price-cost margins should be inversely proportional
to each distributor's own price elasticity of demand. See Willig WDT
]32; 5/2/17 Tr. 2094 (Willig). In setting prices to meet the Ramsey
financial target, ``the Services that should contribute relatively
more, relative to their cost, on a percentage basis are the Services
with the relatively low own price elasticities of demand.'' 5/2/17 Tr.
2095 (Willig).\128\
    \128\ Shorn of economic jargon: For certain distribution
channels, subscribers will be relatively less likely to cancel their
subscriptions if their subscription charge increases, as compared
with other distribution channels.
    When demand for a music service is relatively less sensitive to
price, that suggests that the service is relatively more valuable to
its users. Willig WDT ]33. Accordingly, it follows that Ramsey prices
should be relatively higher for users of that service, to allow for
greater contributions toward compensation to the producers of the
recorded music (i.e., the common asset used by all distribution
channels). Willig WDT ]32. Services with relatively lower elasticities
of demand will lose relatively less downstream revenue, so higher
royalties, even if passed on to subscribers or advertisers, will have
less impact on usage decisions made by those distribution modes and
their consumers, as compared to services with higher elasticities of
demand. See Willig WDT ]]32-33.
    Ramsey pricing reasonably assumes there is a target amount of money
that the producers of the common assets need to realize. In the present
context, Professor Willig identified that financial target as equal to
the monetary value of download sales lost by the labels due to the
increase in streaming. Willig WDT ]31. To identify his Ramsey target,
Professor Willig measured the amount of creator compensation lost as a
result of the movement toward streaming and away from paid downloads
since 2010. Willig WDT ]22. Based on his econometric analysis, he
concluded that substitution of streaming services for downloads has
cost the recording industry about $800 million per year from 2010
through 2016. Willig WDT ]]22-28, & App. B. Professor Willig concluded
that the Ramsey Pricing across distribution channels must be sufficient
to offset these shortfalls, and that, specifically, SDARS royalties
must be increased.
    Professor Willig then estimated the relevant upstream elasticity of
Sirius XM's demand for sound recordings, factoring in both downstream
and upstream effects. He opined that, at current royalty rates, Sirius
XM's upstream demand for sound recordings is much more inelastic than
the upstream demand of interactive services. Given this finding,
Professor Willig concluded that ``even at royalty rates proposed by
SoundExchange, the music input would still be a significantly smaller
percentage of the downstream price for Sirius, meaning that upstream
[price] elasticity is not going to be bigger, probably lower than the
upstream elasticities for the other Services that we're talking
about.'' 5/2/17 Tr. 2099-2100 (Willig). Thus, Professor Willig
estimated that Sirius XM could pay a royalty of $[REDACTED] per
subscriber per month and still achieve the same margin as the
interactive streaming services. Willig WDT ]50. According to Professor
Willig the upshot of that conclusion is that Ramsey pricing principles
suggest that Sirius XM should pay a substantially higher royalty in
order to contribute appropriately (under his Ramsey approach) to meet
the Ramsey revenue target. Willig WDT ]50.
    Sirius XM noted the facial ``theoretical attractions'' of an
appropriately specified Ramsey pricing approach, but finds Professor
Willig's approach not to constitute an actual Ramsey pricing analysis.
Sirius XM found two essential elements of the Ramsey pricing approach
missing from Professor Willig's analysis. First, he did not identify a
financial target sufficient to provide for the creation of the sound
recordings. See 5/2/17 Tr. 2171-72 (Willig); 4/24/17 Tr. 652 (Farrell);
see also 5/2/17 Tr. 2176-77 (Willig) (acknowledging no analysis of
``how much revenue is actually necessary to fund the recording
industry's investment in sound recordings'').\129\
    \129\ While (as noted in the text, supra) Professor Willig did
offer a regression analysis purporting to identify $800 million in
annual losses to the record industry over the past several years
caused by ``streaming'' (not simply satellite radio), Willig WDT
]]22-27, he acknowledged that the figure played no direct role in
any of his calculations, including his ``Ramsey'' analysis. 5/2/17
Tr. 2167:24-2169:18 (Willig).
    Second, Sirius XM asserted that Professor Willig did not identify
all users of the common assets and set prices for each that
collectively would meet the Ramsey financial target, i.e., cover record
industry costs while maximizing consumer welfare. Professor Willig
concedes this point. See 5/2/17 Tr. 2172 (Willig) (did not ``analyze[ ]
all the different modes of distribution that use sound recordings and
determine[ ] the Ramsey prices that would result''); id. at 2177-78
(Willig) (``I have not done a formal financial analysis of impacts of
royalty rates on either creation or what you just called
    In addition, Sirius XM noted that the analysis takes as its
starting point the same measure of opportunity cost used in all of
Professor Willig's approaches, the improper $2.55 opportunity cost
inflated by complementary oligopoly effects. See Farrell WRT ]]90-94;
see also 4/24/17 Tr. 653-54 (Farrell).
    The Judges find Professor Willig's implementation of the Ramsey
pricing approach unhelpful. Professor Willig ultimately neither derived
nor proposed a royalty rate from this analysis.\130\ Nor could he do
so, given that his analysis does not establish a revenue target, and
does not factor in the contribution of other users of the common
assets. To the extent Professor Willig's assertion that his Ramsey
approach has value in this proceeding because it provides
``directional'' evidence has any validity, the Judges note that the
adoption of the 15.5% rate derived from his opportunity cost analysis
is consistent with this directional guidance.
    \130\ Professor Willig stated that one reason he declined to
propose the $[REDACTED] monthly per subscriber royalty (which the
Judges understand to be equivalent to [REDACTED]% of revenue) is
that he could not evaluate how such a substantial increase in the
royalty rate would increase subscription rates and create a loss of
subscribers and subscriber revenue. In economic terms, he could not
opine as to whether, assuming that Sirius XM passed through to
subscribers such a higher royalty rate, the downstream elasticity at
that price point would be so high as to actually reduce Sirius XM's
H. Mr. Orszag's Ratio Equivalency Model
    SoundExchange also presented expert testimony from Mr. Jonathan
Orszag. Mr. Orszag's approach to determining
[[Page 65242]]
SDARS rates was based upon ratio equivalencies. Specifically, he opined
that the royalties in the target market (i.e., those paid by an SDARS)
should be set at a rate that makes the ratio between royalties and
revenues in that target market equal to the ratio between royalties and
revenues in a benchmark market. Mr. Orszag noted that the Judges
``found this assumption to be warranted as a matter of economic
theory'' in Web IV. Amended Written Direct Testimony of Jonathan
Orszag, Trial Ex. 26, ]37(Orszag AWDT).
    Mr. Orszag began his analysis by opining that in this case ``[i]t
is . . . appropriate to use current marketplace agreements in
evaluating the range of reasonable rates for the upcoming licensing
period.'' 4/25/17 Tr. 953 (Orszag) (emphasis added). Marketplace rates
are the appropriate starting points, according to Mr. Orszag, because
``a standard way in which economists estimate a reasonable royalty rate
for the blanket license under consideration in this proceeding is by
examining comparable rates generated through arm's length negotiations
outside the purview of the compulsory license regime for which
satellite radio qualifies,'' i.e., ``[r]ates yielded through . . .
unfettered negotiations . . . .'' Orszag AWDT ]12. Accordingly, Mr.
Orszag utilized a marketplace benchmarking approach.
    Mr. Orszag's first step was to identify what he found to be
comparable benchmark rates that he could adjust, if and as warranted,
to determine the rates that would apply in the target market (SDARS) if
it were unregulated. Orszag AWDT ]13. He looked first at royalty rates
in the interactive music streaming services for data. Then, he analyzed
retail price data for both the interactive and noninteractive music
streaming services. In selecting his benchmarks, Mr. Orszag looked for
agreements entered into by record companies with streaming services
that in his opinion are comparable to satellite radio across pertinent
dimensions. Additionally, he considered whether the benchmark evidence
permitted him to account for material differences, if any, between the
benchmarks and the target market. Orszag AWDT ]28.
1. Mr. Orszag's Benchmark ``Approach One'': Ratio Equivalency With the
Interactive Market
    Applying these considerations, Mr. Orszag identified the market for
the licensing of sound recordings by record companies to interactive
streaming subscription services as the best available benchmark
category for satellite radio, due to what he believed to be ``the
comparability of the two types of service along key dimensions and the
availability of reasonable methodologies with which to adjust for
pertinent differences.'' Orszag AWDT ]29. More particularly, Mr. Orszag
identified the following alleged comparable qualities in the
``downstream market'' \131\ between the target and benchmark markets:
    \131\ The ``downstream market'' is the market in which licensees
of sound recordings offer their services to subscribers or other end
users/consumers. The ``upstream market'' is the market in which
record companies (a/k/a/labels), as licensors, license their
repertoires to services, as licensees, for ultimate dissemination in
the downstream market. See Web IV, 81 FR at 26332 n.69.
     Both categories of services offer a full repertoire of
     both categories of services offer subscription-based
models, thereby demonstrating that their listeners' have a positive
willingness to pay;
     both categories of services face similar downstream
elasticities of demand;
     both categories of services offer products that compete
with each other;
     consumers in both categories of services receive music
     consumers in both categories of services obtain
unlimited usage;
     both categories of services offer mobile functionality,
Sirius XM principally through in-vehicle receivers and interactive
streaming through smartphones and other mobile devices; and
     interactive streaming services increasingly offer a
``lean-back'' \132\ functionality (akin to the functionality of
Sirius XM listening) through playlists generated by the services,
third parties, and subscribers, as well as algorithmic streams.
    \132\ Functionally noninteractive services are generally
described in the industry as ``lean-back'' services, as contrasted
with ``lean forward'' services that have varying degrees of
interactivity. See Web IV, 81 FR at 26336 n.75.
4/25/17 Tr. 968 (Orszag); Orszag AWDT ]32.
    Mr. Orszag further opined that sound recording performance rights
are similarly indispensable inputs in the upstream market for both
interactive streaming services and Sirius XM. From an economic
perspective, he explains that the upstream demand for sound recording
rights is what economists call a ``derived demand,'' i.e., upstream
demand is derivative of downstream consumer demand. Mr. Orszag further
opined that, because of this indispensability, sound recording
copyright holders should receive a material portion of the overall
value of satellite radio service, as reflected in the prices paid by
subscribers, just as they do for interactive music services. Orszag
AWDT ]31.
    To determine the rates actually paid by subscription interactive
services, Mr. Orszag reviewed the monthly royalty rates and royalty
payments set in 27 current license agreements between three major
record labels \133\ and nine interactive streaming services,\134\ from
January 2014 through June 2016. Orszag AWDT ]45; see 4/25/17 Tr. 985
(Orszag) (``So I got the royalty statements from each of the . . .
Services for each of the labels by month, and I went to what they
actually were being paid, which prong was governing.'').\135\
    \133\ Sony Music Entertainment (Sony), Universal Music Group
(UMG), and Warner Music Group (WMG) are the three major record
labels (together, the Majors).
    \134\ The nine services are listed in the table that follows in
the text, infra.
    \135\ The agreements Mr. Orszag studied contain royalty rate
provisions that require the services to calculate royalty
obligations under separate ``prongs'': a [REDACTED] metric and a
[REDACTED] metric, and in some cases a [REDACTED] metric, and then
pay each label its pro rata share of [REDACTED]. A label's pro rata
share of the royalty is based on the share of the total performances
on the service accounted for by sound recordings controlled by that
label. Orszag AWDT ]45.
    The table below presents the actual monthly per-subscriber royalty
payments made by the subscription interactive services to each of the
Majors. These data produce an average monthly per-subscriber payment of
$[REDACTED], weighted by the number of subscribers per service. Orszag
AWDT ]46.
[[Page 65243]]
                                                                              Actual Licensing Fees Per-Subscriber
                                                        Sony                                                   UM                                                    WM
                                      2014              2015              2016              2014              2015              2016              2014              2015              2016
Apple Music...................  ................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED]
Beats.........................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED].....  ................
Google Play...................  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED]
Microsoft.....................  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED]
Rdio..........................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED].....  ................
Rhapsody......................  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED]
Slacker.......................  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED]
Spotify.......................  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED].....  $[REDACTED]
TIDAL.........................  ................  $[REDACTED......  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED].....  ................  $[REDACTED].....  $[REDACTED]
 Source: Royalty payment data from Sony, UMG, and WMG.
Orszag AWDT at 19, Table One.
    For the nine subscription interactive services in the above table,
over the 2014-2016 period covered, individual subscriptions were
offered to consumers at $9.99 per month. At that monthly price, the
weighted average monthly per-subscriber payment of $[REDACTED]
translates to a royalty equal to approximately [REDACTED] % of the
services' revenues ($9.99 x [REDACTED]). Orszag AWDT ] 47.\136\
    \136\ Mr. Orszag did not include in his royalty calculation any
non-rate consideration, such as access to the services' user data
and user email addresses; the services' marketing and promotional
support; and the record companies' right to offer exclusives to
services; including the right to ``window'' certain sound recordings
(i.e., to offer an initial, time-limited exclusivity). Because these
non-pecuniary items are not available under the statutory license at
issue in this proceeding, Mr. Orszag asserts that his omission of
these non-monetary benefits renders his calculated royalty payment
lower than it otherwise would be, thus reducing the royalty rate
derived from his benchmark in favor of Sirius XM. Orszag AWDT ] 106.
See also SE PFF ]] 119-122 (and record citations therein).
    Because Mr. Orszag's interactive data were limited to agreements
with the Majors, he also considered whether the rates paid by
subscription interactive streaming services to the Indies were lower
than those paid to the Majors. He determined that, whether the Indies'
recordings were distributed by a Major or a Major affiliate, or were
distributed by another entity, the terms regarding royalties were
``highly similar'' to the rates paid to the Majors. Consequently, Mr.
Orszag made no adjustment to his interactive benchmark to account for
the rates paid by interactive services to independent record labels.
Orszag AWDT ]] 101-105; see Written Direct Testimony of Jeremy Sirota,
Trial Ex. 36, at 3 (Sirota WDT).
    Mr. Orszag utilized the concept of ``ratio equivalency'' to compare
his benchmark rate for the interactive streaming market to the target
SDARS market. He applied essentially the same ratio equivalency
approach as the Judges applied to the noninteractive subscription
market in Web IV.\137\ Orszag AWDT ] 37. More specifically, Mr. Orszag
relied on the following points from Web IV to identify what he
considered necessary conditions for the application of a ratio
equivalency approach:
    \137\ In Web IV, the Judges stated that the ratio equivalency
concept ``assume[s] equality between two ratios: (1) subscription
revenues to royalties in the interactive market; and (2)
subscription revenues to royalties in the noninteractive market.''
Web IV, 81 FR at 26344.
    (1) Revenues in both markets must be derived from subscription
revenues and thus be reflective of buyers with a positive
willingness to pay (WTP) for streamed music;
    (2) Functional convergence and downstream competition for
potential listeners must indicate a sufficiently high cross-
elasticity of demand as between interactive and noninteractive
services, provided the noninteractive subscription rate is reduced
to reflect the absence of the added value of interactivity; and
    (3) The benchmark market rate must be adjusted downward \138\ to
eliminate the ``complementary oligopoly'' effect arising from the
presence of multiple ``must have'' suppliers, thereby establishing a
rate that is ``effectively competitive.''
    \138\ In Web IV, the Judges applied a ``steering adjustment'' to
reflect noninteractive services' ability to offset the complementary
oligopoly power of the Majors by ``steering'' listeners to sound
recordings licensed from Indies at lower royalty rates.
Id. ] 41 (citing Web IV, 81 FR at 26353). Mr. Orszag posited that all
three of these Web IV conditions are satisfied in this proceeding.
    He noted that in both the interactive streaming and SDARS markets
revenues are derived from subscribers with a positive WTP. More
particularly, subscribers to interactive services typically pay $9.99
per month, Orszag AWDT ] 36, while subscribers to Sirius XM typically
pay at least that amount. Id. at ] 49 & n.40. With regard to the second
condition, Mr. Orszag cites record evidence of functional ``lean-back''
convergence and downstream competition, particularly with regard to the
use of playlists and enhanced mobile technology, which have allowed
interactive streaming services to gain an increasing share of in-car
listening. See 4/24/17 Tr. 605 (Farrell); Orszag AWDT ] 39. Finally,
Mr. Orszag testified that changes in the interactive market after Web
IV had obviated the need for a complementary oligopoly adjustment.
Nonetheless, he provided three alternative potential steering
adjustments in the event the Judges disagreed with his conclusion
regarding complementary oligopoly: (1) A [REDACTED]% steering
adjustment derived from Sirius XM's direct licenses; (2) a 12% steering
adjustment borrowed from Web IV; or (3) a [REDACTED]% steering
adjustment identified in a comparison of two ``Mid-tier'' services
contracts, one with a prohibition on steering and the other
    \139\ These potential steering adjustments are discussed in
detail infra.
    The interactive market benchmark ratio equivalency approach is
well-depicted in algebraic form: \140\
    \140\ The ratios are sometimes expressed reciprocally, with
royalties in the denominator and revenues in the numerator. Because
royalty rates in this proceeding are expressed as a percent-of-
revenue, it is more intuitive to state the ratio as set forth in the
text, supra.
[[Page 65244]]

                                                           Benchmark Ratio                                Target Market Ratio

                                                                       (A)   ............                                   (C)
                                                 Royalty Payment (in $) in   ............           Royalty Payment (in $) in
                                                          Benchmark Market             =                        Target Market
                                       -------------------------------------              -------------------------------------
                                                                       (B)                                                (D)
                                              Downstream Revenue (in $) in                          Downstream Revenue (in $)
                                                          Benchmark Market                                   in Target Market

                                        (Rates and revenues to be calculated on a per-subscriber per month basis.)

    By inserting the known (i.e., calculable) values for (A) and (B),
Mr. Orszag was able to calculate a ratio, or percentage, that--under
the ratio equivalency approach--he opined would also be applicable to
the target market. That is, the royalty payment (C) in the Target
Market would be the same percent of (D) as (A) is a percent of (B) in
the Benchmark Market.
    In this, his ``Approach One,'' Mr. Orszag calculated the royalty
payments of interactive subscription services as a percentage of their
subscription revenues by dividing the effective monthly per-subscriber
royalty payment by the monthly consumer subscription price of the
benchmark services. Orszag AWDT at ] 43. Applying the theory of ratio
equivalency, Mr. Orszag then proposed that the record companies receive
the same percentage of Sirius XM's subscription revenue as they receive
from the interactive services. See 4/25/17 Tr. 985-86 (Orszag).
    Because Sirius XM provides listeners with both music and non-music
content, Mr. Orszag opined that his Benchmark Market Ratio must be
adjusted to be comparable to the Target Market Ratio. Relying
principally on a survey by Stefan Boedeker, Mr. Orszag determined that
the music content on Sirius XM constituted 50% of the value of total
content.\141\ Orszag AWDT ]54. Additionally, SoundExchange asserted
that the pricing structure reflects Sirius XM's understanding that its
customers value music at least as much as non-music content. Id. ] 49 &
n.40 (discussing Sirius XM monthly pricing of $10.99 for News, Sports &
Talk versus $12.52 for Mostly Music). Moreover, as SoundExchange noted,
in the previous SDARS proceeding, Sirius XM itself took the position
that music accounts for more than 55% of Sirius XM's content value.
SDARS II, 78 FR at 23064-65 (Sirius XM's expert Roger Noll attributed
55% of value to music content). Both parties and the Judges agreed on
this issue. See id. at 23063, 23088 (noting SoundExchange's expert Dr.
Ordover conservatively assumed music accounts for at least 50%); id. at
23065, 23089 (Judges finding ``the success of Sirius XM is dependent
upon its access to music'' citing testimony of Sirius XM witnesses).
The Judges take note that Sirius XM provided no evidence or argument to
support a different position that might place in doubt Mr. Orszag's
reliance on the Boedeker survey. Mr. Orszag reasonably and
conservatively utilized an assumption that at least 50% of the value of
a Sirius XM subscription is derived from music offerings. Applying this
assumption, Mr. Orszag divided the benchmark ratio result, [REDACTED]%
of revenue, by two to arrive at a proposed percentage-of-revenue rate
of [REDACTED]% for Sirius XM. Orszag AWDT ] 54.
    \141\ Mr. Boedeker surveyed subscribers to Sirius satellite
radio packages that contain both music and non-music programming,
(i) to measure the degree to which these subscribers value the music
versus non-music content; (ii) to examine subscribers' willingness
to accept a hypothetical Sirius XM package that contains just music
programming or just non-music programming; and (iii) to identify the
discounts they would demand for such a hypothetical product. Written
Direct Testimony of Stefan Boedeker, Trial Ex. 21, ]]7, 19 (Boedeker
WDT); 5/8/17 Tr. 2933, 2947-49 (Boedeker). Mr. Boedeker concluded
from the survey results that Sirius XM subscribers value music
content significantly more than non-music content. Boedeker WDT at
]] 14, 97; 5/8/17 Tr. 2933-34, 2963 (Boedeker). More precisely,
70.1% of all survey respondents said they would no longer subscribe
to Sirius XM satellite radio at their current subscription rates if
music programming were no longer offered, while only 32.4% said they
would no longer subscribe at their current subscription rates if
non-music programming were no longer offered. Boedeker WDT ]77; 5/8/
17 Tr. 2951 (Boedeker). Even if discounts were offered for a non-
music service, 42.7% of respondents still would no longer subscribe
to their Sirius XM package, compared with only 10.0% of respondents
would no longer subscribe to their current package if non-music
programming were no longer offered (even with a discount). Boedeker
WDT ]]83-84; see also 5/8/17 Tr. 2952-53 (Boedeker). In a critique
of Mr. Boedeker's survey, Professor John Hauser, a Sirius XM expert
witness, identified several inconsistencies in Mr. Boedeker's survey
results. Nonetheless, it was undisputed by Sirius XM that Mr.
Boedeker's results are generally consistent with other available
evidence. See SEPFF ]] 252-258 (and record citations therein). Thus,
Mr. Orszag opined that his use of the 50% figure was conservative,
in the sense that it favored Sirius XM rather than the party for
whom he testified, SoundExchange. Orszag AWDT ] 54.
    Mr. Orszag opined that a benefit of his ``Approach One'' is that it
avoids the need to account explicitly for differences between the
target and benchmark services. Rather, he stated that the differences
are implicit in the formula and thus revealed by the market. A
service's retail (subscription) revenues are a direct function of
consumer subscription prices. Those prices should reasonably reflect
consumer valuation of the features and functions of the benchmark and
target services, respectively. In turn, according to Mr. Orszag,
percentage-of-revenue royalty rates should reflect such differences,
because the sound recordings performed by the services in the benchmark
and target markets are identical. Id. ] 55.
    As noted, SoundExchange is proposing a greater-of statutory rate
with a per-subscriber prong as well as a percent-of-revenue prong. To
obtain what Mr. Orszag described as an equivalent per-subscriber rate,
he applied the [REDACTED]% of revenue rate (derived from his benchmark
ratio equivalency analysis) to the ARPU. Mr. Orszag adjusted the Sirius
XM ARPU of $[REDACTED] (as gross revenue is calculated using the
statutory license terms) using the same ratio he applied to reach a
percent-of-revenue rate.\142\ This resulted in a per-subscriber rate of
$[REDACTED] (i.e., $[REDACTED] x [REDACTED]). See id. ] 54.
    \142\ Mr. Orszag calculated ARPU using Sirius XM's regulatory
revenue base for the first six months of 2016. See Orszag AWDT ]]
58-60 and Table Three. Professor Shapiro, on behalf of Sirius XM,
initially identified a monthly ARPU of $[REDACTED] per subscriber,
apparently using Sirius XM's 10-Q filing with the SEC and an
internal Sirius XM planning document. See Lys WRT ]] 151-152 nn.174,
177 & Fig. 18. However, the parties apparently reached agreement
that, under the current definition of ``Gross Revenues,'' the
appropriate monthly ARPU is $[REDACTED]. See SX RPFF ] 392 (``That
$[REDACTED] figure was used directly by economists from both parties
to convert monthly per-subscriber fees into proposed percent-of-
revenue rates.''); see also Lys WRT ]] 149-155.
2. Mr. Orszag's ``Approach Two'': Retail Price Comparison
    Mr. Orszag's Approach One implicitly accounted for the different
values of interactive and noninteractive services by utilizing retail
prices in the denominators that reflected the market-based differences
in those values. In ``Approach Two,'' Mr. Orszag applied an alternative
methodology designed to
[[Page 65245]]
account explicitly for the absence of interactivity in the target SDARS
market. Orszag AWDT ] 56.
    In Approach Two, Mr. Orszag continued to use the interactive market
as his polestar. In this approach, however, he compared the interactive
retail subscription price not to the target SDARS market, but to the
market for noninteractive services, on the assumption that an SDARS
functionally is a noninteractive service.\143\ In this manner Mr.
Orszag was able to isolate explicitly the value of interactivity by
comparing the retail prices of interactive and noninteractive
subscription services. See 4/25/17 Tr. 986 (Orszag). Mr. Orszag opined
that this approach is sensible because these two categories of service
differ only with respect to the distinguishing feature: Interactivity.
See id.\144\
    \143\ A noninteractive service is one that meets the statutory
definition and pays statutory royalties calculated under 17 U.S.C.
114(f)(2)(B). An SDARS service may be described as functionally a
noninteractive service because the listener cannot interact with the
service to select, repeat, skip, or cache specific sound recordings.
See supra, n.73 and accompanying text.
    \144\ Approach Two avoids the need to adjust for non-music
content because streaming services are music-only services. It also
avoids any purported need to adjust for the separate value of a
satellite network because streaming services are internet-based. See
Orszag AWDT at ]56. The Judges address later the question Sirius XM
raises relating to whether its satellite network creates an
additional value that should reduce the statutory royalty rate.
    To determine the monthly retail price in the noninteractive market,
Mr. Orszag used the retail prices of three non-interactive subscription
services: Pandora One, Rhapsody (Napster) unRadio, and Slacker Radio.
He calculated their weighted average monthly retail price to be $4.91.
See Orszag AWDT ] 56 & Table Two.
    As noted before, the monthly retail price for interactive
subscription services was $9.99. Accordingly, the ratio of the
subscription price from the noninteractive market to the subscription
price from the interactive market was $4.91/$9.99, or 0.49. Mr. Orszag
then used the ratio of 0.49 to convert the interactive subscription
services monthly per-subscriber royalty rate of $[REDACTED] to an
equivalent per-subscriber rate for Sirius XM of $[REDACTED] (0.49 x
$[REDACTED]). See id. ] 57.
    The final step in Mr. Orszag's Approach Two is the calculation of a
percentage-of-revenue rate that corresponds to this $[REDACTED] per-
subscriber rate. Applying the same $[REDACTED] ARPU \145\ to the per-
subscriber rate of $[REDACTED], Mr. Orszag derived a percentage-of-
revenue rate of [REDACTED]%. See id. ] 60.
    \145\ See supra note 142 and accompanying text.
3. Adjustment for Lack of Effective Competition in Benchmark Market
    In his attempt to apply the Web IV prerequisites for use of a
``ratio equivalency'' benchmarking approach, Mr. Orszag considered
whether to apply a downward adjustment to reflect any alleged lack of
``effective competition'' in his benchmark interactive market. He
acknowledged that in Web IV the Judges found that the market for
subscription interactive services (i.e., Mr. Orszag's benchmark market
here) was not effectively competitive. The Judges, therefore, adjusted
downward the rate SoundExchange's economic expert calculated using an
interactive services benchmark. Web IV, 81 FR at 26344.
    In this proceeding, however, Mr. Orszag concluded that the record
establishes that more recently the market for subscription interactive
services has become effectively competitive. Mr. Orszag concluded that
he need not adjust to offset a lack of effective competition. Mr.
Orszag's opinion is based on:
     The presence in the market of larger interactive
streaming services, such as Amazon, Apple, Google, and Spotify,
which has injected countervailing ``substantial bargaining power and
leverage'' on the licensee side of the equation, offsetting any
relative disproportionate power that the record companies might have
previously possessed. Written Rebuttal Testimony of Aaron Harrison,
Trial Ex. 49, ]]3-5 (Harrison WRT); 5/16/17 Tr. at 3953-57
     The increasing importance of interactive services as a
revenue source to the record companies, which gives the services
leverage strengthen their bargaining position in negotiations for
sound recording performance licenses. Written Rebuttal Testimony of
David Blackburn, Trial Ex. 39, ]]18, 20 (Blackburn WRT).
     The treatment of Spotify's licensing agreement with the
Majors when it expired, by not [REDACTED], but rather [REDACTED]. 5/
01/17 Tr. 1703-04, 1804-05 (Blackburn).
     The additional bargaining power of individual services
because they have differentiated their offerings, based on platform
preference ([REDACTED]); catalog size ([REDACTED]); and payment
terms ([REDACTED]), meaning that the withdrawal of any
differentiated service from the market would result in customer
``churn'' that would negatively affect record companies financially.
5/16/17 Tr. 3942-45 (Harrison).
     The lack of market evidence of: (1) Suppression of the
output of recorded music; (2) supracompetitive profits achieved by
the record companies; or (3) ready alternatives to which downstream
consumers might turn. SEPFF ]] 305-322 (and record citations
     The inability of the Majors to act as price-setters,
[REDACTED]. 5/16/17 Tr. 3926-27, 3946-47 (Harrison).
     The Majors' agreements in the Mid-Tier limited
interactivity sector to rates as low as [REDACTED] % of revenue when
the licensing agreement includes [REDACTED]. SEPFF ] 356 (and record
citations therein).
     The agreements between Indies and interactive streaming
services that [REDACTED]. SEPFF ]] 335-340 (and record citations
    Mr. Orszag maintained that the interactive streaming rates reflect
an ``effectively competitive'' market. He nonetheless offered three
alternative ``steering adjustments'' to apply to those benchmark rates,
should the Judges find the interactive market to be not effectively
competitive. Mr. Orszag first presented a [REDACTED]% steering
adjustment, reflecting his calculation of an arguable steering effect
arising from Sirius XM's direct licenses with certain Indies.\146\
Next, Mr. Orszag proposed a 12% steering adjustment, simply adopting
the adjustment the Judges made in Web IV. See Web IV, 81 FR at 26404-
05. Finally, he presented a [REDACTED]% steering adjustment, that
reflects the differences in royalty rates in the mid-tier market,
depending upon whether the license agreement has a [REDACTED] (and an
attendant lower rate) or [REDACTED] (with an attendant higher royalty
rate). See 4/25/17 Tr. 1054 (Orszag). The Table below summarizes Mr.
Orszag's alternative rates based on the absence of a steering
adjustment and on all three of the alternative steering adjustments.
    \146\ These direct licenses are discussed in more detail in the
Judges' consideration of Sirius XM's reliance on these licenses as
potential benchmarks.
                                               Approach One                            Approach Two
         Steering adj  %         -------------------------------------------------------------------------------
                                        Rev.  %             Per sub             Rev.  %             Per sub
None............................  28.0..............  $ 3.00............  25.7..............  $ 2.76
[REDACTED]......................  [REDACTED]........  $[REDACTED].......  [REDACTED]........  $[REDACTED]
12..............................  24.6..............  $ 2.64............  22.7..............  $ 2.43
[[Page 65246]]

[REDACTED]......................  [REDACTED]........  $[REDACTED].......  [REDACTED]........  $[REDACTED]
SE PFF ] 361 (Sirius XM did not dispute the accuracy of this summary
table derived from record evidence.).
4. The Mid-Tier Agreements as Corroboration
    According to Mr. Orszag, the applicability of the theory of ratio
equivalency is further supported by agreements between record companies
and Mid-tier services. These Mid-tier Agreements'' are comprised of
recently executed voluntary direct licenses for subscription mid-tier
services, between Pandora and iHeart, respectively, as licensees, and
the Majors and Merlin, a digital rights agency representing Indie
record companies, as licensors.\147\
    \147\ The agreements executed by Pandora and iHeart also covered
fully interactive tiers and, in the case of Pandora, an ad-supported
tier. See, e.g., Trial Exs. 112-114. For ease of exposition, the
Judges use the term ``Mid-tier Agreements'' to refer to the portion
of each agreement that relates to the subscription service offered
to consumers for $4.99 and providing limited on-demand
    The Table below provides a breakdown of rates contained in Mid-tier
Agreements that were admitted into evidence in this proceeding: \148\
    \148\ The agreements in the table were made a part of the
record. See Trial Ex. 112-16B at sec. 11 (SoundX_000107538-39)
(Pandora Plus and Pandora Premium royalty provisions); Trial Ex.
112-16A at Service Schedule #1 sec. 7(a) (SoundX_000107458) (iHeart
Plus royalty provisions); Trial Ex. 112-16A at Service Schedule #2
sec. 7(a) (SoundX_000107492) (iHeart All Access royalty provisions);
Trial Ex. 113-017B at Schedule 1 sec. 3.1(a)(i)-(ii), sec.
3.2(a)(i)-(ii), sec. 4.1 and sec. 4.2 (SoundX_000107051-52, 056);
(Pandora Plus and Pandora Premium Royalty provisions); Trial Ex.
113-017A at Schedule 1 sec. 1.1 and sec. 1.2 (SoundX_000106973)
(iHeart Plus and iHeart All Access royalty provisions). Trial Ex.
113-017B at Schedule 1 sec. 3.2(a)(iii) (SoundX_000107052, 056).
Trial Ex. 114-018B at 11-14 (SoundX_000107127-30) (Pandora Plus and
Pandora Premium royalty provisions); Trial Ex. 114-018A at sec. 3(a)
and sec. 3(b) (SoundX_000107206-07) (iHeart Plus and iHeart All
Access provisions). Trial Ex. 243 at sec. 3(b) and sec. 3(c)
(SoundX_000477169-170) (Pandora Plus and Pandora Premium royalty
provisions); Trial Ex. 272 at Schedule 3 (SoundX_000488916) (iHeart
Plus and iHeart All Access).
                                                                                [Restricted]--Mid-Tier Agreements
                                           Pandora plus ($4.99)                   Pandora premium ($9.99)                   iHeart plus ($4.99)                  iHeart all access ($9.99)
                                     % of Revenue           Per sub          % of Revenue           Per sub          % of Revenue           Per sub          % of Revenue           Per sub
Sony............................  [REDACTED]........  $[REDACTED].......  [REDACTED],         $[REDACTED],-$[RED  [REDACTED],-[REDAC  $[REDACTED],-$[RED  [REDACTED],-[REDAC  $[REDACTED],-$[RED
                                                                           [REDACTED].         ACTED].             TED].               ACTED].             TED].               ACTED]
UMG.............................  [REDACTED]-[REDACT  $[REDACTED]-$[REDA  [REDACTED]-[REDACT  $[REDACTED]-$[REDA  [REDACTED]........  $[REDACTED].......  [REDACTED]........  $[REDACTED]
                                   ED].                CTED].              ED].                CTED].
                                   ED].                CTED].              [REDACTED].         CTED].              TED].               ACTED].             TED].               ACTED]
Merlin..........................  [REDACTED]........  $[REDACTED].......  [REDACTED],         $[REDACTED],-$[RED  [REDACTED],-[REDAC  $[REDACTED],-$[RED  [REDACTED],-[REDAC  $[REDACTED],-$[RED
                                                                           [REDACTED].         ACTED].             TED].               ACTED].             TED].               ACTED]
These Mid-tier Agreements bundled terms for separate tiers offered by
Pandora and iHeart, respectively, including the actual mid-tier
services identified as Pandora Plus and iHeart Plus, respectively, with
limited interactive functionality, and a tier providing fully
interactive functionality. Orszag WDT ] 38.
    Mr. Orszag found confirmation for his benchmarking approach in the
rates at which Pandora and iHeart will license from major and
independent record companies, i.e., at rates ranging from $[REDACTED]-
$[REDACTED] per subscriber per month. These rates are similar to the
per subscriber rates SoundExchange proposes in this proceeding. Trial
Exs. 112-114. Mr. Orszag also noted that these Mid-tier Agreements
include [REDACTED]. See Orszag AWDT at ] 38. When these percent-of-
revenue rates are halved (as in his Approach One) to reflect that 50%
of the value of Sirius XM's service is attributable to non-music
content, the percent-of-revenue rates in these Mid-tier Agreements lie
in the range of [REDACTED]-[REDACTED]%, ``strikingly similar'' to the
23% royalty rate SoundExchange has proposed. See SXPFF ]] 845-847.
    Mr. Orszag found the rates in these Mid-tier Agreements to be
instructive and corroborative of SoundExchange's rate proposal.
SoundExchange conceded that the mid-tier services of iHeart and Pandora
offer some interactivity, whereas Sirius XM's satellite service offers
no interactivity. Mr. Orszag opined, however, that it is not plausible
that the differential would have a significant impact on consumer
valuations and, consequently, on per-subscriber rates. In support of
that argument, he noted that subscriptions to the mid-tier services
offered by Pandora and iHeart are priced at the same $4.99 per month as
Pandora's prior noninteractive offering. See Harrison WDT at ] 19.
Further, Mr. Orszag noted that his highly conservative estimate of the
value of music content on Sirius XM, is even higher, at $[REDACTED].
See Orszag WRT ] 55 & n.68.
    More particularly, Mr. Orszag noted that Pandora's offering of
increased skips, rewind capability, and limited caching to convert its
noninteractive service into a mid-tier service did not cause Pandora to
increase its monthly subscription price above the $4.99 it charged
previously for its noninteractive service. Mr. Orszag testified that
this suggests that consumers' valuation of the increased functionality
is not so high as to allow Pandora to increase its mid-tier retail
subscribership price off the $4.99 per month and closer to the $9.99
monthly price for fully interactive services. 4/25/17 Tr. 1063-64
(Orszag). Mr. Orszag concluded that these facts demonstrate that the
mid-tier services have a value commensurate with a noninteractive
    Finally, Mr. Orszag recognized the hypothetical possibility that,
because these Mid-tier Agreements bundle fully interactive services,
the record companies could have applied their market power in that
segment to extract higher rates and better terms in the mid-tier
segments. To test that hypothetical, Mr. Orszag reviewed the
negotiation documents relating to the Mid-tier Agreements and concluded
that they contained no evidence that the Majors used their alleged
market power in the fully-interactive services market to obtain
concessions on mid-tier terms. Orszag WRT ] 55. To the contrary, the
evidence suggests that Merlin obtained
[[Page 65247]]
rates similar to those negotiated by the Majors in its licensing
agreements with [REDACTED]. Id. Sirius XM did not proffer any evidence
that the record companies leveraged their alleged interactive market
power to obtain better terms in the mid-tier market.
5. Evaluating Orszag Ratio Equivalency Benchmarking Approaches
    Sirius XM asserted that Mr. Orszag incorrectly emphasized an
economically unimportant point, i.e., that ``there is no difference
between interactive streaming services and satellite radio in terms of
the music content they deliver to subscribers.'' See 4/26/17 Tr. 1190-
91 (Orszag) (emphasis added). According to Sirius XM, similarity ``at
this high level of generality'' is meaningless. SXM RPFF ]11.
    The Judges agree. Although markets in which sound recording
performances are licensed (upstream) and delivered (downstream) to
subscribers may be considered as potential benchmarks for each other,
that broad brush of comparability does not indicate whether the
benchmark is suitable on the whole. Mr. Orszag was correct that the
benchmarking approach can commence at a high level of generality, even
though that basic level of comparison is by no means probative or
dispositive.\149\ Not every market in which sound recording
performances are licensed could serve as a benchmark for every other
sound recording performance market.
    \149\ In fact, a market in which some product other than music
is delivered could be a useful benchmark market if it is otherwise
comparable in terms of economic structure. For example, patents, as
a form of intellectual property, may be found to have similar
economic characteristics as copyrights, rendering relevant
information from the market for patent licenses.
    Sirius XM argued that the common use of digital transmissions and
the allowance of unlimited usage by listeners are not illuminating
similarities. SXM PFF ] 12. Once again, the Judges agree; these basic
points are not probative of the usefulness of the interactive market as
a benchmark. Nonetheless, Mr. Orszag's reliance on such common elements
is helpful in identifying and then narrowing the range of potential
    Sirius XM criticized as superficial Mr. Orszag's assertion that the
target and benchmark markets are similar because each offers ``mobile
functionality.'' SXM RPFF at 19; Orszag AWDT ] 32. The Judges find this
criticism to be without merit.\150\ The majority of Sirius XM listening
occurs in the car, Meyer WDT ] 21 n.5, and the improved mobile
functionality of interactive streaming through ``connected cars'' and
more complete cellular coverage allows listeners to access streaming
services in the car. SE PFF ]] 156-159 (and record citations therein).
Thus, the Judges do not agree with Sirius XM that Mr. Orszag's reliance
on the interactive services' mobile functionality is superficial;
indeed, the issue of whether their respective mobile functionalities
are substitutional for each other bears on the Opportunity Cost/ECPR
analysis undertaken by Professor Willig.
    \150\ The Judges address the value of Sirius XM mobile
functionality elsewhere in this Determination.
    Nonetheless, the Judges decline to adopt Mr. Orszag's reliance on
evidence he claimed suggested a ``growing'' use of streaming services,
including interactive services, in the car. Orszag AWDT ] 39(C).
Although the evidence on which he relied is somewhat supportive of this
point, it is not sufficiently persuasive. The Judges are reluctant to
adopt or extrapolate from potential market trends or rates of change
and use them as a basis for a fixed five-year rate. As the Judges have
noted on other occasions, the adoption of market predictions is a
fraught exercise. More probative in the Judges' opinion are the results
from the survey experts who have appeared for both parties. These
experts have attempted to measure present intentions regarding the
substitutability of interactive services (and other services) for
Sirius XM. While their surveys yield starkly different results when
attempting to elicit whether Sirius XM listeners would switch to
interactive services if Sirius XM were nonexistent or too expensive,
none shows anything close to a 1:1 substitutability of interactive
services for Sirius XM.\151\
    \151\ The Judges analyze these survey results in detail, supra,
sections VI.B.1-VI.B.2.
    The survey results highlight a related criticism by Sirius XM of
Mr. Orszag's ratio equivalency approaches. Sirius XM correctly argued
that the economic rationale that supports a ratio equivalency approach
requires ``significant competition, or a high cross-elasticity of
demand, between Sirius XM and subscription services. . . . [A] limited
degree of head-to-head competition . . . will not suffice.'' Shapiro
CWRT at 12; see also Web IV, 81 FR at 26353; 4/26/17 Tr. 1198 (Orszag).
    In Web IV, the Judges stated that the ratio equivalency approach
might be appropriate if the record reflected that
functional convergence and downstream competition for potential
listeners indicate a sufficiently high cross-elasticity of demand as
between interactive and noninteractive services, provided the
noninteractive subscription rate is reduced to reflect the absence
of the added value of interactivity[.]
81 FR at 26353. In the present case, Mr. Orszag did not provide either
qualitative or quantitative evidence of a sufficiently high cross-
elasticity. In fact, it is noteworthy that even the survey results
reported by SoundExchange's own survey witnesses, Professors Ravi Dhar
and Itmar Simonson, indicated that there is no such high
substitutability between subscribership to interactive services and to
Sirius XM. These survey conclusions negate any complete or overwhelming
ratio equivalency Mr. Orszag has posited. Moreover, even Professor
Willig, another SoundExchange economic expert, relied on and adopted
Professor Dhar's survey, which revealed a substitutability of
interactive services for Sirius XM at significantly less than 1:1. See
Willig WDT ] 41.
    Sirius XM also challenged SoundExchange's predicate that there is
``increasing convergence of the interactive services and Sirius XM''
because of ``some `lean back' functionality'' offered by the
interactive services (in the form of pre-programmed playlists). Sirius
XM noted that Mr. Orszag acknowledged on cross-examination that, if the
rate-setting exercise were based solely on his posited convergence, any
increased use of playlists by interactive services would suggest that
interactive services were becoming more like noninteractive services,
rather than vice versa. If any purported convergence is in the
direction of lean-back service, then interactive services' rates should
be falling in an effectively competitive market, rather than
noninteractive or satellite services' rates increasing. 4/26/17 Tr.
1191-92 (Orszag).
    Sirius XM's criticism in this regard is well-taken. There is
insufficient evidence in the record to show that interactive services'
royalty rates have fallen in response to any asserted increase in
listener use of playlists. Indeed, as Sirius XM correctly noted,
[REDACTED]. See, e.g., 5/16/17 Tr. at 3939 (Harrison); 5/15/17 Tr. at
3836 (Walker).\152\
    \152\ Playlists could engender price competition. As the Judges
noted in Web IV, services could lower royalty rates with playlist
steering. Further, the possibility of steering could result in lower
industrywide rates without any actual steering taking place. See Web
IV, 81 FR at 26367. In the present case, there is no evidence of any
such price competition through playlist-based steering in the fully
interactive market.
    Ultimately, the Judges place no weight on the alleged corroboration
of the Mid-tier Agreements identified by Mr. Orszag, for several
reasons. First, as
[[Page 65248]]
a SoundExchange industry witness testified, UMG requires [REDACTED].
Harrison WDT ] 20 (``Even if mid-tier subscription services succeed in
drawing some consumers away from poorly-monetized free ad-supported
streaming services, there is also a danger that they could to a degree
cannibalize the premium on-demand subscription services. [REDACTED].
    Second, the mid-tier services include interactive features which
the record companies recognize are valuable to subscribers. Id. Absent
evidence in this record of an interactivity adjustment specifically
related to the valuable but limited interactive functionality of the
mid-tier services, the probative value of the mid-tier rates in this
proceeding is compromised.
    In sum, the Judges agree with Sirius XM that the record does not
provide sufficient evidence to support Mr. Orszag's ratio equivalency
approaches to rate-setting in this proceeding.\153\
    \153\ Therefore, Mr. Orszag's attempted steering adjustments are
moot with regard to his approaches. The applicability of those
adjustments, vel non, is addressed in connection with the
establishment of effectively competitive rates elsewhere in this
Determination. Also, because Mr. Orszag did not present the mid-tier
royalties as benchmarks in their own right, but rather as
corroborative evidence supporting his (now rejected) ratio
equivalency approach, the Judges do not accept Mr. Orszag's use of
mid-tier royalties as corroborative or probative.
VII. SDARS Performance License--Sirius XM Proposal
    In its specific proposed rate regulations, Sirius XM advocated a
single royalty fee--8.1% of ``Gross Revenues.'' See Second Amended
Proposed Rates and Terms of Sirius XM . . . at Sec.  382.12(a) (2d
APR). However, more broadly, Sirius XM proposed a rate range of 8.1% to
11% of relevant revenue, which it claims is consistent with the
evidence. 2d APR at 1. The existing rate, for 2017, is 11%.
    Sirius XM's expert witness, Professor Carl Shapiro, analyzed three
possible starting points for setting the performance royalty rates in
this proceeding. Professor Shapiro began with an analysis of the
existing rates. He also analyzed two potential benchmarks: Direct
licenses negotiated between Sirius XM and 498 Indie record labels and
the rates determined by the Judges for noninteractive digital
performances over the internet (webcasting).\154\
    \154\ Professor Shapiro did not label the existing rate as a
``benchmark'' per se. Rather, he opined that the existing ``11
percent of revenue rate that Sirius XM will pay in 2017 can be
viewed as an upper bound on the reasonable royalty level for the
2018-2022 period.'' Shapiro WDT at 34. The Judges consider Professor
Shapiro's use of the existing rate as an ``upper bound'' is
functionally similar to a use of that rate as a ``benchmark.'' That
is, he is urging a similarity between: (1) The description of the
SDARS market as it was presented to the Judges in SDARS II in 2012,
and the rates that were set in that Determination (the de facto
benchmark); and (2) the description of the SDARS market (the target
market) as it has been presented to the Judges in this 2017
A. Current Rates
    Professor Shapiro noted that the current statutory rate is 11% of
``Gross Revenues,'' as defined by the relevant regulations. See 37 CFR
part 382, subpart B. The Judges configured the SDARS rates for the
period 2013 to 2017 to increase from 9% to 11% over the five-year
period. Before recommending adoption of the extant rate for the ensuing
rate period, Professor Shapiro analyzed the state of the music industry
to determine whether any changes in the marketplace might warrant a
deviation from the current rate. See Shapiro WDT at 27. Evidence in
this proceeding overwhelmingly supports a finding of increased use of
streaming, both interactive and noninteractive, as the preferred method
of ``consuming'' music. Professor Shapiro's testimony was no exception.
Id. at 28. As Professor Shapiro noted, in 2012, streaming accounted for
approximately 12% of record industry revenues; whereas in the first
half of 2016, streaming accounted for 43% of record industry
revenues.\155\ Id. Analogously, Sirius XM's subscribership grew from
approximately 24.9 million subscribers in 2014 to 28.3 million
subscribers in 2015. Id. at 29.\156\ This growth in subscribers
increased satellite radio's share of music industry revenues during the
period from [REDACTED]% to [REDACTED]%. Id. at 28, Fig. 5.
    \155\ Professor Shapiro defined revenue from streaming services
as that derived from subscription and on-demand services as well as
webcasting. Music industry revenues included those streaming
services, physical sound recording sales, digital downloads,
synchronization royalties, and satellite radio. Shapiro WDT at 28.
    \156\ Sirius XM predicts [REDACTED] during the upcoming rate
period from an estimated [REDACTED] million subscribers in 2018 to
[REDACTED] million subscribers in 2021. Shapiro WDT at 29.
    Professor Shapiro proposed continuing the current percent-of-
revenue rate structure. He concluded that, when using percent-of-
revenue rates, any increase in Sirius XM's relevant revenue would
redound to the benefit of the record companies obviating a need to
change the rate. See Shapiro WDT at 29-30.
    He further argued that the relevant starting consideration for the
Judges would be the rate that would emerge in an effectively
competitive marketplace. 5/3/17 Tr. 2479-80 (Shapiro). Professor
Shapiro asserted that Sirius XM's overall profits would be irrelevant
to the negotiation. Shapiro WRT at 51-52. He opined that, in an
effectively competitive market, the negotiating parties would look only
to the licensee's ``contribution margin''; that is, ``the percentage of
Sirius XM's receipts from a subscriber . . . that drops to their bottom
line.'' Id. This contribution margin is the measure of sales revenue
available for fixed costs and profit after paying variable costs. See
Lys WDT ] 83. According to Professor Lys, Sirius XM includes in
variable costs [REDACTED]. Id. ] 85. Professor Shapiro and Professor
Lys agree that Sirius XM's contribution margin has remained
``remarkably consistent'' over time. See id. ] 87; Shapiro WRT at 5.
    Professor Shapiro focused on the stability of the Sirius XM
contribution margin to argue for a like stability in royalty rates.
Countering that proposition, Professor Lys looked at an economic
bargaining model and concluded that with greater overall profitability,
Sirius XM and any licensor would negotiate to divide those overall
profits, which would result in a higher percentage royalty rate.
    Neither expert's opinion in this regard, however, is persuasive.
Professor Lys may well be correct that record companies, given their
``must have'' status, i.e., in the absence of effective competition,
would seek in unregulated market negotiations to appropriate a portion
of the additional profits (through a rate increase in addition to the
automatic increase from a larger pool of revenue), notwithstanding that
the profits accrued via Sirius XM's scale and growth rather than
through an increase in the contribution margin. On the other hand, as
discussed elsewhere in this Determination, the growth of Sirius XM's
profits allows it to compensate the record companies for the
opportunity costs the latter incur when licensing to Sirius XM. But
neither of these factors is relevant to the appropriateness of adopting
the extant rate in the forthcoming rate period.
    SoundExchange opposed reliance on the current, SDARS II royalty
rates, asserting that the current rates do not capture the effect of
the expansion of music streaming on the labels' opportunity cost.
SoundExchange contended that Professor Shapiro's analysis of current
rates fails to acknowledge or address changes (1) in opportunity cost,
(2) in Sirius XM's financial performance, (3) in the upstream market
for digital sound recording rights,\157\ and (4) in current
[[Page 65249]]
circumstances as opposed to those prevailing at the time of the SDARS
II determination.
    \157\ For this third point of criticism, SoundExchange focused
on the direct licenses Sirius XM negotiated with Indie labels. The
criticism is better directed at the direct license benchmark and the
Judges will discuss it in that portion of the Determination, section
    SoundExchange's arguments regarding opportunity cost relied on the
assumption that Sirius XM and streaming services are closely
substitutable for one another. However, that assumed close
substitutability is contradicted by the survey results as analyzed by
the Judges, supra, in connection with Professor Willig's opportunity
cost analysis. SoundExchange also does not take into account Sirius
XM's unique position of being the only satellite radio provider--
resulting in a remarkable growth in subscribers--as well as the fact
that changes in Sirius XM revenues have resulted from additional
factors, i.e., lower non-music content costs and lower royalty rates in
negotiated direct licenses.\158\ Thus, to the extent discussed above,
changes in the overall market do militate against using current rates
as an appropriate starting point.
    \158\ The Judges do acknowledge Sirius XM's increased
profitability in a different context, i.e., whether Sirius XM should
contribute to the legitimate opportunity costs incurred by the
record companies without disruption that would threaten the
viability of Sirius XM. See infra, section X.D.
    Moreover, the current rates as set in SDARS II were a function of
the deficiencies in the proffered evidence in that proceeding, evidence
that, by comparison, made the then extant rates a relatively superior
guide to an appropriate rate. The Judges were dissatisfied with a
benchmark derived from licenses in the interactive streaming business.
Further, the Judges found it necessary to allow for a downward
adjustment (within the zone of reasonableness) to account for the
enormity of Sirius XM's satellite launch and replacement costs. See
SDARS II, 78 FR at 23069. SoundExchange argued that the ``incredible
financial success'' enjoyed by Sirius XM during the current license
period obviates the need for consideration of Sirius XM's costs of
doing business for the license period at issue in this proceeding. See
Lys WRT ] 56. The Judges agree; in fact, that financial success is a
basis for increasing the royalty rate in this proceeding, as indicated
    For the reasons highlighted by SoundExchange and its experts, the
Judges will not use the extant rates as a starting point (or benchmark
or upper bound) for determination of appropriate rates for the period
2018 through 2022. The SDARS II rates were derived on a record much
less robust than the record in this proceeding. The participants in
this proceeding have presented sufficient facts and analysis to inform
the Judges and to lessen the value of the current rates as a desired
starting point for analysis in these changed circumstances.
B. Current Direct Licenses Negotiated by Sirius XM
    Professor Shapiro proposed a benchmark derived from direct licenses
Sirius XM has negotiated in the market at issue in this proceeding,
i.e., the satellite radio music streaming (upstream) market. In 2012,
when the Judges established rates for the 2013 through 2017 rate
period, direct licensing was in its infancy, with approximately 100
direct licenses executed at the time of the determination. Shapiro WDT
at 34. By 2016, Sirius XM had negotiated almost 500 direct licenses
with record labels. Id. at 35. Because of its direct license effort,
Sirius XM has access to approximately 23,000 music catalogs containing
as many as 5 million tracks, or 6.4% of the tracks on the Sirius XM
playlists. Shapiro WDT at 35 (citing White WDT). Professor Shapiro
promoted the direct licenses as ideal benchmarks, asserting that they
represent market outcomes involving the same sellers (record labels),
the same buyer (Sirius XM), and the same rights (digital performance of
sound recordings) and effectively competitive conditions for the
negotiations. Id. at 37.
    Professor Shapiro reasoned that these negotiations reflect an
effectively competitive marketplace because Sirius XM controls such a
small share of the record industry's overall revenues (approximately
[REDACTED]%). See Shapiro WDT at 37. Measuring Sirius XM's royalties
against the entirety of music industry revenues, however, ignores the
fact that Sirius XM dominates the market for paid services that
listeners use in a vehicle. As the primary alternative to (non-royalty
paying) terrestrial radio in cars, Sirius XM in fact wields tremendous
bargaining power, which would tend to drive down the negotiated rates.
Professor Shapiro contended that, in fact, direct license rates
negotiated in an unregulated market would be lower because based on
recent trends, he believes the statutory license rates act as a
``magnet'' to pull directly negotiated rates up to the statutory rates.
Id. at 45.
    Professor Shapiro's endorsement of direct licenses as a benchmark
ignored the difficulties inherent in determining the effective royalty
rates the parties negotiated. With the direct licenses, Sirius XM
receives the same rights it would under the statutory license and
additional benefits, such as a relaxation of the statutory performance
complement rule, allowing Sirius XM to rely more heavily on the (lower
priced) directly-licensed tracks. Id. at 35-36. Licensors also benefit
from consideration negotiated in direct licenses that is not available
under a statutory license. Licensors might receive more exposure for
their recordings, might benefit from direct payment of both recording
and artist royalties, and could avoid the SoundExchange administrative
fee. No expert in this (or any similar) proceeding has attempted to
value the considerations behind the headline percent-of-revenue rates
in direct licenses, let alone determine which party enjoys the net
    Looking at the upstream market (record labels to streaming
services), Professor Shapiro anticipated more negotiation of direct
licenses influenced by the noninteractive streaming services' ability
to ``steer'' listeners to a particular catalog of music. Id. at 30. As
Professor Shapiro noted, in the webcasting market, the availability of
steering resulted in negotiation of direct licenses with headline rates
below the statutory rates based on the potential benefits of greater
streaming frequency of the labels' music. Id. at 30.
    SoundExchange was critical of Professor Shapiro's reliance on
direct licenses primarily because more recent direct license agreements
have omitted steering incentives or have included anti-steering
alternatives that recognize the prospect of steering but muddy the
analytical waters with regard to the effect steering might have on
their negotiated rates. Even Professor Shapiro conceded that he could
not ``quantify the value of steering.'' 4/20/17 Tr. 488 (Shapiro).
Furthermore, the direct licenses involve exchanges of consideration
apart from the headline royalty rate that no party has attempted to
    \159\ The record labels also derive benefit from the direct
licenses. See Shapiro WDT at 36. Notably, Sirius XM is able to
distribute both the label's share and the artists' share of
performance royalties directly to the contracting label. Sirius XM
provides administration of the royalties without charging the fee
that would be payable to SoundExchange under the statutory scheme.
Under the direct license agreements with Sirius XM, some licensors
also benefit from a more generous methodology for calculating the
label's royalty pool. Id.
    The Judges do not accept Sirius XM's direct licenses as
sufficiently probative of the relevant market to accept them as a
meaningful benchmark. Direct licenses cover only a small portion of the
sound recordings on Sirius XM's playlists.
[[Page 65250]]
They are uninformative of any effect of steering on royalty rates
because none of them contain steering guarantees or economic incentives
to promote (or avoid) steering. There is no basis for the Judges to
segregate consideration in these licenses that is properly attributed
to elements that are unavailable under the compulsory license.
C. Web IV Rates
    Professor Shapiro offered as a final benchmark the rates
established by the Judges in Web IV. The Judges used benchmarks in Web
IV, including direct licenses,\160\ and considered interactive market
(non-statutory) negotiated direct license rates to determine the Web IV
rates. See Shapiro WDT at 49. Professor Shapiro converted the Web IV
per-performance rate of $0.0022 to derive a percentage-of-revenue rate
applicable in this proceeding of 8.1%. Id. at 55. Professor Shapiro
used a figure of 469 performances per subscriber per month for his
conversion. Id. at 54.\161\
    \160\ Professor Shapiro opined that the direct licenses, such as
the Pandora/Merlin agreement, ``reflected the forces of competition
at work,'' namely the leveling power of steering. Shapiro WDT at 49.
    \161\ In a similar exercise, Professor Willig used a weighted
average figure of [REDACTED] performances per subscriber in his
calculation of creator compensation cannibalization (opportunity
cost). The higher opportunity cost would result in a higher
percentage-of-revenue rate. See Willig WDT at B-7.
    Anticipating questions regarding whether webcasting and satellite
radio are too different to warrant this benchmark, Professor Shapiro
analyzed the Web IV benchmark to resolve the differences. According to
Professor Shapiro, there are two key differences to examine. First is
the possible difference between a label's full marginal cost of a
Sirius XM satellite performance and a webcast performance.
Specifically, Professor Shapiro defined the marginal cost difference,
if any, as one of relative promotional or substitutional effects.
Second, Professor Shapiro looked at differences in the ability to steer
as between Sirius XM and a webcaster. Noting that Sirius XM relies on
human programmers while webcasters rely more heavily on algorithms,
Professor Shapiro felt Sirius XM might be more able to steer without
losing listeners. On the other hand, he noted that webcasters (using
Pandora as an example) have the ability to and the practice of allowing
listeners to create individualized ``stations'' giving Pandora greater
flexibility to steer without alienating listeners. See Shapiro WDT at
    In the end, Professor Shapiro concluded that Sirius XM and
webcasters are ``quite comparable along both dimensions.'' Shapiro WDT
at 50. When he combined the favorable comparison of satellite radio and
webcasting with the fact that the sellers in both markets are the same,
the rights at issue are the same, and that the Web IV benchmark
accounts for the forces of competition, ``it becomes clear that the Web
IV benchmark is a very good benchmark for rate setting in this
proceeding.'' Id.
    Sirius XM witness, Steven Blatter, detailed anecdotal evidence of
the promotional effects of sound recording plays on Sirius XM. See
Written Direct Testimony of Steven Blatter, Trial Ex. 5, passim
(Blatter WDT). Mr. Blatter touted Sirius XM's subscription model as
supportive of its ability to broaden the listening (and presumably
consumption) habits of its subscribers. Freed of the commercial demands
of ad-supported radio, Mr. Blatter contended, Sirius XM can cultivate a
broader audience than the ``Top-40'' stations. Listeners to Sirius XM's
curated playlists and niche channels thus discover music that might
otherwise have gone unnoticed. Id. ] 2. Mr. Blatter recited ``thank-
you'' letters from artists and labels, trade publication reporting and
analysis, and sales statistics on selected titles as evidence of Sirius
XM's promotional value to licensors. In addition to artist testimonials
and press coverage, Mr. Blatter noted that ``many musicians and record
labels'' grant Sirius XM waivers of statutory limitations relating to
frequency of play under a statutory license (i.e., the ``sound
recording performance complement'') in order to enjoy the benefits of
promotion on Sirius XM. Id. ] 36.
    Countering Mr. Blatter's assertions, SoundExchange expert, Dr.
George Ford, opined that promotional effects of a particular platform
are irrelevant to the Judges' task in this proceeding. See Written
Direct Testimony of George S. Ford, Trial Ex. 23, at 3-4 (Ford WDT).
Dr. Ford pointed out most notably that no ``broad inter-platform
analysis'' of promotion and substitution is in evidence. Id. Further,
he asserted promotional effect is meaningless unless it is net of
substitutional effects. In the current music marketplace, Dr. Ford
asserted, given the dramatic decline in sales of permanent music media,
a streaming service's promotion of CD sales and downloads is outdated.
Id. at 4. Professor Willig actually performed econometric analyses
looking at all streaming services (including Sirius XM) and found a net
substitutional effect when compared to permanent sales. Willig WDT ]]
24-27. According to Professor Willig, the substitution of streaming for
permanent sales contributed to a dramatic drop in creator compensation,
meaning the opportunity cost to artists and labels of streaming is
significant. Id. ] 30.
    Mr. Orszag likewise disputed Professor Shapiro's reasoning relating
to the relative ability to steer in satellite radio and webcasting. As
Mr. Orszag reasoned, the Judges relied on direct licenses and their
steering provisions to make an adjustment to bring the webcasters'
marketplace in line with a hypothetical effectively competitive market.
See Orszag WDT ]] 64-66. Direct licenses negotiated by Sirius XM are
[REDACTED], however. Id. ] 67. Nor is there any record evidence of any
actual steering by Sirius XM. As the Judges noted elsewhere in this
Determination, [REDACTED].
    The most salient criticism of Professor Shapiro's Web IV benchmark
came from Professor Willig. Professor Willig discounted use of the Web
IV rates, specifically the Pandora noninteractive rates, for various
reasons, but the most telling was his uncontradicted assertion that not
even [REDACTED] uses the statutory rates. After the Web IV
determination, [REDACTED] negotiated direct licenses with [REDACTED].
Using the renegotiated rates as a benchmark, Professor Willig
calculated the SDARS rate resulting from Professor Shapiro's
methodology would be [REDACTED]% of revenue, approximately [REDACTED]
the 8.1% of revenue proposed by Professor Shapiro. See Willig WRT ] 57.
    The Judges are troubled by the implicit assumption in Professor
Shapiro's use of the Web IV per play rate, given that Sirius XM, as
opposed to noninteractive streaming, is listened to predominantly in
the car. As Mr. Orszag testified, any per play analysis implicitly
starts with the questionable assumption that each play has an
equivalent value in both distribution channels. Orszag WRT ] 53.
Further diminishing the value of a per play analogy, the Judges note
that the parties' use of a percent-of-revenue form of royalty is
inconsistent with the idea that there is a single per play value that
cuts across all distribution channels
    Further, the Judges agree with Mr. Orszag that there is no valid
reason--and certainly no proof in the record--that would permit the
Judges to conclude or presume an equal per play value for a Sirius XM
play--usually in the car--and a play of a noninteractive song. In fact,
the Judges find that, as a matter of common sense, there is likely
greater utility in a sound recording played in an automobile. A driver
(in particular) has a limited set of options for entertainment, given
his or her need to remain attentive to the road and to
[[Page 65251]]
traffic. In the car, therefore, radio listening is a scarce form of
entertainment and therefore more valuable product than it is elsewhere,
where it competes with all other forms of utility and diversion (market
and non-market).\162\
    \162\ Sirius XM asserted that Mr. Orszag did not undertake any
empirical analysis in support of this argument. SXM RPFF ]] 273-274.
However, Mr. Orszag explained sufficiently that this value is a
particular form of ``access'' value, whereby the driver knows he or
she has the option of listening to music on Sirius XM in the car, a
particular value given the limited alternatives for entertainment
and diversion behind the wheel. See SE PFF ]] 1228-1229 (and record
citations therein). Moreover, the limited nature of alternatives for
entertainment and diversion for a driver are matters of common
knowledge, and that point is not dependent upon expert testimony.
Further, because Sirius XM advanced the argument that the per play
values are equivalent across these two distribution channels, it
should have proffered evidence to support the assertion that
consumers value access and per play values provided by Sirius XM the
same as they value such benefits when provided by a noninteractive
service, given the greater use of Sirius XM in the car.
    The participants have not provided evidence sufficient for the
Judges to reach any conclusions regarding a conversion of the Web IV
per-play rates to a Sirius XM percent-of-revenue rate. Even if the
parties had provided sufficient evidence to make the conversion, the
Judges are unconvinced that the characteristics of webcasting and
satellite radio are sufficiently similar to transfer, without
adjustment, the royalty rate from one platform to the other.
D. Lenski Survey Data
    Sirius XM engaged Mr. Joe Lenski of Edison Research to collect
empirical data regarding the sources of Sirius XM satellite radio
listeners and to evaluate where those listeners might turn for music
consumption if Sirius XM were unavailable. See Written Direct Testimony
of Joe Lenski, Trial Ex. 7, 2 (Lenski WDT). Sirius XM also asked Mr.
Lenski to develop similar data for Pandora listeners.\163\ See id. Mr.
Lenski conducted a national random digit dial telephone survey, using
both landline and cellular telephone contacts (Lenski Survey). He
employed a survey methodology ``widely recognized as the most reliable
form of survey research and . . . used by most major polling
organizations. . . .'' Id. at 3. The survey queried 983 Sirius XM
listeners and 1,323 Pandora listeners. Of the total respondents, 350
identified themselves as listeners to both Sirius XM and Pandora. The
surveyors asked respondents in the two groups (Sirius XM and Pandora)
separate sets of questions. Respondents identifying as listeners to
both Sirius XM and Pandora answered both sets of questions.
    \163\ At the time of the Lenski Survey, Pandora had not yet
launched its fully interactive subscription service. It operated
only lean-back or Mid-tier services that were not fully interactive.
    A large majority--62%--of Sirius XM listeners responded that they
migrated from terrestrial radio, with 20% of respondents answering that
before Sirius XM they listened to ``CDs or your own music downloads.''
See id. at 5. Online streaming services, AM/FM stations streaming on
the internet, and interactive streaming services in the aggregate
accounted for 7% of Sirius XM's current listeners. Id. As for
alternatives to Sirius XM, survey respondents indicated they would turn
to terrestrial radio (74%), CDs or music downloads (65%), online
streaming services (49%) and interactive streaming services (32%).\164\
Id. Once survey respondents identified all possible alternatives to
Sirius XM, the surveyors asked respondents to distribute their possible
alternatives by frequency. In this cut, a plurality of respondents'
listening time, 40.8%, would be to terrestrial radio. Id. at 6. CDs and
digital downloads would capture 23.1% of former Sirius XM listening
time. In the aggregate, 22.1% of listening time would be to
noninteractive (14.3%) and interactive (7.8%) streaming services.
    \164\ At this juncture, listeners could choose more than one
potential alternative to Sirius XM; hence the percentages exceed
100%. Notably, 28% of survey respondents answered they would listen
to less audio overall if Sirius XM were unavailable. See Lenski WDT
at 5.
    By contrast, Pandora listeners reported migrating slightly more
frequently from ``CDs or your own music'' (35%) than from terrestrial
radio (33%).\165\ As alternatives, if Pandora were no longer available,
survey respondents chose CDs or music downloads (67%), terrestrial
radio (59%), interactive streaming services (47%), noninteractive
streaming services (46%), and Sirius XM (23%).\166\ When asked to
allocate their time among the alternatives, Pandora listeners allocated
their listening time to CDs or music downloads (26.3%), terrestrial
radio (24.4%), interactive streaming services (16.6%) and other
noninteractive streaming services (11.7%). Id. at 7.
    \165\ Sixteen percent of Pandora respondents answered that their
Pandora listening was new listening time, not diverted from other
sources. Lenski WDT at 6-7.
    \166\ As they did with Sirius XM, the surveyors first
established all alternatives (adding to more than 100%) before
having respondents allocate their time by preference. Id. at 6.
    These survey results showed that Sirius XM competes most directly
with terrestrial radio, whereas Pandora's noninteractive service
competes almost equally with CDs and downloads, interactive streaming
services, and terrestrial radio. Professor Shapiro applied these
conclusions to support his assertion that Sirius XM is mostly
substitutional for terrestrial, non-royalty paying, radio. See Shapiro
WRT at 14. In other words, Sirius XM is not cannibalizing creator
compensation from other sources; it is augmenting creator compensation
with an alternate source of royalties. Id. at 37. Professor Shapiro
pointed out that, using the Modified Dhar Survey, Professor Farrell
calculated a much lower opportunity cost than Professor Willig, viz.,
$1.35 per subscriber per month as compared with $2.55 per subscriber
per month. See id. The Farrell conclusions, he testified are ``notably
closer'' to the results Professor Shapiro obtained using the Lenski
Survey. Id.
    Professor Dhar criticized the Lenski Survey as having ``no
scientific value.'' Dhar WRT ] 9. Professor Dhar criticized the
methodology, the response order, and the word choices in the Lenski
Survey. See Dhar WRT passim. In essence, Professor Dhar concluded the
Lenski Survey could not be of any value in reflecting ``marketplace
reality.'' See, e.g., id. ] 16. The thrust of the Dhar criticisms
revealed the differences in the assignments the parties gave their
survey experts. Sirius XM asked Professor Lenski to gather listener
preference information, whereas SoundExchange tasked Professor Dhar
with looking at a defined, limited marketplace.
    Professor Willig acknowledged that the ``the structures of these
two surveys [Dhar and Lenski] are fundamentally different: they ask
fundamentally different questions.'' Willig WRT ] 41. Professor Willig
also criticized the Lenski Survey because it purported to measure
listeners' assessments of their use of time whereas the Dhar Survey
measures listeners' assessment of their spending, or more precisely,
their willingness to pay. See Willig WRT ]] 13, 46. Professor Willig
asserted that the latter would be a more appropriate measure to
determine creator compensation cannibalization. Id. Professor Willig,
at bottom, criticized Professor Shapiro's reliance on the Lenski Survey
data to evaluate relative substitutional effects of webcasting and
satellite radio because the Lenski Survey did not give Professor
Shapiro a basis to quantify the effects. Professor Shapiro testified in
response to that criticism that, nonetheless, ``switching behavior
that's not price-based is quite useful in terms of how [economists] . .
. see things,'' yet he cautioned that ``I would accept that because Mr.
Lenski is
[[Page 65252]]
asking about where would you move your listening, that could give a
different answer than what would you subscribe to if Sirius XM were
more expensive.'' 4/20/17 Tr. 3765-76 (Shapiro).
    The Judges accept that the Lenski Survey and the Dhar Survey (and
even the Modified Dhar Survey) were not aimed at establishing the same
empirical evidence. The Judges do not agree with Professor Dhar's
criticism of the Lenski Survey methodology. Without parsing every
question in the Lenski Survey for ambiguity or order bias, the Judges
also accept that both the Lenski Survey and the Dhar Survey were
faulty. Those surveys are, however, sources of empirical evidence
available in this proceeding. The Modified Dhar Survey resulted in
adjustment of Professor Willig's analyses and conclusions. The Lenski
Survey supported Professor Shapiro's analyses and conclusions. But in
addition, the Judges understand the Lenski Survey to be of limited use
in comparing the opportunity cost analyses conducted by Professors
Willig and Farrell, as discussed supra.
VIII. GEO Music Rate Proposals for PSS and SDARS
A. Rate Structures and Proposals
    Mr. George D. Johnson testified \167\ on behalf of GEO Music and
proposed that the Judges bridge what he described as a ``gap'' in
creator compensation. See 5/2/17 Tr. at 2203, 2209-10 (Johnson). The
premise upon which GEO relied is that each performance of a copyrighted
work should be compensated. See (Corrected) Testimony of George D.
Johnson (GEO), Trial Ex. 60, at 24-25 (Johnson CWDT). GEO acknowledged,
however, that for some digital services, including the two services
seeking licenses in this proceeding, measurement of individual
performances might not be possible. Consequently, GEO sought rate
structures that could provide
    \167\ As adjuncts to his testimony, Mr. Johnson proffered
numerous exhibits. Sirius XM and Music Choice filed objections to
GEO's exhibits, citing lack of foundation, hearsay, and relevance
objections. The Judges grant those objections in their entirety. The
GEO exhibits are not admitted for the truth of the matters asserted
therein, but are nonetheless permitted to remain in the record as
illustrative of Mr. Johnson's testimony.
a livable music royalty rate . . . [with which creators] can be sure
in our royalty payments, real payments, that are guaranteed, at a
rate we would get if there were no `shadow' of a compulsory license
. . . .
Id. at 5; see id. at 14 (``to know that they are secure in their
royalty income . . . .'').
    The solutions GEO proposed appeared to arise from a per-work
formula.\168\ He began his analysis with reference to the history of
``mechanical'' royalties paid to license musical works.\169\ Mechanical
royalties for physical phonorecords and permanent digital downloads
have and continue to be structured on a per-unit basis. To capture a
value he considered equivalent to a per-unit royalty for streaming
services, Mr. Johnson proposed four different rate structures: A per-
subscriber rate, a percentage of revenue rate, a per-play rate, and a
permanent download rate.
    \168\ Mr. Johnson has advocated in each of his appearances
before the CRB a holistic approach to licensing music performances.
See id. at 2209. In his approach he asked the Judges to take into
account royalties for all uses of musical works embodied in sound
recordings: Royalties for the publishers, songwriters, record
companies, and artists.
    \169\ Section 115 of the Copyright Act creates the compulsory
license to make and distribute phonorecords of musical works. See 17
U.S.C. 115; 106(1) and (3) (exclusive right to reproduce in copies
or phonorecords and to distribute reproductions of musical works).
The definition of phonorecords has evolved to include digital
reproductions of embodied musical works.
    GEO proposed a per-subscriber SDARS rate ranging from $ 4.96 per
subscriber per month in 2018 to $ 5.58 per subscriber per month in
2022. GEO would have this rate apply to all subscribers except those
that receive channels with no, or incidental, music content and free
trial period subscribers (limited to 30 days royalty free). Proposed
Rates and Terms of George D. Johnson . . . at 10 (GEO Rates). GEO
proposed PSS per-subscriber rates ranging from $ 0.10 in 2018 to $ 0.20
in 2022. Id. at 14.
    GEO proposed a SDARS percentage of revenue rate within a ``current
marketplace'' range of 25% to 40% of ``Gross Revenues.'' He proposed
defining ``gross revenues'' in a manner similar to the current
regulations, but to include payments or payments in kind to key
executives or shareholders. See id. at 12. For PSS, GEO proposed using
the same definition of ``gross revenues'' and calculating the royalty
rate at 45% of gross revenues.
    For per-play rates for SDARS, GEO relied on ``anonymous, but
actual'' Sirius XM royalty rates and adjusted those rates by varying
the percent-of-revenue target. Id. at 13.
    As an additional revenue stream for both the services and the
copyright owners, GEO proposed requiring both Sirius XM and Music
Choice to create a ``BUY button.'' In this proposal, GEO envisioned
listeners acquiring (1) a permanent download to the listener's device
of choice, (2) a ``cloud locker'' stored sound recording, or (3) a
permanent download to a purchased content locker or paid locker
service.\170\ GEO proposed a royalty range of $1.00 in 2018 to $2.50 in
2022 per purchase. Id. at 15.
    \170\ GEO did not clarify how a paid locker service or purchased
content locker service might be different from a ``cloud locker.''
    The economic underpinnings of Mr. Johnson's proposals are that
streaming and broadcasting music, i.e., the access models of music
consumption, have substituted for (``cannibalized'') music sales. With
this shift in music consumption, Mr. Johnson opined, users and
exploiters of the artists' work have continued to prosper as the
artists' revenue streams have declined. See Johnson CWDT at 36-40.
    Sirius XM did not rebut directly the GEO proposals, but filed
replies to GEO's proposed findings and conclusions. See generally
Sirius XM . . . Reply to George Johnson's Proposed Findings . . .
(Sirius XM Reply to GEO). With one exception,\171\ Sirius XM disputed
all of GEO's proposed findings and conclusions. With respect to all
other proposed findings and conclusions, Sirius XM did not uniformly
dispute the content of GEO's cited material, but argued that the
citations were inapposite or irrelevant to the SDARS/PSS rate
proceeding or without factual or legal support. Id., passim. Sirius XM
argued that GEO's proposals conflated with SDARS the rate
configurations for different licenses, e.g., Phonorecords and
Webcasters, without regard for the differences in rate setting
standards for those configurations and without acknowledging the
separateness of the record evidence supporting those different rates.
    \171\ Sirius XM did not dispute GEO's Proposed Conclusion of Law
number 24, to wit: ``George D. Johnson is an individual pro se
singer/songwriter, music publisher and independent sound recording
creator.'' Sirius XM Reply to GEO at 27.
    Music Choice addressed directly the GEO proposals. Mr. David Del
Beccaro, President and CEO of Music Choice, testified that he could not
parse the GEO proposals. See Del Beccaro WRT at 65. Mr. Del Beccaro
pointed out that the GEO rate proposals lacked explanation,
``benchmark, model, or any other evidence . . . .'' Id. at 66.
    Further, Mr. Del Beccaro took issue with the GEO proposal that
Music Choice be required to offer a digital download service. As Mr.
Del Beccaro observed, the digital performance sound recording license
at issue in this proceeding does not extend to sales of sound
recordings--physical or digital. Id. Music Choice has not licensed the
[[Page 65253]]
rights necessary to sell phonorecords. Further, Music Choice provided
retail sales of physical phonorecords (CDs), a business that did not
require a license from record companies. Ultimately, Music Choice
abandoned that service because it was not profitable. Id. at 66-67.
    The Judges agree with Sirius XM that GEO's proposed rates and terms
are unsupported by record evidence. The Judges also agree with the
Music Choice criticisms of GEO's presentations. GEO's arguments are
primarily policy arguments beyond the scope of this proceeding. The GEO
proposed findings, conclusions, and rate proposals are inadequately
supported in the record.
B. Statutory and Constitutional Considerations
    GEO referred to the constitutional provision giving Congress the
power to provide for copyrights.\172\ He acknowledged that Congress
provided for certain ``exclusive rights'' for copyright holders in
section 106 of the Act. He argued unconvincingly, however, that the
statutory licenses inappropriately infringe on the exclusive rights
Congress created. He also questioned whether the Judges, or their
predecessors whose precedent the Judges consider, were at worst
confiscating, or at best marginalizing, copyright owners' rights by
failing to provide for fair compensation. See Johnson CWDT at 6, 13.
GEO asserted that current statutory royalty rates are ``extremely low
below-market'' rates. Id. at 13.
    \172\ Article I, Section 8, clause 8 of the Constitution gives
Congress the power ``to promote the progress of science and useful
arts, by securing for limited times to authors and inventors the
exclusive right to their respective writings and discoveries.'' U.S.
Const., Art. I, sec. 8, cl. 8.
    GEO made much of the ``full independence'' of the Judges. See,
e.g., Johnson CWDT at 7; 5/3/17 Tr. at 2244 (Johnson). Mr. Johnson
appeared to equate judicial independence for the Copyright Royalty
Judges with disconnection from the dictates of the law. His arguments
failed to analyze the separate licenses created by Congress or the
differing standards by which the Judges must set those rates. By
focusing unduly on ``fair market'' considerations, Mr. Johnson ignores
the policy factors Congress established for certain licenses in section
801(b)(1) of the Act. Further, in every rate setting or rate adjustment
proceeding, the Judges hear testimony from economists and other market
experts to determine a fair rate for each license under the
circumstances extant at each license period.
    Notwithstanding the import of Mr. Johnson's (and other's) evidence
of economic imbalances in the present-day music industry, nothing in
the Constitution or the Copyright Act empowers the Judges to create new
law or fill in legislative ``gaps'' arising by the course of commerce.
Only Congress has that power.
IX. Adjustment for Promotional or Substitutional Effect
    Neither SoundExchange nor Sirius XM proposed an adjustment to the
rates that they advocated to account for any promotional effect.
Compare Shapiro WDT at 56 (``good reason'' to conclude promotional
value from performances on Sirius XM greater than promotional value of
performances by webcasters, ``I am not able to precisely quantify just
how much lower the royalty rate would be, so I make no downward
adjustment to the rate) with Orszag AWDT ]] 97-100 (considered whether
adjustment was required between target market (Sirius XM) and benchmark
market (interactive services) with respect to promotion and concluded
no adjustment necessary).
    Additionally, as the Judges explained in Web IV:
    To the extent that the Judges adopt a rate based on benchmark
evidence, it is not necessary to make additional adjustments to
benchmarks to reflect the promotion and substitution factors. The
Judges hold in this determination, as they have held consistently in
the past, that the use of benchmarks ``bakes-in'' the contracting
parties' expectations regarding the promotional and substitutional
effects of the agreement.
Web IV, 81 FR at 26326
    The Judges have also repeatedly found that relative promotion, not
absolute promotion/substitution, is the relevant factor in their
consideration of statutory rates. See SDARS II, 78 FR at 23066-67
(``Because only the relative difference between the benchmark market
and the hypothetical target market would necessitate an adjustment, the
absence of solid empirical evidence of such a difference obviates the
need for such further adjustment''). Testimony from a SoundExchange
economic expert in the present proceeding re-confirmed the logic of
these conclusions in more formal economic terms. See 5/1/17 Tr. 1827
(Ford); see generally Ford WDT; Written Rebuttal Testimony of George
Ford, Trial Ex. 41 (Ford WRT). In the present case, the parties'
position is consistent with these pronouncements regarding relative
promotion, in that they do not propose a rate adjustment on the basis
of any relative promotional differences.
    Accordingly, the Judges do not adjust the rates they establish in
this proceeding to reflect any hypothetical, absolute or relative
promotional effects arising from performances on Sirius XM.\173\
    \173\ There is anecdotal evidence in the record regarding
promotional effect. The Judges have previously noted the
insufficiency of anecdotal evidence to support a rate adjustment. In
this proceeding, however, they find that issue to be moot given that
the parties' respective experts have not proposed a rate adjustment
to reflect promotional effect.
    Further, as discussed elsewhere in this Determination, the
substitution effects arising from record company licensing of sound
recordings to Sirius XM is a lynchpin for the setting of the rate in
this proceeding.
X. The Itemized Section 801(b) Policy Considerations
    As detailed in this Determination, the Judges find that the 15.5%
of revenue rate arising from the Opportunity Cost approach represents a
market-based rate that, in its entirety, mitigates the complementary
oligopoly effects of certain positive opportunity costs embedded within
it and reflects the parties' existing market power. Further, the record
in this proceeding does not support any adjustment to the resulting
rate to account for performances on Sirius having a promotional or
substitutional effect. Accordingly, the Judges find this 15.5% of
revenue rate to be an effectively competitive rate, and therefore a
``reasonable rate'' under 17 U.S.C. 801(b)(1) before consideration of
the policy factors within that statutory section.
    The Judges now analyze each of the itemized 801(b)(1) policy
considerations to determine whether they should make any upward or
downward adjustment in this proceeding and, if so, the magnitude of any
such adjustment. In this and prior proceedings, the Judges have
concluded that these four factors cannot necessarily be considered
separately from one another. See, e.g., SDARS I, 73 FR at 4094.
Moreover, in the process of identifying the ``reasonable rate'' before
specifically applying these four itemized factors, the Judges may have
already considered issues that overlap with the four factors, such that
any further application of the same considerations would constitute
improper double-counting of those considerations.
    SoundExchange argued that the first three statutory objectives
promote policies that are generally advanced through market
transactions. According to its economic expert, Mr. Orszag, ``market-
based rates are consistent with
[[Page 65254]]
the first three of the 801(b) factors.'' 4/25/17 Tr. 954 (Orszag). If
that were true, then any attempt by the Judges to adjust a market-based
rate would be improper or, to the extent the Judges had already
considered market principles, a form of double-counting, were they to
use those factors again to adjust the rate.
    By contrast, Sirius XM asserted that, as a matter of law, ``it is
well established that reasonable 801(b)(1) rates need not correspond to
market rates.'' SXPFF ] 87 (citing SoundExchange v. Librarian of
Congress, 571 F.3d 1220, 1224 (D.C. Cir. 2009) (any ``claim that
[section 801(b)] clearly requires the use of market rates is simply
wrong''); Recording Indus. Ass'n of Am. v. Librarian of Congress, 176
F.3d 528, 533 (D.C. Cir. 1999) (same).
    Thus, Sirius XM further asserted that in a proceeding governed by
the section 801(b)(1) rate standard, ``market-approximating rates''
must be further evaluated against the Section 801(b)(1) policy
objectives in order to arrive at ``reasonable rates'' that comport with
the statutory command. Id. The Judges do not agree that this
construction of their statutory charge is legally mandated or otherwise
necessary. The Judges understand that they may establish ``reasonable
rates,'' and only thereafter decide whether or how to apply the four
itemized factors. If the Judges find that a market-based rate \174\ is
consistent with a ``reasonable rate,'' they may adopt that rate and
apply the four factors to that rate. And, if the record does not
support a further adjustment based on an application of the four
itemized factors, or any of them individually, then the Judges may
allow their market-based reasonable rate to stand as the new statutory
    \174\ Although the Judges are not required to utilize market-
based rates, they surely are not prohibited from doing so, as
discussed supra.
    As the foregoing analysis of the parties' proposals makes clear,
the Judges have found that the 15.5% rate is a ``reasonable rate''
derived from a combination of market-based opportunity costs, survey
evidence and countervailing considerations. Thus, the Judges do not
consider the four itemized factors in section 801(b)(1) as bearing upon
the reasonableness of the market-based rate they have already
identified as ``reasonable.'' Rather, in this case, the Judges consider
whether these four factors, separately or in combination, require any
policy-based adjustments of the 15.5% rate and whether the Judges have
already incorporated those factors into the analysis that led them to
identify the 15.5% rate.
    Before embarking on an analysis of the parties' separate factor
arguments, the Judges note an overarching theme in many of those
discrete arguments. The parties argue broadly that their costs and
investments are significant (Factor C issues) and that they are
entitled to a ``fair'' income or return (Factor B issues) that is not
disruptive of their businesses (Factor D issues), in order to maximize
the distribution of sound recordings to the public (Factor A issues).
These arguments echo the historical ambiguity in the creation of the
itemized section 801(b)(1) factors and the debate between Messrs.
Nathan and Arnold prior to the adoption of those factors, as discussed
supra.\175\ Because the historic antecedent of the factors is the
traditional public utility rate-setting process, the Judges cannot
easily apply the factors to a determination of rates that is not based
on a rate of return that accounts for specified costs, invested
capital, a delineated rate base and a return on invested capital.
Rather, the arguments in this context are by necessity more directional
in nature. With this caveat, the Judges examine the parties' evidence
regarding the need for any adjustment pursuant to the four itemized
factors in section 801(b)(1).
    \175\ See supra, section II.B
A. Factor A: Maximizing the Availability of Creative Works to the
    SoundExchange construed Factor A as calling for royalty rates that
are sufficiently high to foster the creation of new content, but not so
high as to jeopardize the ongoing viability of a licensee-service
``that has gained acceptance among consumers in the marketplace.'' SE
PFF ]1435. Based on this understanding of Factor A, SoundExchange
asserts that the market-based rates it has proposed do not require
adjustment to satisfy the objectives of Factor A.
    In support of this point, SoundExchange first relied on an
explanation by Professor Willig as to why Factor A is consistent with
his Ramsey Pricing approach. Willig WDT ] 13 (``The defining objective
of Ramsey pricing is the maximization of consumer welfare, and this is
an economic concept fully consistent with the portion[ ] of the Section
801(b)(1) criteria that call[s] for the maximization of the
availability of creative works to the public.''); see also 5/2/17 Tr.
1981 (Willig) (``Ramsey pricing by definition . . . says the price has
got to be high enough to be financially sustainable on the supply side,
but balanced across uses in a way that maximizes consumer welfare.'').
However, because Professor Willig did not identify a proposed rate
under his Ramsey pricing approach, the Judges do not find that this
approach compels a Factor A adjustment.
    Nevertheless, Professor Willig did indicate that his Ramsey Pricing
approach generally demonstrated that the statutory rate should increase
from the present rate of 11%. Because the reasonable market-based rate
identified by the Judges of 15.5% is 41% higher than the present rate,
the Judges see no need to make an additional increase in order to be
consonant with Professor Willig's directional recommendation arising
from his Ramsey pricing approach.
    SoundExchange's other economic expert, Mr. Orszag, provided a
separate reason why SoundExchange's rate proposal was consistent with
the Factor A principles. He stated that rates that are ``market-based''
meet the Factor A criteria because they cause rates to be
``sufficiently high to incentivize copyright holders to create content,
as reflected in content distributors' [licensees']--and by extension
consumers'--willingness to pay for sound recordings.'' Orszag WDT ] 15.
In addition, Mr. Orszag opined that the presence of streaming services
operating under market-based rates demonstrates that those ``market-
based rates are not so high as to prevent content distributors from
earning economic returns sufficiently attractive to induce the
investments required to transmit content to consumers, to broaden their
distribution networks, and to develop quality enhancements and a richer
menu of features and functionality.'' Id. Thus, he concluded that
``market-based rates will produce rates that are high enough to incent
artists and labels to create their product,'' and ``are high enough for
the content distributors to earn sufficiently high returns that they
will want to distribute that content.'' 4/25/17 Tr. 956-57 (Orszag).
    In support of this argument, SoundExchange noted the many specific
costly ways in which labels must invest in their businesses, incurring
repeated ``sunk costs,'' in order to provide a continuing flow of
recorded music. As SoundExchange noted, the testimony and evidence
highlight these specific risky and costly investments incurred to sign
artists, create and produce recordings, manufacture product, market and
distribute the music, build an audience and fan base, and license the
copyrighted content to services such as Sirius XM for listening by end
users. See, e.g., Written Direct Testimony of Jason Gallien, Trial Ex.
30, at 2 (Gallien WDT).
[[Page 65255]]
    In opposition, Sirius XM correctly argued that Mr. Orszag ``merely
offers truisms,'' such as that higher revenue encourages record
companies to make sound recordings available to the public. However,
Sirius XM noted that SoundExchange does not go beyond this truism to
``elucidate how properly determined market rates fail to ensure that
record companies are fairly compensated.'' SXRPFF ] 340.
    The mere (and obvious) fact that record companies incur substantial
costs is not illuminating, because that fact simply begs the question
whether rates are sufficient in light of those costs. Moreover, the
Judges do not acknowledge that SoundExchange's position even rises to
the level of a ``truism.'' An increase in the royalty rate will not
necessarily result in an increase in revenue, if the increase causes a
downstream retail percentage reduction in quantity demanded that is
greater than a percentage increase in subscription prices.
    The Judges find that a rate properly crafted to reflect an
effectively competitive market rate will maximize the availability of
creative works to the public by providing appropriate market
incentives. Lower rates, ceteris paribus, would result in increased
distribution but less incentive to produce sound recordings. Higher
rates, ceteris paribus, would encourage increased production of sound
recordings but discourage distribution. Nothing in the record indicates
that, on balance, either an increase or a decrease in the reasonable
rate of 15.5% would increase the availability to the public of sound
    Further, because the 15.5% rate identified by the Judges is market-
based, the Judges are advancing the general proposition asserted by
SoundExchange, that the market, properly construed, will balance the
interests of producers (licensors) and distributors (licensees),
without an increase in that rate under Factor A. See 5/2/17 Tr. 1956-57
(Willig) (from economic perspective, factor will ``require rates,
royalty rates and terms generally that perform the economic function of
motivating the record companies and the artists to create desirable
sound recordings . . . [and] at the same time, . . . those rates and
those terms should motivate . . . the distribution Services, to
distribute those recordings to the public in a way that reflects
consumer preferences.'').
    Sirius XM suggested that the record supports a reduction in the
royalty rate below the present 11% rate. In support of that point, it
relied on the testimony of Professor Shapiro, who noted that an element
of providing proper economic incentives to both the creators of sound
recordings and to Sirius XM to make the necessary investments to
``maximize the availability of creative works to the public'' is the
extent to which plays on Sirius XM's satellite radio service promote or
substitute for other record label revenue streams. Shapiro WDT at 57-
58. The Judges find this argument to be as much a ``truism'' as
SoundExchange's argument emphasizing the incentivizing effect of higher
royalty rates, and thus equally unavailing. Moreover, the Judges have
already incorporated into their rate analysis survey evidence that
demonstrates the substitution patterns between Sirius XM and other
distribution channels. In that analysis, the Judges relied on an
evidentiary roadmap provided to them by Sirius XM, through Professors
Hauser and Farrell, for the identification and valuation of the
substitutability of other distribution channels for Sirius XM.
    Finally with regard to Sirius XM's argument, although Professor
Shapiro asserted that a downward adjustment is warranted because Sirius
XM is more promotional and less substitutional than non-interactive
webcasters for other record label revenue streams, he found it too
difficult to measure the magnitude of such an adjustment. Id. at 58 &
App. D. Accordingly, he declined to propose such an adjustment. Shapiro
WDT at 58. The Judges, therefore, have no evidentiary basis to make
such a downward adjustment, even if they had found that a reduction was
    The Judges interpret the ``maximize'' directive more broadly than
either party to this proceeding. SoundExchange interpreted maximization
as an upstream supply issue while Sirius XM interpreted maximization as
a downstream distribution issue. The Judges must look at both steps in
the process. Aside from the economic issues the parties argued, there
is also simply no record evidence that indicates a shortfall in the
overall production of sound recordings, or in the dissemination of
sound recordings through Sirius XM or other distribution channels. For
all these reasons, the Judges find no basis in the record for a policy
adjustment to the 15.5% ``reasonable rate'' based on Factor A.
B. Factor B: Fair Income/Fair Return Under Existing Market Conditions
    Factor B requires the Judges to balance fair return to licensors
and fair income to licensees. There is an inherent tension within this
factor. Further, economic analysis cannot identify royalty rates or a
division of revenue that is ``fair.'' See 4/25/17 Tr. 957 (Orszag)
(``Fairness is not a well-defined term in the economics literature.'').
Economics can, however, provide a framework for a fair process. Id. at
958 (``Market-based rates are fair in the sense of, as long as they are
being determined in a workably competitive environment, they are going
to produce outcomes that are efficient.''). Thus, the Judges analyze
the Factor B issues with an understanding of the inherent subjectivity
of the endeavor, and an appreciation for the nuanced distinction
between a ``fair outcome'' and a ``fair process.''
    Equating the market rate with a rate that provides a fair return,
SoundExchange argued that the current rate does not afford a fair
return to copyright owners because it is lower than a market rate.
Exacerbating this problem, according to SoundExchange, is the decline
in sales of downloads and physical products, which have made royalty
revenues from Sirius XM (and other services that offer ``access''
rather than ``ownership'') even more important than in the past. See
Gallien WDT at 3-6. To the extent this argument is simply a plea by
SoundExchange for rates that subsidize declining business segments, it
is rejected. As the Judges have said previously with regard to
services' business models, rates are not set merely to support a
particular business model. See Web IV, 81 FR at 26329 (the statute
``neither requires nor permits the Judges to protect any given business
model proposed or adopted by a market participant''). Likewise in this
proceeding, the Judges are not obliged to offset, mitigate, or
subsidize a decline in physical or download sales by setting higher
royalty rates for satellite radio. Moreover, as Sirius XM correctly
argued, in this proceeding there is no record evidence that the decline
in revenues from other distribution channels can be laid at the
doorstep of Sirius XM and, further, any such decline cannot
automatically mean that the current level of income received by the
record companies is not ``fair.'' See SXM RPFF ] 344.
    SoundExchange refined its argument by reformulating its
substitution/cross-elasticity argument as a basis to raise rates
pursuant to Factor B. More particularly, noting the self-evident fact
that consumers have a limited amount of time to listen to music,
SoundExchange pointed out that, when subscribers tune in to Sirius XM,
they forego other direct revenue generating services, like Apple Music
or Spotify, and that may also diminish their purchases of physical
product and downloads because they spend their music-listening time
tuned in to Sirius
[[Page 65256]]
XM. Gallien WDT at 7. The Judges reject this argument as a basis to
adjust the rates pursuant to Factor B. In setting the ``reasonable
rate'' of 15.5% for an effectively competitive market, the Judges
examined the survey evidence that demonstrated the relevant
substitution patterns. The Judges cannot gainfully pursue that same
issue a second time by reconfiguring it as a basis for making
adjustments under Factor B.
    Approaching the Factor B issue from the other side of the ledger,
so to speak, SoundExchange argued that ``Sirius XM earns far more than
a fair income under the current 11% rate, and will continue to do so
under SoundExchange's rate proposal.'' SE PFF at 605. In support of
this conclusion SoundExchange pointed to several facts proffered by
Professor Lys that demonstrate how and why Sirius XM has realized
substantial and profitable growth: \176\
    \176\ The summary of Professor Lys's exhaustive analysis of
Sirius XM's financial success lays out SoundExchange's Factor B
analysis and also demonstrates that the 15.5% rate set by the Judges
cannot be construed as ``unfair.'' The rate provides the record
companies with their opportunity costs, a form of return that
Professor Willig acknowledged to be appropriate, while allowing
Sirius XM to realize ongoing profits.
    (a) At the time of SDARS I, Sirius and XM were two separate
companies competing for subscribers based on price, and likewise
engaged in price competition for non-music content such as sports
leagues and talk show personalities. However, in July 2008, Sirius
and XM merged, and the merged entity, Sirius XM, became the sole
provider of satellite radio in the United States, holding a monopoly
in this market segment. Lys CWDT ] 43.
    (b) The merger eliminated price competition between the two pre-
merger satellite radio services for subscribers and for non-music
content, and also allowed the combined company to take advantage of
the economies of scale that are central to its business model. Lys
CWDT ] 44.
    (c) Sirius XM's operating costs are predominantly fixed with
respect to subscriber revenue. These fixed costs include programming
and content, satellite and transmission, sales and marketing,
engineering and design, subscriber acquisition costs, and general
and administrative costs.\177\ Id. ] 45.
    \177\ Alternatively, these costs are the same whether one person
is listening to a Sirius XM broadcast, or millions.
    (d) Sirius XM's variable operating costs (i.e., costs that do
vary with subscriber revenue) are small in comparison, and include
royalties, customer service, and cost of equipment. See id ] 46; see
also id. ] 46 n.17 (citation omitted).
    (e) Because of its largely fixed cost structure and its post-
merger market share growth, Sirius XM's profits increased
dramatically once its sales reached its ``break-even point,'' i.e.,
the point at which its fixed costs are covered. Id. ] 47.
    (f) This growth in profits is reflected in Sirius XM's high
contribution margin (i.e., the fraction of each additional revenue
dollar that covers fixed costs or increases profits). Specifically,
by 2015, Sirius XM achieved a contribution margin of [REDACTED] %,
meaning that each additional dollar of revenue increases pre-tax net
income and cash flows by $[REDACTED]. Id.
    (g) Sirius XM's ``free cash flow'' (FCF) (a metric commonly used
to assess a company's performance and value), captures the amount of
cash that is available, after necessary business investment
(including satellite investments), that can be used to pay dividends
and repurchase shares. In 2012, Sirius XM's FCF was [REDACTED]% of
EBITDA,\178\ a higher percentage than other large entertainment-
media companies. That is, Sirius XM can distribute [REDACTED]% of
its EBITDA to its shareholders without affecting its operations.
    \178\ EBITDA means earnings before interest, taxes, depreciation
and amortization.
    (h) Looking at FCF over a longer period, over the past decade
Sirius XM has generated $2.6 billion in such FCF. Since the merger,
starting in 2009 Sirius XM has recorded seven straight years of
positive FCF and has over that seven-year period generated $4.91
billion of FCF. Lys CWDT ]] 91-92. Sirius XM's FCF has increased
from a deficit of $1.23 billion in 2006 (meaning that the company
was not generating sufficient cash and needed to rely on external
funding sources for its operations and investments) to a positive
$1.32 billion in 2015. This means that after it satisfied its
investment needs, its operations generated $1.32 billion in cash
that it could distribute to its investors. Id. ] 55. Cumulatively,
from 2006-2015, Sirius XM earned $5.9 billion in operating cash
flows. Id. ] 90.
    (i) Sirius XM's executives trumpet the company's more recent
performance as ``one of the best growth stories in media,'' and
conclude that its ``business is thriving''--a claim confirmed by
Professor Lys's analysis. Id. ] 52 & nn.24-25.
    (j) In the 2009-2015 post-merger period, Sirius XM earned a
total of $5.6 billion in EBIT.\179\ Similarly, in the period since
the merger, Sirius XM has generated over $7 billion in adjusted
EBITDA, an increase from negative $690 million in 2006 to positive
$1.66 billion in 2015. Lys CWDT ] 54.
    \179\ EBIT means earnings before interest and taxes.
    (k) Turning from financial to volume metrics, over the past
decade, Sirius XM has substantially increased its number of
subscribers, even as it has increased the prices and fees it
charges. Lys CWDT ] 57; see also 4/26/17 Tr. 1323 (Lys) (Sirius XM's
historic revenue base). Specifically, over the past decade Sirius
XM's subscriber base has grown on average [REDACTED]% per year, more
than doubling from [REDACTED] million subscribers in 2006 to
[REDACTED] million subscribers at the end of 2015. Lys CWDT ] 59. As
of March 2016, Sirius XM had over 30 million subscribers. Id. ] 58 &
    (l) Sirius XM's total revenue has grown even faster than the
growth in the number of its subscribers--from $1.57 billion in 2006
to $4.57 billion in 2015--a 12.6 percent compounded annual growth
rate (CAGR). Lys CWDT ] 65. This higher revenue growth resulted from
Sirius XM's increase in its subscription prices and fee charges that
occurred contemporaneous with the growth of its subscriber base,
allowing Sirius XM to realize a 15.8% increase in its ARPU between
2008 and 2015, corresponding to a compounded annual growth rate of
1.6%. See id. ] 66, Fig. 11. The table below presents the increase
in the total effective monthly cost of subscribing to Sirius XM's
most popular subscription package, the ``Select'' package), i.e.,
combining the subscription fee and the U.S. Music Royalty Fee:
                         Sirius XM Historical Effective Monthly Total Subscription Cost
                                          [Select subscription package]
              Date                    Nominal       U.S. Music       effective       Increase       % Increase
                                   subscription     royalty fee    subscription
1-Jan-06........................          $12.95           $0.00          $12.95             n/a             n/a
29-Jul-09.......................           12.95            1.98           14.93            1.98            15.3
6-Dec-10........................           12.95            1.40           14.35          (0.58)            -3.9
1-Jan-12........................           14.49            1.42           15.91            1.56            10.9
1-Feb-13........................           14.49            1.81           16.30            0.39             2.5
1-Jan-14........................           14.99            1.81           16.80            0.50             3.1
5-Jan-15........................           14.99            2.08           17.07            0.27             1.6
27-Apr-16.......................           15.99            2.22           18.21            1.14             6.7
[[Page 65257]]
Id. ] 70, Fig. 12; see also Summary of U.S. Music Royalty Fees by
Package, Trial Ex. 321 (excerpt from Sirius XM website). As this figure
shows, Sirius XM's pricing on its Select subscription package has
increased by 41% over the past decade, from $12.95 in 2006 to $18.21 as
of April 2016, corresponding to a total increase of $5.26 or a
compounded annual increase of 3.5%. Lys CWDT ] 71.
    According to Professor Lys, Sirius XM's pricing increases appear to
have had little effect on demand for its services, as evidenced by the
essentially non-existent impact of the price increases on subscriber
``churn'' (defined by Sirius XM as ``the monthly average of self-pay
deactivations for the period divided by the average number of self-pay
subscribers for the period''). Sirius XM Holdings, Inc., Proxy
Statement & 2015 Annual Report, Trial Ex. 372, at 21 (Sirius XM 2015
Annual Report).\180\
    \180\ The only noticeable bump is an increase in churn from 1.8%
to 2.0% in 2009 when Sirius XM introduced the U.S. Music Royalty
Fee, resulting in the largest percentage increase in the effective
subscription price, and coinciding with the 2008-09 recession. Lys
CWDT ] 73.
    Sirius XM's most recent annual performance has been consistent with
its past post-merger growth and profitability, as evidenced by the
following points.
    In 2016 Sirius XM set records for subscribers, revenue, adjusted
EBITDA, and free cash flow, beating its guidance on all of those
metrics. 5/15/17 Tr. 3759-60 (Meyer).
    In 2016, Sirius XM added more than 1.7 million net subscribers,
outperforming expectations. It added 1.66 million ``self-pay net
subscribers,'' also exceeding expectations (Sirius XM's original
guidance was 1.4 million). Trial Ex. 25, Figs. 43 at 56.
    In 2016, Sirius XM's subscriber level increased by 6%, raising
its subscribership level to 31.346 million. Lys WRT ] 164.
    In 2016, Sirius XM's 2016 revenue grew by 10% compared to 2015,
to more than $5 billion; EBITDA grew by 13% to $1.9 billion; FCF per
share grew 26% to $0.30; and net income grew 46% to $746 million.
Lys WRT ] 166.
    In sum, SoundExchange argued that there is abundant and undisputed
evidence that Sirius XM's profitability has grown dramatically--and
significantly faster than its revenue--indicating an improved ability
to monetize the operational gains and scale.
    Accordingly, SoundExchange's critical conclusion from Professor
Lys's exhaustive analysis was this: Sirius XM has been facing a
relatively inelastic demand, enabling it to increase prices to
consumers without causing a loss of subscribers. Lys CWDT ] 74.\181\
    \181\ Professor Lys also opined that Sirius XM will continue to
grow across these metrics for all of 2017 and into the foreseeable
future. See Lys CWDT ]] 152-198. As the Judges have stated
previously, they are less than sanguine about projections and
forecasts given the inherent speculative nature of such a process.
However, as Professor Lys pointed out, his projections in SDARS II
regarding the future financial performance of Sirius XM were
accurate, and prior financial forecasts, as well as Sirius XM's own
internal forecasts, [REDACTED]. See id. These facts suggest that
there is no present reason to project a scenario in which Sirius
XM's current level of profitability will fall or will not be
    Sirius XM did not challenge the wealth of evidence demonstrating
its economic, market, and financial success. Rather, Sirius XM
contended that these measures of Sirius XM's economic position are
``entirely irrelevant to the rate-setting task at hand.'' Shapiro CWRT
at 5. More specifically, Sirius XM argued that ``Professor Shapiro has
demonstrated'' through his direct and rebuttal testimony ``that Sirius
XM's overall profitability would not be among the variables impacting
the outcome of a license negotiation in a workably competitive
market.'' See Shapiro CWRT, App. D. & 24-26.
    Professor Shapiro explained that, in his opinion, it is not the
overall profits that are relevant in a Factor B analysis, but ``the
incremental profit f[rom] additional Sirius XM customers, as measured
by the contribution margin (which takes into account only variable
costs) that enters the analysis. Shapiro CWRT at 52 (emphasis added).
Sirius XM noted that its ``contribution margin'' has been essentially
unchanged over time, and that even Professor Lys acknowledged that the
contribution margin had ``remained remarkably consistent over time.''
See Lys WDT ] 87) (emphasis added).
    Sirius XM sought to impeach Professor Lys with excerpts from his
testimony in Web IV:
    From the standpoint of economics, a company's ability to pay
royalties, while still remaining profitable, and the ``willing
buyer/willing seller'' standard are two very distinct concepts.
See 4/27/17 Tr. 1592-93 (Lys).
    It ``is wrong to suggest that [a service's] current or past
profitability should be used to determine the royalty rate a willing
buyer and a willing seller would agree upon.''
See 4/27/17 Tr. 1593 (Lys).
    It was ``incorrect'' to ``suggest[ ] that Pandora's current
profitability and financial performance determine its ability to pay
royalties, and that Pandora's ability to pay determines the rates
the Judges should adopt.''
See 4/27/17 Tr. 1592 (Lys).
    Sirius XM also pointed out that its non-music content costs have
declined, demonstrating that there is no positive correlation between
its profitability and its content costs. See Shapiro CWDT at 52-53 &
Fig. 4.
    In sum, Sirius XM concluded that its potential ability to pay
higher royalties out of increasing profits is simply irrelevant to the
question whether it is receiving a ``fair return'' pursuant to Factor
    The Judges find that Sirius XM's increased profitability does not
provide an independent basis to adjust the 15.5% identified by the
Judges. Sirius XM earns sufficient profits, as the only satellite radio
provider, to allow it to pay the opportunity costs of its service to
the record companies. Those opportunity costs, properly weighted,
constitute the building blocks for the 15.5% rate. The evidence, again,
as detailed by Professor Lys, makes it abundantly clear that Sirius XM,
through its monopoly of the satellite radio distribution channel, has
the financial capacity to pay higher rates and still maintain a high
level of profitability.
    The Judges find no inconsistency with regard to Professor Lys's Web
IV testimony and his testimony in this proceeding. If a service were
operating at a loss rather than a profit, the record companies would
not consider that fact relevant, especially if the service did not add
new (i.e., non-cannibalizing) listeners who could be monetized by
subscription or advertising revenues. However, when a service is
profitable, in an unregulated market, the record companies, empowered
by their ``must have'' status, can and will seek to acquire as much of
the surplus (profits) as they can through the bargaining process. As
explained in this determination (and in Web IV), though, the Judges
reject a division of profits based on the ``must have'' power of the
record companies, absent application of an appropriate offsetting
factor, such as identified in the steering analysis in Web IV or in the
opportunity cost analysis in this determination.
    Beyond Professor Lys's financial analysis, SoundExchange made
additional arguments with regard to Factor B that do not aid in the
Judges' analysis. SoundExchange argued essentially that a fair
allocation of the revenue attributable to satellite radio will arise
either from: (1) A Ramsey pricing approach as described by Professor
Willig; or (2) arm's-length negotiations in a benchmark market such as
the interactive market suggested by Mr. Orszag. Neither of these points
supports a Factor B analysis. First,
[[Page 65258]]
Professor Willig did not identify a rate pursuant to his Ramsey pricing
approach, and he argued that this approach counseled generally for an
increase in the existing rate (which the Judges have found to be
appropriate pursuant to their reasonable rate analysis). Mr. Orszag's
assertion that arm's length negotiations in the interactive market
demonstrate a fair process (if not necessarily a fair outcome) is
belied by the fact that: (1) The survey results reached by all survey
experts demonstrates the inapposite nature of the interactive
benchmark; and (2) the interactive benchmark is tainted by a
complementary oligopoly effect that cannot be mitigated, on the present
record, by a fact-based steering adjustment.
    SoundExchange, again relying on Mr. Orszag, cautioned that the
Judges should not apply Factor B so as to provide an unjustified
ceiling on the royalty rate, which could constitute a subsidy to Sirius
XM. The Judges' 15.5% reasonable rate does not constitute an arbitrary
ceiling or a subsidy, because it is derived pursuant to the
``opportunity cost'' approach that, according to Professor Willig,
resulted in a reasonable rate.\182\
    \182\ To be sure, Professor Willig calculated a higher rate
because he used the diversion ratios in the Dhar Survey, but the
Modified Dhar Survey (as corrected), with its superior diversion
ratios, applies the same opportunity cost approach advocated by
Professor Willig, and even applied his ``Creator Contribution''
walk-away values.
    Sirius XM found no basis under Factor B to change its proposed
rate. Shapiro WDT at 58. Of course, the Judges' 15.5% rate is above
Sirius XM's proposed rate range that extends to 11% (the current rate).
However, Sirius XM made no arguments that would support a reduction of
the 15.5% rate pursuant to Factor B. See Shapiro WDT at 58. Sirius XM
limited its Factor B analysis to the bald assertion that its
benchmarking analysis (rejected by the Judges) led to a fair return for
copyright owners and a fair income for copyright users.
C. Factor C: Relative Roles of the Parties
    SoundExchange asserted that, pursuant to Factor C, the statutory
rate should be above its proffered benchmark, or at least at the high
end of its benchmark range. In support of this argument, SoundExchange
pointed to testimony that record companies and artists make substantial
contributions through their search for artistic talent, a process that
is long, competitive, and often unsuccessful. See 5/11/17 Tr. at 3542-
43 (Kushner). More particularly, SoundExchange explained that Artist &
Repertoire (A&R) representatives from labels go to clubs and concerts
worldwide, listen to thousands of demonstration (demo) recordings, and
search the internet to identify emerging and undiscovered artists.
According to SoundExchange, these tasks are labor-intensive, because
finding musical talent requires people with sufficient industry
knowledge and experience. Gallien WDT at 8. SoundExchange pointed to a
2015 RIAA study that found the major labels spent $13.4 billion between
2003 and 2012 to find new artists and help them reach an audience.
Written Direct Testimony of Michael Kushner, Trial Ex. 34, ] 77
(Kushner WDT).
    SoundExchange noted that after record companies incur the foregoing
costs, they must also incur costs to shape the artists' music and image
in order to maximize their commercial appeal. Those investments can
include the costs of dance and vocal lessons, personal stylists, makeup
artists, trainers, and media training. Many of those investments do not
yield a financial return. See 5/11/17 Tr. 3542-43 (Kushner) (``[I]f you
look at the totality of the number of artists we sign and the numbers
that are successful, clearly the unsuccessful ones outweigh the
successful ones'').
    SoundExchange further noted that recording companies incur
substantial additional costs to create recorded works, and to market,
manufacture, and distribute recorded music.\183\ SoundExchange avers,
for example, that in 2015 alone, UMG spent $[REDACTED] million on
recording costs, mastering costs, producer and sampling fees, royalty
advances, and overhead funding to contracting parties who provide A&R
services. Gallien WDT at 8. Mr. Kushner testified for Atlantic Records
that, on an album basis, the recording costs for a maiden album from a
new artist typically range from $[REDACTED] to $[REDACTED]--and can
exceed $[REDACTED] for an established artist. Kushner WDT ] 36. If the
record companies cannot recoup these expenditures and advances from
sales revenue, they--not the artists or the music services--bear the
unrecouped cost and foregone profits.\184\
    \183\ These costs typically may include the additional expense
of a producer's salary, studio rental, hiring a sound engineer,
paying musicians to play with the featured artist, and preparing a
master recording. See Written Direct Testimony of Bruce Iglauer,
Trial Ex. 33, at 10-11 (Iglauer WDT); Kushner WDT at ]] 48-50.
    \184\ For example, SoundExchange proffered UMG's 2015 income
statement, which reflects $[REDACTED] in (1) advances and recording
costs for new unproven artist signings and (2) write offs of
investments in established artists, net of recoveries. Gallien WDT
at 10.
    As to marketing costs, Mr. Kushner testified that for Atlantic
Records, the typical initial U.S. marketing budget for an album cycle
for a new artist is in the range of $[REDACTED] to $[REDACTED]. Id. ]
68. In fiscal year 2015, UMG alone spent $[REDACTED] specifically on
gross marketing costs, as well as $[REDACTED] in overhead costs for its
various departments that also provided marketing services. For UMG,
marketing costs included over $[REDACTED] in advertising, $[REDACTED]
in artists' press and TV appearances, over $[REDACTED] in internet
marketing & advertising, over $[REDACTED] in radio promotion, and over
$[REDACTED] in video production costs. With specific reference to
streaming and playlisting efforts, UMG has also invested in the setup
costs and personnel to establish a team dedicated to streaming
marketing and playlisting efforts. Gallien WDT at 13-14.
    Regarding recording companies' manufacturing and distribution
costs, they remain substantial in spite of the industry's transition
away from physical media. Because of declining physical product sales,
physical manufacturing has been declining, but it still carries high
costs. UMG reported that its manufacturing costs for physical records,
including costs they advance for pressing and distribution deals, were
$[REDACTED] in fiscal year 2015. Id. at 14. Digital distribution has
been increasing, and there is misperception that it is costless to the
record companies. The reality is that digital distribution is highly
complex and requires expensive investments. UMG reported that since the
early 2000s, it has invested over $[REDACTED] in IT infrastructure and
operating costs, as well as the professionals that today distribute the
thousands of digital files it provides to hundreds of music services
and to handle the processing of billions of micro transactions related
to recognizing digital revenues and calculating the associated royalty
obligations. Id. at 14. And in 2016 and throughout 2017, UMG will be
investing in its 3rd generation of digital supply systems and digital
revenue processing systems at an estimated cost of over $[REDACTED].
Id. at 15-16.\185\
    \185\ Indies' costs differ in magnitude from those of the
Majors, but the categories are similar, according to SoundExchange.
Mr. Iglauer provided qualitative testimony stating that his Indie
label, Alligator Records, spends substantial time seeking out
recording artists to sign--listening to demos, attending shows and
music festivals, reading the music press, and taking referrals from
other bands, labels, managers, and booking agents. It also devotes
significant resources to promoting the music and touring of artists
they have signed, including the payment of recording costs and
advances. Iglauer WDT at 9.
[[Page 65259]]
    In sum, SoundExchange asserted that major labels spend billions of
dollars finding and developing new artists, helping them reach an
audience, and creating and marketing recorded music.
    Sirius XM gave short shrift to these lengthy descriptions of the
record companies' various expenses. First, Sirius XM claimed that
SoundExchange's request for an upward adjustment pursuant to Factor C
is inconsistent with the latter's prior broad proclamation that the
first three itemized 801(b)(1) factors are satisfied by market rates.
Second, Sirius XM noted that the categories of costs that SoundExchange
has itemized ``have long prevailed in the recording industry,'' and
that nothing set forth in SoundExchange's Factor C argument provided
specific reasons to suggest that those costs have changed in a manner
to support an adjustment upward in the statutory rate. Third, Sirius XM
noted that SoundExchange did not measure ``the investments made by the
record companies'' against ``Sirius XM's investments ``and thus did not
perform the ``relative'' analysis of costs, risks, and other factors
expressly required by the statutory language. In this criticism, Sirius
XM also noted parenthetically that SoundExchange did not explain how or
why particular portions of the record industry's costs should be
allocable to Sirius XM, rather than other distribution channels.
    Additionally, relying on Professor Shapiro's testimony, Sirius XM
argued that when the emphasis is placed properly on the ``relative''
contributions of the parties, the record companies' cost of creating
sound recordings, ``is almost certainly significantly less than the
contribution that Sirius XM plans to make over the 2018-2022 license
period,'' including the launching of two new satellites and improving
its repeater network.'' Shapiro WDT at 58. Although he concluded that
this relative difference points toward reducing the statutory rate, the
relative balancing ``does not readily lend itself to quantifying'' an
appropriate downward adjustment. Id.
    Sirius XM also claimed that it contributes additional value through
its ``delivery network.'' As Professor Shapiro argued: ``[B]y combining
music, non-music, curation, and a delivery platform all into one
bundle, Sirius XM is creating significant value for consumers, with
each piece of the bundle contributing to the overall value of the
service.'' 4/20/17 Tr. 398-99 (Shapiro) (emphasis added).\186\
    \186\ Sirius XM did not address its contribution of this
additional network value in its Factor C argument. However, the
Judges find that this issue is best considered in the context of
Factor C, which broadly addresses relative contributions.
    In response, SoundExchange, through Mr. Orszag, asserted that
Sirius XM's ``delivery platform'' does not add separate value, because
any value created by that platform flows principally to Sirius XM; that
is, even under the SoundExchange proposal the record companies receive
only 23% of Sirius XM's revenue. Therefore, he noted, most of the gains
flow to Sirius, ``but there is a portion that goes to the labels which
[provide] a necessary input,'' 4/25/17 Tr. 1034 (Orszag), which is
``consistent with sound economics.'' Id. at 1034-35 (Orszag) (emphasis
    In reply, Sirius XM argued that Mr. Orszag's justification for the
labels' sharing of any value added (via revenue) from Sirius XM's
unique inputs begs the question as to ``what the split should be,'' and
fails to ``address whether an adjustment to the [interactive] benchmark
is warranted to account for Sirius XM's independent contributions to
the value of its service offerings.'' SXM RPFF ] 62.
    The Judges agree with Sirius XM that the value of its unique inputs
(relative to interactive and other services), such as its expensive
satellite and ancillary technical equipment \187\ and its use of live
``on-air'' talent and other specialized personnel,\188\ are intended
to--and do--create a product that is differentiated from interactive
services. However, SoundExchange is correct that inputs do not have
independent value per se.\189\
    \187\ See, e.g., Written Direct Testimony of James E. Meyer, ]
12 (Meyer WDT); Written Direct Testimony of Bridget Neville, passim
(Neville WDT); Written Direct Testimony of Terrence Smith, passim
(Smith WDT).
    \188\ See Blatter WDT ]] 9-10.
    \189\ As Professor Orszag asserted, David Frear, Sirius XM's
CFO, conceded this point during the SDARS II proceeding: [REDACTED]
See Orszag WRT ] 53 n.65.
    Rather, Sirius XM incurs the cost of these inputs to create a
differentiated and thus more profitable service. If it succeeds, the
benefits will be evidenced by higher revenues (in excess of those input
costs) and will, therefore, result in higher profits. A separate
accounting of the costs of the Sirius XM satellite radio platform would
constitute a clear double-counting of value.
    By contrast, if the cost of Sirius XM's investments in its unique
inputs failed to differentiate its output (i.e., its service) from,
say, interactive services, then there would be no justification for
Sirius XM to obtain any recompense for its investments, either through
an adjustment to the revenue (royalty) base or to the royalty rate. As
the Judges noted previously, a party is not entitled to a rate simply
to preserve its particular business model. See, e.g., Web IV, 81 FR at
26329 (``the statute neither requires nor permits the Judges to protect
any given business model proposed or adopted by a market
participant.''). If Sirius XM's unique and expensive inputs have
marketplace value, those inputs will differentiate its service in an
attractive manner, resulting in relatively low cross-elasticities and
own-elasticities, lower opportunity costs for the labels in licensing
to Sirius XM, and higher profits for Sirius XM. It is through this
economic transmission mechanism that Sirius XM may extract value from
its unique inputs--not from a separate valuation of the inputs.
    This argument does not fully address Mr. Orszag's point that the
labels, as providers of a ``necessary input'' would, in an unregulated
market, command a portion of the value created by these unique Sirius
XM inputs. Again, Mr. Orszag concluded that such ``sharing'' is simply
``sound economics.'' However, that reasoning is ``sound'' only to the
extent the Judges would find it appropriate to reject Professor
Willig's opportunity cost approach and adopt instead his Nash
Bargaining Solution model. For the reasons set forth at length supra,
the Judges have done precisely the opposite: Accepting his opportunity
cost approach and rejecting his Nash Bargaining Solution approach.\190\
    \190\ The Judges' finding appears consistent with Sirius XM's
position: ``SoundExchange's attempt to expropriate a portion of the
value that Sirius XM alone creates is entirely at odds with the
section 801(b) factors.'' SXM RPFF ] 64. However, Sirius XM's claim
of expropriation is hyperbolic. By its logic, Sirius XM's use of the
labels' music likewise would constitute expropriation--of the sound
recording value that the labels created. The difficult issue is the
application of the statutory and economic factors to allocate the
value of the output created by a production function (containing
sound recordings and a delivery network) that utilizes these
separate inputs in combination, to cover all costs (including
opportunity costs) while rewarding the investment in technology that
leads to innovative product differentiation. The Judges' 15.5% rate
addresses these various and competing factors in a reasonable
D. Factor D: Minimizing Disruptive Impact on Structure of the
Industries Involved and Generally Prevailing Industry Practices
    The Judges' long-standing test for whether a rate is ``disruptive''
pursuant to Factor D provides that a rate change would be disruptive if
it ``directly
[[Page 65260]]
produce[s] an adverse impact that is substantial, immediate, and
irreversible in the short-run because there is insufficient time for
either the SDARS or the copyright owners to adequately adapt to the
changed circumstances produced by the rate change and, as a
consequence, such adverse impacts threaten the viability of the music
delivery service currently offered to consumers under this license.''
SDARS II, 78 FR at 23069 (quoting SDARS I, 73 FR at 4097). Accordingly,
the Judges apply this standard to the 15.5% rate they have found to be
reasonable in this proceeding to determine whether the 15.5% rate would
be disruptive.\191\
    \191\ Nothing in the record indicates that the reasonable rate
of 15.5% identified by the Judges (a 41% rate increase from 11% to
15.5%) would be disruptive to the record companies, even though it
is below the 23% rate sought by SoundExchange. See Shapiro WDT at
Fig. 5 and p. 59 (noting industry data showing that Sirius XM
accounted for only about 4% of overall record industry revenue in
2016). Given the 4% figure identified by Professor Shapiro, 23% of
that percentage equals .9%, and 15% of that 4% equals .6%. The
difference in revenue to SoundExchange between its percent-of-
revenue proposal and the rate set in this Determination therefore is
approximately .3% of overall record industry revenue, and thus not
disruptive within the applicable standard. Accordingly, the Judges
focus on whether this increase would be disruptive for Sirius XM.
    SoundExchange relied on the testimony of Professor Lys, who
demonstrated that Sirius XM would still earn substantial returns
(compared to other companies in closely-related industry sectors), even
if the Judges were to increase the statutory royalty rate to 24%. See
4/26/16 Tr. 1321-23 (Lys).\192\ First, Professor Lys calculated that
the pre-tax incremental impact of even a 24% royalty payment (based on
2015 figures available to him when preparing his direct testimony) was
$[REDACTED] million and the net after-tax impact would be $[REDACTED]
million. Lys CWDT ]] 129-30. At those levels, Sirius XM would obtain
the following financial results:
    \192\ The Judges provide this detailed summary of Professor
Lys's exhaustive analysis of Sirius XM's financial picture not only
to demonstrate the proper application of an itemized factor, but
also to underscore that Sirius XM can easily afford to pay the
market-based reasonable rate of 15.5% crafted by the Judges.
  Sirius XM 2015 Performance Metrics Under 24% Royalty Rate vs. SIC 483
                                                                for SIC
                                                    SXM (@24%     483
                Performance metric                   royalty)  broadcast
Return on Assets..................................        5.5        3.0
EBITDA Margin.....................................       27.4       19.9
Free Cash Flow Margin.............................       23.0        6.1
Lys CWDT ]] 132-42 & Fig. 33; see 4/26/17 Tr. 1321-22 (Lys).
    Professor Lys also analyzed Sirius XM's forecasted performance,
again assuming arguendo that the Judges set the statutory rate at 24%
of revenue. His analysis shows that, at that rate, the incremental
after-tax impact on Sirius XM would range between $[REDACTED] million
in 2018 and $[REDACTED] million in 2021. Professor Lys noted that
Sirius XM is expecting to perform so well in the future that it could
easily absorb this higher rate for the SDARS III period, 2018 through
2022. Lys WRT ] 219. More particularly, under this scenario, Professor
Lys testified that Sirius XM:
    Would earn between $[REDACTED] and $[REDACTED] in EBITDA in
every year of the forecast, and would continue growing. Id. ] 220.
    Would earn over $[REDACTED] in net income each year of the
forecast, and would continue growing. Id. ] 221.
    Would generate over $[REDACTED] in free cash flow almost every
year of the forecast and would continue growing. Id. ] 222.
    Professor Lys further noted that, even under Sirius XM's own
internal forecasts, with a royalty rate of 24%, it would remain
extremely profitable throughout the SDARS III term (2018-22), earning
$[REDACTED] in EBITDA, $[REDACTED] in net income, and $[REDACTED] in
free cash flow. Id. ] 223. Additionally, Sirius XM's 2016 [REDACTED]
indicates that, at the end of the forecasted period (2022), it would
have a strong balance sheet, with $[REDACTED] in cash and equivalents,
total assets of $[REDACTED], and shareholder equity of $[REDACTED]. Id.
] 224.\193\
    \193\ The Judges place much less emphasis on projections
compared with current facts, absent additional proof that the entity
making the projection has a track record that makes its projection
credible. However, the Judges note that these projections are
consistent with [REDACTED].
    For these reasons, SoundExchange argued that Sirius XM can
comfortably afford a rate increase from the current 11% to its proposed
23% of revenue. As Professor Lys colorfully and emphatically opined:
[REDACTED]. 4/27/17 Tr. 1391-92 (Lys).
    Professor Lys also examined in great detail Sirius XM's growth in
equity value compared to broader market metrics such as the S&P 500 and
the Dow Jones Industrial Average, and noted that Sirius XM far outpaces
those indices. He further noted that Sirius XM outperforms other firms
in the noninteractive markets. From these facts, Professor Lys
concluded that Sirius XM enjoys an ``unfair advantage over competing
digital music services that pay higher royalty rates.'' SE PFF ] 1584;
see Lys CWDT ]] 117-124.
    To provide yet another perspective on the financial success of
Sirius XM, Professor Lys calculated how its performance would have
changed in 2015 if the statutory rate had been increased above the 10%
applicable in that year. His calculations demonstrated that:
    Sirius XM could have afforded to have its 2015 statutory royalty
rate increased from 10.0% to up to 41.9%, 35.9% or 31.4% and still
earned an average EBITDA level of $735.7 million . . . , $909.5
million . . . , or $1.037 billion . . . , respectively. While this
level of the royalty rate would have reduced Sirius XM's EBITDA
profitability by $921 million, $747 million, and $620 million,
respectively (from the actual $1,657 million), would have only
equated Sirius XM's performance with its industry peers' EBITDA
profitability levels.
Lys CWDT ] 136.
    Sirius XM could afford to have its 2015 statutory royalty rate
increased from 10.0% to 65.1% and still earn a free cash flow level
commensurate with SIC 483 of $278.8 million.
Id. ] 138.
    Sirius XM could afford to have its 2015 statutory royalty rate
increased from the actual 10.0% to 35.0% and still earn an average
SIC 483-level (in terms of return on assets) net income level of
$39.6 million.
Id. ] 142.
    In sum, SoundExchange made a compelling case that an increase in
rates far greater than the 15.5% identified as a reasonable rate by the
Judges would be easily sustainable for Sirius XM, and therefore not
disruptive under the Factor D standard as quoted supra. Moreover,
Sirius XM did not provide any evidence sufficient to question Professor
Lys's analysis, which indicated that Sirius XM could afford a much
larger rate increase. Accordingly, the Judges find that, a fortiori,
Professor Lys's analysis indicates that Sirius XM could also afford a
smaller increase, to the 15.5% rate determined by the Judges.\194\
    \194\ SoundExchange also asserts that Sirius XM has paid less
than an appropriate rate in previous rate terms. See SEPFF ]] 1598-
1606 (and record citations therein). However, the Judges do not
conclude that, as a matter of law, they can set rates for a
forthcoming period that reimburse a licensor for any alleged
underpayments caused by a purported error in the statutory rate for
a past rate period.
XI. Terms
    Besides seeking a revision of the royalty rates for the 2018-22
rate period, the participants proposed certain additional changes to
the extant regulations. The final regulations appended to this
determination reflect
[[Page 65261]]
the Judges' decisions on points that were in controversy. For the
reasons detailed below,\195\ the Judges adopted some of the proposed
changes and declined to adopt others, as indicated in the final
regulatory language.
    \195\ The Judges do not provide narrative discussion about every
detail of the regulatory changes; rather, they concentrate on the
areas of legally significant controversy.
A. Generally Applicable Terms
1. Advance Payment and Minimum Fee
    SoundExchange did not propose any substantive change to the current
ephemeral royalty minimum fee of $100,000 per year, which is creditable
to ephemeral royalty payments for the relevant year (37 CFR 382.3(b)).
SE PFF 85. SoundExchange sought to designate the $100,000 annual
advance payment as the minimum fee for use of the section 112 Ephemeral
License by SDARS and PSS. Under SoundExchange's proposal, the advance
payment would be applied first to section 112 royalties due, and the
balance, if any, would be nonrefundable and not applicable to a
subsequent year's license. Music Choice argued rightly that section 114
does not provide for a minimum fee for SDARS or PSS. Compare 17 U.S.C.
114 (f)(1)(A) with section 114(f)(2)(A).\196\ Section 112 does,
however, require the Judges to set a minimum fee for ``each type of
service offered by transmitting organizations'' using the ephemeral
license. See 17 U.S.C. 112(e)(3).
    \196\ The extant regulation setting the PSS advance payment does
not mention a ``minimum fee'' but it does limit application of the
advance payment to ephemeral license royalties and prohibit rollover
of any portion of the advance payment to subsequent royalty years.
See 37 CFR 382.3(b) (2016). Perversely, the current regulation
establishing the $100,000 advance payment by SDARS is entitled
``Ephemeral Recordings Minimum Fee.'' See 37 CFR 382.12(c) (2016).
Nothing in the subsection mentions a minimum fee, however. Id.
    By agreement of the parties and in conformity with prior rate
periods, the section 112 ephemeral license royalty fee is set at a five
percent portion of the total bundled royalty for both section 112 and
section 114 and is included in that bundled royalty payment. Music
Choice contended that in SDARS II, SoundExchange and Music Choice
stipulated to advance payment language that would have allowed the full
advance payment to be creditable to the PSS's entire royalty payment,
rather than to its ephemeral payment only. MC PFF ] 554. According to
Music Choice, the stipulated language was changed in the final rule
(i.e., the advance payment is creditable only to the ephemeral royalty
payment) with no explanation or justification. Music Choice asserted
that the language SoundExchange and Music Choice stipulated to in SDARS
II should be adopted and SoundExchange provided no rationale for
retaining the current language. MC PFF ]] 556-57. SoundExchange did not
appear to dispute Music Choice's assessment that the extant recoupment
provision differs from what the parties had stipulated to and has not
provided a compelling reason to retain the current offset provision for
PSSs. See, e.g., 5/10/17 Tr. 3308-13 (Bender). Therefore, the Judges
adopt the minimum fee language Music Choice proposes.
    It would seem incongruous to require an advance payment for section
114 and section 112 royalties in the aggregate but to require the
entirety of that payment to be applied as a ``minimum fee'' for the
ephemeral license. No participant objects to the $100,000 advance
royalty payment. The Judges have no basis upon which they could
allocate 100% of that payment to the ephemeral license.
    To comply with the statutory requirement that they set a minimum
fee for use of the section 112 ephemeral license by transmission
services, viz., SDARS and PSS, the Judges set the section 112 minimum
fee at five percent of the advance payment, or $5,000, for each type of
SDARS or PSS service for which the Judges establish a different section
114 performance royalty. SoundExchange must, thereafter, apply the
remaining amount of the advance payment, after application of $5,000
per type of service to ephemeral licenses, to section 114 royalties.
2. Definitions
    Music Choice objected to the placement of ``Definitions'' at the
end of each subpart of the regulations. The Judges agree with Music
Choice that the placement seems counterintuitive. Definitions will
migrate to the beginning of each subpart. In addition, Gross Revenues
calculations will migrate from the Definitions section to the services'
respective subparts.
    The parties were in essential agreement regarding imposing a U.S.
geographical limitation in the definition of GAAP. Sirius XM asked the
Judges to apply a temporal element to the definition requiring
application of the version of GAAP in effect ``during the month when
the performances giving rise to a Licensee's royalty payment obligation
were transmitted.'' SoundExchange countered that a more definite time
limit would be preferable, viz., ``on the last day of the accounting
period to which the subject payment relates'' or ``the date payment
[was] due.'' The Judges adopt the definitive date for choosing GAAP
principles as the date payment was due.
b. Qualified Auditor
    In prior iterations of royalty rate regulations relating to various
licenses, the Judges noted the repetition of the phrase ``independent
and qualified auditor.'' In their Web IV determination, the Judges cut
the verbiage by 50% by defining a Qualified Auditor to be one that is
independent. In this proceeding, the parties have proposed language to
assure both the qualification and the independence of any auditor
working to verify royalty payment and distribution.
    In a slight departure from the Web IV language, the Judges
eliminate the Web IV requirement for an auditor to be licensed in the
state in which the audit is conducted. In this proceeding, the Judges
accept that Certified Public Accountants are governed by a code of
ethics that permits them the ``mobility'' to practice across state
lines. To remove any doubt, the Judges refer to the Code of
Professional Conduct adopted by the American Institute of Certified
Public Accountants.
c. Additional Definitions
    On their own motion, the Judges added ``Payor'' and ``Verifying
Entity'' as defined terms. These terms were added during the revamping
of regulations following the Web IV proceeding because they clarified
that auditing rights did not reside exclusively in the Collective. In
this iteration, the Judges clarify the terms they added to convey this
reciprocal audit right.
    The Judges also amended SoundExchange's proposed definition of
``Licensee'' for clarity.
3. Regulatory Terms
a. Section 382.3(a) 197--Payment to the Collective
    \197\ Section references are to the new numbering system that
results from reorganizing the regulations in part 382.
    In general, any due date in federal litigation that falls on a
Saturday, Sunday, or federal holiday is tolled until the next following
business day. The Judges regulations currently adopt this convention as
a general procedural rule when discussing litigation filing deadlines.
See 37 CFR 350.5. The Judges see no reason not to adopt the suggestion
of Sirius XM to enunciate the same rule when referring to royalty
payment due dates.
[[Page 65262]]
b. Section 382.4(a)(3)--Signature
    In updating the royalty regulations after the Web IV proceeding,
the Judges clarified the capacity of signers of Statements of Account.
Music Choice objected to reconfiguration of the Web IV language
suggested by SoundExchange. The Judges agree with Music Choice that the
language in the Web IV regulation is more appropriate for these
    \198\ The Judges are not swayed by Music Choice's plaint that it
could not have an authorized signer because Music Choice is a
partnership made up of corporations. Music Choice's sophisticated
representatives can figure out how the partnership may designate an
authorized signer.
c. Section 382.5(a)(2)--Best Efforts
    SoundExchange is obliged to use ``best efforts'' to locate
Copyright Owners and Performers entitled to receive a distribution of
royalty payments. The Judges' regulations need not specify the
specifics of those ``best efforts.''
d. Section 382.5(b)--Unclaimed Funds
    At the conclusion of the Web IV proceeding, the Judges adopted
language proposed by one of the Licensees directing SoundExchange to
treat unclaimed funds in accordance with federal, state, or state
common law. SoundExchange argued against this provision seeking to
retain permission to apply unclaimed funds to administrative expenses.
The Judges conclude that governance of applicable law will provide more
transparency regarding the disposition of unclaimed funds.
e. Section 382.6(c)(3)--Outside Counsel
    SoundExchange proposed a change to the rule regarding dissemination
and use of confidential information relating to royalty collection and
distribution. Music Choice objected to the additional language
SoundExchange proposed and the Judges agree with Music Choice.
SoundExchange is required to use and analyze sensitive business
information in its administration of royalty collection and
distribution. On occasion, SoundExchange might employ consultants or
experts to assist in that effort or in the auditing of the
administrative systems.
    SoundExchange sought to allow outside counsel access to
confidential information ``for the purpose of performing their duties
during the ordinary course of their work.'' This dissemination of
confidential information is not sufficiently constrained to limit it to
collection and distribution of royalty payments. The notion of outside
counsel obtaining the sensitive information ``in the ordinary course of
their work'' is too broad. The Judges will not grant that privilege.
Outside counsel has express authority to see confidential information
when acting on behalf of the Collective for ``verification of a . . .
statement of account'' or on behalf of a Copyright Owner or Performer
for purposes of ``verification of royalty distributions . . . .'' This
permission is sufficient.\199\
    \199\ Further, in a litigated rate proceeding, outside counsel
are entitled to obtain confidential information without signing a
non-disclosure agreement pursuant to a Protective Order specific to
each proceeding.
f. Section 382.7(c)--Notice of Intent To Audit
    SoundExchange requested that the Judges change the requirement that
a Verifying Entity ``deliver'' a copy of its filed Notice of Intent to
Audit to the Payor to a requirement that the Verifying Entity ``send''
the notice. Music Choice defended the term ``delivery'' because it
provides ``protections'' to the PSS. See MCRFF at 323. The Judges
conclude that this language issue is a solution in search of a problem.
The language will remain unchanged.
g. Section 382.7(d)--The Audit
    Music Choice and SoundExchange disagreed regarding language
SoundExchange sought to add to the provision that permits a licensee to
perform its own, independent audit.\200\ SoundExchange asked the Judges
to add the qualifier ``with respect to the information that is within
the scope of the audit'' to describe an acceptable ``defensive audit.''
This qualifying language is in the current regulation relating to
audits of SDARS and webcasters. The Judges see no reason not to make it
equally applicable to PSS. A report of a Qualified Auditor will include
a description of the scope of the audit and if the scope of the
defensive audit is too narrow to meet the specific needs of
SoundExchange, then SoundExchange should be permitted to round out the
findings with its own audit, limited to the points omitted from the
scope of the defensive audit.
    \200\ Music Choice uses the term ``defensive audit'' for this
h. Section 382.7(f)--Issuance of Audit Report
    On their own motion, the Judges change the word ``rendering'' to
the word ``issuing'' for clarity.
i. Section 382.7(g)--Interest on Underpayments Discovered by Audit
    The current regulations do not provide for a specific interest
accrual on underpayments discovered by audit. Sirius XM requested that
the Judges add a provision setting interest on underpayments discovered
by audit at the federal post-judgment rate in 28 U.S.C. 1961.
SoundExchange urged applying the late payment interest rate of 1.5% per
month, compounded monthly. Sirius XM requested that the federal post-
judgment rate that it seeks to be applied to late payments also be
applied to underpayments and overpayments discovered by audit. However,
Sirius XM opposed as punitive the use of SoundExchange's proposed 1.5%
per month interest rate, noting that audits may be delayed by up to
three years, while interest accrues. Barry WDT ] 8.
    The proposed regulations the Judges adopt in this proceeding
utilize the federal post-judgment rate rather than the more punitive
1.5% per month rate. Audits can uncover good faith errors as well as
bad faith manipulations, and the Judges do not find that a punitive
interest rate, spanning up to three years on underpayments, is
appropriate in such a circumstance.
j. Section 382.7(h)--Cost Shifting
    Current SDARS/PSS regulations provide that the Verifying Entity
bears the cost of an audit, unless the auditor finds an underpayment of
sufficient magnitude to justify shifting responsibility for payment to
the Payor. For PSS, the underpayment that triggers cost-shifting
currently is 5%. For SDARS, the underpayment that triggers cost-
shifting is 10%.\201\ Music Choice sought to equalize the cost-shifting
threshold, making all services liable if an audit discrepancy reaches
10%. SoundExchange argued that cost-shifting should occur when an
auditor discovers underpayment of 5% for PSS or SDARS. The rationale is
that the absolute value of SDARS royalty payments justifies reducing
the trigger.
    \201\ For Webcasters, the costs of the audit shift to the Payor
when an underpayment equals 10% or more.
    The Judges are unconvinced that absolute payment amounts are a
sufficient basis to change the cost-shifting trigger. Further, the
Judges can find no evidentiary basis to change the cost-shifting
threshold when all participants in this proceeding indicate that cost-
shifting has yet to occur at the current thresholds.
[[Page 65263]]
k. Sections 382.23(a) and (b)
    SoundExchange proposed changes to the methodology for Sirius XM to
calculate the direct license share and the pre-1972 license share.
Besides inserting language relating to Aggregate Tuning Hours (ATH)
data, SoundExchange sought to impose a requirement on Sirius XM to
report that usage data for every eligible track it claims as a directly
licensed or pre-1972 sound recording for which Sirius XM seeks a
royalty adjustment. Sirius XM contended that current reporting
requirements, based on Reference Channel metrics are sufficient to
support the royalty adjustments it takes for these exempt sound
    As the Judges decline to adopt the additional ATH language
requested by SoundExchange, they see no basis to impose the additional
reporting requirements on Sirius XM at this time.
l. Proposed Section--Distribution of SDARS Royalties
    SoundExchange proposed a new section 382.22 adding language to the
regulations that would permit it to adjust its distribution model by
reference to ATH if and when Sirius XM becomes able to track listener
usage of its satellite radio service. Sirius XM countered that it
anticipates offering next-generation technology within the rate period
at issue, but that this developing technology will not be sufficiently
reliable or have sufficient market saturation to make any reports of
its usage reliable. See 5/17/17 Tr. 4358 (Barry).
    Given the contingent nature of both the launch and the saturation
of Sirius XM's anticipated technological advances, the Judges decline
to adopt contingency regulations at this time.
m. Proposed Section--Finality of Audit Results
    Sirius XM proposed an additional subsection for the audit
provisions to establish the finality of disputed audit reports. Sirius
XM sought to establish a two-year statute of limitations for disputed
audit findings after which the Licensee's calculations would be deemed
binding and final, unless the Collective initiated a legal action
before the running of that proposed limitations period.
    SoundExchange objected to the creation of this statute of
limitations, asserting that the change Sirius XM requests would have
the effect of overriding the three-year statute of limitations provided
for in the Copyright Act. As SoundExchange argued, the Judges do not
have the authority to overrule a statutory provision by regulation.
    The Judges see no reason to establish a statute of limitations in
the context of rate setting proceedings where the Act does not provide
for one. Further, any pursuit of remedies relating to audit findings
would be outside the Judges' jurisdiction and the Judges would be
overstepping to attempt to impose a limitation of actions over which
they have no authority.
B. Gross Revenues
    In this proceeding, SoundExchange proposed a per-subscriber rate
structure for PSS and proffered PSS regulations consistent with its
proposed rate structure. Accordingly, SoundExchange proposed to place
its definition of ``Gross Revenues'' only in ``Subpart C,'' the subpart
regarding SDARS. The Judges have determined that PSS rates shall
continue to be calculated on a percent-of-revenue basis. Because the
business models of SDARS and PSS are different, however, the Judges
maintain separate elements for the calculation of the respective Gross
Revenues bases for PSS and SDARS.
    Neither Music Choice nor SoundExchange proposed a change to the
current definition of Gross Revenues applicable to PSS. The Judges
adopt that term to describe the method of calculating PSS royalties for
the 2018 to 2022 period.
    Sirius XM and SoundExchange proposed essentially the same
definition to establish the SDARS base for Gross Revenues. Their
substantive differences arose in the nature and explication of
permissible exclusions from that base.\202\ In adopting the definition
applicable to the license period at issue in this proceeding, the
Judges modified SoundExchange's proposed language to eliminate
ambiguity \203\ and to effect the decisions detailed below.
    \202\ Both SoundExchange and Sirius XM presented proposals to
resolve long-standing controversies that were brought into focus by
the primary jurisdiction referral of the questions from the D.C.
District Court. The need for the referral arose in SoundExchange v.
Sirius XM, 65 F. Supp. 3d 150 (D.D.C. 2014). In September 2017, the
Judges issued their amended ruling on the referred questions. See
Amended Restricted Ruling on Regulatory Interpretation Referred by
the United States District Court for the District of Columbia, No.
2006-1 CRB DSTRA (20017-12) (Sept. 11, 2017). (Ruling on Referred
Questions). The Judges resolve the same controversies in this
proceeding in conformity with that Ruling.
    \203\ In constructing its proposed definition of Gross Revenues,
SoundExchange began with a limited definition of what to include in
the base: Subscription revenues and ad revenues including those
categories of revenues if they were paid to a parent, subsidiary, or
division of the Licensee. SoundExchange then listed types of revenue
that should be excluded from the base ``to the extent otherwise
included'' in the definition of the base. The result is in the
nature of a double-negative configuration. For example, equipment
sales income is NOT included in the revenue base, but the exclusion
of equipment sales revenues would apply only ``to the extent [those
equipment sales revenues were] otherwise included'' in the base. The
better approach is to retain the current regulatory language, which
states simply, ``Gross Revenues shall exclude . . . .''
    SoundExchange proposed to amend the definition of ``Gross
Revenues'' currently found in 37 CFR 382.11 to confirm that revenue
from non-music offerings ``offered for a separate charge'' shall be
excludable only when those offerings are ``provided on a standalone
basis.'' Bender WDT at 22. SoundExchange did not view this new proposed
language as a substantive deviation from the existing regulations, but
rather made the proposal ``[p]urely [as] a clarification to language
that we had previously thought was sufficient.'' 5/10/17 Tr. 3184
    SoundExchange recounted that, since SDARS I, it has consistently
understood that the references to a ``separate charge'' in current
paragraph (3)(vi)(A) and (B) were unambiguous. See SDARS I, 73 FR at
4087 (explaining that the ``gross revenues'' definition ``excludes
monies attributable to premium channels of nonmusic programming that
are offered for a charge separate from the general subscription charge
for the service.''). See id. at 4081 (noting that, with regard to
``data services,'' the ``separate charge'' language was added by the
Judges ``to make clear that this portion of the definition dealing with
data services does not contemplate an exclusion of revenues from such
data services, where such data services are not offered for a separate
charge from the basic subscription product's revenues.'').
Additionally, SoundExchange pointed out that, in SDARS II, the Judges
reiterated the necessity of a ``separate charge,'' ``stress[ing] that
the exclusion is available only to the extent that the channels,
programming, products and/or other services are offered for a separate
charge.'' SDARS II, 78 FR at 23072 n.45.\204\
    \204\ SoundExchange asserted that its auditor alerted it to the
fact that, throughout the SDARS I period (at least), Sirius XM was
[REDACTED]. Trial Ex. 101 at 5-6, Schedule 3. As of the time
SoundExchange filed its direct case in the present proceeding,
Sirius XM continued to assert that the ``separate charge'' language
permitted deduction of an allocated part of its Premiere package.
Ruling on Referred Questions at 17.
    Subsequent to the filing of direct cases in this proceeding, the
Judges decided that ``the language in the revenue exclusion described
in subsection (vi)(B) did not permit Sirius XM to exclude from the
Gross Revenues
[[Page 65264]]
royalty base the price difference, i.e., the Upcharge, between the
Premier package and the Basic package.'' Amended Restricted Ruling on
Regulatory Interpretation Referred by the United States District Court
for the District of Columbia at 17, No. 2006-1 CRB DSTRA (2007-12)
(Ruling on Referred Questions). Given that decision, SoundExchange
noted that its proposed clarification may be unnecessary. Nonetheless,
in the interest of clarity, SoundExchange urged the Judges to ``confirm
again'' their position as to the meaning of the regulatory language
concerning exclusions to gross revenues. Bender WDT at 22.
    Sirius XM, conversely, criticized the current regulatory language
that limits the exclusion to revenue recognized for the provision of
data services and non-music channels, programming, products and/or
other services to those instances in which the subject programming is
offered for a ``separate charge.'' Sirius XM proposed to strike the
longstanding ``separate charge'' requirement and add new language to
the Gross Revenues definition allowing allocation of all bundle revenue
regardless of whether the components of the bundle are offered for a
separate charge. That proposed language specifies that the exclusion to
be taken in the case of any bundle is ``the difference between: (a) the
stated sale price of the bundle, minus (b) the stated sale price of the
bundle multiplied by a fraction, the numerator of which is the publicly
stated retail price of the standard music/non-music package when sold
on a standalone and undiscounted basis, and the denominator of which is
the publicly stated retail price of the bundle when sold on a
standalone and undiscounted basis.'' Sirius XM First Amended Proposed
Rates and Terms at 3 (Feb. 17, 2017); 5/17/17 Tr. 4342-48 (Barry);
Barry WRT ] 21.
    Sirius XM had no choice but to acknowledge that its proposal fails
to address the ``economic indeterminacy'' of its bundling approach. In
the Ruling on Referred Questions, the Judges held that--to use Sirius
XM's own words--``the difference between the larger bundle price and
the Select package price may not in all cases reliably measure the
economic value of the additional programming to consumers, at least
absent some objective evidence of the market value of that additional
programming.'' SXM PFF ] 440.
    Sirius XM sought to minimize the importance of this acknowledged
economic indeterminacy by noting the importance of bundling to Sirius
XM's business model and by pointing out the ubiquity of bundling by
many major businesses. Barry WRT ]] 12-18 & n.6. The Judges recognize
the importance of product bundling as described by Mr. Barry, both for
Sirius XM and numerous retailers of multiple products. As the Judges
explained at length in the Ruling on Referred Questions, such bundling
is a common form of price discrimination that increases revenue. That
is, sellers can induce buyers/subscribers to reveal their Willingness
to Pay (WTP) and pay more through bundling.
    In a context in which the retailers pay for their inputs on a per
unit basis, bundled retail pricing is benign, because input suppliers
would be indifferent to downstream pricing and bundling. However, when
the input suppler, as here, is paid as a percent of retail revenue, and
the bundled revenue consists of some revenue attributable to the
royalty base and other revenue excluded from the royalty base, the
economic indeterminacy of the revenue attributable to each bucket
creates a measurement problem, absent further information regarding the
WTP of buyers/subscribers to the bundle.
    Nonetheless, Sirius XM urged that the ``practical benefits'' of its
proposal outweigh such economic indeterminacy. The Judges disagree and
reaffirm their conclusions in the Ruling on Referred Questions arising
from the SDARS I proceeding. As Mr. Barry made clear, such bundling was
undertaken to increase Sirius XM's revenues and it would be reasonable
to assume that Sirius XM has information relevant to the economic
allocation of the bundled revenue. However, Sirius XM presented no such
evidence at the hearing. Sirius XM must bear the burden of providing
evidence that might mitigate the acknowledged ``economic
indeterminacy'' problem inherent in bundling, because any such evidence
would be in its possession, not in the possession of SoundExchange or
the record companies. If Sirius XM lacks allocation information and
prices its bundles without that data, it cannot assert ``practical
benefits'' as grounds for subjecting licensors to the acknowledged
economic indeterminacy of the revenue split.
    For all of the reasons stated, and based upon the Judges' analysis
in the Ruling on Referred Questions, the Judges reject Sirius XM's
attempts to rewrite the regulations to reach a contrary result. Because
the Judges are reaffirming here their Ruling on Referred Questions,
which confirmed the meaning of the present regulatory definition of
Gross Revenues, they find (as SoundExchange itself anticipated) no need
to amend the text of the regulatory definition. Accordingly,
SoundExchange's request for a change in that definitional language is
rejected as moot.
    Finally, Sirius XM proposed a change to the prefatory language in
the exclusion from ``Revenues recognized by Licensee for the provision
of'' to the simpler ``Licensee revenues for the provision of.'' (That
language is set forth in forthcoming Sec.  382.22(b)(7)). As Mr. Barry
explained, this is not meant to imply that Sirius XM can exclude
revenues that have not been recognized. Rather, it is merely intended
to avoid SoundExchange's perpetuating the argument (as addressed and
rejected by the Judges in the recent litigation regarding the SDARS I
period) that Sirius XM could not exclude revenue for portions of a
bundle because those items were not separate units of accounting under
GAAP (and the revenue for those items therefore was not
``recognized''). Barry WRT ] 20 n.8.
    SoundExchange argued that there is no reason to delete the
reference to ``[r]evenues recognized'' in the preamble, and some risk
in doing so. SE Response to SXM PFF ] 442. However, SoundExchange did
not cite to the record for this assertion of risk, nor did it identify
that alleged risk. SoundExchange also noted that, at the hearing, Mr.
Barry acknowledged his understanding that revenue would need to be
``recognized'' to be excluded. 5/17/17 Tr. 4401-02 (Barry). Thus,
SoundExchange concluded that deleting the reference to revenue
recognition would create the implication that that is not the case.
    The Judges find that these differences can be bridged. The language
at 382.22(b)(7) will read, ``Revenues recognized by Licensee (or
otherwise received by Licensee if no GAAP ``recognition'' principles
are applicable) for the provision of . . . .''
C. Ephemeral License Terms
    The participants in the present proceeding raised two issues
relating to the section 112 Ephemeral Recordings license. The first
issue was raised by Music Choice regarding the valuation of the
ephemeral license. The second controversy between SoundExchange and
Music Choice came to light in response to SoundExchange's proposed
revisions to Sec. Sec.  382.3(b) and 382.12(c) regarding advance
payments and minimum payments and is discussed supra, section XI.A.1.
SoundExchange contended that the record in the proceeding
``unanimously'' supports SoundExchange's proposal of a bundled rate for
both the Section 112(e) and 114 rights, 5% of which should be allocated
as the Section 112(e) royalty for the
[[Page 65265]]
making of ephemeral copies and the remaining 95% of which should be
allocated as the Section 114 performance royalty. SoundExchange stated
that ``[t]he parties agree in substance concerning this matter.'' SX
PFFCL ] 2369. SoundExchange contended that ``it appears that both
SoundExchange and Music Choice agree that the Judges should set some
kind of an overall royalty payment and allocate it 95%/5%.'' SX PFFCL ]
    Sirius XM mirrored SoundExchange's proposal. See SXM PFFCL at 1.
Music Choice argued, however, that SoundExchange did not demonstrate
that ephemeral copies have any independent value. See Del Beccaro WDT
at 46-47 (``I am unaware of any marketplace context in which the record
labels seek, or get, a separate payment just for ephemeral copies.'').
Nevertheless, Music Choice acknowledged that the ephemeral license has
been and can be bundled with the sound recording performance license,
and took no position on SoundExchange's proposal to continue the
current apportionment between the performance and ephemeral copying
license. MC PFF ]551.
    SoundExchange, Sirius XM, and Music Choice agreed that a portion of
the overall PSS royalties should be attributed to the ephemeral copying
license. None of them suggested that the overall PSS royalty rate
should be increased to account for ephemeral copying royalties.
SoundExchange and Sirius XM proposed that the current 5% allocation of
overall royalties to the section 112(e) license should continue in the
upcoming rate period, and Music Choice took no position on the
allocation. The only apparent issue concerning the ephemeral
reproduction license is that Music Choice asserted that that license
has no ``independent value,'' MC PFF at ]550 (emphasis added), while
SoundExchange contended that ephemeral copies do ``have economic value
. . . .'' Designated Web III Written Direct Testimony of Dr. George S.
Ford, Trial Ex. 51, at 9 (Ford Web III WDT). Music Choice did not
contend that the ephemeral copies have no economic value--only that the
ephemeral copies have no economic value independent of the Section 114
license. Music Choice's position was inconsistent with neither
SoundExchange's contention that the ephemeral copying does have
economic value, nor a bundled rate allocated between the two licenses.
    To support both the bundled rate and the proposed 5% allocation to
the ephemeral license, SoundExchange relied on the designated testimony
of Dr. George Ford from the Web III proceeding. See generally Ford Web
III WDT; see also Web III, 76 FR at 13042 (``The testimony offered by
SoundExchange supports this proposal and we adopt it.''). According to
Dr. Ford, ``ephemeral copies have economic value to services that
publicly perform sound recordings because these services cannot as a
practical matter properly function without those copies.'' Ford Web III
WDT at 9. Dr. Ford noted that ``marketplace benchmarks show that the
royalty rate for ephemeral copies, if directly established, is almost
always expressed as a percentage of the overall royalty rate for
combined activities under Section 112 and 114.'' Id. at 9-10.
    As to the specific allocation between the two licenses, Dr. Ford
noted that it is not the services, but the ``[r]ecord companies and
artists [who] care about what portion of royalty payments are allocated
to ephemerals because the higher the portion allocated to ephemerals,
the lower the portion paid directly to artists per the terms of the
Section 114 license.'' Id. at 4. Dr. Ford concluded that, in light of
the purported disinterest by the willing buyer (or licensee) in the
allocation between the Section 112(e) and 114 licenses, an agreement
between the artists and the copyright owners (i.e., the licensors) is
the best measure of how a willing buyer and willing seller would
allocate royalties between the performance and ephemeral licenses. Id.
at 10. As evidence of such an agreement, Dr. Ford was informed that
``the recording artists and the record companies have reached an
agreement that five percent (5%) of the payments for activities under
Section 112(e) and 114 should be allocated to Section 112(e)
activities.'' Id. at 15. He concluded that ``that appears to be a
reasonable proposal.'' Id. Upon examination in Web III, Dr. Ford
clarified that he was informed by counsel for SoundExchange that the
SoundExchange board, which includes representatives from record labels
and artists, had approved a recommendation that 5% of royalties should
be allocated to the ephemeral license. Designated Hearing Testimony of
George S. Ford, Trial Ex. 51, at 434 (Ford Web III Hrg. Test.).\205\
    \205\ Dr. Ford represented that he reviewed the minutes of the
board meeting that referenced the agreement, and it appears that the
Judges in Web III admitted the board minutes into evidence. Ford Web
III Hrg. Test. at 434, 438. Those minutes were not introduced into
evidence in the current proceeding, rendering hearsay Dr. Ford's
testimony concerning the agreement between artists and record
companies. The Judges exercise their discretion under 37 CFR
351.10(a) to admit Professor Ford's hearsay testimony.
    The Judges find SoundExchange's proposals concerning the bundling
of performance and ephemeral Royalties, as well as the 95%/5%
allocation of royalties between the two licenses, to be reasonable and
supported by the evidence, and therefore adopt them for both PSS and
XII. Conclusion
    For all of the foregoing reasons, the Judges issue this
Determination of Rates and Terms in the captioned proceeding. The
Register of Copyrights may review the Judges' Determination for legal
error in resolving a material issue of substantive copyright law. The
Librarian shall cause the Judges' Determination, and any correction
thereto by the Register, to be published in the Federal Register no
later than the conclusion of the 60-day review period.
    Dated: October 11, 2018.
    Corrected: October 15, 2018.
Suzanne M. Barnett,
Chief Copyright Royalty Judge.
Jesse M. Feder,
Copyright Royalty Judge.
David R. Strickler,
Copyright Royalty Judge.
List of Subjects in 37 CFR Part 382
    Copyright, Digital audio transmissions, Performance right, Sound
Final Regulations
    For the reasons set forth in the preamble, the Copyright Royalty
Judges revise 37 CFR part 382 to read as follows:
Subpart A--Regulations of General Application
382.1 Definitions.
382.2 Scope and compliance.
382.3 Making payment of royalty fees.
382.4 Delivering statements of account.
382.5 Distributing royalty fees.
382.6 Handling Confidential Information.
382.7 Auditing payments and distributions.
Subpart B--Preexisting Subscription Services (PSS)
382.10 Royalty fees for the digital performance of sound recordings
and the making of ephemeral recordings by preexisting subscription
382.11 Calculation of gross revenues for PSS.
[[Page 65266]]
Subpart C--Preexisting Satellite Digital Audio Radio Services (SDARS)
382.20 Definitions.
382.21 Royalty fees for the public performance of sound recordings
and the making of ephemeral recordings by SDARS.
382.22 Calculation of Gross Revenues for SDARS.
382.23 Adjustments to royalty fee.
    Authority: 17 U.S.C. 112(e), 114 and 801(b)(1).
Subpart A--Regulations of General Application
Sec.  382.1  Definitions.
    In this subpart:
    Collective means the collection and distribution organization that
is designated by the Copyright Royalty Judges.
    Copyright Owners means sound recording copyright owners who are
entitled to royalty payments made under part 382 pursuant to the
statutory licenses under 17 U.S.C. 112(e) and 114.
    Digital Audio Transmission has the same meaning as in 17 U.S.C.
    Eligible Transmission means a Digital Audio Transmission made by a
Licensee that is subject to licensing under 17 U.S.C. 114(d)(2) and the
payment of royalties under 37 CFR part 382.
    Ephemeral Recording has the same meaning as in 17 U.S.C. 112.
    GAAP means generally accepted accounting principles in effect in
the United States on the date payment is due.
    Licensee means the provider of an Satellite Digital Audio Radio
Service (SDARS) or Preexisting Subscription Service (PSS) that has
obtained a license under 17 U.S.C. 114 to make eligible transmissions
and a license under 17 U.S.C. 112(e) to make Ephemeral Recordings to
facilitate those Eligible Transmissions.
    Payor means the entity required to make royalty payments to the
Collective or the entity required to distribute royalty fees collected,
depending on context. The Payor is:
    (1) A Licensee, in relation to the Collective; and
    (2) The Collective in relation to a Copyright Owner or Performer.
    Performers means the independent administrators identified in 17
U.S.C. 114(g)(2)(B) and (C) and the parties identified in 17 U.S.C.
    Preexisting Subscription Service (PSS) has the same meaning as in
17 U.S.C. 114(j)(11). A service's offering on the internet that is
available to a subscriber outside the subscriber's residence is not a
Preexisting Subscription Service for purposes of this part.
    Qualified Auditor means a Certified Public Accountant independent
within the meaning of the American Institute Certified Public
Accountants Code of Professional Conduct.
    Satellite Digital Audio Radio Service (SDARS) means the preexisting
satellite digital audio radio services as defined in 17 U.S.C.
    Transmission has the same meaning as in 17 U.S.C. 114(j)(15).
    Verifying Entity means the party requesting an audit and giving
notice of intent to audit. For audits of SDARS and PSS, the Verifying
Entity is SoundExchange, Inc. For audits of SoundExchange, Inc. the
Verifying Entity is any Copyright Owner or its authorized
Sec.  382.2  Scope and compliance.
    (a) Scope. This part codifies rates and terms of royalty payments
for the public performance of sound recordings in certain Digital Audio
Transmissions by certain Licensees in accordance with applicable
provisions of 17 U.S.C. 114 and for the making of Ephemeral Recordings
by those Licensees in accordance with the provisions of 17 U.S.C.
112(e), during the period January 1, 2018, through December 31, 2027.
    (b) Legal compliance. Licensees relying upon the statutory licenses
set forth in 17 U.S.C. 112(e) and 114 must comply with the requirements
of 17 U.S.C. 112(e) and 114, this part and any other applicable
    (c) Voluntary agreements. Notwithstanding the royalty rates and
terms established in any subparts of this part, the rates and terms of
any license agreements entered into by Copyright Owners and Licensees
may apply in lieu of these rates and terms.
Sec.  382.3  Making payment of royalty fees.
    (a) Payment to the Collective. A Licensee must make the royalty
payments due under subparts B and C of this part to SoundExchange,
Inc., which is the Collective designated by the Copyright Royalty Board
to collect and distribute royalties under this part. If any payment due
date is a weekend or a federal holiday, then the payment is due on the
first business day thereafter.
    (b) Advance payment. Licensees must pay the Collective an annual
advance payment of $100,000 by January 31 of each year. The Collective
must credit 5% of the advance payment as payment of the minimum fee for
Ephemeral Recordings and credit the remaining 95% to section 114
royalties. The funds are nonrefundable. Any uncredited portion of the
funds shall not carry over into a subsequent year.
    (c) Minimum payments. A Licensee must make any minimum annual
payment due under subpart B or C of this part by January 31 of the
applicable license year.
    (d) Monthly payments. A Licensee must make royalty payments on a
monthly basis. Payments are due on or before the 45th day after the end
of the month in which the Licensee made Eligible Transmissions.
    (e) Late fees. A Licensee must pay a late fee for each payment and
each Statement of Account that the Collective receives after the due
date. The late fee is 1.5% (or the highest lawful rate, whichever is
lower) of the late payment amount per month. The late fee for a late
Statement of Account is 1.5% of the payment amount associated with the
Statement of Account. Late fees accrue from the due date until the date
that the Collective receives the late payment or late Statement of
    (1) Waiver of late fees. The Collective may waive or lower late
fees for immaterial or inadvertent failures of a Licensee to make a
timely payment or submit a timely Statement of Account.
    (2) Notice regarding noncompliant Statements of Account. If it is
reasonably evident to the Collective that a timely-provided Statement
of Account is materially noncompliant, the Collective must notify the
Licensee within 90 days of discovery of the noncompliance.
Sec.  382.4  Delivering statements of account.
    (a) Statements of Account. Any payment due under this part must be
accompanied by a corresponding Statement of Account that must contain
the following information:
    (1) Information as is necessary to calculate the accompanying
royalty payment;
    (2) The name, address, business title, telephone number, facsimile
number (if any), electronic mail address (if any) and other contact
information of the person to be contacted for information or questions
concerning the content of the Statement of Account;
    (3) The signature of:
    (i) The Licensee or a duly authorized agent of the Licensee;
    (ii) A partner or delegate if the Licensee is a partnership; or
    (iii) An officer of the corporation if the Licensee is a
    (4) The printed or typewritten name of the person signing the
Statement of Account;
    (5) If the Licensee is a partnership or corporation, the title or
official position held in the partnership or corporation by the person
signing the Statement of Account;
[[Page 65267]]
    (6) A certification of the capacity of the person signing;
    (7) The date of signature; and
    (8) An attestation to the following effect:
    I, the undersigned owner/officer/partner/agent of the Licensee have
examined this Statement of Account and hereby state that it is true,
accurate, and complete to my knowledge after reasonable due diligence
and that it fairly presents, in all material respects, the liabilities
of the Licensee pursuant to 17 U.S.C. 112(e) and 114 and applicable
regulations adopted under those sections.
    (b) Certification. Licensee's Chief Financial Officer or, if
Licensee does not have a Chief Financial Officer, a person authorized
to sign Statements of Account for the Licensee, must submit a signed
certification on an annual basis attesting that Licensee's royalty
statements for the prior year represent a true and accurate
determination of the royalties due and that any method of allocation
employed by Licensee was applied in good faith and in accordance with
Sec.  382.5  Distributing royalty fees.
    (a) Distribution of royalties. (1) The Collective must promptly
distribute royalties received from Licensees to Copyright Owners and
Performers that are entitled thereto, or to their designated agents.
The Collective shall only be responsible for making distributions to
those who provide the Collective with information necessary to identify
and pay the correct recipient. The Collective must distribute royalties
on a basis that values all performances by a Licensee equally based
upon the information provided under the Reports of Use requirements for
Licensees pursuant to Sec.  370.3 or Sec.  370.4 of this chapter, as
applicable, and pursuant to this part.
    (2) Identification of Copyright Owners. The Collective must use its
best efforts to identify and locate copyright owners and featured
artists to distribute royalties payable to them under section 112(e) or
114(d)(2) of title 17, United States Code, or both. Such efforts must
include, but are not limited to, searches in Copyright Office public
records and published directories of sound recording copyright owners
when consulting those records and directories is likely to be helpful.
    (b) Unclaimed funds. If the Collective is unable to identify or
locate a Copyright Owner or Performer who is entitled to receive a
royalty distribution under this part, the Collective must retain the
required payment in a segregated trust account for a period of three
years from the date of the first distribution of royalties from the
relevant payment by a Licensee. No claim to distribution shall be valid
after the expiration of the three-year period. After expiration of this
period, the Collective must handle unclaimed funds in accordance with
applicable federal, state, or common law.
    (c) Retention of records. Licensees and the Collective shall keep
books and records relating to payments and distributions of royalties
for a period of not less than the prior three calendar years.
    (d) Designation of the Collective. (1) The Judges designate
SoundExchange, Inc., as the Collective to receive Statements of Account
and royalty payments from Licensees and to distribute royalty payments
to each Copyright Owner and Performer (or their respective designated
agents) entitled to receive royalties under 17 U.S.C. 112(e) or 114(g).
    (2) If SoundExchange, Inc. should dissolve or cease to be governed
by a board consisting of equal numbers of representatives of Copyright
Owners and Performers, it shall be replaced for the applicable royalty
period by a successor Collective according to the following procedure:
    (i) The nine Copyright Owner representatives and the nine Performer
representatives on the SoundExchange board as of the last day preceding
SoundExchange's cessation or dissolution shall vote by a majority to
recommend that the Copyright Royalty Judges designate a successor and
must file a petition with the Copyright Royalty Judges requesting that
the Judges designate the named successor and setting forth the reasons
    (ii) Within 30 days of receiving the petition, the Copyright
Royalty Judges must issue an order designating the recommended
Collective, unless the Judges find good cause not to make and publish
the designation in the Federal Register.
Sec.  382.6  Handling Confidential Information.
    (a) Definition. For purposes of this part, ``Confidential
Information'' means the Statements of Account and any information
contained therein, including the amount of royalty payments and any
information pertaining to the Statements of Account reasonably
designated as confidential by the party submitting the statement.
Confidential Information does not include documents or information that
at the time of delivery to the Collective is public knowledge. The
party seeking information from the Collective based on a claim that the
information sought is a matter of public knowledge shall have the
burden of proving to the Collective that the requested information is
in the public domain.
    (b) Use of Confidential Information. The Collective may not use any
Confidential Information for any purpose other than royalty collection
and distribution and activities related directly thereto.
    (c) Disclosure of Confidential Information. The Collective shall
limit access to Confidential Information to:
    (1) Employees, agents, consultants, and independent contractors of
the Collective, subject to an appropriate written confidentiality
agreement, who are engaged in the collection and distribution of
royalty payments hereunder and activities related directly thereto who
require access to the Confidential Information for the purpose of
performing their duties during the ordinary course of their work;
    (2) A Qualified Auditor or outside counsel who is authorized to act
on behalf of:
    (i) The Collective with respect to verification of a Licensee's
statement of account pursuant to this part; or
    (ii) A Copyright Owner or Performer with respect to the
verification of royalty distributions pursuant to this part;
    (3) Copyright Owners and Performers, including their designated
agents, whose works a Licensee used under the statutory licenses set
forth in 17 U.S.C. 112(e) and 114 by the Licensee whose Confidential
Information is being supplied, subject to an appropriate written
confidentiality agreement, and including those employees, agents,
consultants, and independent contractors of such Copyright Owners and
Performers and their designated agents, subject to an appropriate
written confidentiality agreement, who require access to the
Confidential Information to perform their duties during the ordinary
course of their work;
    (4) Attorneys and other authorized agents of parties to proceedings
under 17 U.S.C. 112 or 114, acting under an appropriate protective
    (d) Safeguarding Confidential Information. The Collective and any
person authorized to receive Confidential Information from the
Collective must implement procedures to safeguard against unauthorized
access to or dissemination of Confidential Information using a
reasonable standard of care, but no less than the same degree of
security that the recipient uses to protect its own Confidential
Information or similarly sensitive information.
[[Page 65268]]
Sec.  382.7  Auditing payments and distributions.
    (a) General. This section prescribes procedures by which any entity
entitled to receive payment or distribution of royalties may verify
those payments or distributions with an independent audit. The
Collective may audit a Licensee's payments of royalties to the
Collective and a Copyright Owner or Performer may audit the
Collective's distributions of royalties to the Copyright Owners or
Performers. Nothing in this section shall preclude a Verifying Entity
and the Payor under audit from agreeing to verification methods in
addition to or different from those set forth in this section.
    (b) Frequency of auditing. A Verifying Entity may conduct an audit
of each Payor only once a year and the audit may cover any or all of
the prior three calendar years. A Verifying Entity may not audit
records for any calendar year more than once.
    (c) Notice of intent to audit. The Verifying Entity must file with
the Copyright Royalty Judges a notice of intent to audit the Payor,
which notice the Judges must publish in the Federal Register within 30
days of the filing of the notice. Simultaneously with the filing of the
notice, the Verifying Entity must send a copy to the Payor.
    (d) The audit. The audit must be conducted during regular business
hours by a Qualified Auditor who is not retained on a contingency fee
basis and is identified in the notice. The auditor shall determine the
accuracy of royalty payments or distributions, including whether the
Payor made an underpayment or overpayment of royalties. An audit of
books and records, including underlying paperwork, performed in the
ordinary course of business according to generally accepted auditing
standards by a Qualified Auditor, shall serve as an acceptable
verification procedure for all parties with respect to the information
that is within the scope of the audit.
    (e) Access to third-party records for audit purposes. The Payor
under audit must use commercially reasonable efforts to obtain or to
provide access to any relevant books and records maintained by third
parties for the purpose of the audit.
    (f) Duty of auditor to consult. The auditor must produce a written
report to the Verifying Entity. Before issuing the report, unless the
auditor has a reasonable basis to suspect fraud on the part of the
Payor, the disclosure of which would, in the reasonable opinion of the
auditor, prejudice any investigation of the suspected fraud. The
auditor must review tentative written findings of the audit with the
appropriate agent or employee of the Payor in order to remedy any
factual errors and clarify any issues relating to the audit; provided
that an appropriate agent or employee of the Payor reasonably
cooperates with the auditor to remedy promptly any factual error[s] or
clarify any issue raised by the audit. The auditor must include in the
written report information concerning the cooperation or the lack
thereof of the employee or agent.
    (g) Audit results; underpayment or overpayment of royalties. If the
auditor determines the Payor underpaid royalties, the Payor shall remit
the amount of any underpayment determined by the auditor to the
Verifying Entity, together with interest at the post-judgment rate
specified in 28 U.S.C. 1961, accrued from and after the date the
payment was originally due. In the absence of mutually-agreed payment
terms, which may, but need not, include installment payments, the Payor
shall remit promptly to the Verifying Entity the entire amount of the
underpayment determined by the auditor. If the auditor determines the
Payor overpaid royalties, however, the Verifying Entity shall not be
required to remit the amount of any overpayment to the Payor, and the
Payor shall not seek by any means to recoup, offset, or take a credit
for the overpayment, unless the Payor and the Verifying Entity have
agreed otherwise.
    (h) Paying the costs of the audit. The Verifying Entity must pay
the cost of the audit, unless the auditor determines that there was an
underpayment of 10% or more, in which case the Payor must bear the
reasonable costs of the audit, in addition to paying or distributing
the amount of any underpayment.
    (i) Retention of audit report. The Verifying Entity must retain the
report of the audit for a period of not less than three years from the
date of issuance.
Subpart B--Preexisting Subscription Services (PSS)
Sec.  382.10  Royalty fees for the digital performance of sound
recordings and the making of ephemeral recordings by preexisting
subscription services.
    (a) Royalty fees. Commencing January 1, 2018, and continuing
through December 31, 2027, Licensees must pay royalty fees for all
Eligible Transmissions of sound recordings at the rate of 7.5 percent
of Gross Revenues.
    (b) Ephemeral recordings royalty fee. (1) The fee for all Ephemeral
Recordings is part of the total fee payable under this section and
constitutes 5% of it. All Ephemeral Recordings that a Licensee makes
that are necessary and commercially reasonable for making
noninteractive Digital Audio Transmission as a PSS are included in the
    (2) The minimum fee is $5,000 per year.
Sec.  382.11  Calculation of gross revenues for PSS.
    (a) Gross revenues are monies derived from the operation of the
programming service of the Licensee and are comprised of the following:
    (1) Monies received by Licensee from Licensee's carriers and
directly from residential U.S. subscribers for Licensee's programming
    (2) Licensee's advertising revenues (as billed), or other monies
received from sponsors, if any, less advertising agency commissions not
to exceed 15% of those fees incurred to a recognized advertising agency
not owned or controlled by Licensee;
    (3) Monies received for the provision of time on the programming
service to any third party;
    (4) Monies received from the sale of time to providers of paid
programming such as infomercials;
    (5) Where merchandise, service, or anything of value is received by
Licensee in lieu of cash consideration for the use of Licensee's
programming service, the fair market value thereof or Licensee's
prevailing published rate, whichever is less;
    (6) Monies or other consideration received by Licensee from
Licensee's carriers, but not including monies received by Licensee's
carriers from others and not accounted for by Licensee's carriers to
Licensee, for the provision of hardware by anyone and used in
connection with the programming service;
    (7) Monies or other consideration received for any references to or
inclusion of any product or service on the programming service; and
    (8) Bad debts recovered regarding paragraphs (a)(1) through (7) of
this section.
    (9) Revenues described in paragraphs (a)(1) through (8) of this
section to which Licensee is entitled but which are paid to a parent,
subsidiary, division, or affiliate of Licensee, in lieu of payment to
Licensee but not including payments to Licensee's carriers for the
programming service.
    (b) Gross Revenues exclude affiliate revenue returned during the
reporting period and bad debts actually written off during reporting
[[Page 65269]]
Subpart C--Preexisting Satellite Digital Audio Radio Services
Sec.  382.20  Definitions.
    In this subpart:
    Directly-Licensed Recording means a sound recording for which the
Licensee has previously obtained a license of all relevant rights from
the sound recording Copyright Owner.
    Pre-1972 Recording means a sound recording fixed before February
15, 1972, that is not a restored work as defined in 17 U.S.C.
104A(h)(6) or otherwise subject to protection under title 17, United
States Code.
    Reference Channels means internet webcast channels offered by the
Licensee that directly correspond to channels offered on the Licensee's
SDARS that are capable of being received on all models of Sirius radio,
all models of XM radio or both, and on which the programming consists
primarily of music.
Sec.  382.21  Royalty fees for the public performance of sound
recordings and the making of ephemeral recordings by SDARS.
    (a) Royalty fees. Commencing January 1, 2018, and continuing
through December 31, 2027, Licensees must pay royalty fees for all
Eligible Transmissions of sound recordings at the rate of 15.5% of
Gross Revenues.
    (b) Ephemeral recordings royalty fees. (1) The fee for all
Ephemeral Recordings is part of the total fee payable under this
section and constitutes 5% of it. All Ephemeral Recordings that a
Licensee makes that are necessary and commercially reasonable for
making noninteractive Digital Audio Transmissions as an SDARS are
included in the 5%.
    (2) The minimum fee is $5,000 per year.
Sec.  382.22  Calculation of Gross Revenues for SDARS.
    (a) Gross Revenues are:
    (1) Revenue recognized by the Licensee in accordance with GAAP from
the operation of an SDARS and comprised of the following:
    (i) Subscription revenue recognized by Licensee directly from U.S.
subscribers for licensee's SDARS; and
    (ii) Licensee's advertising revenues, or other monies received from
sponsors, if any, attributable to advertising on channels other than
those that use only incidental performances of sound recordings, less
advertising agency and sales commissions.
    (2) Revenues set forth above to which Licensee is entitled but
which are paid to a parent, wholly-owned subsidiary, or division of
    (b) Gross Revenues exclude:
    (1) Monies or other consideration attributable to the sale and/or
license of equipment and/or other technology, including but not limited
to bandwidth, sales of devices that receive the Licensee's SDARS and
any shipping and handling fees therefor;
    (2) Royalties paid to Licensee for intellectual property rights;
    (3) Monies or other consideration received by Licensee from the
sale of phonorecords and digital phonorecord deliveries;
    (4) Sales and use taxes;
    (5) Credit card, invoice, activation, swap and early termination
fees charged to subscribers and reasonably related to the Licensee's
expenses to which they pertain;
    (6) Bad debt expense; and
    (7) Revenues recognized by Licensee (or otherwise received by
Licensee if no GAAP ``recognition'' principles are applicable) for the
provision of:
    (i) Current and future data services offered for a separate charge
(e.g., weather, traffic, destination information, messaging, sports
scores, stock ticker information, extended program associated data,
video and photographic images, and such other telematics and/or data
services as may exist from time to time);
    (ii) Channels, programming, products and/or other services offered
for a separate charge where such channels use only incidental
performances of sound recordings;
    (iii) Channels, programming, products and/or other services
provided outside of the United States; and
    (iv) Channels, programming, products and/or other services for
which the performance of sound recordings and/or the making of
Ephemeral Recordings is exempt from any license requirement or is
separately licensed, including by a statutory license and, for the
avoidance of doubt, webcasting, audio services bundled with television
programming, interactive services, and transmissions to business
Sec.  382.23  Adjustments to royalty fee.
    (a) Reduction for Direct License Share. The royalty fee specified
in Sec.  382.21(a) may be reduced by the percentage of Eligible
Transmissions comprising the Direct License Share.
    (1) The Direct License Share reduction is available to a Licensee
only if--
    (i) The Reference Channels constitute a large majority of and are
generally representative of the music channels offered on the
Licensee's SDARS; and
    (ii) The Licensee provides the Collective, by no later than the due
date for the relevant payment under Sec.  382.3(d), a list of each
Copyright Owner from which the Licensee claims to have a direct license
of rights to Directly-Licensed Recordings that is in effect for the
month for which the payment is made and of each sound recording for
which the Licensee takes the reduction, identified by featured artist
name, sound recording title, and International Standard Recording Code
(ISRC) number or, alternatively to the ISRC, album title and copyright
owner name. Notwithstanding Sec.  382.6, the Collective may disclose
such information as reasonably necessary for it to confirm whether a
claimed direct license exists and claimed sound recordings are properly
    (2) To arrive at the percentage allocable to the Direct License
Share for each month, the Licensee shall divide the internet
Performances of Directly-Licensed Recordings on the Reference Channels
by the total number of internet Performances of all sound recordings on
the Reference Channels. In no event shall the Direct License Share be
an amount greater than the result of dividing the number of plays of
Directly-Licensed Recordings on the SDARS by the total number of plays
of all sound recordings on the SDARS.
    (3) The Licensee may not credit use of a Directly-Licensed
Recording under this paragraph if that use is credited as a use of a
Pre-1972 Sound Recording for purposes of claiming the Pre-1972
Recording Share reduction to the royalty fee.
    (b) Reduction for Pre-1972 Recording Share. The royalty fee
specified in Sec.  382.21(a) may be reduced by the percentage of
Eligible Transmissions comprising the Pre-1972 Recording Share.
    (1) A Pre-1972 Recording Share reduction is available to a Licensee
only if--
    (i) The Reference Channels constitute a large majority of and are
generally representative of the music channels offered on the
Licensee's SDARS; and
    (ii) The Licensee provides to the Collective, by no later than the
due date for the relevant payment under Sec.  382.3(d), a list of Pre-
1972 Recordings for which the Licensee takes the reduction, identified
by featured artist name, sound recording title, and International
Standard Recording Code (ISRC) number or, alternatively to the ISRC,
album title and copyright owner name.
    (2) To arrive at the percentage allocable to the Pre-1972 Recording
Share for each month, the Licensee shall divide the internet
Performances of Pre-1972 Sound Recordings on the
[[Page 65270]]
Reference Channels by the total number of internet Performances of all
sound recordings on the Reference Channels.
    (c) Definition of Performance. For purposes of this section,
Performance means:
    (1) Except as discussed in paragraph (c)(2) of this section, a
Performance is an instance in which any portion of a sound recording is
publicly performed to a listener within the United States by means of a
Digital Audio Transmission (e.g., the delivery of any portion of a
single track from a compact disc to one listener).
    (2) An instance in which a portion of a sound recording is publicly
performed to a listener within the United States by means of a Digital
Audio Transmission is not a Performance if it both:
    (i) Makes no more than incidental use of sound recordings
including, but not limited to, brief musical transitions in and out of
commercials or program segments, brief use during news, talk and sports
programming, brief background use during disk jockey announcements,
brief use during commercials of sixty seconds or less in duration, or
brief use during sporting or other public events; and
    (ii) Does not contain an entire sound recording and does not
feature a particular sound recording of more than thirty seconds (as in
the case of a sound recording used as a theme song), except for ambient
music that is background at a public event.
Suzanne M. Barnett,
Chief Copyright Royalty Judge.
Jesse M. Feder,
Copyright Royalty Judge.
David R. Strickler,
Copyright Royalty Judge.
    Approved by:
Carla D. Hayden,
Librarian of Congress.
[FR Doc. 2018-26922 Filed 12-18-18; 8:45 am]