Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III)

AGENCY:

Copyright Royalty Board, Library of Congress.

ACTION:

Final rule and order.

SUMMARY:

The Copyright Royalty Judges announce their final determination after remand of the rates and terms for making and distributing phonorecords for the period beginning January 1, 2018, and ending on December 31, 2022.

DATES:

Effective date: August 10, 2023.

Applicability date: The regulations apply to the license period beginning January 1, 2018, and ending December 31, 2022.

ADDRESSES:

The final determination after remand is posted in eCRB at https://app.crb.gov/. For access to the docket to read the final determination after remand and submitted background documents, go to eCRB and search for docket number 16–CRB–0003–PR (2018–2022) (Remand).

FOR FURTHER INFORMATION CONTACT:

Anita Brown, CRB Program Assistant, (202) 707–7658, crb@loc.gov.

SUPPLEMENTARY INFORMATION:

Final Determination After Remand

On October 26, 2020, the United States Court of Appeals for the D.C. Circuit (D.C. Circuit) issued its mandate vacating and remanding in part the original Determination issued by the Copyright Royalty Judges (Judges) in the captioned proceeding. See Johnson v. Copyright Royalty Board, 969 F.3d 363 (D.C. Cir. 2020). In its ruling on appeal, the D.C. Circuit found that in the original Determination, the Judges (1) failed to give adequate notice to participants of their overhaul of the royalty rate structure combined with significantly increased and uncapped rates for section 115 licenses; (2) failed to explain why they rejected a benchmark based on a past settlement agreement in lieu of overhauling of the rate structure and significantly increasing rates; and (3) failed to identify their legal authority to redefine a material term after they promulgated a definition of that term in the original Initial Determination circulated to the participants. See Johnson, 969 F.3d at 367, 381; Initial Determination, Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 16–CRB–0003–PR (2018–2022) (Jan. 27, 2018).

After receipt of the D.C. Circuit's ruling and mandate, the Judges consulted with the parties to the appeal and established procedures for the remand proceeding. See Order Adopting Schedule for . . . Remand (Dec. 23, 2020). Each side offered opening submissions, responsive submissions, additional evidentiary filings, and further supplemental briefing requested by the Judges. The parties' submissions included legal briefing and incorporated evidence from the original proceeding as well as evidence newly developed for the remand proceeding. After preliminary deliberations, the Judges asked for supplemental briefing from the parties responsive to a proposed alternative rate structure. See Notice and Sua Sponte Order Directing the Parties to Provide Additional Materials (Dec. 9, 2021). With respect to redefinition of the material term Bundled Revenue, the Judges also sought legal analysis from the parties relating to the D.C. Circuit's directive that the Judges either provide “a fuller explanation of the agency's reasoning at the time . . .” or take “new agency action accompanied by the appropriate procedures.” See Johnson, 969 F.3d at 392 (citing Department of Homeland Security v. Regents of the Univ. of Cal., 140 S. Ct. 1891, 1908). On February 9, 2022, the Judges invited additional briefing on the Bundled Revenue definition issue, specifically permitting the parties to offer additional analysis of possible characterization of the Copyright Owners' motion for clarification following the Determination as a motion for rehearing under the Copyright Act, title 17, United States Code at sec. 803(c)(2). See Sua Sponte Order Regarding Additional Briefing (Feb. 9, 2022).

At the request of the parties, the Judges agreed to forego live testimony. On March 8, 2022, all parties were afforded an opportunity to present oral argument on all remand issues. On July 1, 2022, the Judges issued an Initial Ruling and Order after Remand (Initial Ruling) —applying Johnson and considering the entire record developed pre-remand and post-remand.

In the Initial Ruling, the Judges directed the parties to attempt to submit jointly agreed-upon regulatory provisions implementing the Initial Ruling for the Judges to consider. The Judges further ruled that, if the parties could not agree on all the regulatory language, they should make separate submissions regarding regulatory provisions in dispute. See Initial Ruling at 114.

The parties agreed to many regulatory provisions but disagreed as to several such provisions. Accordingly, they filed separate submissions and respective replies regarding the regulatory provisions. Services' Joint Submission of Regulatory Provisions (July 18, 2022); Copyright Owners' Submission of Regulatory Provisions to Implement the Initial Ruling (July 18, 2022); Services' Joint Response to Copyright Owners' Submission of Regulatory Provisions (Aug. 5, 2022); Copyright Owners' Response to Judges' July 27, 2022 Order Soliciting Responses Regarding Regulatory Provisions (Aug. 5, 2022).

The Judges considered those submissions and entered an order addressing the disputed regulatory provisions. See Corrected Order regarding Regulatory Provisions Following Initial Ruling and Order (after Remand) (Nov. 10, 2022) (November 10th Order).

On November 30, 2022, the parties filed a Joint Submission in which they provided joint regulatory language no longer in dispute that applied the binding rulings of the Judges and the D.C. Circuit. However, the parties identified the single issue in dispute that relates to the “Total Content Cost” (“TCC”) rates for nine offerings made by interactive streaming services. Joint Submission . . . Regarding Regulatory Provisions Following Initial Ruling and Order (after Remand) (Nov. 30, 2022) (Joint Submission) (eCRB no. 27337).

Having considered the parties' submissions (including the Joint Submission), the Initial Ruling, and all other pertinent material, the Judges adopted the several TCC rates set forth in the Phonorecords II-based benchmark as proposed by the Services. See Order 43 on Phonorecords III Regulatory Provisions (eCRB no. 28210).

Based on the entirety of the record, the Judges adopt in toto the Initial Ruling and the Order 43 on Phonorecords III Regulatory Provisions which are set out in this document. Accordingly, those two documents are adopted by reference in this Final Determination After Remand. Additionally, the regulatory terms that will codify this Final Determination After Remand are set out in this document.

On the basis of the foregoing, the Judges propound the rates and terms described in this Final Determination After Remand for the period January 1, 2018, through December 31, 2022. No participant having filed a timely petition for rehearing, the Judges have made no substantive alterations to the body of the Initial Determination After Remand. The Register of Copyrights reviewed the Judges' Final Determination After Remand for legal error in resolving a material issue of substantive law under title 17, United States Code, and has closed her review. Non-substantive typos have been corrected and non-substantive formatting changes have been made to the version reviewed by the Register in order to accommodate the Federal Register 's formatting standards. The Librarian shall cause the Judges' Final Determination After Remand, and any correction thereto by the Register, to be published in the Federal Register no later than the conclusion of the Register's 60-day review period.

Related Rulings and Orders

A. Initial Ruling and Order After Remand (Redacted Version With Federal Register Naming and Formatting Conventions)

On October 26, 2020, the United States Court of Appeals for the D.C. Circuit (D.C. Circuit) issued its mandate vacating and remanding in part the Determination issued by the Copyright Royalty Judges (Judges) in the captioned proceeding. See Johnson v. Copyright Royalty Board, 969 F.3d 363 (D.C. Cir. 2020). In its ruling on appeal, the D.C. Circuit found that in the Determination, the Judges (1) failed to give adequate notice to participants of their overhaul of the royalty rate structure combined with significantly increased and uncapped rates for section 115 licenses; (2) failed to explain why they rejected a benchmark based on a past settlement agreement in lieu of overhauling of the rate structure and significantly increasing rates; and (3) failed to identify their legal authority to redefine a material term after they promulgated a definition of that term in the Initial Determination circulated to the participants. See Johnson, 969 F.3d at 367, 381; Initial Determination, Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III), 16–CRB–0003–PR (2018–2022) (Jan. 27, 2018).

After receipt of the D.C. Circuit's ruling and mandate, the Judges consulted with the parties to the appeal and established procedures for the remand proceeding. See Order Adopting Schedule for . . . Remand (Dec. 23, 2020). Each side offered opening submissions, responsive submissions, additional evidentiary filings and further supplemental briefing requested by the Judges. The parties' submissions included legal briefing and incorporated evidence from the original proceeding as well as evidence newly developed for the remand proceeding. After preliminary deliberations, the Judges asked for supplemental briefing from the parties responsive to a proposed alternative rate structure. See Notice and Sua Sponte Order Directing the Parties to Provide Additional Materials (Dec. 9 Order). The Judges also sought legal analysis from the parties relating to the D.C. Circuit's directive that the Judges either provide “a fuller explanation of the agency's reasoning at the time . . .” or take “new agency action accompanied by the appropriate procedures.” See Johnson, 969 F.3d at 392 (citing Dep't of Homeland Sec. v. Regents of the Univ. of Cal., 140 S. Ct. 1891, 1908 ( Regents)). On February 9, the Judges invited additional briefing on the service bundle definition issue, specifically permitting the parties to offer additional analysis of possible characterization of the Copyright Owners' motion for clarification following the Determination as a motion for rehearing under the Copyright Act, title 17, United States Code (Act) at sec. 803(c)(2).

At the request of the parties, the Judges agreed to forego live testimony. On March 8, 2022, all parties were afforded an opportunity to present oral argument on all remand issues. Following oral argument, the Judges deliberated and now issue this Initial Ruling after Remand.

After due consideration of all of the evidence and oral argument of counsel, the Judges determine:

(1) With regard to the applicable rates and rate structure, the percent-of-revenue all-in headline royalty rate for the mechanical license shall be set at 15.1%, phased-in, as set forth below:

In all other respects, the rates and rate structure of the Phonorecords II -based benchmark proposed by the Services (as that benchmark is defined herein) shall constitute the rates and rate structure for the Phonorecords III period.

To be clear: the 15.1% headline percentage rate substitutes for the headline percentage rates in subparts B and C of the Services Phonorecords II -based benchmark, and the definition of “Service Revenue” for bundles shall be the definition contained in 37 CFR 385.11 (paragraph (5) for the “Service Revenue” definition) as proposed in the Services' Phonorecords II -based benchmark.

(2) The Services' Phonorecords II -based benchmark is the better of the benchmarks proposed by the parties and satisfies the requirements of 17 U.S.C. 801(b)(1) in all respects. However, as noted supra, to be consistent with this statutory section and the decision in Johnson, the royalty rate of 10.5% in that benchmark shall be replaced with the 15.1% rate set forth in paragraph (1) above.

(3) To reiterate for clarity, consistent with the adoption of the Phonorecords II -based benchmark, and for the reasons more fully developed herein, the Judges adopt the definition of “Service Revenue for Bundled Services” as it appeared in the Initial Determination in the underlying proceeding. Following are the Judges' analysis and ruling after remand.

I. Preliminary Issue: Burden of Proof

As a preliminary matter, the Judges address the issue of burden of proof raised by both parties. Pursuant to the Administrative Procedure Act (APA), “the proponent of a rule or order has the burden of proof.” 5 U.S.C. 556(d). See also Initial Remand Submission of Copyright Owners at 48 (Apr. 1, 2021) (“CO Initial Submission”) (citing section 556(d) of the APA as setting forth “a basic rule of these rate-setting proceedings that a participant is required to provide evidence establishing the propriety of all aspects of its own proposed rates and terms, including all aspects of the participant's proposed rate structure.”). Accordingly, it is clear to the Judges that the Services should continue to bear the burden of proof regarding the sufficiency of their proffered Phonorecords II-based benchmark in this remand proceeding. And, in like fashion, because on remand Copyright Owners have assumed the mantle of pursuing the vacated rate structure and rates, they bear the burden of proof with regard to their proposal.

However, Copyright Owners assert that it is the Services who bear the burden of proof as to Copyright Owners' proposal regarding the appropriateness, vel non, of an uncapped TCC rate prong. According to Copyright Owners, this burden falls on the Services because “only the Services . . . proposed TCC prongs at the hearing,” in the form of the mix of capped and uncapped TCC prongs contained in the Services' Phonorecords II benchmark. Id. at 47. The Judges find that the fact that the Phonorecords II-based benchmark advanced by the Services contains this mix of capped and uncapped TCC prongs does not bear on Copyright Owners' duty, under 5 U.S.C. 556(d), to satisfy the burden of proof with regard to the rates and rate structure they are advancing on this remand. Moreover, the D.C. Circuit has already held that the fact that some of the Streaming Services' proposals contemplated continued use of an uncapped total content cost prong for some categories “does not mean they anticipated that the [Judges] would uncap the total content cost prong across the board. . . [which] is quite different.” Johnson, 369 F.3d at 382. The difference, according to Johnson, is that “[u]ncapping the total content cost prong across all categories leaves the Streaming Services exposed to potentially large hikes in the mechanical license royalties they must pay.” Id.

Accordingly, the Judges find that Copyright Owners indeed do bear the burden of proof with regard to the appropriateness of uncapped rate structure and rates they are proposing on remand and the Services bear the burden of proof with regard to the appropriateness of the Phonorecords II-based benchmark they are continuing to advance on remand.

II. Rate Structure and Rates

A. Relevant Rulings in Johnson

In establishing a royalty rate structure and the rates within it in the context of this remand proceeding, the Judges are guided by the rulings in Johnson.

1. Percent of Revenue Prong

The D.C. Circuit noted that the Judges found the royalties in the Phonorecords II period were too low and that record companies were receiving a disproportionate share of the sum of the mechanical and sound recording royalties. Johnson, 969 F.3d at 384–85. The D.C. Circuit acknowledged that “[t]he Judges . . . then carefully analyzed the competing testimony and drew from it rates that were grounded in the record and supported by reasoned analysis.” Id. at 385. The D.C. Circuit found that the Judges acted well within their discretion and not arbitrarily, relying on substantial evidence in establishing the “zone of reasonableness” for the rates. Id. As the D.C. Circuit noted, the Judges' process was “the type of line-drawing and reasoned weighing of the evidence [that] falls squarely within the [Judges'] wheelhouse as an expert administrative agency.” Id. at 385–86 (emphasis added).

2. Uncapped TCC Prong

The D.C. Circuit found fault, however, in the Judges' determination to establish an uncapped and increased percentage-based total content cost (TCC). Id. at 380. This approach “removed the only structural limitation on how high the [TCC] . . . can climb.” Id. The D.C. Circuit reasoned that uncapping the TCC alternative rate prong across all categories of service exposed the Services to potentially large hikes in the overall mechanical royalties they must pay. Id. at 382. The D.C. Circuit noted: “As the [Judges] acknowledge, sound recording rightsholders have considerable market power vis-à-vis interactive streaming service providers . . . . The interactive streaming services are . . . exposed to the labels' market power and record companies could, if they so chose, put those services out of business entirely . . . . [B]y virtue of their oligopoly power, the sound recording copyright holders have extracted `inflated' royalties. . . .” Id. (cleaned up).

While the Services had advocated uncapping the TCC alternative rate prong for some categories of service, that “does not mean they anticipated that the [Judges] would uncap the total content cost prong across the board. That is quite different.” Id. at 382. The D.C. Circuit found that the Judges “failed to provide adequate notice of the drastically modified rate structure [they] ultimately adopted.” Id. at 381. The D.C. Circuit emphasized that the failure to provide adequate notice of their intentions “is no mere formality [because] [i]nterested parties' ability to provide evidence and argument . . . not only protects the parties' interests, it also helps ensure that the [Judges'] ultimate decision is well-reasoned and grounded in substantial evidence.” Id. at 381–82.

To support their adoption of an uncapped TCC rate prong, the Judges “predicted that the sound recording copyright owners' royalty rates would naturally decline in the course of their negotiations with interactive streaming services.” Id. at 372. The Judges found persuasive the rebuttal testimony of one of Copyright Owners' economic expert witnesses, Professor Watt, that an increase in mechanical royalties payable by the Services would lead to a corresponding decrease in the Services' sound recording royalty obligations. See Determination at 73–74 (“[S]ound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works [and] Professor Watt's bargaining model predicts that the total of musical works and sound recordings royalties would stay “almost the same” in response to an increase in the statutory royalty.”). The Services painstakingly criticized this “see-saw” theory.

The D.C. Circuit concluded that, on remand, if and when the Judges consider the “uncapped” rate structure, they shall address all substantive challenges to that approach raised by the Services, including the issue of whether “an increase in mechanical license royalties would lead to a decrease in sound recording royalties.” Id. at 383.

Thus, the D.C. Circuit held, the Judges erred procedurally in adopting an uncapped TCC alternative rate prong. The D. C. Circuit therefore instructed the Judges to provide the parties with the opportunity to fully address the issues regarding the uncapped TCC prong, and for the Judges to address the “substantive challenges” raised by the Services.

3. Four Itemized Statutory Objectives

The statutory standard found in section 801(b)(1) instructs the Judges to set rates that are not only “reasonable,” but also reflective of four itemized objectives, or factors, which, as the D.C. Circuit stated, set forth “competing priorities.” 17 U.S.C. 801(b)(1)(A)–(D); Johnson, 969 F.3d at 387. With regard to these four priorities, the D.C. Circuit found that the Judges properly analyzed and applied the first objective (Factor A). Id. at 387–88. In particular, the D.C. Circuit did not disturb the Judges' ruling that an increase in the royalty rates for mechanical licenses was necessary in order to satisfy Factor A. Johnson, 369 F.3d at 387–88. According to Johnson, in making this finding, the Judges had engaged in a “reasonable reading of the record” and had relied on “substantial evidence.” Id. at 388. Thus, Factor A (when considered without regard to the other three objectives) indicated that the statutory rate needed to be higher than it was during the Phonorecords II period.

With regard to the other three objectives, Johnson stated that “[t]he question whether the [Judges] adequately addressed factors B through D . . . is intertwined with the nature of the rate structure ultimately imposed by the [Judges].” Id. at 389. Accordingly, the D.C. Circuit concluded that it “need not . . . address whether the [Judges] adequately considered these remaining factors.” Id.

Within the parameters of the holdings in Johnson, the Judges consider the record facts and the arguments made in this remand proceeding, together with the pertinent facts and arguments made in the original proceeding.

B. Rate Evidence for the 33-Months From January 2018 Through September 2020

After the Determination was issued, from its effective inception on January 1, 2018, through September 30, 2020—a 33-month period—the parties operated under the rates and rate structure set forth in that ruling. In light of the D.C. Circuit's decision in Johnson, as of October 1, 2020, the parties reverted to the Phonorecords II rates. The Services have asserted in this remand proceeding that, during the 33-month period when the Majority's new and higher Phonorecords III rates were in effect, [REDACTED]. By contrast, Copyright Owners, on remand, looking at the same data over this 33-month period, aver that they prove the existence of the seesaw theory.

1. Services' Position

According to the Services, [REDACTED]. [REDACTED]. Moreover, according to the Services, [REDACTED]. The Services further maintain that, [REDACTED].

The Services make the [REDACTED]. And, [REDACTED]. Id. ¶¶ 5, 9–13, 16–19, 22–23, 26–27.

The Services claim that [REDACTED]. More particularly, [REDACTED].

[REDACTED]. [REDACTED].

The Services' economic experts rushed to judgment upon learning of these facts, claiming that they disproved the seesaw theory. See Katz WDRT ¶¶ 25–27 (relying on testimonies cited supra and concluding that seesaw theory was disproved, based on [REDACTED]); Marx WDRT ¶¶ 48–51 (relying on same testimonies and likewise finding because [REDACTED]); Leonard WDRT ¶ 17. ([REDACTED]).

2. Copyright Owners' Position

Copyright Owners analyzed the royalty data over the same 33-month period (January 2018 through September 2020) and reach the opposite conclusion. One of their economic expert witnesses, Dr. Jeffrey Eisenach, testified that [REDACTED]. Moreover, he opined that [REDACTED]. See Eisenach RWRT sec. 2(A) & appx. C.

Based on this analysis, Professor Watt declares empirical vindication of his seesaw theory. Watt RWRT ¶¶ 41–42, 46 (“The [Judges'] bargaining theory insights about the relationship between royalty rates were correct. . . . [REDACTED]. . . .”).

3. Analysis and Decision Regarding Evidence of Post-Determination Rates

The Judges are perplexed by the willingness of the expert economic witnesses on both sides to opine that the rate changes from January 2018 through September 2020 can serve as confirmation of their clients' respective positions. The issue to be considered empirically was whether the sound recording rate would decrease in response to the increase in the mechanical rate. That is, if the record labels had previously set royalties at a level that would allow the Services merely to survive, would the record labels agree to lower their sound recording rate if more of the Services' surplus were acquired by Copyright Owners? To answer this question, the economists on both sides applied sophisticated bargaining models and critiques to explain the nature of the negotiations that would ensue.

In the process, the economists lost track of an obvious, elementary point: The Phonorecords III rates were being challenged by the Services' appeal, and might not persist. Indeed, the rates were ultimately vacated and the parties returned in October 2020 to the Phonorecords II rates. Now, the rates will be changed again by this post-remand Determination, and going forward may be subject to further potential change, consistent with the provisions of title 17. In light of such ongoing fundamental uncertainty, why would any economist or businessman assume that the sound recording companies would agree to adjust their rates in response to a change in the mechanical rate? The Judges are amazed that the economic experts neglected even to raise this uncertainty as a complicating issue, let alone a dispositive one.

Moreover, no party called as a witness any representatives of the Majors, or subpoenaed their testimony or documents, to provide the Judges with evidence of how these record companies perceived the seesaw issue, whether as a permanent phenomenon or as an uncertain matter, given the pendency of the legal proceedings regarding the ultimate mechanical rate. Any of the parties could have requested that the Judges subpoena a sound recording industry witness to give testimony and produce documents as to this issue, pursuant to 17 U.S.C. 803(b)(6)(C)(ix), but none did so. Further, Copyright Owners, who are representing the music publishing interests of inter alios, Sony, Universal, Warner, and Merlin, likely could have produced such sound recording witnesses without the need for a subpoena. Witnesses from these entities who negotiated with the Services after the Phonorecords III rates and rate structure became effective certainly would have knowledge relevant to the testimony of the Services' witnesses [REDACTED] who claimed that [REDACTED].

Simply put, the period from period from January 2018 through September 2020 was a time the Judges construe as “33-months of uncertainty,” see 3/8/22 Tr. 87, 91 (Closing Argument) when no party could ascertain with any assuredness the ultimate Phonorecords III rates and rate structure. Thus, for the economists and the parties to claim vindication for their arguments by reliance on how the record labels did or did not respond to the challenged and ever-shifting rates during this “33 months of uncertainty” reflects the elevation of adversarial zeal over objective judgment.

Accordingly, the Judges place no weight on the purported changes or stability of the sound recording rates during the Phonorecords III rate period.

C. Percent-of-Revenue Rate Prong

1. Copyright Owners' Position

In their initial remand submission, Copyright Owners provided no new evidence to support any aspect of the 15.1% revenue-based rate (or for that matter, any new evidence to support the rates or rate structure in the Determination), and elected to rely on the pre-remand record. In fact, in their initial remand submission, Copyright Owners do not so much as mention the 15.1% revenue rate derived by the Judges. However, in their reply remand submission (which the Judges found also to constitute, in part, a substantive initial submission ) Copyright Owners do address the 15.1% revenue rate. In the reply submission, Copyright Owners simply stated: “[T]he Circuit affirmed the Board's derivation of rate percentages, including raising the revenue rate to 15.1%.” Copyright Owners' Reply Brief on Remand (in Reply Remand Submission of Copyright Owners, Vol. 1) at 64, n.48 (July 2, 2021) (“CO Reply”). In a subsequent submission, Copyright Owners added that “ [t]he narrow mandate on this Remand does not allow for reopening the rate percentage determination in the [ ]Determination. ” Copyright Owners' Motion for Reconsideration or Clarification at 15 & n.10 (Dec. 17, 2021) (emphasis added) (Dec. 17th Motion).

Thereafter, Copyright Owners asserted that the D.C. Circuit's affirmance of the [Judges'] revenue percentage rate calculation was “strong[ ]” and “detailed.” Copyright Owners' Reply in Further Support of Motion for Reconsideration or Clarification at 4 (January 5, 2022). Moreover, Copyright Owners took note that the Services had relied on substantively identical language in Johnson to support their argument that other statements in that D.C. Circuit decision should be deemed affirmed. See id. at 4–5 (noting Services' reliance on Johnson's description of the Judges' rulings regarding student and family discounts (“ grounded in substantial record evidence . . . based on the weight and credibility of the evidence [and] squarely within the Judges' expertise) ” as demonstrating that the D.C. Circuit had affirmed those rulings) (emphasis added); see also Copyright Owners' Brief in Response to the Additional Materials Orders at 2, 6–7 (Jan. 24, 2022) (“CO Additional Submission”) (again asserting that “the 15.1% revenue rate . . . was specifically affirmed in detail by Johnson. ”).

2. Services' Position

In their initial submission after the remand, the Services objected to any continued application by the Judges of the 15.1% revenue rate because, “as the Majority acknowledged, this particular division of revenues will never happen in the real world because of the complementary oligopoly power of the record labels.” Services' Joint Opening Brief (in Services' Joint Written Direct Remand Submission at Tab D) at 52 (“Services' Initial Submission”) (Apr. 1, 2021). More particularly in this regard, the Services note that Professor Marx's Shapley Value Model, which served as an input for the generation of the 15.1% revenue rate, also indicated that only [REDACTED]% of the interactive streaming revenue should be paid out as royalties to the sound recording rightsholders, with the remaining [REDACTED]% of these revenues retained by the interactive streaming services. Id. (“Both Professor Marx's and Professor Watt's models show lower combined royalties being paid by the services than are currently paid in the marketplace. . . The discrepancy in total royalties between the models and the real world is explained, in part, by the absence of supranormal complementary oligopoly profits in the Shapley model, and the presence of those profits in the actual market.”). Id. (quoting Phonorecords III,84 FR 1952).

By this approach, the Services maintain, “the Majority awarded the Copyright Owners the full 15.1% of revenue dictated by its model (phased in over time), and left it up to the Services to convince the complementary oligopolist major labels to dramatically lower sound recording rates.” Id. at 54–55. The Services argue that, instead, the Majority should have applied to Professor Marx's [REDACTED]% total royalty obligation what they characterize as “any of the[ ] real-world ratios in place of the [REDACTED] ratio taken from “Professor Gans' “Shapley-inspired” model. Id. at 54. According to the Services, these lower ratios would have reduced the revenue percentage rate well below 15.1%. Id.

Alternatively, the Services propose, through Professor Marx's post-remand written testimony, that the Judges now adopt “a more balanced, burden-sharing approach” to address what she described as the Majority's “imbalance” problem. Id. at 57; see also Marx WDRT ¶¶ 52–63. Essentially, her proposal begins with an assumption, based on record evidence, that labels typically take specific shares of service revenue, including shares of [REDACTED]%, [REDACTED]% and [REDACTED]%. These shares are significantly higher than the [REDACTED]% that Professor Marx generated from her Shapley model. Next, Professor Marx's post-remand burden-sharing approach uses as inputs the 15.1% of service revenue and the [REDACTED]% of service revenue that would be retained by the musical works owners and the Services respectively. Putting these two factors together, she sets forth the basic math: Using her [REDACTED]% sound recording share as an example, she notes that there is not enough revenue for the labels to take this [REDACTED]% share, if the musical works owners also receive 15.1% and the Services also retain the [REDACTED]% derived from her model ([REDACTED]% + 15.1% + [REDACTED]% = [REDACTED]%, an irrational result). See Services' Joint Opening Brief at 57.

Professor Marx engages in an analysis based on the following math and logic (again, using the [REDACTED]% sound recording rate as an example of the fixed amount taken by the labels): (1) [REDACTED]% of the streaming revenues remain available to be split between the services and the musical works copyright owners; (2) adding the 15.1% revenue rate and her [REDACTED]% revenue retention percentage equals [REDACTED]%; and (3) the 15.1% revenue rate, as a percent of this [REDACTED]%, is [REDACTED]%; and (4) [REDACTED]% of the [REDACTED]% available for splitting between the services and the musical works copyright owners is [REDACTED]% (rounded). Id. at fig.8.

Thus, she identifies her version of a “fair” result: The Services and Copyright Owners would split the residual revenue remaining after the labels have exercised their complementary oligopoly power to take an outsized fixed share—with the split proportional to the 15.1%-to-[REDACTED]% revenue amounts calculated respectively by the Judges (the 15.1% musical works rate) and Professor Marx (the [REDACTED]% service revenue retention). Id. 59, table. 8.

In their final post-remand submission, the Services also flatly state: “[T]he D.C. Circuit did not “affirm” the 15.1% rate—it vacated that rate.” Services' Joint Rebuttal Brief Addressing the Judges' Working Proposal at 2 (Feb. 24, 2022) (“Services' Additional Submission”). However, the Services do not support that quoted statement with any citation to Johnson. See id. Further, the Services assert that the 15.1% revenue rate is not immune from post-remand review and reduction because “the D.C. Circuit withheld judgment “on whether that final percentage satisfies factors B through D of Section 801(b)(1). . . .” Id. at 3.

3. Analysis and Decision Regarding 15.1% Revenue Rate Prong

The Judges determine that they are clearly bound by the D.C. Circuit's decision in Johnson to maintain the 15.1% revenue rate, as phased-in by the Determination. Several reasons support this decision.

First, the Judges conclude that the D.C. Circuit's decision in Johnson is conclusive and unambiguous regarding the revenue percentage rate. The D.C. Circuit rejected the Services' assertion that the Judges acted “arbitrarily” as to this particular issue, noting that the Services had misstated the relevant facts. Johnson, 969 F.3d at 385–86 (responding to Services' misdescription of Judges' analysis and explaining what Services described as “not what happened.”). Moreover, the D.C. Circuit held that with regard to the construction of the 15.1% revenue rate, the Judges had “engaged in the type of line-drawing and reasoned weighing of the evidence [which] falls squarely within the [Judges'] wheelhouse as an expert administrative agency.” Id. at 386. The D.C. Circuit further noted that the Judges “proceed[ed] cautiously” to set the 15.1% revenue rate by establishing a “zone of reasonableness” for the revenue rate. Id. at 385. Indeed, with regard to each aspect of this revenue rate analysis, the D.C. Circuit found that the Judges' decision making was “grounded in the record and supported by reasoned analysis” and that “[s]ubstantial evidence supports [their] judgment.” Id. at 385.

Second, when the D.C. Circuit reviewed the Determination, it applied “the same standards set forth in the Administrative Procedure Act, 5 U.S.C. 706.” Id. at 375 (noting that 17 U.S.C. 803(d)(3) cross-references 5 U.S.C. 706); see also id. (“[W]e will set aside the [ ] Determination `only if it is arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law, or if the facts relied upon by the agency have no basis in the record.”).

Here, the D.C. Circuit explicitly found that the Judges' analysis and findings in connection with the 15.1% revenue rate are not arbitrary and capricious, and that the facts relied upon by the Judges have a sufficient basis in (are “grounded in”) the record. It seems beyond dispute that the D.C. Circuit affirmed the Judges in their setting of the 15.1% revenue rate as a rate that is reasonable, and thus satisfies that aspect of the section 801(b)(1) standard. Indeed, it would border on the Orwellian to misconstrue the D.C. Circuit's unequivocal and obvious affirmance of the reasonableness of the 15.1% revenue rate as a vacating of that finding.

Third, the Judges note that Johnson conspicuously declines to identify the Judges' setting of the 15.1% percent-of-revenue rate as one of the findings to be revisited on remand. Rather, Johnson states that the three overarching issues for resolution on remanded are the Majority's failure: (1) “to provide adequate notice of the rate structure it adopted,” (2) “to explain its rejection of a past settlement agreement as a benchmark for rates going forward; and (3) “[to] identif[y] the source of its asserted authority to substantively redefine a material term after publishing its Initial Determination.” Johnson, 369 F.3d at 367. The Majority's finding that the 15.1% royalty rate is “reasonable” was not identified by the D.C. Circuit as a finding that was vacated and subject to further review and, indeed, as noted supra, the appellate panel credited what it characterized as the Majority's careful analysis and line-drawing in arriving at that finding.

The clarity of the D.C. Circuit's affirmance of the royalty rate of 15.1% for the percent-of-revenue prong moots the issue of whether Professor Marx's attempt, described supra, to correct the so-called “imbalance” problem has merit. However, the Judges note that, even if this issue had not been conclusively decided in Johnson, they would reject her approach as futile. That is, Professor Marx fails to acknowledge that any surplus that her approach would appear to provide to the Services would be siphoned off by the Majors, given their complementary oligopoly power.

More particularly, the sound recording royalty rates she posits ([REDACTED]%, [REDACTED]% and [REDACTED]%) are all functions of the sound recording companies' understanding of the Services' non-content costs (costs that the Services must recover out of retained revenues in order to remain in operation, i.e., to “survive”) and the then-existing musical works content (royalty) costs (comprised of the mechanical rate and the performance rate). If, as Professor Marx contemplates, the mechanical rate is reduced so that Copyright Owners “share the burden” of the complementary oligopoly effect on sound recording rates, that “burden sharing” would increase the revenues retained by the Services (that is the purpose of Professor Marx's approach!). But such an increase would raise the Services' revenue above their “survival” rate, as understood by the record labels. Thus, the record labels, given their complementary oligopoly power, would increase the Services' royalty rate above what it otherwise would have been.

Alternately stated, when Professor Marx hypothesizes a given sound recording royalty rate in column 1 of Figure 8 in her WDRT, that rate is assumed, by the logic of the complementary oligopoly theory, to have already allowed the services to cover only their non-content costs and musical works royalties, as understood by the record labels. So, her assumed rate in column 1 is not a fixed parameter, but rather an independent variable, which is a function of, inter alia, the costs incurred by the services, i.e., their non-content costs plus their musical works royalty costs. If those service costs decreased (for example, in an attempt to reduce the services' burden of bearing the full brunt of the labels' complementary oligopoly power as in Professor Marx's attempt to correct the imbalance problem), the percentage in column 1 of Figure 8 would increase, as the labels siphoned off that surplus over the services' survival revenue requirements. To find otherwise would be to refute the logic of the dynamics of the complementary oligopoly effect.

Moreover, the defect in Professor Marx's attempt to remedy the so-called “imbalance” problem is a consequence of the statutory licensing and royalty scheme. To recap, the licensing of content used by the interactive services is bifurcated. The sound recording royalties paid by the interactive services to the record labels are not regulated, and complementary oligopoly power exists in that market, inflating sound recording royalty rates above an effectively competitive level. See Determination at 73 (“[T]he existence of complementary oligopoly conditions in the market for sound recordings” is the basis for “the record companies' ability to obtain most of the available surplus” generated by interactive streaming.) However (and to state the obvious), the mechanical rate paid by the interactive services for musical works is regulated, pursuant to 17 U.S.C. 115 and, until the 2018 enactment of the Music Modernization Act, according to the rate standards in 17 U.S.C. 801(b)(1). Thus, there is no statutory or regulatory impediment to prevent record labels from responding to a decrease in the mechanical rate by increasing the unregulated sound recording rate if such an increase is in their economic interest.

Accordingly, any attempt by the Judges to reduce the mechanical royalty rate in order to allow the Services to retain more of the surplus would fail; it would be like pouring water into a bucket with a siphon at its base. More water would not remain in the bucket, but rather would accumulate wherever the siphon leads—in this case, to the record labels. The Judges could keep mechanical royalty rates depressed and allow this to occur, but that would harm Copyright Owners while providing no relief to the Services. And despite the old adage that “misery loves company,” the Judges detect no directive under section 801(b)(1) that they harm Copyright Owners without providing a gain for the interactive streaming services—and that they provide a windfall for the record labels, to boot.

Although Professor Marx's attempt to reduce the Services' “misery” by sharing it with Copyright Owners is unavailing, the statutory scheme and market forces do appear to combine to mitigate the burden created by the complementary oligopoly power of the sound recording companies. If interactive streaming revenue were to grow over the rate period, then the phase-in to the 15.1% rate will reflect fixed annual percentages of a larger base, allowing services to retain a higher dollar level of the interactive streaming revenues. [REDACTED]. See, e.g., Diab WDRT ¶¶ 10–11 (Google agreements); Mirchandani WDRT ¶¶ 16–17 (Amazon agreements); Bonavia WDRT ¶¶ 8; 14–19 (Spotify agreements); White WDRT ¶¶ 6; 8–14; 19; 24; 27–28 (Pandora agreements). Additionally, the Services' headline sound recording rates [REDACTED]. Services' Joint Remand Reply Brief at 40 (and record citations therein). Thus, assuming no increase in non-content costs (or increases smaller than the increases in streaming revenue), the Services will realize increased revenue above and beyond what they needed to survive.

The Services and Copyright Owners recognize the mitigation of harm to the Services generated by these facts (although they may well disagree with the Judges' application of these facts). During colloquy with counsel for Pandora and Spotify during closing arguments on remand, the Judges asked why they should in essence apply the “misery loves company” adage:

[JUDGE STRICKLER] [T]he problem is . . . the sound recording [rates] are unregulated in the interactive market . . . . Congress did not want that to be controlled at all. So every time I see . . . the services' argument about how we have [to] set a rate that's fair even though there's this ability of the sound recording [companies] to take more, my margin note is always this: “Are they arguing that ‘misery loves company?’ ” [W]hy shouldn't that misery be shared with Copyright Owners? . . . Isn't that really Professor Marx's argument in her proposed split . . . using the 15.1 percent figure . . . ?

[COUNSEL] [Regarding] Judge Strickler['s] . . . “misery loves company” issue. . . . I think . . . the way [Judge Strickler] put it during the trial was, even if I thought rates needed to come down, how would that help you; wouldn't the labels just take all that surplus for themselves based on their complementary oligopoly power? . . . . I want[ ] to address it right off the bat . . . . in open session.

Relat[ed] to . . . the seesaw . . . our point is that these label rates are sticky in both directions. If you see an increase in musical works rates, you do not see a quick decrease in label rates, and the opposite is true. These rates are sticky.

. . .

There's a lot of friction with respect to the ability of label rates to change quickly in response to the dynamic marketplace or the dynamic for business reasons or because of regulatory changes in musical works rate. These are multi-year contracts. They take a long time to negotiate. They are complex, et cetera.

So, I do think it's right that at a minimum you can buy time where the ratio is more aligned with the 801(b) factors. In other words, you don't have to worry that the labels will take it all right away, even if you believe they will ultimately take that.

[JUDGE STRICKLER] So you are saying we have something that reduces misery for a period of time until the misery returns?

[COUNSEL] That's right. And I think that would have been true in 2018 when you were sitting drafting the decision. It's even more true today in 2022 when the label rates, as I mentioned, are effectively set, bought and paid for.

3/8/22 Tr. 29–30, 43–46 (Closing Argument) (emphasis added).

Similarly, on this topic, Copyright Owners' counsel accurately characterized the Judges' adoption of the static 15.1% Shapley-based rate as the inevitable consequence of “regulatory lag,” that requires a regulator to keep a rate constant over the statutory term because there is no sufficient data to project future rates. Id. at 273–75; see generally A. Kahn, 2 The Economics of Regulation at 48 (1971) “The regulatory lag [is] the inevitable delay that regulation imposes in the downward . . . [and] upward adjustments” to rate levels, and “thus is to be regarded not as a deplorable imperfection of regulation but as a positive advantage [because] companies can for a time keep the higher profits they reap from a superior performance. . . .”).

4. Consideration of Factors A–D in Section 801(b)(1)

Finally, the Judges consider the impact of Factors A–D of section 801(b)(1) in connection with the setting of the revenue percentage rate of 15.1%. Regarding Factor A, it cannot be gainsaid that the D.C. Circuit has left this issue unresolved. Rather, Johnson unambiguously affirmed the Majority's finding that an increase in the mechanical royalty rate was warranted. Specifically, Johnson states that the Majority's decision in this regard met the “test” that it be “supported by substantial evidence [and] reflect a reasonable reading of the record.” Johnson, supra, at 388. Moreover, with regard to the level of the increase, the D.C. Circuit did not disturb the finding by the Majority that “[t]he rates determined by the Judges represent a 44% increase over the current headline rate, and thus satisfies the Factor A objective. . . .” Determination at 85.

With regard to Factors B and C, even if Johnson were construed as permitting the Judges to revisit this issue, they would not adjust the 15.1% revenue rate on the basis of these two factors. In this regard, the Judges note that the Majority found that the 15.1% revenue rate was not only “reasonable,” but also a “ fair allocation of revenue between copyright owners and services.” Determination at 87 (emphasis added). The Majority thus found explicitly that “with regard to Factors B and C . . . there is no basis to depart from [its] determination of the reasonable . . . rate structure and rates as set forth supra.” Id. More particularly, the Majority calculated the 15.1% rate by utilizing the total royalty percentage revenue of only [REDACTED]% as calculated by Spotify's economic expert witness, Professor Marx, whose economic modeling intentionally reflected a conception of fairness by reducing the effect of the labels' complementary oligopoly market power. See Determination at 67–68 (noting that Professor Marx testified that this aspect of her model “represents a fair outcome in the absence of market power [and] . . . eliminates . . . market power” which . . . if left in the economic analysis would “render[ ] . . . the analysis incompatible with the objectives of Factors B and C of section 801(b)(1).)”) (emphasis added).

Accordingly, the Judges find it would be substantively unwarranted to engage in any new consideration on remand of the impact, if any, of Factors B and C on the otherwise reasonable 15.1% revenue rate.

The final itemized statutory factor—Factor (D)—instructs the Judges to consider the “competing priority” of “minimiz[ing] any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.” 17 U.S.C. 801(b)(1)(D). As with Factors B and C, even if Johnson were construed to allow the Judges to revisit this issue on remand with respect to the 15.1% revenue rate, the Judges would not change the Majority analysis or findings. In the Determination, the Judges adopted the following interpretation of this standard set forth in previous determinations:

[T]he Judges reiterated their understanding of Factor D, concluding that a rate would need adjustment under Factor D if that rate directly produces an adverse impact that is substantial, immediate and irreversible in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance 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.

Determination at 86 (emphasis added).

Also, in order to minimize any economic disturbance to the Services' businesses, the Majority decided to phase-in the 15.1% rate over the five-year rate term, setting annual percent-of-revenue rates as follows: 11.4% in 2018; 12.3% in 2019; 13.3% in 2020; and 14.2% in 2021, before the full 15.1% rate became effective in 2022 the final year of the rate term. Id. at 87–88.

On remand, the Services have not made any argument that the rate structure or rates set by the Majority were “disruptive under this standard.” In sum, there is insufficient basis for the Judges to change the Majority's application of Factor (D) to the 15.1% revenue rate finding by the Majority.

5. Conclusion Regarding the 15.1% Revenue Rate

For the forging reasons, the Judges do not disturb the Majority's finding that the percent-of-revenue rate at 15.1%, phased-in annually over the rate period, constitutes a “reasonable” rate under section 801(b)(1) to be used as the statutory rate for the 2018 to 2022 period.

D. Uncapped TCC Rate Prong

1. Two Post-Remand Rationales for Uncapped TCC Rate Prong

The Determination set forth the following two primary reasons for adopting a “greater-of' rate structure that also included an uncapped TCC rate prong:

First, the use of an uncapped TCC metric is the most direct means of implementing a key finding . . . by the experts for participants on both sides in this proceeding: the ratio of sound recording royalties to musical works royalties should be lower than it is under the current rate structure. Incorporating an uncapped TCC metric into the rate structure permits the Judges to influence that ratio directly.

Second, an uncapped TCC rate prong effectively imports into the rate structure the protections that record companies have negotiated with services to avoid the diminution of revenue.

Determination at 35–36.

2. Copyright Owners' Position

Copyright Owners claim that the uncapped TCC prong should be adopted. They contend that the D.C. Circuit remand was merely “procedural” rather than substantive, and the Judges thus are not precluded from readopting the uncapped TCC prong in this remand proceeding. CO Initial Submission at 35–38 (and record citations therein).

They further contend that the uncapped TCC prong was adopted to provide protection against revenue deferment and displacement occasioned by the Services choosing to elevate the growth of subscribers and other listeners over revenue maximization. Id. at 38–43 (and record citations therein). The uncapped TCC prong was first proposed by Google to persuade the Judges to reject Copyright Owners' proposed “greater-of” rate structure containing a per-play prong and a per subscriber prong. Id. at 43–46 (and record citations therein).

Copyright Owners argue that the uncapped TCC prong should be adopted because: (1) the Services have not shown any actual or threatened “disruption” or other harm resulting from the uncapped TCC prong during the 33-month period; (2) the Services actually experienced “unprecedented growth and profit” during this period; and (3) the Services paid lower percentages of revenues in mechanical and total royalties when the uncapped TCC prong was in effect. Copyright Owners' Reply Brief on Remand at 34–48 (and record citations therein).

Relatedly, according to Copyright Owners the Services' argument that the “see-saw” effect is unsupported by empirical evidence has collapsed, given the evidence relating to market performance. Further Copyright Owners maintain that this argument is irrelevant to the rate structure issue. Id. at 48–50 (and record citations therein).

3. Services' Position

The Services argue on remand that the uncapped TCC rate prong must be rejected. The Services reject the “seesaw” theory claiming it is disproved by the experience of the parties during the 33-month period. Services' Joint Opening Brief at 48–49; Services' Joint Supplemental Brief at 7–13 (Nov. 15, 2021) (and record citations therein). The Services further contend that Copyright Owners have disavowed the “seesaw” theory as understood by the Majority. The Services allege that Copyright Owners now claim that the theory was nothing more than “a nod” to certain “core principles” of bargaining theory, rather than a specific prediction of a commensurate inverse relationship between increases in the mechanical royalty rate and decreases in the sound recording royalty rate. Services' Joint Supplemental Brief at 2, 5–7 (and record citations therein).

With regard to the uncapped TCC rate prong, the Services assert that Copyright Owners have not even attempted to demonstrate—nor could they demonstrate—that the uncapped TCC rate prong is consistent with all four statutory objectives set forth in section 801(b)(1). Services' Joint Reply Brief at 1, 3–4, 33–34 36 (July 2, 2021) (“Services' Reply”); see also Services' Joint Opening Brief at 44–64 (and record citations therein). The Services claim that “yoking” the mechanical rate to the “complementary oligopoly rates extracted by the labels is plainly unreasonable.” Services' Joint Opening Brief at 44–46. The Services argue that the existence, vel non, of any “disruptive impact” arising from the uncapped TCC rate prong, is misguided and not dispositive, because it is only one of the four separately itemized factors and, as this factor relates to Copyright Owners' proposed uncapped TCC prong, they bear the burden of proof. Services' Reply at 35–37.

Finally, the Services contend that Copyright Owners have failed to explain their self-contradictory pre-remand argument that “an uncapped TCC prong `does nothing to protect Copyright Owners from the Services' revenue displacement and deferment.' ” Services' Reply at 43.

4. Application of Johnson Findings Regarding Uncapped TCC Rate Prong

The Judges conclude that the D.C. Circuit affirmed the Majority's derivation and calculation of the 26.1% TCC rate, but vacated and remanded the Judges' application and inclusion of that rate prong in the rate structure. The D.C. Circuit noted that, on appeal, the Services contended that “it was arbitrary and capricious for the [Judges] to rely on information drawn from different expert analyses in calculating the mechanical royalty rates.” Johnson, 969 F.3d at 384. Thus, the Services were making the same “information”-based argument in opposition to the calculation of both aspects of the mechanical royalty rates—the revenue percentage prong and the TCC prong. See also id. (“the Streaming Services separately leveled objections to the particular percentages adopted by the Copyright Royalty Board to calculate the revenue and total content cost prongs. ”) (emphasis added)

In fact, both rate prongs were indeed derived from the same analyses. See Determination at 75 (table) (showing that both 15.1% revenue rate and 26.2% TCC rate derived from same data—Professor Marx's model showing total royalties as high as [REDACTED]% [Majority's lower bound] and Professor Gans's “Shapley-inspired” model showing TCC percent should be [REDACTED]%.)

It is also clear from Johnson that the D.C. Circuit found that the Majority had reasonably derived and calculated the 26.2% TCC rate:

When it came to . . . the ratio of sound recording to musical work royalties that Gans derived from his analysis the [CRB Judges] specifically found . . . reasonable Gans' equal value assumption [for dividing the Shapley surplus . . . between sound recording and musical works owners] and his reliance on Goldman Sachs' profit projections. That type of line-drawing and reasoned weighing of the evidence falls squarely within the Board's wheelhouse as an expert administrative agency.

See Johnson, 969 F.3d at 385–86 (cleaned up) (emphasis added). Accordingly, because the identical analysis was performed by the Judges to derive the 26.2% TCC rate as was done to derive the 15.1% revenue rate, the Majority's finding with regard to the derivation and calculation of the TCC rate likewise is not subject to further consideration on remand by the Judges.

However, it is equally clear that the D.C. Circuit vacated and remanded the Majority's application and inclusion of the 26.2% TCC rate in a separate “greater-of” TCC prong. The defect that generated the vacating on this issue was procedural— “the Streaming Services had no notice that they needed to defend against and create a record addressing such a significant, and significantly adverse, overhaul of the mechanical license royalty scheme . . .” Id. at 382. The consequence of the D.C. Circuit's action, however, was substantive. The D.C. Circuit stated:

This is no mere formality. Interested parties' ability to provide evidence and argument bearing on the essential components and contours of the [Judges'] ultimate decision not only protects the parties' interests, it also helps ensure that the [Judges'] ultimate decision is well-reasoned and grounded in substantial evidence. . . .

The Streaming Services separately challenge the uncapped rate structure as arbitrary and capricious. In particular, they argue that the rate structure formulated by the [Judges] failed to account for the sound recordings rightsholders' market power. They also object that the [Judges] failed to provide a `satisfactory explanation, or root in substantial evidence, [their] conclusion that an increase in mechanical license royalties would lead to a decrease in sound recording royalties [the “inverse relationship” a/k/a the “seesaw” effect].

Id. at 381–83 (cleaned up) (emphasis added). Thus, the D.C. Circuit explicitly declined to address these substantive issues, because of the deficient procedure. Instead, the D.C. Circuit remanded these substantive issues back to the Judges. Id. Simply put, Johnson found that the absence of notice here could be outcome-determinative. Thus, the Judges categorically reject Copyright Owners' assertion that the remand as to the uncapped TCC rate structure was merely “procedural.” The Judges do not accept the notion that the Majority simply committed some ministerial faux pas that could be summarily corrected so that the uncapped TCC rate structure could be rubber-stamped on remand. Rather, the Judges' error rendered it impossible for them to consider the pros and cons of such a rate structure without the necessary input from the Services (and, for that matter, Copyright Owners as well).

Because the procedural infirmity precluded the D.C. Circuit from deciding whether the Majority's decision was “well-reasoned and grounded in substantial evidence,” there also can be no substantive presumption of the appropriateness of the uncapped TCC rate prong, as suggested by Copyright Owners. To the contrary, the D.C. Circuit's opinion makes it clear that on remand the Judges must engage in a fresh consideration of the statutory appropriateness, vel non, of the uncapped TCC rate prong, by weighing and contextualizing the competing evidence and testimony entered into the record both before and after the remand.

Accordingly, although Copyright Owners correctly assert that Johnson did not find the uncapped TCC rate structure to be “unfair, unreasonable or inequitable,” Johnson just as clearly did not find that structure to be “fair, reasonable or equitable.” Rather, the purpose of the remand was for the Judges to make these determinations. Accordingly, the Judges next examine whether setting the statutory mechanical rate as an uncapped TCC rate is “reasonable,” as required by section 801(b)(1).

5. Determining Whether Uncapped TCC Rate Prong is “Reasonable”

a. Rejection of First Rationale for Including Uncapped TCC Rate

Two substantive issues are implicated raised with regard to the issue of reasonableness: (1) whether the “seesaw” theory is valid; and (2) if it is valid, whether there exist sufficient data to support the phased-in 26.2% uncapped TCC rate. To demonstrate that this uncapped TCC rate prong and the (phased-in) 26.2% rate are reasonable, Copyright Owners rely on the combined application of two economic models—the Shapley Value model and a Nash Bargaining Model. Accordingly, it is necessary to consider how these two models relate to each other and how these models and their interrelationship impact the setting of the statutory rate.

The D.C. Circuit described the Shapley Value Model methodology:

The Shapley methodology is a game theory model that seeks to assign to each market player the average marginal value that the player contributes to the market. This methodology first determines the costs that each player should recover, then divides the “surplus” among the players in proportion to the value of their contributions to the worth of the hypothetical bargain that would be struck.

Johnson, 969 F.3d at 372. The Judges provided a consistent but more detailed definition:

The Shapley value gives each player his average marginal contribution to the players that precede him, where averages are taken with respect to all potential orders of the players. The Shapley value approach models bargaining processes in a free market by considering all the ways each party to a bargain would add value by agreeing to the bargain and then assigns to each party their average contribution to the cooperative bargain. The idea of the Shapley value is that each party should pay according to its average contribution to cost or be paid according to its average contribution to value. It embodies a notion of fairness. The Shapley model is a game theory model that is ultimately designed to model the outcome in a hypothetical `fair' market environment. It is closely aligned to bargaining models, when all bargainers are on an equal footing in the process.

Determination at 62–63 (cleaned up).

To apply a Shapley Value Model in a rate proceeding, the economic modeler must obtain usable cost and revenue data to be inputted into the model. More particularly for this proceeding, the modeler must identify the parties' input costs, including the Services' non-content costs, and the revenue derived from interactive streaming. The difference between these revenues and the Services' noncontent costs represents the Shapley “surplus” that can be shared among the Services, the sound recording companies and Copyright Owners.

(i) The Shapley Approach of the Parties' Economic Expert Witnesses

(a) Professor Gans's “Shapley-Inspired” Model

Professor Gans, Copyright Owners' expert, utilized royalty and profit interactive streaming data for record companies and music publishers that he obtained from “a [then] recent music industry equity analysis report,” namely, a Goldman, Sachs Equity Research report dated October 4, 2016 entitled “Music in the Air, Stairway to Heaven.” Gans WDT ¶ 76 & n.39. As the Majority summarized Professor Gans's approach, “[h]e found that, for the music publishers to recover their costs and achieve profits commensurate with those of the record companies under his approach, the ratio of sound recording royalties to musical works royalties derived from his Shapley-inspired analysis was [REDACTED] (which attributes equal profits to both classes of rights holders and acknowledges the higher costs incurred by record companies compared to music publishers).” Determination at 69 (citing Gans WDT ¶ 77 tbl.3) (emphasis added).

Regarding Professor Gans's Shapley-inspired analysis, the Majority stated:

[T]he Judges find the ratio of sound recording to musical work royalties that Professor Gans derived from his analysis to be informative. Professor Gans computed this ratio based on an assumption of equal Shapley values between musical works and sound recording copyright owners. The Judges find this assumption to be reasonable . . . .

Determination at 70. This is part and parcel of the “line-drawing” undertaken by the Majority that the D.C. Circuit affirmed. Thus, on remand, the Judges do not find cause to reconsider the Majority's limited adoption of Professor Gans's Shapley-inspired analysis.

(b) Professor Marx's Shapley Value Model

Professor Marx constructed two Shapley Value Models, one of which was relied upon by the Majority. In the model credited by the Majority, Professor Marx assumed one collective owner of sound recording copyrights and one collective owner of musical works. She also assumed the presence of a single interactive service. See Determination at 64–68. That approach yielded a total royalty obligation for sound recordings and musical works ranging between [REDACTED]% and [REDACTED]% of the hypothetical service's revenue. Dissent at 133.

Copyright Owners criticized Professor Marx's decision to assume in her model only one interactive streaming service, rather than the multiple services that actually existed. They contend that assumption reduced the market power of the licensors in her model. According to Copyright Owners' economic experts, Professor Marx's approach was a misuse of the Shapley Value Model. They aver that the Shapley Value approach is intended only to eliminate from the rate derivation the bargaining ability of a “Must Have” input supplier (like the sound recording companies and Copyright Owners) to “hold-out” and thus squeeze licensees for higher royalties. By modeling every possible “arrival ordering,” they contend, the “hold-out” problem is avoided. They further contend that Professor Marx misconstrued the purpose of the Shapley approach by wrongly modeling market participants in a manner that significantly reduced the actual market power of these “Must Have” input suppliers. Determination at 66–67.

The Majority agreed with Professor Marx. The two Judges in the Majority found that her modeling reasonably “attempts to eliminate a separate factor—market power—that she asserts renders a market-based Shapley Analysis incompatible with the objectives of Factors B and C of section 801(b)(1).” Id. at 68.

Although the Majority ultimately relied upon Professor Marx's modeling in this regard, the Majority found that her data inputs were problematic. Determination at 65. Specifically, Professor Marx relied on 2015 data from Warner/Chappell and Warner Music Group for music publisher sound recording company noncontent costs, respectively. The Majority found that 2015 data was less probative than 2016 data and understated the percentage of revenue to be paid to the two classes of content providers. However, the Majority ultimately found only that this one-year older data served to “understate” the allocation of surplus to the upstream content providers, and thus rejected only her lower [REDACTED]% bound for total royalties, The Majority did decide to adopt her upper bound of [REDACTED]% value for total royalties, which could (and ultimately did) “constitute a lower bound for total royalties in computing a royalty rate,” applied by the Majority in order to make a downward adjustment to offset the complementary oligopoly effect of “Must Have” inputs. Id. at 73, 75.

(c) Professor Watt's Criticisms of and Adjustments to Professor Marx's Shapley Modeling

Professor Richard Watt was called by Copyright Owners as a rebuttal witness at the hearing, for the purpose of reviewing Professor Marx's WDT. Watt WRT ¶ 3. He concluded that Professor Marx's Shapley Value Model contains important methodological and data flaws which, in his opinion, caused her to significantly understate the mechanical and overall (musical works + sound recording) royalty rates to be paid by interactive services pursuant to a proper Shapley analysis. Id. at ¶ 5.

Professor Watt also criticized her Shapley Value Model for failing to incorporate the fact that “the different interactive streaming companies—Spotify, Apple Music, Rhapsody/Napster, Google Play Music, Amazon, etc.—do all compete (and rather fiercely) among themselves, offering (perhaps perfectly) substitutable services.” Id. at ¶ 25. Even more strongly in this vein, Professor Watt relied on the following description of the substitutability of the streaming services, inter se:

Each [interactive streaming] service in the increasingly crowded field is working frantically to overcome the perception that the main distinction among the uniformly priced $9.99 a month offering is little more than font style, quirky playlist title and color scheme. . . . [M]usic platforms have long fought against the perception that they're . . . selling a nearly interchangeable product . . . You're getting sold the same car [with] just got a different lick of paint on it.”).

Id. at ¶ 32 n.19.

Professor Watt claimed that incorporating this downstream competition into the model would reduce the Shapley values of the Services and increase the Shapley values for the input suppliers, by recognizing which players provide “essential inputs” and which are in competition with other suppliers of substitutable inputs. Id.

He further criticized Professor Marx for including in her model “other distributors” who are not interactive streaming services. Id. at ¶ 27. According to Professor Watt, these other distributors “do not belong in a properly constructed Shapley Value Model because their presence would “show up” in the model as lower revenues for interactive services as their subscribers or listeners left for these other distributors (such as noninteractive services). Id.

Additionally, because he criticized Professor Marx's use of 2015 data (as noted supra), Professor Watt re-worked Professor Marx's model by examining how the use of 2016 data, as opposed to her 2015 data, would “better reflect[ ] . . . the reality of the market. Id. at ¶ 37 ; see also id. at ¶ 44. When using the (higher) 2016 revenues (and making some relatively more minor adjustments he found necessary), Professor Watt estimated that the share of streaming revenues that would be paid out in total royalties (for musical works + sound recordings) in Professor Marx's model would range from [REDACTED]% to [REDACTED]%. Id. at ¶¶ 50–52.

After analyzing these Shapley analyses, the Majority found that the mechanical royalty rate needed to be increased in order to provide Copyright Owners with a reasonable rate as required by section 801(b)(1). As a matter of arithmetic though, if the mechanical rate increased and the sound recording rate did not decrease by a corresponding amount, then the total royalties paid by the Services would increase. That issue brings the Judges to consideration of Professor Watt's bargaining model, on which the Majority relied to posit an inverse relationship (the seesaw effect), by which an increase in the mechanical rate would result in a commensurate reduction in the sound recording rate.

(ii) Professor Watt's Bargaining Model

Professor Watt's Nash Bargaining Model is the linchpin that connects: (a) the higher mechanical royalty rates generated by the Shapley Value results relied upon by the Majority with (b) the assumed lower sound recording rates—a connection that the Majority found to render “reasonable” and “fair” its uncapped TCC prong. See Determination at 73–74 (“As to the issue of applying a TCC percentage to a sound recording royalty rate that is artificially high as a result of musical works rates being held artificially low through regulation, the Judges rely on Professor Watt's insight (demonstrated by his bargaining model) that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works.”). Alternately stated, Professor Watt's bargaining model result, i.e., the seesaw effect, if sufficiently supported in the record, is the phenomenon that would allow the Judges on remand to apply the Shapley results by increasing the mechanical rate, without unduly exposing the Services to the risk of higher total royalties.

More particularly, the Majority recognized a potential problem that those Judges would have to resolve before utilizing the Shapley Value approach to create an uncapped TCC prong: “This is problematic because the sound recording rate against which the TCC rate would be applied is inflated . . . both by . . . complementary oligopoly [market] conditions . . . and the record companies' ability to obtain most of the available surplus due to the music publishers' absence from the bargaining table.” Determination at 73. But the Majority found that Professor Watt had provided a rationale which permitted them to resolve the second problem:

As to the issue of applying a TCC percentage to a sound recording royalty rate that is artificially high as a result of musical works rates being held artificially low through regulation, the Judges rely on Professor Watt's insight . . . that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works. 3/27/17 Tr. 3090 (Watt) (“[T]he reason why the sound recording rate is so very high is because the statutory rate is very low. And if you increase the statutory rate, the bargained sound recording rate will go down.”).

Determination at 73–74; see also Watt WRT ¶ 23 n.13 (“[I]in my Appendix 3, I show that . . . if the musical works rate is increased to what would be a realistically fair and reasonable rate, then the negotiated fee for sound recordings would decrease almost dollar for dollar . . . .”); see also id. at ¶ 36 (“The statutory rate for mechanical royalties . . . is significantly below the predicted fair rate, and the statutory rate effectively removes the musical works rightsholders from the bargaining table with the services. Since this leaves the sound recording rightsholders as the only remaining essential input, bargaining theory tells us that they will successfully obtain most of the available surplus.”).

To repeat: This inverse relationship is what has been described as the “seesaw” effect. The question in this regard on remand is whether the record proves that the seesaw theory is valid and measurable going forward. Alternately stated, does the record prove that Professor Watt's bargaining model serves as the linchpin that would allow the Judges to apply the Shapley results by increasing the mechanical rate, without unduly exposing the Services to the risk of higher total royalties?

To resolve this issue, the Judges examine this bargaining model dispute in detail, as it bears on whether the uncapped TCC rate structure can be incorporated into the statutory rate.

(a) Bargaining Model Dispute

Professor Watt utilized a general Nash Bargaining Model. In his particular application, Professor Watt modeled the streaming services and the labels each as a “single unit,” asserting (as is common in Shapley analyses) that this single-unit modeling was done “for simplicity.” Watt WRT, appx. 3 at 10. Applying this and other modeling assumptions, Professor Watt posited: “If there were to be no successful deal, then each of these two bargainers [the assumed “single” interactive service and “single” label] would earn 0, since in that case the interactive streaming service could not operate.” Id.

In his oral testimony at the hearing, Professor Watt did not opine as to whether changes in variables other than musical works royalties would also have an impact on the level of sound recording royalty rates, even as higher musical works rates would otherwise place virtually 1:1 downward pressure on the sound recording rate. However, in his written rebuttal hearing testimony, i.e., his WRT, Professor Watt did make varying assumptions regarding the changes in the Services' non-content costs, by which he did change the total revenue share for content providers. Watt WRT ¶¶ 50–52. He concluded from this varying replication of Professor Marx's Shapley model “ that the results that it delivers are very dependent upon the amount of total interactive streaming revenue and the fraction of that revenue that is taken up by downstream non-content costs. ” Id. at ¶ 53 (emphasis added).

The Services had no procedural right under part 351 of the Judges' regulations to proffer surrebuttal written testimony from economic witnesses to challenge Professor Watt's assertion, made for the first time in rebuttal, of the seesaw relationship between changes in the musical works royalty rate and the sound recording royalty rate paid by interactive services. Moreover, the Services and their economists also had no opportunity to weigh in on the Majority's application of same (which was not revealed until the Judges rendered their decision). See Johnson, 969 F.3d at 381 (“Streaming Services had no notice that they needed to defend against and create a record addressing such a significant, and significantly adverse, overhaul of the mechanical license royalty scheme.”). Now though, on this remand, the Services have been afforded the opportunity to present these criticisms, through their expert witnesses.

(b) Professor Katz's Principal Criticism

Pandora's economic expert, Professor Michael Katz, levied several criticisms of the bargaining model proffered by Professor Watt and applied by the Majority. The most important problem with Professor Watt's analysis, according to Professor Katz, is that the former's model assumes an “extremely unrealistic” zero payoff to the label in the absence of an agreement with a streaming service—an assumption which is “far from . . . innocuous.” Written Direct Remand Testimony of Professor Michael Katz (Katz WDRT) ¶¶ 16, 20.

Professor Katz opines that this zero payoff assumption is equivalent to assuming, contrary to undisputed market facts, that: (1) subscribers and listeners to an interactive service would not switch to other interactive services if that service failed to reach an agreement with the labels; and (2) the interactive service is a “Must-Have” input supplier. Katz WDRT ¶¶ 17–18. In terms of Nash modeling, according to Professor Katz, Professor Watt's assumption is thus equivalent to “assum[ing] that the sound recording copyright owners have no outside option. ” Katz WDRT ¶ 127 (app. A) (emphasis added).

Moreover, not only does Professor Katz assert the indisputability that such substitution would occur, he points out that Professor Watt himself acknowledged in his own testimony that such substitution would occur. Katz WDRT ¶ 19.

Beyond this purported inconsistency, Professor Katz finds Professor Watt's no-substitution assumption to be a serious modeling error because, in order to quantify accurately each Nash bargainer's contribution to the net surplus to be divided, the extent of substitutability on each side of the market must be captured by the modeling. Katz WDRT ¶ 20. That is, he opines that “Professor Watt's assumption that there is no substitution dramatically biases his model toward finding a large seesaw effect and renders his analysis unreliable . . . lead[ing]to a prediction that the share of an increase in musical works royalties that will fall on the streaming services is approximately eight times larger than Professor Watt's prediction. Id. at ¶ 21.

As a matter of music business dynamics, Professor Katz interprets Professor Watt's substitutability error as follows.

The assumption that a label receives a zero payoff if it does not reach agreement with a streaming service is equivalent to assuming that, if a streaming service shut down, none of the consumers who would otherwise have used that streaming service will switch to alternative streaming services or other sources of licensed music. The two forms of the assumption are equivalent because, when the services are substitutes, failure to reach an agreement with one service will not drive a label's payoffs from interactive streaming to zero. It will not result in the loss of all of the benefits that could be enjoyed by reaching an agreement. Instead, many consumers would engage in substitution and choose other streaming services, which will allow the label to earn profits from the additional royalties that would be paid to it by those other services.

Id. at ¶ 18.

Professor Katz attempts to adjust Professor Watt's Nash Bargaining Model to account for this substitution effect. In his Appendix A, Professor Katz—acknowledging the reality of multiple interactive services—changes Professor Watt's assumed single label's payoff (designated as parameter “A” in the Nash Bargaining Model) from a value of zero to a value equal to “the share of revenues that would be diverted to other streaming services” multiplied by “the royalty rate that the label receives from the other interactive streaming services.” Id. ¶¶ 119, 127. Professor Katz asserts that the diversion to other streaming services represents an “outside option” available to a label. Id. ¶ 127. Professor Katz incorporates this “outside option” in his revised version of Professor Watt's Nash Bargaining Model.

In addition, Professor Katz asserts that Professor Watt's modeling is unreliable because “his prediction of the size of the see-saw effect is very sensitive to the assumed values of various other parameters.” Id. at ¶ 23. For example, Professor Katz asserts that a change in the royalty rate paid to the labels could materially affect the balance or even the existence of the seesaw effect. Id. at ¶ 127. As further support for his opinion, Professor Katz relies on the testimony of one of Copyright Owners' own economic expert witnesses, who gave testimony clearly indicating that the “seesaw” effect was not at all likely to occur. Id. ¶ 24, n.16 (citing Gans WRT ¶ 32).

In sum, Professor Katz finds Professor Watt's Nash Bargaining Model to be unusable as a foundation to set royalty rates because, although “there are theoretical reasons to believe that a see-saw effect may occur, . . . there are complications and it is difficult to predict how big the effect will be.” Id. ¶ 24 (emphasis added).

(c) Professor Watt's Rebuttal to Professor Katz

In rebuttal to Professor Katz's criticisms, Professor Watt states that “the record needs to be straight on Nash bargaining theory,” in order to explain “the foundational error” committed by Professor Katz. Watt RWRT ¶ 52. This basic mistake, according to Professor Watt, is Professor Katz's erroneous assertion that the bargaining model must account for a label's “outside option.” Id. ¶ 53. Relying on economic authority regarding bargaining theory, Professor Watt defines an “outside option” as “the best alternative that a player can command if he withdraws unilaterally from the bargaining process.” Id. ¶ 59 (emphasis added); see also id. ¶ 53 (“An outside option is a payoff that the label would receive if negotiations with the service do not result in an agreement. ”) (emphasis added).

Connecting this principle of bargaining theory to economic theory, Professor Watt explains his understanding of the relationship of the “outside option” to the more familiar economic concept of “opportunity cost”:

An outside option could also be referred to as an “opportunity cost,” since it is the value of what would be foregone should a deal with the service actually be struck. It is . . . useful to recognize the equivalence between an outside option and an opportunity cost, because economics in general has a very long history of understanding how opportunity costs weigh in on economic decision making.

Id.

Professor Watt then opines how Professor Katz confused the “outside option” with the disagreement (a/k/a threat) point in the Nash Bargaining Model:

[Professor] Katz claim[s] that the outside option value that the labels would enjoy should they not reach an agreement with the services should be included as part of the “disagreement point” within the bargaining model and reimbursed like a cost prior to bargaining. Doing this can dramatically alter the results of the model. It is also definitively not how such an option should be modelled. [Professor] Katz [is] guilty of misunderstanding the Nash bargaining model, and concretely, the meaning of a “disagreement point,” and the way that an outside option should be brought into the model.

Id. ¶ 55.

More particularly, according to Professor Watt, these outside options/opportunity costs do not belong in a Nash Bargaining Model, because they are “not the types of status quo actual financial payments that may be modelled as disagreement points.” Id. ¶ 57. Rather, he asserts that, as Professor Katz essentially acknowledged, they are “payoffs from substitution, [ i.e.,] an option instead of the deal, and they are not actual financial payments, but opportunity costs. Id.

Professor Watt then explains that an outside option/opportunity that by definition exists as an alternative to a bargain between two parties lies outside the two parties' bargain, and is thus out-of-place within a proper Nash Bargaining Model:

In the case at hand, if the parties never stop negotiating and never take up substitute options, then no joint enterprise is offered and there is no surplus to share, so each necessarily gets a payoff equal to 0, just as I assumed in my model.

. . .

[A]gainst this backdrop, an outside option (a potential payoff that is not directly related to a share of the surplus that is being negotiated) . . . comes in [to the model] as a constraint upon the set of feasible deals that could be struck, exactly as an opportunity cost would be treated.

Id. ¶¶ 57–58.

(d) Dr. Leonard's Criticisms of Professor Watt's Bargaining Model

According to Google's economic expert witness, Dr. Gregory Leonard, the Majority wrongly relied on Professor Watt's bargaining model because it is “highly stylized” and theoretically “simplified” in ways that make it unable to predict that “an increase in the musical works royalty would be offset nearly dollar-for-dollar by a decrease in the sound recording royalties (the “seesaw effect”), thus leaving the services virtually unaffected by the proposed increase in musical works royalties.” Leonard WDRT ¶ 8.

Pointedly, Dr. Leonard criticizes Professor Watt's bargaining model as comprised of a “veneer of `complexity' . . . mathematical formulas and [a] reference to John Nash,” adopted to provide a rationalization for adoption of his Shapley Value modeling that would significantly increase the mechanical royalty rate.” Id. ¶ 16. These modeling deficiencies, Dr. Leonard asserts, are not merely “simplifying assumptions [that] better focus on the specific question the model is meant to address,” but rather “simplify away economic characteristics . . . entirely abstract[ing] away economic characteristics . . . central to the question at hand.” Id. ¶ 18.

In particular, Dr. Leonard avers that Professor Watt's bargaining model materially abstracts away from, inter alia: (1) the nature of consumer demand for streaming services and competing forms of music; (2) how services decide to enter or exit the streaming market; (3) the nature of the oligopolistic interaction among the labels; (4) the nature and timing of the bargaining between each label and each service; (5) the potential for “hold-up” by labels that perceive the services to be in a vulnerable bargaining position due to their previous industry-specific investments made under their assumption that the pre-existing statutory structure would be maintained; and (6) the failure of Professor Watt's bargaining model to grapple with the complementary oligopoly structure of the sound recording market. Id. ¶¶ 18, 20.

These factors, he posited, are “important for determining how sound recording royalties would actually change in response to a change in the statutory musical works royalty.” Id. Professor Leonard concludes that, by not modeling these factors, Professor Watt's “prediction of a virtual dollar for dollar decrease in sound recording royalties is unreliable as a basis for formulating policy.” Id. ¶ 20.

Regarding the complementary oligopoly structure of the market and its impact on the bargaining process, Professor Leonard emphasizes that an important “real-world hurdle” assumed away by Professor Watt's modeling of a single label entity is that “each label would prefer to have the other labels lower their sound recording royalties while maintaining its own royalties at pre-existing levels . . . .” Id. ¶ 21. More particularly, Dr. Leonard explains that “even if a label were to recognize that it is more efficient for overall sound recording royalties to be lower, the label may not be willing to lower its royalty rate without assurance that the other labels will do the same,” a result which he asserts “is unlikely to happen absent some form of collusive behavior.” Id. Thus, Dr. Leonard maintains that the existence and size of any “seesaw”-induced decrease in sound recording royalties remains indeterminate, and it remains “within the realm of theoretical possibility that the labels do not agree to any reduction in sound recording royalties even if a reduction in overall royalties would be economically efficient. Id.

(e) Professor Watt's Rebuttal to Dr. Leonard's Criticisms

Professor Watt replies with a spirited defense of economic modeling in general and his economic bargaining model in particular. He begins by pointing out that models are not supposed to be “perfect representations of reality [but rather] are intended to isolate what is important, in order to expose a useful insight on some issue of relevance.” Watt RWRT ¶ 105. He adds that economic models (not merely his bargaining model) “ do not necessarily deliver predictions of situations that are immune to changes in variables outside the model, but rather the results inform conclusions about the relationships between the variables and parameters within the model, [which is] by nature a crude representation[ ] of reality, but the lessons and insights that they provide can be very relevant to real-world applications.” Id. ¶¶ 106–07 (emphasis added).

With particular regard to his bargaining model, Professor Watt takes issue with Dr. Leonard's assertion that in the former's model the surplus is a “fixed constant.” See Watt RWRT ¶¶ 110–111. Rather, Professor Watt avers that his bargaining model assume[s] that when the surplus . . . whatever value it takes. . . is to be shared, the parties understand that the amount to be shared is, at that moment, given.” Id. ¶ 111 (emphasis added).

Turning to Dr. Leonard's critique regarding the purported distortionary effect of Professor Watt's modeling assumption of a single label and a single interactive service, Professor Watt responds by acknowledging that, if he had modeled multiple labels and services in the bargaining process, that would be “not particularly enlightening vis-à-vis the single bargain setting, as it will not lead to different insights than those distilled by the [Majority].” Id. ¶ 113. Further, Professor Watt characterizes this criticism as “empty,” because under either his two-player Nash model or Dr. Leonard's posited multi-player (Nash-in-Nash) model, the labels will not respond to a musical works royalty increase ipso facto with a reduction in the sound recording royalty ( i.e., the seesaw effect will not occur if there is “a change in some other variable.”). Id. ¶ 114.

(f) Professor Marx's Criticisms of Professor Watt's Bargaining Model

Professor Marx criticizes Professor Watt's application of the Nash Bargaining Model because, in her opinion, its “precise prediction” of the nearly one-to-one seesaw relationship “depends critically on the assumptions that he makes and the numerical inputs that he uses.” Marx WDRT ¶ 33. First, criticizing his modeling assumptions, like Professor Katz, she criticizes his decision to abstract from reality by positing a single label and a single interactive streaming service. She opines that his one label/one service modeling assumption ineluctably leads to his conclusion that each of these two parties “has a `disagreement payoff' of zero [meaning that] each party ends up with nothing in the absence of a deal.” Id. ¶ 34. But this zero “disagreement payoff” is merely a product of Professor Watt's abstraction from reality, according to Professor Marx, because “[i]n reality, if interactive streaming went away, a share of the music listening that had occurred through interactive streaming services would migrate to other forms of music distribution, generating revenues for the label . . . meaning that the disagreement payoff would be positive for the label). Id. (emphasis added). Consistent with Professor Katz, she maintains that Professor Watt himself acknowledged the presence of this substitution effect when he testified that “[t]he existing interactive streaming companies do not hold an essential input, as first they compete with the non-interactive services . . . .” Id. ¶ 35, n.43 (citing Watt WRT, app. 3).

More particularly, Professor Marx maintains, a record label's disagreement payoff must be considered realistically “in any accounting of what would happen if record labels and interactive streaming services failed to reach an Agreement . . . .” Marx RWDT ¶ 35. And, she opines, when this real-world substitution effect is taken into account, the seesaw effect that Professor Watt estimates is reduced dramatically, because “[t]he greater . . . the substitution between streaming and other forms of distribution, the greater is the revenue that the record label can capture in the event of disagreement and the lower is the estimated see-saw effect.” Id.

Professor Marx opines that modeling the bargaining process without these real-world particulars diminishes the value of Professor Watt's Nash model in several significant ways. First, because his model fails to incorporate the presence of three major record labels, “each with substantial complementary oligopoly power,” it fails to capture the fact that “each record label does not fully internalize the impact of its rates on the viability of the industry.” Id. ¶ 39. She points to the Judges' Final Determination in Web IV, where the Judges note how this aspect of complementary oligopoly compromises the value of a rate as a useful benchmark. Id. ¶ 39 n.45 (quoting Web IV Final Determination). More particularly, she opines that when, as here, “there are multiple negotiations between multiple record labels and multiple services,” sound recording rates can be affected “by the order of negotiations” among the several label:service negotiating pairs—a factor that Professor Watt's bargaining model fails to capture. Marx WRDRT ¶ 41.

Next, Professor Marx avers that Professor Watt's bargaining model “does not explain how or over what time frame the market would move to a new equilibrium.” Id. ¶ 40. More particularly, she testifies, because interactive services' “agreements with record labels often contain multi-year terms and can take many years to negotiate . . . there may be little incentive or practical ability for both sides to move to a new rate before the contract expires”. Id. ¶ 41. She takes note that this point was established at the hearing during questioning of Professor Watt from the bench:

JUDGE STRICKLER: What of the situation . . . that the . . . time period for the existing agreements between the . . . labels and the interactive streamers is such that they've already locked in a particular rate and then we set a rate that's higher for the mechanical to reflect the fact that the sound recording royalty should drop, but it's locked in for a period of time? Are we running the risk, then, of disrupting the market by having a total royalty that's greater than what is indicated by your Shapley testimony, simply because of the disparity of times in which the rates are . . . implemented?

PROFESSOR WATT: That's a very fair point. And I didn't even think of that until you've mentioned it . . . [T]he model I have done is . . . assuming that . . . the bargained thing happens at the same time as the—or in the same general period of time as a change in the statutory rate. You're absolutely correct.

3/27/17 Tr. 3091–92 (Watt); see Marx WRDRT ¶ 42, n.46

Third, Professor Marx points out that Professor Watt's Nash model does not attempt to capture the effects of the heterogeneous and asymmetric distribution of information relevant to the bargain available to each party at the time of negotiation. Id. ¶ 41.

Lastly, Professor Marx avers that Professor Watt's Nash Bargaining Model fails to address, on a more general basis beyond informational issues, other “asymmetries among record labels and among services.” Marx WDRT ¶ 41.

In sum, Professor Marx concludes that these foregoing real-world points all preclude the Judges from relying on Professor Watt's testimony to identify a stable relationship between changes in the mechanical royalty rate and the sound recording royalty rate because they all share a common defect—they “lie outside Professor Watt's model.” Marx WRDT ¶ 41.

To be clear, Professor Marx does not criticize Professor Watt for neglecting to include these points in his bargaining model; rather, she acknowledges that “[t]hese are difficult features to capture in a tractable equilibrium model.” Id. Indeed, she urges the Judges to appreciate that relying on such a necessarily limited model, as the Majority did, can have “dramatic effects” on the royalty rates derived. Id. Professor Marx emphasizes that all of these inherent modeling deficiencies are especially pernicious, if the bargaining model is applied yet again on remand, to set specific rates over a five-year period, when other variables will have independent effect on royalty rates. Id.

(g) Professor Watt's Rebuttal to Professor Marx

Because Professor Marx's criticisms are of a similar nature to Professor Katz's criticisms, Professor Watt responds to Professor Marx as he did to Professor Katz. To summarize, Professor Watt responds to Professor Marx's points as follows:

• Her criticism is centered on what he characterizes as her “bogus” argument that he supposedly had predicted almost a “dollar for dollar” sound recording rate reduction in response to an increase in the musical works rate (the seesaw effect). Watt RWRT ¶ 19. Professor Watt finds this argument “particularly disheartening,” because Nash bargaining theory explains why the seesaw would apply to the splitting of the surplus based on the available data, and that “there are quite apparent reasons why available surplus may not decrease even if the musical works rate increased, because of simultaneous changes to other variables in the model.” Id. ¶ 34 (emphasis added).

• Professor Marx implicitly contradicts her own reliance on the complementary oligopoly power of the Major labels by modifying his bargaining model through the insertion of a lower value for their bargaining power. Id. ¶¶ 19, 22–24, 26.

• Professor Marx misconstrues the purpose of his Nash model, which was to serve “as a reply” to Professor Marx's direct testimony, and “to show bargaining insights that bore upon aspects of the case.” Id. ¶ 29.

• Professor Marx, like Professor Katz, improperly includes in her bargaining model a potential payoff for the label arising from an “outside option,” i.e., from an alternative that the label can choose only if the Nash bargaining terminates. Id. ¶¶ 53—68.

(h) Professor Marx's Reply to Professor Watt's Criticism

In her supplemental remand testimony, Professor Marx challenged several of Professor Watt's criticisms contained in his remand testimony. First, she takes issue with what he identified as two “core” economic principles of bargaining: (1) that all of the available net surplus will be shared; and (2) that neither of the two bargainers will demand a share such that more than the total net surplus is shared. Marx WSRT ¶¶ 7–8.

As an initial matter, she disputes the notion that these are “core” principles of bargaining. Id. ¶ 8. More particularly, she states that, in the present case, because “the label does not know with exactitude the precise maximum that a service would be willing to pay ( i.e., its “survival” rate), and the service likewise does not know the exact minimum that the label would be willing to accept,” the simple bargaining model must be expanded to address “the potential for delay and/or bargaining breakdown.” Id.

As a further criticism, Professor Marx avers that “[i]n the real world, the negotiated royalty outcomes do not involve just two parties, but rather a sequence of overlapping, interrelated, bilateral bargains involving multiple competing services and multiple record labels with complementary oligopoly power.” Id. ¶ 12. This complication, she opines, exacerbates the informational deficit noted in the immediately preceding paragraph, such that negotiations within the several pairings of labels and services “are affected by uncertainty and private information and . . . Professor Watt's discussion of bargaining theory [thus] does not support any particular real-world see-saw outcome.” Id.

(iii) Resolution of the Bargaining Dispute

(a) Professor Watt's Nash Bargaining Model Does Not Support Adoption of Uncapped TCC Rate

The purpose of Professor Watt's Nash Bargaining Model was to allay the Judges' concern that increasing the mechanical rate would lead to higher total royalties for the Services. His bargaining model was understood by the Majority to show that such higher total royalties would not result, because the model demonstrated the “seesaw” effect, whereby the sound recording rate would fall almost dollar-for-dollar with the increase in the mechanical rate. See Determination at 73–74 (“[T]he Judges rely on Professor Watt's insight . . . demonstrated by his bargaining model that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works. . . . Professor Watt's bargaining model predicts that the total of musical works and sound recordings royalties would stay `almost the same' in response to an increase in the statutory royalty.”) (emphasis added).

On the surface, the economic experts on both sides appear to be at loggerheads regarding the existence and applicability of the seesaw relationship. However, as discussed below, on further analysis of their respective positions, in light of Professor Watt's remand testimony regarding a key assumption in his bargaining model, their disagreement narrows considerably and—in an important respect—vanishes completely.

To recap: In his WRT, Professor Watt stated

[W]ith an appropriately modelled bargaining analysis . . . in my Appendix 3 . . . I show that for every dollar that the statutory rate for musical works undercuts a fair and reasonable rate, the freely negotiated rate for sound recordings will increase by an estimated [REDACTED] cents.

That is, if the musical works rate is increased to what would be a realistically fair and reasonable rate, then the negotiated fee for sound recordings would decrease almost dollar for dollar, with only a minor change in the total royalty rate for all copyrights combined.

Watt WRT ¶ 23 & n.13. But nowhere in his WRT did he qualify this statement by explicitly acknowledging that in his bargaining model there are certain assumptions lurking, i.e., that his “concrete” analysis is subject to the “ ceteris paribus ” constraint—that all other things are held constant ( i.e., equal before and after the change in the musical works rate) other things being equal).

It is only in his later remand testimony—after the D.C. Circuit's remand had compelled him to confront criticism from adverse economists—that Professor Watt expresses this assumption overtly, making explicit the “understanding” that he had theretofore only tacitly assumed:

In other words, a model in which only the two copyright rates are permitted to change . . . as was the understanding in my original model, allows the system to derive a clear relationship between those two rates, and that relationship is that an increase in one leads to a decrease in the other, that is, the `see-saw effect.' But if . . . something else changes along with the musical works rate . . . then the net effect does not predict that the negotiated rate of the labels will decrease.”

Watt RWRT ¶ 35 (emphasis added).

Indeed, as noted supra, Professor Watt did give a nod to the relaxing of his implied ceteris paribus assumption in his WRT, by identifying varying “scenarios” in which he considered the impact of potential changes in service revenues and service non-content costs, leading to different percentages of royalties paid to content providers. Watt WRT ¶¶ 45–52. Professor Watt then used these several assumptions and scenarios to opine as follows: “The message that should be taken from this exercise . . . is that the results . . . are very dependent upon the amount of total interactive streaming revenue and the fraction of that revenue that is taken up by downstream non-content costs.” Id. ¶ 53.

Professor Spulber, on behalf of Copyright Owners, likewise emphasizes on remand the importance of the ceteris paribus assumption in economic modeling:

[A]long with an increase in the compulsory license rate, all other things being equal, we would expect to see a decrease in sound recording royalty rates.

. . .

“All other things being equal” ( ceteris paribus in Latin), is a central principle for economic modelling. This economic analysis of bargaining highlights an important relationship between two content cost variables. However, that relationship does not exist in a vacuum. Many other variables affect the bargaining situation and, for any given period, the net effect of all of the different variables may be different than the effect of the modeled variable alone. Thus, this economic analysis of bargaining will not assure that a streaming service will not face disruption in the real world for any reason.

. . .

Economic modeling is supposed to simplify the situation in order to distill useful principles and teachings.

Spulber RWRT ¶¶ 26–28 (emphasis added).

The Judges agree that the ceteris paribus principle is a fundamental principle in economic analysis and modeling. Professor Watt succinctly makes this point, quoting the Nobel laureate economist James Buchanan, for the following proposition:

At the heart of any analytical process lies simplification or abstraction, the whole purpose of which is that of making problems scientifically manageable. In the economic system we recognize, of course, that `everything depends on everything else,' and also that `everything is always changing'.

Watt RWRT ¶ 32 ( quoting J. Buchanan, Ceteris paribus: Some Notes on Methodology, 24 So. Econ. J. 259, 259 (1958).

However, Professor Watt does not quote another portion of Professor Buchanan's article that makes a point that looms large in the present proceeding, to wit, the limitations inherent in applying the necessary ceteris paribus condition:

Real problems require the construction of models, and the skill of the scientist is reflected in the predictive or explanatory value of the model chosen. We simplify reality to construct these models, but the fundamental truth of interdependence must never be forgotten. . . . [However,] [f]ew, if any, meaningful results may be achieved by using ceteris paribus to eliminate the study of large numbers of variables. If such variables are closely related, they must be studied simultaneously; there is no escape route open.

Id. at 259–60 (emphasis added); see also A. Rubinstein, Comments on Economic Models, Economics, and Economists: Remarks on Economics Rules by D. Rodrik, 55 J. Econ. Lit.162, 167 (2017) “[W]hat matters to the empirical relevance of a model is the realism of its critical assumptions”) (emphasis added).

This is not to say that Professor Watt was unaware of this caveat. As noted supra, he recognizes the difficulty of extrapolating from a ceteris paribus world to the real world. The present panel of Judges likewise recognizes this. However, the Majority missed this distinction in the Determination when it applied Professor Watt's correct but ceteris paribus “insight” for a constant real-world relationship between sound recording and musical works royalty rates. Again, not a single economist made this improper analytical leap or proposed an uncapped TCC rate in order to set a TCC ratio across the entire rate term. Indeed, on careful inspection, no economist states in his or her remand testimony that Professor Watt's bargaining model provides economic support for the uncapped TCC rate prong.

With the foregoing testimony in mind, the Judges see particularly relevant several additional points in Professor Watt's remand rebuttal testimony that pertain to the appropriateness, vel non, of a TCC rate prong. Referring to the application of his bargaining model to the present case, Professor Watt made these crucial statements regarding the lack of a seesaw effect that would generate decreases in sound recording rates when the mechanical rate is increased:

[T]he actual effects one would expect to see several years later would be based on the actual data at that time. Moreover, I would expect many other variables to have a larger effect on the bargains than the relatively small changes in the musical works rate. . . . [U]nderstanding actual market outcomes requires understanding these variables.

. . .

[A]n attempt to capture all aspects of the real world is too complex for a simple statistical exercise involving an econometric regression. There is no obvious data to actually use for some of the independent variables, such as consumer demand equations, costs of entry and exit, a measure of oligopolistic interaction, different timings of different rate bargains, and the actual values of outside options.

Watt WRWT ¶¶ 6(iv), 118.

Although Professor Watt was hardly transparent in disclosing his ceteris paribus assumption in his original testimony, it seems clear that he always understood its presence, and that, when this assumption was relaxed, “the actual effects . . . several years later would be based on the actual data at that time [and] many other variables [with] a larger effect on the bargains than the relatively small changes in the musical works rate.” Id. ¶ 6(iv) (emphasis added).

Professor Spulber likewise opined that the absence of an explicit statement of these assumptions in Professor Watt's testimony was unremarkable and appropriate:

[A]ll other things being equal'. . . should be generally read into economic modeling conclusions or predictions, whether or not the words are repeated in each instance. Economists do not typically repeat these words in each place where they apply, since it would lead to constant repetition.

Spulber RWRT ¶ 46, n.8.

Regardless of whether economists invariably identify the existence of implicit assumptions lurking in each other's models, Professor Watt overlooked a cardinal rule of communication: Know your audience. Here, his audience is comprised of three Judges, only one of whom is also an economist. Failing to appreciate Professor Watt's implied ceteris paribus assumption, the Majority transformed his limited (albeit important) “insight” regarding the equal split of the Shapley surplus between the two classes of rights holders—and the seesaw effect that would have if the mechanical rate were increased when the split was imposed— into a justification for the imposition of an uncapped TCC rate prong over the five-year rate term. The Majority's language reveals this point clearly:

As to the issue of applying a TCC percentage to a sound recording royalty rate that is artificially high as a result of musical works rates being held artificially low through regulation, the Judges rely on Professor Watt's insight . . . demonstrated by his bargaining model that sound recording royalty rates in the unregulated market will decline in response to an increase in the compulsory license rate for musical works. See 3/27/17 Tr. 3090 (Watt) (“[T]he reason why the sound recording rate is so very high is because the statutory rate is very low. And if you increase the statutory rate, the bargained sound recording rate will go down.”)

Professor Watt's bargaining model predicts that the total of musical works and sound recordings royalties would stay “almost the same” in response to an increase in the statutory royalty. Id. at 3091.

Determination at 73–74 (emphasis added).

Making the point ever so plainly, Professor Watt now expressly acknowledges that his “ `see-saw effect' was never really a ` prediction ' ” at all! Watt RWRT ¶ 117. Rather, he now cautions the present panel of Judges, that, “to make the jump from the model to the actual real-world effects, one cannot ignore the words that are omnipresent in all economic modeling, that predictions about causal relationships are understood to be “all else equal.” Id. ¶ 32.

Without the benefit of these caveats regarding an extrapolation of the “seesaw” theory to the real-world, and with absence of an explicit statement of the ceteris paribus assumption, the Majority misapplied his testimony as a basis to adopt a fixed TCC rate, based upon data from a snapshot in time (2016) to cement that rate relationship for the entire five-year period. The Majority misapplied Professor Watt's correct insight from bargaining theory regarding the use of a fixed ratio for the equal division by two “Must Have” input suppliers of the Shapley surplus to set royalty rates in a period, by using that insight incorrectly to establish a fixed ratio of royalty rates over the rate term.

Additionally, an examination of the expert economists' testimony reveals that their facial disagreements vanish once the necessary assumptions are laid bare. Professor Watt and the Services' three economists all identify the following independent variables that will impact the relative levels of sound recording and musical works rates paid by interactive services:

(1) the level of downstream consumer demand;

(2) entry costs;

(3) exit costs;

(4) oligopolistic interaction;

(5) the timing of sound recording agreements vis-à-vis statutory rate setting; and

Professor Watt and the three Service economists agree with regard to the relevancy of these six independent variables. Compare Watt RWRT ¶¶ 6(iv), 118 (identifying all five independent variables) with Leonard WDRT ¶ 18 (identifying independent variables 1–4 above); Marx WDRT ¶¶ 4–5, 42; (identifying independent variables 1–5 above); Katz WDRT ¶¶ 127, 134 n.115 (identifying independent variables 4 and 6 above). Accordingly, the remand record shows a consensus as to the lack of modeling of independent variables that would be important to estimate an uncapped TCC royalty ratio that could be utilized by the Judges to lock-in a ratio over the rate term.

Indeed, as noted supra, a careful reading of the remand testimony by Copyright Owners' economists, Professors Watt and Spulber, reveals that neither of them actually testifies that there is sufficient theoretical and empirical evidence to support the uncapped TCC rate prong and the 26.2% TCC rate phased in on that prong. Rather, those two witnesses testify to something far narrower: the alleged correctness of Professor Watt's “seesaw” theory as demonstrating an equal splitting of the surplus between the two “Must Have” input suppliers, and the effect of that split when all other relevant independent variable are held constant.

In this regard, it is noteworthy that none of Copyright Owners' several economic experts in this proceeding (Dr. Eisenach, Professor Gans, Dr. Rysman, or Professor Watt) ever proposed an uncapped TCC rate prong in any form, let alone within a greater-of formulation. Such a proposal would have been improper, because, as the expert testimony described above makes clear, the ceteris paribus assumption, reasonable for modeling purposes to provide insight as to the surplus split, lacks the input of the omitted variables that the experts on both sides find relevant to the application of economic modeling in this proceeding. A further review of Copyright Owners' economic expert witness testimony on remand—the first time any of them had occasion to weigh-in on the appropriateness of the uncapped TCC prong—reveals that they also have not endorsed the uncapped TCC rate prong as a proper form of rate setting. To be sure, they strongly endorse the insight first described by Professor Watt in his WRT that the Nash surplus would be split essentially evenly between the two suppliers of essential content, given his simplifying assumptions. But such endorsement is hardly the same as endorsement of the uncapped rate prong itself.

For these reasons, the Judges find erroneous the Majority's identification of a fixed relationship between the sound recording and mechanical royalty rates that could serve as a basis for the Majority's first rationale for yoking the mechanical rate to an uncapped TCC rate prong.

(b) The Services Have Not Rebutted Copyright Owners' Prima Facie Showing That Professor Watt's Model Demonstrates a More Limited “Seesaw” Effect

The foregoing analysis and decision related to the absence of a fixed relationship between the sound recording and mechanical royalty rates. A separate fixed relationship—the one Professor Watt has clarified he was demonstrating all along—is that if the Judges increase the mechanical royalty rate, the Shapley surplus realized by the labels will decrease almost dollar-for-dollar with the increase in the mechanical rate. The Services' economists aver that even this version of the seesaw is defective.

According to Professors Katz and Marx, the Nash Bargaining Model constructed by Professor Watt is deficient because it fails to properly characterize the “disagreement payoff” to the sound recording company when it and an interactive service fail to reach an agreement. More particularly, as explained supra, they assert that Professor Watt's model omits the value of “outside options” available to the sound recording company. This criticism relates to the issue of whether the seesaw effect would occur as posited in Professor Watt's model. That is, the increase in the sound recording company's “disagreement payoff” (a/k/a “threat point”) would lead to a higher royalty in the Nash bargain between the sound recording company and the interactive service than needed to generate the seesaw effect to offset the higher mechanical royalty rate.

As the several experts' positions in this regard, discussed supra, make clear, however, each side has a different understanding of whether an “outside option” is properly included in the definition and calculation of the “disagreement payoff.” On the one hand, Professors Katz and Marx claim that the existence and value of “outside options” should be included in the “disagreement payoff.” However, they provide no economic authority for that assertion.

By contrast, Professor Watt cites to multiple economic game theory publications and authorities for the proposition that the presence and value of “outside options” are not to be included in the “disagreement payoff” contained in a Nash Bargaining Model. See A. Muthoo, Bargaining Theory with Applications at 105 (1999) (“I thus emphasize that the outside option point does not affect the disagreement point. ”); M. Osborne & A. Rubinstein, Bargaining and Markets at 88 (1990) (“it is definitely not appropriate to take as the disagreement point an outside option. . . .”); K. Binmore, A. Rubinstein & A. Wolinsky, The Nash Bargaining Solution in Economic Modeling, 17 RAND J. Econ. 176, 185 (1986) (“An outside option is defined to be the best alternative that a player can command if he withdraws unilaterally from the bargaining process.”).

According to Professor Watt and these authorities, the reason for excluding “outside options” from the Nash Bargaining Model is fundamental to the nature of the model itself. In the Nash approach, the negotiating parties are bargaining with each other only over the surplus their deal can generate, and they are attempting to agree upon an allocation of that surplus that exists within the bounds of their respective “disagreement payoffs.” Each may have “inside options,” which are alternatives available to them while bargaining is ongoing and they temporarily disagree. See Muthoo, supra, at 137. However, “outside options” are available to a Nash bargaining party only in lieu of continuing the Nash bargaining with the original counterparty if it “withdraws” from the Nash bargaining process. See Binmore et al., supra. Professor Watt characterizes the distinction as follows:

[T]he Nash bargaining model [is] designed as [a] self-contained portrayal[ ] of negotiating behavior. . . . Given a surplus to share, the Nash model . . . provide[s] allowance for financial payments that a party is actually receiving, only while negotiations are ongoing, without walking away for another option, and that would cease as a result of the deal, to be factored into modelling as a cost in some situations.”)

. . .

[A]n outside option (a potential payoff that is not directly related to a share of the surplus that is being negotiated) . . . comes in as a constraint upon the set of feasible deals that could be struck. . . .”

Watt RWRT ¶¶ 56, 58.

The Services never sought to introduce further testimony regarding this important dispute. This is particularly striking because the Services filed a motion to strike certain portions of the CO Reply, or for leave to file supplemental testimony responsive to those itemized portions. The portions the Services identified in their motion did not include Professor Watt's criticisms as to the inclusion of “outside options” in their experts' Nash modeling. Further, after the Judges granted the Services' motion by providing them leave to file supplemental testimony—consistent with the designations in their motion—the supplemental testimonies did not address this “outside options” issue.

In the course of discussions among the parties and the Judges regarding remand procedures, the Judges invited the parties to produce witnesses for a hearing, at which one or more of the Services' economic expert witnesses could have addressed this “outside options” issue. However, the Services (and Copyright Owners) waived the opportunity to produce witnesses at a hearing. Rather, they offered, and the Judges agreed, that they would stand on their written testimonies and proceed to closing arguments by counsel.

In the closing arguments, each side argued numerous points of controversy and provided the Judges with dozens of demonstrative aids summarizing record evidence and the parties' arguments, but none of those arguments or demonstrative aids so much as mentioned this “outside options” dispute. Moreover, when the Judges inquired during closing arguments as to whether Services' counsel would be addressing any of the experts' “modeling disputes,” counsel said that they were resting on their papers. 3/8/22 Tr. 86–87 (Closing Argument). Similarly, when the Judges inquired of Copyright Owners' counsel whether he would be addressing the modeling “dust-up” between Professors Watt and Katz, counsel demurred, stating that although he would “love to engage on it but . . . “there would be too many slides. . . .” Id. at 262–64.

Simply put, the Services' economic experts made an assertion regarding the need for Professor Watt to have included “outside options” in his Nash Bargaining Model, but Professor Watt presented authority clearly stating that such inclusions would be improper. Thus, Copyright Owners made a prima facie showing that in a Nash Bargaining Model, the surplus generated by the streaming surpluses acquired by the content providers would be split equally as between the sound recording licensors and musical works licensors, and that, ceteris paribus, an increase in the mechanical rate to provide Copyright Owners more of the surplus (per the Shapley-based results relied on by the Majority) would be essentially offset through a nearly 1:1 reduction in the sound recording rate. In response to Copyright Owners' prima facie case, the Services stood mute in response to the rebuttal argument claiming that their experts misapprehended the Nash modeling distinctions between “inside options” and “outside options.”

Accordingly, the Judges find that the Services' criticisms in this regard are insufficient to rebut Copyright Owners' prima facie showing that Professor Watt's Nash Bargaining Model properly identified and valued the “disagreement payoff.”

b. Rejection of Second Rationale for Including Uncapped TCC Rate Prong

In the Determination, as noted supra, the Majority also justified the adoption of the uncapped TCC rate prong because it had the effect of “import[ing] into the rate structure the protections that record companies have negotiated with services to avoid the undue diminution of revenue through the practice of revenue deferral.” Determination at 36; see also Johnson, 369 F.3d at 372 (“By pegging the mechanical license royalties to an uncapped total content cost prong, the Board sought to ensure that owners of musical works copyrights were neither undercompensated relative to sound recording rightsholders, nor harmed by the interactive streaming services' revenue deferral strategies.. . .”) (emphasis added).

(i) Parties' More Specific Arguments

Copyright Owners likewise argue that the uncapped TCC rate structure should be “adopted to provide protection against revenue deferment and displacement in a revenue-based rate structure.” CO Initial Submission at 38; see also id. at 40 (describing uncapped TCC rate prong as “critical backstop in a revenue-based rate structure.”).

Whereas Copyright Owners echo the Majority, the Services adopt the reasoning of the Dissent. They argue as follows:

[A] rate structure with a capped TCC prong, like the Phonorecords II settlement, achieves the same goal of protecting the Copyright Owners from any potential revenue deferral through a “structure that provides alternate rate prongs and floors, below which the royalty revenue cannot fall,” . . . and does so without allowing Copyright Owners to impermissibly share in the labels' complementary oligopoly power.. . . [T]he streaming industry has twice concluded, after extensive negotiations, that the appropriate way to address any concerns regarding revenue deferral is to have a rate structure that includes a capped TCC prong. Phono I,74 FR 4510; Phono II,78 FR 67938.

Services' Joint Opening Brief at 62 (quoting Dissent,84 FR 1990) (emphasis added).

In their Reply, Copyright Owners argue that the Majority maintained the benefits of price discrimination contained in the prior Phonorecords II framework, but balanced that goal with added protection against Service revenue deferral and displacement. Copyright Owners' Reply Brief on Remand at 49 (“In adopting a rate structure with [an uncapped] TCC for all service offerings, the [Majority] balanced its concerns about fostering price discrimination while also protecting against proven revenue diminution by the Services.”).

The Services, in their Reply, take note that pre-remand, Copyright Owners had strenuously objected to any yoking of the mechanical royalty rate to the sound recording rate, maintaining that, although the Copyright Owners now advocate for an uncapped TCC rate to protect against revenue displacement and diminution:

[I]n their [pre-remand] reply proposed findings, the Copyright Owners had expressed a very different view, arguing that an uncapped TCC prong “does nothing to protect Copyright Owners from the Services' revenue displacement and deferment” [and] Copyright Owners have not even tried to explain away their complete about-face on this issue.

Services' Reply at 43.

(ii) Analysis and Decision Regarding Revenue Diminution or Deferral

The Judges find that the second rationale put forth to support an uncapped TCC rate does not justify the adoption of that rate prong. Several reasons support this finding.

First, there is insufficient evidence to show how the sound recording companies contractually structure their own royalty rates, which would constitute the rate base for an uncapped TCC rate for the mechanical royalty. The sound recording royalty rate, when proffered for use as a mechanical royalty rate base, is analogous to pegging the value of a foreign currency to the U.S. dollar. That is no mere benchmark. The Judges must have the benefit of sufficient record evidence to demonstrate that the pegging (or, to use the D.C. Circuit's word in Johnson, “yoking”) of a statutory rate to an unregulated rate serves the statutory purposes for the rate at issue, here, the mechanical rate.

But Copyright Owners presented virtually no evidence regarding how the sound recording companies structure their interactive service royalties. Indeed, in the hearing, Dr. Eisenach acknowledged that the “relative value of sound recording [to] musical works licenses may depend on a variety of factors,” but he intentionally eschewed unnecessary “assumptions, complexities and uncertainties associated with theoretical debates” as to why the particular market ratios existed. See Determination at 44. Indeed, the Majority found fault with Dr. Eisenach's willful ignoring of these issues, agreeing with the Services' criticism that Dr. Eisenach's “use of sound recording royalties paid by interactive services embeds within his analysis the inefficiently high rates that arise in that unregulated market through the complementary oligopoly structure of the sound recording industry and the Cournot Complements inefficiencies that arise in such a market. See Determination at 47. The uncapped TCC rate advocated now by Copyright Owners suffers from the same affliction.

The only reference to such sound recording rate formulae in Copyright Owners' voluminous PFF after the hearing was its statement that the effective revenue calculations in two of the Major labels' agreements with the services was based on [REDACTED]. See Copyright Owners' PFF ¶¶ 72, 91 (cited post-remand at Copyright Owners' Motion for Reconsideration or Clarification at 25, n.14). On remand, the Services have provided a further summary of the types of [REDACTED]. See White WDRT ¶¶ 6–7, 14–15, 20, 24–26, 28–29 ([REDACTED]); Bonavia WDRT ¶¶ 15–17 ([REDACTED]); Mirchandani WDRT ¶¶ 16, 21–24 ([REDACTED]). Clearly, the levels of [REDACTED] would have to be weighed and the impact of complementary oligopoly power would need to be identified in order to adjust the rate prongs to account for that power. But the record is devoid of such details.

Second, compounding this problem, because the uncapped TCC rate is embedded in a “greater-of” rate structure, the labels can exploit their complementary oligopoly power when creating the switching points that toggle royalty payments between and among rate prongs. As the Judges have explained previously, in declining to import a “greater of” structure from the unregulated interactive market, this structure[it] is based on “agreements [which] 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.” SDARS III,83 FR 65210, 65228 (Dec. 19, 2018) (emphasis added). Further, the Judges held therein that the “advantageous” nature of a “greater-of” structure to sound recording licensors “may well represent an example of what licensors can and would obtain when they exploit their “must have” status for a special competitive advantage.” Id.; see also Dissent at 47 (in absence of testimony explaining how greater-of structure is consonant with effective competition, use by licensor suggests a game of “heads I win tails you lose.).”

Thus, there is insufficient evidence or testimony that would permit the Judges to make any adjustment for the complementary oligopoly power that may be built into each prong of the sound recording royalty rate structures.

Third, as the Services note, Copyright Owners pre-remand, opposed the identical rate structure —consisting of a percent-of-revenue prong and an uncapped TCC prong—before Copyright Owners were in favor of it, post-remand. Although Copyright Owners took a 180-degree turn on this issue, they never stated they were wrong to oppose it previously. Indeed, the Dissent relied upon Copyright Owners' strenuous objection to an uncapped TCC rate, quoting it verbatim:

Copyright Owners rightly note that they obtain no legal protection under such a TCC prong. In making this argument regarding displacement and deferral of revenue, Copyright Owners lay out comprehensively all the problems inherent in an uncapped TCC prong set in a greater of rate structure, such as adopted in the majority opinion:

The notion that [the] TCC prong will provide protection from revenue gaming, deferral and displacement, and other revenue prong problems is unsupported and speculative. Relying on just the TCC to solve those admitted problems leaves the Copyright Owners' protection from such problems entirely outside the statute. . . . the per-user rates in the label deals are what protects the Copyright Owners from price-slashing by the services. What is left unanswered . . .is . . . how can it be reasonable to ask the Judges to set a rate that does not itself provide for a fair return . . . but simply puts the Copyright Owners' fair return in the hands of the labels to negotiate terms that will adequately protect the publishers and songwriters as well? The labels do not have a mandate to ensure that the Services provide a fair return to the Copyright Owners, and cannot be directed to ensure such. Indeed, labels may not have the same incentives as songwriters and publishers to negotiate such protections in their deals. To wit, a label could make an agreement with a service that includes only a revenue prong in exchange for equity or some other consideration that it may never include in the applicable revenue subject to the TCC. . . . [W]hat if Google purchased one or more record labels and did not have to pay any label royalties? Or what if Spotify chose to avail itself of the compulsory license to create its own master recordings embodying musical works—which it is already doing . . . and chose to compensate itself for its use of the master recordings on a sweetheart basis (or not at all)? Or what if one or more labels decided to enter the interactive streaming market and did not have to pay themselves royalties? In each case, the Copyright Owners' protection—the protection that the Services admit the Copyright Owners need and is provided by the TCC—would be gone.

Dissent at 5–6 (quoting Copyright Owners' RPFF-Google at 39–41) (emphasis added). To make the identical point post-remand, but from the Services' perspective, Pandora's economic expert witness, Professor Katz, simply utilizes Copyright Owners' verbatim language (bolded above), but substitutes the word “Services” for “Copyright Owners” (and “income” for “return”) to highlight how reliance on the sound recording royalty rate is improper:

What is left unanswered . . . is . . . how can it be reasonable to ask the Judges to set a rate that does not itself provide for a fair income . . . but simply puts the Services' fair income in the hands of the labels to negotiate terms that will adequately protect the Services as well? The labels do not have a mandate to ensure that the Copyright Owners provide a fair income to the Services, and cannot be directed to ensure such.

Katz WDRT ¶ 71.

The Judges find this argument persuasive, both in its own right and in the fact that it has been advanced by Copyright Owners and the Services alike.

Fourth, the Judges note that the Majority did not find that revenue diminution, via displacement, deferral, or otherwise was pervasive, as Copyright Owners aver. Compare CO Initial Submission at 40 (“The record overwhelmingly established that the percent of revenue prong often results in musical works royalties that are too low . . . drive[n] [by] . . . . revenue deferral [and] revenue displacement”) with Determination at 21 (“The Judges agree that there is no support for any sweeping inference that cross-selling has diminished the revenue base. ”) (emphasis added) and 36 (“The Judges find that the present record indicates that the Services do seek to engage to some extent in revenue deferral in order to promote their long-term growth strategy.”) (emphasis added).

Given that the Majority found revenue diminution through displacement and/or deferral exists only “to some extent” and is not a “sweeping” issue, the Judges on remand find that the uncapped TCC rate structure creates the potential for unbalanced harm. As noted supra, the only protection against runaway mechanical rates, the seesaw hypothesis, cannot justify yoking the mechanical rate to a fixed ratio with the unregulated sound recording rate. By contrast, and as discussed infra, the Phonorecords II -based benchmark approach, despite its own imperfections, is superior in this regard, because its series of alternate rate prongs and floors represents a negotiated compromise (negotiated by trade associations with countervailing power) between the potential for revenue diminution that would harm Copyright Owners, on the one hand, and the potential for runaway mechanical rates (yoked to the sound recording companies' complementary oligopoly power) that would injure the Services, on the other.

(iii) Distinction Between the “Reasonable” Rate Statutory Standard and the Factor (D) Objective To Minimize “Disruptive Impact”

The Judges next consider an issue emphasized by Copyright Owners: whether the Services have demonstrated that the uncapped TCC rate prong would cause a “disruptive impact” as set forth in Factor (D) of section 801(b)(1).

Section 801(b)(1) provides that one of the competing priorities of the Judges in setting the mechanical rate is “[t]o minimize any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.” 17 U.S.C. 801(b)(1)(D). In Johnson, the D.C. Circuit did not identify any argument by the Services that was predicated on a claim that this statutory form of “disruption” had occurred, or was likely to occur, as a consequence of the Majority's rates and rate structure. Additionally, the D.C. Circuit did not ground its decision to vacate and remand the Judges' uncapped TCC rate and rate structure rulings based on the potential that these rulings would be disruptive to the Services, let alone would cause a statutory “disruptive impact.”

After the D.C. Circuit's ruling, an argument regarding “disruption” was first made by Copyright Owners, not the Services. Copyright Owners argued that the vacated rates should nonetheless be maintained as interim rates, during the pendency of the remand proceeding. Motion of Copyright Owners to Adopt Interim Rates and Terms Pending the Remand Determination, passim (Nov. 2, 2020). Copyright Owners argued that reverting to the rates that existed before the Determination would constitute a “disruption” and self-servingly predicted that the Services would attempt to argue that the uncapped TCC rate and rate structure were themselves “disruptive.” Copyright Owners opined that such an argument would be a “hollow exercise.” Id. at 12, n.5; see id. at 2–3, 9 (claiming absence of disruption from uncapped TCC rate and structure despite absence of such argument by Services).

In response to that motion, the Services did not assert that the Majority's uncapped TCC rates and rate structure would constitute disruption or have disruptive impact, whether under statutory Factor D or otherwise. See Services' Opposition to the National Music Publishers' Association (NMPA) and Nashville Songwriters Association International's (NSAI) “Interim Rates Motion” (Nov. 18, 2020). In reply, Copyright Owners shifted from anticipating a “disruption” argument to misinterpreting Johnson, asserting, without citation: “With respect to the TCC prong, the remand directs only that services be given opportunity to offer evidence of disruption from rates that have now been in effect for three years without any disruption.” Copyright Owners' Reply in Support of Motion to Adopt Interim Rates at 7–8 (Nov. 25, 2020) (emphasis added).

On December 10, 2020, the Services submitted to the Judges their Proposal for Remand Proceedings, in which they made no argument that the uncapped TCC rates and rate structure (or, for that matter, any aspect of the Determination) would cause disruption or have a disruptive impact, whether under statutory Factor D or otherwise. By contrast, in their remand proposal, Copyright Owners reference twelve times that, for the Judges to reject the uncapped TCC rates and structure, the Services must show the presence of “disruption” arising from the Majority's uncapped TCC rates and structure. Copyright Owners made this argument notwithstanding that the “reasonable” rate standard is separate from the “disruptive impact” issue, which is an itemized objective (one of four) to be considered as an adjustment to what would otherwise constitute a “reasonable” rate. See Proposal of Copyright Owners for the Conduct and Schedule of the Resolution of the Remand at 2, 7–8, 22–24 (Dec. 10, 2020).

In the CO Initial Submission, Copyright Owners assert, without citation to any of the Services' filings: “The Services contend that, had they been given such an opportunity [at the hearing], they supposedly could have established that an “uncapped” TCC is disruptive because the market for sound recordings is not effectively competitive.” Id. at 5. Copyright Owners further aver that the Services must “provide evidence, consistent with the [CRB Judges'] well-established disruption standard, that because of the labels' supposed market power, the TCC structure adopted by the Board has actually, substantially, immediately and irreversibly threatened the continued viability of the interactive streaming industry” in a manner that will “threaten the viability of the music delivery service currently offered to consumes under [the] license.” Id. at 7, 56 (citations omitted).

Copyright Owners then assert that the Services bear the burden of proving disruption under Factor D from the uncapped rates and rate structure embodied within the rate proposal (even though only Copyright Owners are pursuing this approach on remand). Further, Copyright Owners assert that the Services' objection to the uncapped rates and rate structure must fail unless they can show that such a disruptive impact occurred during the 33-month period (from January 2018 through September 2020) when the Phonorecords III rates were in effect. Id. at 56.

In their initial substantive remand briefing, the Services once more did not assert that the Determination's uncapped TCC rates and structure would cause disruption pursuant to Factor D of section 801(b)(1), or even assert a non-statutory disruption arising therefrom. Rather, the Services directly attacked this rate approach as inconsistent with the statutory “reasonable” rate requirement, maintaining that “[t]ying the mechanical rates directly to the complementary oligopoly sound recording rates in the manner of the Majority's uncapped TCC rates and rate structure is plainly unreasonable. ” Services' Joint Opening Brief at 46 (Apr. 1, 2021) (emphasis added). The Services also asserted that the uncapped TCC rates and rate structure are “unreasonable” because they do not promote the statutory objectives of Factor B (“fair income” to the copyright user) and Factor C (reflecting the copyright users' itemized role in making the musical works “available to the public.”). Id. a t 45, 50–51, 55.

In the Services' Reply, the Services attack Copyright Owners' “singular focus on the disruptive impact of the uncapped TCC prong.” Services' Reply at 35. In particular, the Services argue:

1. they have maintained and demonstrated that Copyright Owners' uncapped rates and rate structure are “ unreasonable, ” separate and apart from demonstrating that this uncapped approach also fails to satisfy the four itemized statutory factors;

2. the burden of proof with regard to Factor D disruption lies with Copyright Owners, because they are the ones who are advocating for the uncapped TCC rates and rate structure;

3. the presence of Factor D disruption, vel non, is not dispositive, because section 801(b)(1) and Johnson require the Judges to apply the entirety of the statutory standard (which consists of the “reasonable rate” requirement and consideration of all four itemized Factors; and

4. the “full extent of the disruption to the Services from an uncapped TCC prong was never tested in the marketplace [because] [t]he Majority set escalating rates, and the [ ] Determination was vacated before the significant hike in rate levels was fully implemented.”

Id. at 35–36.

In their Remand Reply, with regard to the issue of “disruption,” Copyright Owners assert:

1. The Services have “completely abandoned” their appellate argument asserting disruption, and admit to having no evidence that the Board's adopted rate structure has any materially disruptive impact. Copyright Owners' Reply Brief on Remand at 5 (July 2, 2021).

2. The Services have not even attempted to show any Factor D related effect or other disruption from the adopted rates and structure. Id. at 15, n.9.

3. The failure of the Services to provide evidence of disruption or to pursue the argument that disruption had occurred was inconsistent with their prior assertions that the uncapped TCC rates and rate structure created “a real risk of economic harm” and the “impact” or “harm' that the uncapped approach generated. Id. at 35.

4. Each of the Services, in response to Copyright Owners' discovery requests, acknowledges that it was not offering new evidence regarding the “impact” of the Phonorecords III rates and rate structure. Id. at 36–38.

5. The Services did not merely suffer no disruption, they experienced unprecedented growth and profit under the uncapped TCC rate prong. Id. at 45.

6. The Services on remand have attempted to replace their prior “disruption” assertion with a claim of “unreasonableness.” Id. at 50, n.36.

(iv) Analysis and Decision Regarding “Disruption” Issue

The full Factor D “disruption” standard, as set forth by the Judges, states that an adjustment is warranted by Factor D if the rate analysis made by the Judges would otherwise:

directly produce[ ] an adverse impact that is substantial, immediate and in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance 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.

Determination at 87. Factor D is not applicable, particularly as proposed by Copyright Owners. Thus, the Judges reject Copyright Owners' assertion that the uncapped TCC prong should be adopted because of the absence of evidence of “disruptive impact” proffered by the Services. This rejection is based on several findings of fact and conclusions of law.

First, the issue of “disruptive impact” pertains here to the proposal advanced by Copyright Owners, not the Services. Thus, the burden of proving that this uncapped TCC rate prong proposal satisfies the elements, including Factor D, of the section 801(b)(1) standard in a sufficient manner lies with Copyright Owners, not the Services. See5 U.S.C. 556(d). Accordingly, the fact that the Services did not affirmatively assert an argument of “disruptive impact” is of no consequence. Moreover, as the review of the Services' filing makes clear, the Services never abandoned that argument, because they never made it. Rather, they have consistently argued that the uncapped TCC rate prong was unreasonable, not that it was statutorily “disruptive” as that standard has been applied by the Judges.

Second, Copyright Owners did not demonstrate with sufficient evidence or testimony that the uncapped TCC rate would be consistent with Factor D. To be clear, by this the Judges do not mean that Copyright Owners were obliged to prove a negative. Rather, they needed to prove, and indeed attempted to do so, that it was unlikely that their rates would cause a “disruptive impact.”

In this regard, as an empirical matter, Copyright Owners proffered the testimony of an economic expert witness, Dr. Eisenach, who opined that the Services' [REDACTED]. Eisenach WRT ¶¶ 12–41 ([REDACTED]) CO Reply at 40–41. However, as the Judges discuss supra, that period reflected “33 months of uncertainty,” during which no one could predict the final mechanical rate and structure that would be adopted by the Judges and/or the D.C. Circuit after appeals. Accordingly, that factual evidence is unpersuasive.

Further, as a theoretical matter, Copyright Owners rely on Professor Watt's testimony regarding the “seesaw” effect. In that regard, and as discussed supra, the Majority took comfort in what it understood to be Professor Watt's “prediction” that increases in mechanical royalties would be offset almost dollar-for-dollar by reductions in the sound recording royalty. However, as also discussed supra, Professor Watt has now clarified on remand that he never made such a “prediction,” and that his testimony regarding the so-called “seesaw” was limited to shifts in the share of the surplus to Copyright Owners and from sound recording companies as a consequence of an increase in the mechanical rate, holding all other factors unchanged (the ceteris paribus assumption).

Moreover, Professor Watt further explained that many other factors would likely impact the sound recording rate together with an increase in the mechanical rate, including “a measure of oligopolistic interaction, different timings of different rate bargains, and the actual values of outside options.” Watt RWRT ¶ 118. Professor Watt candidly acknowledged that he has not modeled these independent variables, and he further notes that the data may not exist to allow for such modeling. Id. But the inability to model the impact of independent variables does not mean that their potential to cause disruption can be ignored.

In particular, the purpose of the “seesaw” contention was that it prevented economic harm to the Services in connection with a rise in the mechanical rate. Although not of Professor Watt's design, that connection is intentionally built into the Majority's uncapped TCC rate. See Determination at 35 (“Incorporating an uncapped TCC metric into the rate structure permits the Judges to influence that ratio directly.”) But the “measure of oligopolistic interaction” referenced by Professor Watt was the very concern expressed by the Dissent, which cautioned that there was no evidence that the sound recording companies would be compelled to maintain the same industry structure and accept the loss of substantial royalty income. See Dissent at 4 (“[T]he record companies may decide to keep their rates high despite the increase in mechanical rates, or decide it is in their interest to avoid a reduction in royalty revenue by creating a completely different paradigm for streaming, by which the record companies move the streaming service in-house and effectively destroy the existing services.”).

Also, the “different timings of different rate bargains,” another independent variable identified in Professor Watt's remand testimony, was an issue raised to him at the hearing by Judge Strickler. Professor Watt candidly agreed that the Judge was “absolutely correct” that there is a “risk, then, of disrupting the market by having a total royalty that's greater than what is indicated by your Shapley testimony, simply because of the disparity of times in which the rates are . . . implemented.” 3/27/17 Tr. 3091–92 (Watt) (emphasis added). However, this admitted risk of disruption was not addressed by sufficient record evidence.

Third, disruption in the narrow sense of Factor D as applied by the Judges previously is not relevant to the present problem. An increase in total royalties is not a short-run immediate issue, but rather an ever-present possibility that the seesaw analysis does not sufficiently address. Rather, the uncapped nature of the TCC rate prong renders it unreasonable rather than narrowly disruptive.

Balancing the foregoing considerations, the Judges find that Copyright Owners' disruption-based argument lacks merit.

6. Conclusion Regarding Uncapped TCC Rate Prong

For the foregoing reasons, the Judges decline to adopt the uncapped TCC rate tier proposed on remand by Copyright Owners.

III. Rejection of Phonorecords II Settlement as a Benchmark

A. D.C. Circuit Ruling

Each of the Streaming Services advanced somewhat different rate plans, but all four proffered a benchmark that “broadly sought to maintain the Phonorecords II rate structure,” while lowering or eliminating the mechanical floor. Johnson, 969 F.3d at 371. With regard to the Services' proposed benchmark based on the Phonorecords II rates, rate structure, and terms (hereinafter, PR II-based benchmark), the Judges are guided by several rulings in Johnson.

In particular, the D.C. Circuit found the Judges' treatment of the PR II-based benchmark to be “muddled.” Johnson, 969 F.3d at 387. The D.C. Circuit emphasized that the Judges “failed to explain” their rejection of the PR II-based benchmark. Id. at 367. See also id. at 376 (Judges “failed to “reasonably explain” rejection).

In the appeal, Copyright Owners attempted to defend the Judges' reliance on the absence of evidence of the settling parties' subjective intent in reaching the Phonorecords II terms. Id. at 387. The D.C. Circuit dismissed Copyright Owners' post hoc attempt, noting that “nowhere does the [ ] Determination explain why evidence of the parties' subjective intent in negotiating the Phonorecords II settlement is a prerequisite to its adoption as a benchmark.” Id. at 387 (emphasis added).

The D.C. Circuit also criticized the attempt by the Judges' appellate counsel to “change tack” and argue that their rejection of the PR II-based benchmark was reasonable because: (1) evidence showed that the prior rates had been set far “too low” and (2) it was “outdated”. The D.C. Circuit found that those arguments also were “nowhere to be found in the [ ] Determination's discussion” of the appropriateness of the Phonorecords II settlement as a potential benchmark. Id. at 387 (emphasis added). In the end, the D.C. Circuit agreed with the Streaming Services that, inter alia, the Judges failed to reasonably explain their rejection of the benchmark and, for all of the reasons cited, vacated and remanded the adopted rate structure and percentages for further proceedings. Id. at 381.

B. Remand Procedure Regarding the PR II-Based Benchmark

On December 15, 2020, subsequent to the D.C. Circuit's decision, the Judges entered an Order Regarding Proceedings on Remand, in which the Judges stated:

The Judges accept the parties' proposals to resolve the issues concerning the use of the Phonorecords II settlement as a benchmark. . . .

. . .

The Services and Copyright Owners also agree that the Judges should resolve this issue based on the existing record, after receiving two rounds of additional briefing from the parties.

Remand Order at 1–2.

Based on the ruling in Johnson the Judges reject Copyright Owners' position that they need not engage in a full analysis of the issue. The Judges conclude that they must engage in, and fully articulate, a reasoned analysis that adequately addresses “the issues concerning the use of the Phonorecords II settlement as a benchmark.” Id. (emphasis added). If the Judges determine that the Majority properly rejected the Services' proposed use of the PR II-based benchmark, the rejected portions will play no part in the Judges' remand ruling. On the other hand, if the Judges find, after engaging in that analysis, that the PR II-based benchmark was not properly rejected then, as a matter of law and logic, the Judges must weigh the Services' PR II-based benchmark for application, in whole or in part.

The Judges reject Copyright Owners' reading of Johnson as holding that the Judges cannot fully consider the PR II-based benchmark on remand. Copyright Owners argue that the D.C. Circuit “did not suggest the [Judges] substantively erred” in rejecting that benchmark, or that they “needed to reconsider [their] decision,” but had “merely remanded for a `reasoned analysis' . . . as to why it did so.” CO Initial Submission at 10; see also Copyright Owners' Reply Remand Brief at 7–8. Because Johnson ruled that the Majority's reasoning was muddled, indiscernible, unexplained and lacking in reason, the D.C. Circuit obviously neither accepted nor rejected the Majority's disregard for the PR II-based benchmark—thus requiring the CRB Judges to take a comprehensive look at that benchmark. In this regard, the Judges agree with the Services that, pursuant to apposite case law, if the outcome of the remand as to this issue was preordained pending the further “reasoned analysis,” the D.C. Circuit would have expressed a desire simply to remand without vacating as to this issue. Services' Joint Remand Reply Brief at 7–8 (citing Allied-Signal, Inc. v. NRC, 988 F.2d 146, 150–51 (D.C. Cir. 1993) (“The decision whether to vacate depends on the seriousness of the order's deficiencies (and thus the extent of doubt whether the agency chose correctly) and the disruptive consequences of an interim change that may itself be changed.”)).

Because Johnson held that the Majority's reasoning was muddled, indiscernible, unexplained, and lacking in reason, the D.C. Circuit obviously neither accepted nor rejected the Majority's disregard for the PR II-based benchmark. Thus, the Judges take a comprehensive look at that benchmark's rates and rate structure to evaluate its usefulness in this proceeding.

Relatedly, the Judges also reject Copyright Owners' assertion that the Judges can only consider on remand the Phonorecords II rates, and cannot consider on remand the relative strengths and weaknesses of the structure in which those rates are embedded. See Copyright Owners' Reply Brief on Remand at 14. This distinction is impractical and unworkable. If the (non-“headline” rates ) themselves can be reviewed and found acceptable (as they are infra) into what structure would they be placed? There are multiple provisions in the Phonorecords II rate structure providing for different rates, designed to balance (1) the ability of services to attract consumers with a low Willingness-to-Pay and/or a low Ability-to-Pay (the price discriminatory and differentiated features ) with (2) the revenue diminution protections for which Copyright Owners had successfully negotiated. Moreover, the D.C. Circuit has vacated the Determination, and in doing so did not make any rulings critical of the rate structure in the Phonorecords II -based benchmark that would suggest the cramped review advocated by Copyright Owners. Indeed, the D.C. Circuit explicitly stated, without distinguishing between rates and structure, that it “agree[s] with the Streaming Services that the [Judges] . . . failed to reasonably explain [their] rejection of the Phonorecords II settlement as a benchmark . . .” See Johnson, 969 F.3d at 376; see also id. at 389 (issues relating to “rates” and “rate structure” are “intertwined”).

Further, the Judges emphasize that the rate structure of the PR II-based benchmark provides protection sought by Copyright Owners against revenue diminution by the Services— protection they would otherwise lose— because in this Initial Ruling the Judges are not adopting the vacated uncapped TCC prong for which Copyright Owners are now advocating, and which they claim would have protected them in that regard. Cf. CO Additional Submission at 4–6 (acknowledging PR II-based benchmark provided some TCC provisions, allowing for protection against revenue diminution). Thus, the Judges' remand rulings on the PR II-based benchmark rates and on the uncapped TCC rate prong are inextricably interlaced. See Johnson, 969 F.3d at 381 (absence of “reasoned explanation” for rejecting PR II-based benchmark was problematic because it occurred “ when ” Judges adopted an alternative proposal that called for “setting . . . total content cost and revenue rates.”) (emphasis added).

The Judges weigh each benchmark's intrinsic strengths and weaknesses, as well as its comparative advantages and disadvantages vis-à-vis other proffered benchmarks. On remand, the interrelationships of the competing benchmarks are of particular importance, given Copyright Owners' need for the aforementioned protections against revenue diminution via price discrimination.

Through this approach, the Judges ultimately may adopt only one of the parties' benchmarks or other methodologies, or they may modify the proposals by combining them, provided such a modification is “within a reasonable range of contemplated outcomes . . . piecing together a rate structure, the economic and policy consequences of which had already been explored and developed by the parties in the record.” Johnson, 369 F.3d at 382.

In their consideration of the PR II-based benchmark, the Judges are not suggesting that this benchmark is the optimal tool to use in order to identify rates and terms among all approaches that might have been proffered (but were not). But the Judges are cabined by the evidence they receive. See17 U.S.C. 803(a)(1) (“the Judges shall act . . . on the basis of a written record . . . .”); see also P. Wald, supra, (noting that parties' economic proposals made in an action “impose[ ] a practical constraint” on judge who will, “for the most part, be limited by what the parties serve up to her.”). Based upon the available record evidence, the Judges find that the Services' PR II-based benchmark—although not necessarily perfect—is more than sufficient to satisfy the legal requisites for application, as well as a practical benchmark, when used in conjunction with the 15.1% headline revenue rate advocated by Copyright Owners. See generally Nat'l Cable Television Ass'n v. Copyright Royalty Tribunal, 724 F.2d 176, 182 (D.C. Cir. 1983) (rate-setting is an intensely practical affair).

C. Parties' Remand Arguments Regarding PR II-Based Benchmark

1. Services' Arguments

The Services maintain that their PR II-based benchmark satisfies the “reasonable” rate requirement and is consistent with the four itemized factors set forth in section 801(b)(1). They make several arguments in favor of this position.

First, they aver that their PR II-based benchmark possesses all the characteristics of an “ideal” benchmark. Services' Joint Opening Brief at 19. In this regard, they argue that their proffered benchmark “involves the same sellers, the same or similar buyers, and the same rights as at issue in this proceeding,” and that there has been “no material change in the economic circumstances of the marketplace that would warrant adjusting the rate levels or rate structure in the benchmark.” Id. at 20.

Applying the facts to these benchmark characteristics, the Services assert that the first three elements—same sellers (here, licensors), same buyers (here, licensees) and same rights (the mechanical license for interactive streaming) are satisfied. In particular, they note that the majority of the participants in the present proceeding either directly participated in the Phonorecords II settlement process or were active in the market contemporaneous with that settlement. Id. at 20–21.

Turning to the next benchmark characteristic—the absence of a “material change in the economic circumstances of the marketplace that would warrant adjusting the rate levels or rate structure in the benchmark”—they emphasize that the PR II-based benchmark contains different rate levels for different product offerings, to account for (a) consumers' varying willingness-to-pay (WTP) and (b) the zero marginal physical cost of digital reproductions of sound recordings containing musical works. Id. at 21–22 (citing multiple experts).

Next, the Services point to the fact that the “headline” royalty rate is based on a percent-of-revenue, so that revenue growth (or decline) on this rate prong allows for royalty payments to directly adjust in tandem. Id. Further, the Services assert that the importance of streaming as “the future of the music industry” was known to the Phonorecords II negotiators, as evidence by the then-recent launch in the United States of the popular Spotify service. Id. at 23.

Beyond these benchmark requisites, the Services also emphasize that the PR II-based benchmark is the product of a settlement whose negotiated features burnish the value of this benchmark as reflective of effective competition. Specifically, they note:

• The settlement was negotiated in the same statutory context, concerning the identical rate standard and factors as applicable to the present proceeding.
• Neither side would have accepted a deal materially worse than what it expected from a section 115 proceeding applying the section 801(b)(1) considerations.
• The statutory alternative diminishes any additional licensor-side negotiating power arising from “Must-Have” complementary oligopoly of the licensors of the musical works publishers.

Id. at 22. Moving from the negotiating context to market performance under this standard, the Services aver that this approach has borne fruit for the industry as a whole. They point to the evidence of the licensors' consistent profitability and the licensees' ability to “benefit” from the Phonorecords II approach. Id. at 23.

The Services also maintain that the Phonorecords II structure “addresses any concerns with bundling and the potential for revenue deferment.” Id. at 24. They assert that these issues were specifically addressed by Copyright Owners during the Phonorecords II negotiation, because “multiproduct firms such as Yahoo and Microsoft” that offered streaming services had the capacity to make bundled offerings to consumers. These concerns were addressed in the Phonorecords II rate structure, the Services note, through the use of “multiple rate prongs, minima and floors,” ensuring that “the total musical works royalty for certain types of offerings does not fall below a specified level,” thereby “mitigat[ing] the effect of any potential revenue deferrals and appropriately address[ing] any concerns with bundling.” Id.

Finally, the Services maintain that “[d]irect agreements between Copyright Owners and Services also support adoption of the PR II-based benchmark.” Id. at 34. In particular, they note that many of the royalty rates (and terms) in these direct agreements apply the Phonorecords II rates. Moreover, the Services maintain, because these direct agreements are in the nature of blanket license of a publisher's entire catalog, they provide an added “access” value in the form of full-repertoire licensing. These direct agreements do not include a rate above Phonorecords II levels; thus, the Services contend, they underscore the reasonableness of the Phonorecords II rates. Id.

Finally, the Services aver that the PR II-based benchmark satisfies the itemized four section 801(b)(1) factors. With regard to Factor A, they maintain that: (1) the Phonorecords II framework has corresponded with an increase in the supply of musical works; (2) the PR II-based benchmark will increase the likelihood that the Services will increase subscriber counts, generating profitability, which will make streaming available to more listeners; and (3) the price discriminatory aspects of this royalty rate structure allows the Services to afford to offer streamed music to listeners with a low willingness (or ability)-to-pay, at lower rates or through ad-supported services. Services' Joint Opening Brief at 25–27.

Regarding Factors B and C (the “fair return” and “relative contributions” objectives), the Services emphasize that the PR II-based benchmark satisfies these statutory elements because it: (1) was the result of negotiations between industrywide representatives who had every incentive to obtain a “fair” return and to receive recompense for their “contributions” to streaming; and (2) allowed interactive streaming to become “a significant means for consumers to listen to music” while simultaneously generating growth in annual royalties for Copyright Owners.” Id. at 27–29.

Lastly on the subject of the statutory factors—regarding Factor D (minimizing disruptive impact)—the Services make a succinct argument: “By renewing the rate levels and structure of Phonorecords II, there is minimal risk of disruption.” Id. at 29–30.

The Services also address several further criticisms of the PR II-based benchmark contained in the Determination. Focusing first on an issue specifically addressed in Johnson, they assert the irrelevancy of the “subjective intent” of the parties that negotiated the Phonorecords II settlement—a factor on which the Majority relied in deciding not to adopt the PR II-based benchmark. In this regard, the Services are also responding to the D.C. Circuit's concern regarding this issue. See Johnson, 969 F.3d at 387 (“In rejecting that settlement as a possible benchmark, the [Judges] faulted the Streaming Services for failing to explain why the parties to the Phonorecords II settlement agreed to the rates in that settlement . . . [b]ut nowhere does the [ ] Determination explain why evidence of the parties' subjective intent in negotiating the Phonorecords II settlement is a prerequisite to its adoption as a benchmark.”).

The Services note that no benchmark evidence presented by any party is proffered with supporting evidence of the subjective intent of the bargainers who negotiated the benchmark. Moreover, they note that the Majority in fact acknowledged that “[r]elying on a benchmark as objectively useful without [the need for] further inspection” is “typical and appropriate for the benchmarking method.” Id. at 35 (quoting Determination at [55] & n.106 (emphasis added)).

With regard to other criticisms of the Majority's failure to use the PR II-based benchmark, the Services argue that the Majority misapplied their previous rulings that they “cannot and will not set rates to protect any particular streaming service business model.” Id. at 37 (quoting Phonorecords III,84 FR 1945). The Services find this principle inapposite, because their point is that the multiple price-discriminatory aspects of the Phonorecords II approach made it “a valuable benchmark . . . because it had allowed for different service types to emerge and grow, which benefits the entire market.” Id. at 37. The Services also take issue with the Majority's assertion that the Phonorecords II rate structure was too complex, deriding it as a “Rube-Goldberg-esque” contraption. Id. at 38. Rather, the Services maintain that the structure was as complex as necessary to effectuate the parties' needs, particularly the price discriminatory features and the protections against revenue diminution. Id. at 38–39. Further, the Services note that the record is devoid of any testimony or evidence indicating any actual confusion caused by the Phonorecords II rate structure. Id. at 39. Finally in this regard, the Services maintain that the rate structure adopted by the Majority is essentially as complex as the structure in Phonorecords II, with the only major change being the replacement of the capped TCC rates with uncapped TCC rates. ¹⁰⁶ Id.

The Services address another criticism—that the rates in the PR II-based benchmark are too low. This issue is largely moot, as the D.C. Circuit's affirmance of the Majority's expert “line-drawing” and “reasoned weighing of the evidence” confirmed that a rate increase was necessary. In this Initial Ruling, the Judges have acknowledged specifically the appropriateness of the 15.1% revenue rate—a 44% increase over the 10.5% headline rate in the PR II-based benchmark.

2. Copyright Owners' Arguments

Copyright Owners assert that the record evidence overwhelmingly supports the Judges' rejection of the PR II-based benchmark. At the outset, they maintain that the Judges found—and the D.C. Circuit affirmed—that a rate increase was required in the Phonorecords III terms. CO Initial Submission at 13. (As noted, an increase in the headline rate by 44%, to 15.1%, is adopted in this Initial Ruling.)

Next, Copyright Owners maintain that the evidence established that “market conditions” were “radically different” at the time of the Phonorecords III proceeding compared with when the parties entered into their 2012 industrywide agreement in Phonorecords II. Id. at 17. In particular, Copyright Owners point to testimony describing the streaming industry as “nascent” in 2012, with fewer streams, subscribers, services, and choices of music; operating in a consumer environment when download purchases and Pandora's noninteractive service were the predominant means for consumers to listen digitally to music. Id. at 18–21. In sum, Copyright Owners maintain, that streaming was “economically insignificant” to the music industry when the PR II provisions were adopted. Id. at 20.

Copyright Owners particularly emphasize the substantial increase in streaming revenue during the Phonorecords II period. They point out that while “total streaming revenue had ranged from approximately $150 million in 2005 to $212 million in 2010, . . . after 2012[,] annual [streaming] revenue exploded to reach approximately $1.6 billion by 2015.” Id. at 23. Further, they note there is no evidence that the music publishers or anyone else had predicted this substantial rise in streaming and the revenues it generated, and that in no way could it be inferred that those rates had “baked-in” future growth. In fact, Copyright Owners assert at the hearing that the PR II rates were merely “experimental”—consistent with the relatively nascent stage of the streaming industry. Id. at 25.

Additionally, Copyright Owners maintain that the identities of the parties involved in the Phonorecords III proceeding are different from those who established the Phonorecords II framework. Although they acknowledge the presence of current interactive services Spotify and Rhapsody in this market prior to the Phonorecords II framework agreed to by the trade associations for the interactive services and the music publishers, they point out that “[n]one of the other participants in this proceeding even entered the streaming business until after the Phonorecords II settlement.” Id. at 21.

Next, Copyright Owners assert that the Services' evidence is inadequate to support a finding that the rates in their PR II-based benchmark are suitable for use in setting royalty rates in this proceeding. First, they echo the Determination, which stated that the Services (1) did not examine in detail the particular rates within the existing rate structure; (2) relied on the 2012 rates as objectively useful without further inspection; and (3) did not call witnesses to testify regarding the 2012 settlement negotiations. Id. at 27 (citing Determination, 84 FR 1944 & n.106). Because of the absence of the foregoing evidence, Copyright Owners assert that the Services were left with “no evidence explaining how the particular rates and percentages in those settlements were calculated or derived, how they were negotiated, or how they were reasonable in light of the explosive growth in the streaming marketplace between the time of those settlements and the Phonorecords III proceeding.” Id. at 28. The absence of such evidence, according to Copyright Owners, meant that the Services had failed to carry their burden of proof under 5 U.S.C. 556(d) with respect to their proposal, a burden Copyright Owners assert the Services acknowledged they bore. Id. at 29–30.

Additionally, Copyright Owners claim that the D.C. Circuit found “validity” in Copyright Owners' assertion that the subjective intent of the parties to the Phonorecords II settlement is relevant because it would have revealed whether the agreed-upon rates were based on economic realities or instead were driven by other considerations. Id. at 30–31 (citing Johnson, 969 F.3d at 387). However, Copyright Owners acknowledge that, because this was not a reason given by the Majority, it carried no weight with the D.C. Circuit on appeal. Id. at 31.

3. Analysis and Decision Regarding PR II-Based Benchmark

a. PR II-Based Benchmark Meets Most of the Requisites for a Useful Benchmark

The four classic characteristics of an appropriate benchmark are:

(1) the degree of comparability of the negotiating parties to the parties contending in the rate proceeding,

(2) the comparability of the rights in question,

(3) the similarity of the economic circumstances affecting the earlier negotiators and the current litigants, and

(4) the degree to which the assertedly analogous market under examination reflects an adequate degree of competition to justify reliance on agreements that it has spawned.

In re Pandora Media, 6 F.Supp.3d 317, 354 (S.D.N.Y. 2014, aff'd sub nom Pandora Media Inc. v. ASCAP, 785 F.3d 73 (2d. Cir. 2015). As discussed below, the PR II-based benchmark meets criteria (1), (2) and (4), but requires adjustment to fully satisfy criterion (3).

First, the PR II-based benchmark obviously pertains to the same rights at issue in this proceeding, as it reflects the licensing provisions from the immediately preceding mechanical license proceeding.

Second, the licensors (songwriters and music publishers) and licensees (interactive streaming services) are comparable (albeit not identical). While Copyright Owners emphasize the different identities and market involvement of the licensees, particularly the greater market penetration of Amazon, Apple, and Google, the Services note that even prior to the more significant entry of these three entities, similar multiproduct firms, such as Yahoo and Microsoft, were active licensees. The Judges find that the changing identities of the large multiproduct technology firms does not demonstrate the absence of comparability between and among such firms in the Phonorecords II and Phonorecords III rate periods. The shifting market entries, exits, strategies, successes and setbacks of otherwise comparable firms are expected occurrences in a dynamic capitalist market system and are not factors that materially diminish the necessary comparability of the parties for benchmarking purposes.

Third, important economic fundamentals of the marketplace are sufficiently similar in crucial respects. First, the heterogeneity of the willingness-to-pay among subscribers and listeners in the downstream market continues to support price discrimination and thus differentiated royalty rates upstream pursuant to the concept of “derived demand.” See Determination at 19 (and record citations therein) (“Weighing all the evidence and based on the reasoning in this Determination, the Judges conclude that a flexible, revenue-based rate structure is the most efficient means of facilitating beneficial price discrimination in the downstream market.”); Dissent at 32, 51, 86, 121, 126 (and record citations therein). Second, the items being licensed for transmission—“second copies” of sound recordings (with embedded musical works)—have a marginal physical cost of zero, a critical economic point on which the experts for both parties concur, and as to which the Majority and the Dissent repeatedly and significantly rely. See Determination at 18, 21, 36, 59, 80 (and record citations therein); Dissent at 30–31, 33–34, 37, 47, 49–50, 59, 122, 127–128 (and record citations therein).

Copyright Owners are clearly correct, however, in noting a substantial change in economic circumstances that distinguished the Phonorecords II negotiations from the current proceeding; viz., the dramatic growth of interactive streaming revenues. The economic impact of this revenue growth is incorporated into the experts' Shapley Value Models and the Judges' analysis of same. This analysis has generated the 44% increase in the headline royalty rate, from 10.5% to 15.1% (as phased-in by the Majority and again in this Initial Ruling).

Simply put, three economic principles co-exist. First, the downstream interactive streaming market remains differentiated among listeners with different willingnesses and abilities to pay, based on varied preferences (utility) and disparities in income. Second, streaming of the “second copy” of the sound recordings (with embedded musical works) remains physically costless (but generates potential “opportunity costs”). But, third, streaming revenues have grown substantially. There is no incompatibility or inconsistency in the simultaneity of these economic principles. Each of them must be taken into account and they are in this Initial Ruling.

This economic context refutes the arguments made during oral argument at the D.C. Circuit that the PR II-based benchmark should be rejected in toto because it was supposedly “outdated.” The heterogeneity of the downstream demand of listeners and the zero physical cost of “second copies” are enduring features that affect the upstream market via the principle of derived demand. The substantial growth of streaming revenues, however, necessitated an increase in the headline rate from 10.5% to 15.1% (as phased-in), for the reasons discussed in the Judges' analysis in this Initial Ruling of the interrelationship among: (1) Shapley Value modeling; (2) Nash Bargaining; (3) complementary oligopoly power; and (4) effective competition.

Further, the foregoing analysis also undermines the pre-remand argument made by Copyright Owners that the PR II-based benchmark reflects a market that was not yet “mature,” or was only “experimental.” Markets are not “mature” as opposed to, say, “adolescent.” Indeed, the metaphor is strained because all economic models are subject to revision if the salient facts have changed, without rendering the prior models mere “experiments.” Markets simultaneously exhibit enduring characteristics—here, heterogeneous customers and zero marginal physical costs and dynamic change—here, significant revenue increases.

And yet, Copyright Owners seek to deny the idea that these principles could exist simultaneously. In an attempt to disqualify the application of the PR II-based benchmark, Copyright Owners complain:

[W]hile streaming activity and revenues grew under the Phonorecords II royalty rates, the [REDACTED]. For example . . . [REDACTED].

CO Initial Submission at 15–16 (emphasis added).

But as the Services explained, the economic defect in Copyright Owners' analysis, is that it ignores the principle of price discrimination and its beneficial effects:

[A]s [Professor] Hubbard explained, it is “meaningless” to compare growth in streams to growth in royalties in the context of Prime Music in particular because the record showed that Prime Music brings “new people into the market.” . . . If not for the flexibility (and beneficial price discrimination) the existing Service Provider Revenue definition and rate structure facilitated, the Copyright Owners “would have gotten zero” from those new listeners. . . . “So they're better off by that amount” of royalty growth. . . . The undisputed fact that [REDACTED]—reflects that the existing rule enables beneficial price discrimination that expands the total royalty pool and benefits Copyright Owners.

Services' Reply at 58–59.

This rebuttal by Professor Hubbard is an example of the important distinction between “increases in demand” (when the demand curve shifts outward) and movements “down the demand curve” (when sellers use price discrimination to generate more revenue without additional cost to attract buyers with a lower willingness or ability to pay). The parties' otherwise dueling economists agreed on this point. Compare 4/3/17 Tr. 4373–74 (Rysman) (Copyright Owners' witness acknowledging that under the current rate regime overall revenues might be increasing because of movements “down the demand curve” ( i.e., changes in quantity demanded in response to lower prices), rather than because of, or in addition to, an outward shift of the demand curve ( i.e., increase in demand at every price)) with 3/13/17 Tr. 701 (Katz) (the Services' witness who likewise noted that the present structure enhances variable pricing that allows streaming services “to work[ ][their]way down the demand curve.”).

Moreover, Copyright Owners baldly cherry-pick the data they present. [REDACTED] CO Initial Submission at 15–16. So, by their own data, presented in their own brief, they acknowledge that [REDACTED]. See Services' Reply at 57–58 (Copyright Owners have proven the “opposite” of what they intended). This is precisely what beneficial price discrimination is designed to accomplish.

The appropriateness of adopting the price discriminatory rate provisions of the PR II-based benchmark is further underscored by Copyright Owners' candid acknowledgement at the hearing that they were essentially urging the Judges to adopt what is known as the “Bargaining Room” approach to rate setting. See Dissent at 24 (and record citations therein).

In the present proceeding, the appropriateness, vel non, of the Bargaining Room approach boils down to the following:

Copyright Owners emphasize the inability of the Judges (or anyone) to identify present market rates precisely, let alone over the five-year rate period because the compulsory license set by the Judges cannot possibly contemplate every single business model that may develop in the ensuing time. . . . If the statutory rate is set below market rates, then the parties will never negotiate upward toward the market rates, because the licensees will always prefer to invoke the right to use the licensed work at the below-market statutory rates. However, if the Judges set the statutory rate above what they find to be market rates, different licensees who each have a maximum willingness to pay (WTP) below such a statutory rate would seek to negotiate lower rates with the licensors. In response to such requests to negotiate, according to this argument, Copyright Owners would respond by negotiating various lower rates for those licensees, provided lower rates were also in the self-interest of Copyright Owners.

Dissent at 24–25 (and record citations therein).

The Judges find no reason to depart from the policy decision in Phonorecords I that the rate setting policies made explicit in section 801(b)(1) are best discharged if the Judges eschew the Bargaining Room approach and continue to identify rate structures and rates that reflect the standards set forth in the statutory provision. To supplant the statutory factors with a Bargaining Room approach would essentially be to adopt a purely market-based rate-setting approach that is inconsistent with section 801(b)(1) and with the Judges' application of that statute to set rates, rate structures, and terms consonant with effective competition.

With this background in mind, the Judges turn specifically to the interrelationship between the price discrimination aspects of the rates in the PR–II benchmark and the Bargaining Room approach.

Copyright Owners have demonstrated (albeit tacitly) their understanding that, if the statutory provisions did not contain a price discriminatory rate structure to reflect the varying WTP, they would have to invent it. This finding is apparent from their advocacy for the adoption of a Bargaining Room approach to rate-setting. See, e.g., 4/3/17 Tr. 4390, 4431 (Rysman) (lauding bargaining room approach as reflecting “economical element of price discrimination. . . the [licensor] is picking its prices carefully.”) (emphasis added); id. at 4431 (explaining that under this approach, when negotiating with Spotify regarding a rate for ad-supported service, “Must Have” music publishers would “have the right . . . to set that price.”); 4/4/17 Tr. 483–45 (Eisenach) (acknowledging Copyright Owners' approach was consistent with Bargaining Room theory because they were seeking rates so high as to force would-be licensees to negotiate for the “Must Have” mechanical license.).

Thus, the Judges find there to be no real dispute as to whether there is a market-based need for an upstream discriminatory rate structure. Rather, the parties appear to be in disagreement as to who shall be in control of the setting of rates, the Judges, through their application of law, or Copyright Owners, through the exercise of their complementary oligopoly power. The resolution of this choice is clear; the Judges, not the licensors, are statutorily-charged with establishing provisions that are reasonable and otherwise properly reflect the itemized objectives of section 801(b)(1).

Fourth, the PR II-based benchmark reflect a rate structure with an adequate degree of competition, because there was a balance of bargaining power between the two negotiating industrywide trade associations, offsetting the complementary oligopoly effects in place when a “Must Have” licensor bargains separately with each licensee. Recently, the Judges discussed in detail how the presence of countervailing bargaining power generates royalty rates at effectively competitive levels. See Web V,86 FR 59452, 59457 (Oct. 27, 2021).

Further with regard to this fourth point, the parties have been operating over the past ten years under this basic rate structure, with profits accruing to the licensors and admittedly tolerable losses befalling the licensees. Moreover, after experience with these rates and this rate structure in the Phonorecords I period, they renewed and expanded this structure for use in the Phonorecords II period, when the alternative of a statutory rate proceeding was available to licensors and licensee alike. Their mutual willingness to continue in this manner is important evidence of the workability and reasonableness of this approach.

b. Evidence of Subjective Intent Not Prerequisite to Partial Adoption of the PR II-Based Benchmark

The Judges rely on the PR II-based benchmark as an objective benchmark. Thus, the absence of testimony regarding what went through the minds of the negotiators of the Phonorecords II agreement (and the predecessor Phonorecords I agreement) does not diminish the objective value of this benchmark. The Judges view the provisions of the PR II-based benchmark as they would any benchmark, in the context of the requisite benchmarking elements identified and discussed supra. This approach allows the factfinder to analyze the benchmark through the lens of its service in the marketplace as an objective model for the market at issue, the Phonorecords III market. See, e.g., 3/13/17 Tr. 550–51, 566 (Katz) (knowledge of why parties negotiated specific provisions is unnecessary, because objective results demonstrate satisfactory performances of market).

Both Professors Katz and Hubbard noted that the current rate structure remains useful, not based on consideration of the parties' subjective understandings at the time of its creation, but because the market has not since changed in a manner that would create a basis for departure. Katz WDT ¶ 80 (“My analysis has identified no changes in industry conditions since then [2012] that would require changing the fundamental structure of the percentage-of-revenue prong.”); 4/13/17 Tr. 5977–78 (Hubbard) (changes in market are “not uncorrelated with the structure that was in place” in 2012).

In this regard, it bears emphasis that Copyright Owners' own witness, Dr. Eisenach, relied on several potential approaches that the Majority characterized as benchmarks for his rate analysis, without attempting to examine the subjective intent of the parties who negotiated those agreements. Indeed, the Majority found that the PR II Rates were properly considered as an objective benchmark, in the same manner as Dr. Eisenach's proffered benchmarks:

The Services do not examine in detail the particular rates within the existing rate structure. Rather, they treat the rates within that structure as benchmarks, i.e., generally indicative of a sufficiently analogous market that has “baked-in” relevant economic considerations in arriving at an agreement. Dr. Eisenach did not analyze why he chose the levels for the rates and ratios on which he relied as benchmarks or consider the subjective understandings of the parties who negotiated his benchmarks. Similarly, the Services' economists elected to rely on the 2012 rates as objectively useful without further inspection.

This point is not made to be critical of Dr. Eisenach's approach, but rather to show that the Services' reliance on the 2012 settlement as a benchmark shares this similar analytical characteristic, typical and appropriate for the benchmarking method. (The factual wrinkle here is that, hypothetically, the Services could have called witnesses and presented testimony regarding the negotiations that led to the 2012 (and 2008) settlements, but did not, rendering the 2012 benchmark similar to other benchmarks taken from other markets.)

Determination at 55 & n.106.

Copyright Owners also aver that they entered into the Phonorecords II settlement simply to avoid litigation costs. Copyright Owners' Reply Brief on Remand at 29. At the hearing, this assertion was presented by David Israelite, NMPA's President. Israelite WRT ¶ 28; 3/29/17 Tr. 3649–52 (Israelite) (claiming NMPA lacked financial position to fund rate litigation). The Services countered by noting that there was no evidence to support Mr. Israelite's testimony in this regard, or how it may have impacted the NMPA decision to participate. And, the Services pointed out, notwithstanding his testimony regarding financial constraints, NMPA had incurred the expense of a year-long negotiation with the Services to seek higher rates, create new service categories in subpart C, and change the TCC calculations. Id. at 159, 161–64; 3/29/17 Tr. 3856 (Israelite).

Further, as a general principle, a party's mere assertion that the Phonorecords II approach was the product of a settlement that was predicated on the avoidance of litigation costs savings does not invalidate its use as a benchmark in proceedings before the Judges, especially because, by statute, the Judges are authorized to consider such agreements. See Music Choice v. Copyright Royalty Board, 774 F.3d 1000, 1014–15 (D.C. Cir. 2014) (testimony alleging agreement was reached to avoid litigation costs does not invalidate evidentiary use of that agreement for rate setting purposes, absent other evidence demonstrating settlement was involuntary or otherwise unreasonable.). Thus, the Judges find that the evidentiary record does not support Copyright Owners' position that this “litigation cost avoidance” assertion constituted a separate, idiosyncratic value that diminishes the Judges' partial reliance on the PR II Rates in this Initial Ruling.

Copyright Owners also mistakenly rely on the fact that the Services bore the burden of proof regarding the absence of any subjective idiosyncratic factors that hypothetically could have diminished the useful value of the PR II-based benchmark. Id. at n.21. The Services indeed bore the burden of proof ( i.e., persuasion) with regard to their proffered benchmark PR II Rates, and they presented adequate objective evidence and testimony that this approach has worked in the marketplace to serve as prima facie proof to support the Judges'(partial) use of this benchmark in this remand proceeding. And, as explained above, such subjective intent was not a necessary element of their benchmark proofs. But, with regard to Copyright Owners' rebuttal to those proofs, Copyright Owners bore the burden of production, to present sufficient evidence and/or testimony that the Judges could rely on to reject the (partial) use of the PR II-based benchmark. This Copyright Owners failed to do.

In fact, given Copyright Owners' reliance on the subjective intent of the parties to a benchmark, the Judges attempted to identify potential subjective evidence of how the capped TCC rates in the PR II-based benchmark were derived, during the examination of Dr. Eisenach at the hearing:

[JUDGE STRICKLER] Do you discuss, Dr. Eisenach, . . . in your written direct or written rebuttal testimony how the parties arrived . . . at the ratios for sound recording to musical works in [witness interrupts]

[DR. EISENACH] That process is opaque to me, Your Honor.

[JUDGE STRICKLER] Did you [witness interrupts]

[DR. EISENACH] I know—I know there was a 2008 negotiation. I know there was a 2012 negotiation. I wasn't . . . present, and I'm not privy to any of the details.

[JUDGE STRICKLER] You were not informed by your client or by any other source of information as to how they arrived at those particular ratios?

[DR. EISENACH] When I've asked the question, I've found people chuckle and—and there doesn't seem to have been too much system—systematic thought that went into it, but I don't really know that. I just—when I ask the question, people say: Nobody really knows.. . . Someone may know, but that's what I've been told.

4/4/17 Tr. 4611 (Eisenach) (emphasis added). The Judges find it perplexing, to say the least, that Copyright Owners would “chuckle” when asked by their expert witness for the very subjective evidence which they claim to be relevant. But of perhaps greater relevance is Dr. Eisenach's further testimony, quoted above, that he was also told by Copyright Owners that “nobody really knows” how the parties arrived at those rate ratios. Copyright Owners' “chuckle,” in response to its expert's critical inquiry as to the derivation of rates—and that expert's understanding that his client simply did not know how those rates were derived— undercut Copyright Owners' claim that subjective understanding of those rates could undermine their usefulness in the benchmark.

c. Substantial Evidence Demonstrates That PR II Rates, Other Than the Headline Rate, Are Not “Too Low”

As noted supra, one reason the D.C. Circuit vacated and remanded the Determination was because it declined to entertain the argument made only by appellee's counsel that “the prior rates had been set far too low, thus negating the usefulness of the prior settlement as a benchmark.” Johnson, 969 F.3d at 387. The Judges have noted throughout this Initial Ruling their adoption of the Shapley Value modeling analysis undertaken by the Majority, and raised the headline royalty rate by 44% from 10.5% to 15.1% (as phased-in), rendering moot appellate counsel's suggestion regarding the rate level.

Here, the Judges further consider whether other rates within the PR II-based benchmark are reasonable, not only because they are part and parcel of the workable structure of that benchmark, but also to determine if they are supported by record evidence. To put this issue in context, those rates would apply on the second prong of the “greater-of” rate structure in the PR II-based benchmark. The first prong in the PR II-based benchmark rates is the 10.5% revenue rate—increased to 15.1% (as phased-in) by this Initial Ruling. The second prong consists of the “lesser of” a TCC rate or a per subscriber rate. For certain delivery configurations, these rates also cannot fall below any applicable Mechanical Floor. See Johnson, 969 F.3d at 370.

The Services describe the key feature of these non-headline rates as the fostering of beneficial price discrimination, i.e., the adoption of “different rate levels for different product offering,” in order [t]to account for consumers' different willingness to pay [WTP] for music. Services' Joint Opening Brief (on Remand) at 21. As an example of how these price discriminatory rates impacted the market, the Services compare and contrast two Amazon offerings, Amazon Music Unlimited (for Echo) and Amazon Prime Music.

Amazon Music Unlimited, with more than 30 million available songs as of the Phonorecords II proceeding period, see Mirchandani WDT ¶ 41, [REDACTED]. By contrast, Amazon Prime Music, calculated as a “bundled subscription” configuration, makes available only an abridged repertoire of 2 million songs, see Mirchandani, supra, and [REDACTED]. See id. at § 385.13(a)(4).

Thus, Amazon pays [REDACTED] for listening by the more casual consumers who use the limited catalog Prime Music service at no additional charge beyond their Prime membership fee, compared to consumers who want the full repertoire provided by Amazon Music Unlimited on their Echo devices. See Services' Joint Opening Brief at 71. These royalty obligations demonstrate the combination of price discrimination, product differentiation and “derived demand” in action; that is, the [REDACTED] are derived from the lower demand of consumers of the limited Amazon Prime Music service compared with subscribers to Amazon Music Unlimited on their Echo devices, which in turn drive higher revenues.

It is also important to note that these differential rates on the second prong of the “greater-of” structure of the PR II Rates are overridden by the revenue percentage rate on the first prong if that first prong rate generates more revenue. For example, [REDACTED], see Dissent at 29 (Table) and 116; see also [REDACTED]. With the headline rate now increased on a phased-in basis, the price discriminatory royalty generated by this [REDACTED].

It is noteworthy that Johnson affirmed the Majority's setting of other price discriminatory features, e.g., the family and student plan provisions, based on the Judges' reliance on the Services' expert testimony regarding the benefits of “having a way . . . where low willingness to pay consumers can still access music in a way that still allows more monetization of that provision of that service.” Johnson, 969 F.3d at 392–93. In similar fashion, the multi-tiered rates in the PR II-based benchmark likewise were supported by the same type of testimony; indeed, from expert testimony proffered by both parties, as considered below.

First, Professor Katz notes that the existing rate structure captures two important aspects of the economics of the interactive streaming market: (1) the variable WTP among listeners; and (2) the corollary variable demand for streaming services. See 3/13/17 Tr. 586–87 (Katz); see also Marx WRT ¶ 239 et seq.; 4/7/17 Tr. 5568 (Marx) (noting that the present structure serves differentiated products offered to customer segments with a variety of preferences and WTP). In more formal economic terms, Professor Katz notes that the present structure enhances variable pricing that allows streaming services “to work [their] way down the demand curve,” i.e., to engage in price discrimination that expands the market, providing increased revenue to the Copyright Owners as well as the Services. 3/13/17Tr. 701 (Katz).

Second, in similar testimony, Professor Hubbard captures the interrelationship between the economics of this market and the existing rate structure:

[F]rom an economic perspective, you can think of this market and this industry as being composed of different customer segments by tastes and preferences and willingness to pay. And so no rate structure can really work without understanding that, and no business model can really work without understanding that.

[I]n terms of rate structures, the Phonorecords II framework from the previous proceeding does offer a benchmark to start because it provides for differences in distinct product categories in terms of music service offerings, pricing possibilities, and so on. And it has encouraged a very diverse digital music offering set from actual competitors.

3/21/17 Tr. 2175–76 (Hubbard). Moreover, Professor Hubbard [REDACTED] 4/13/17 Tr. 5978 (Hubbard); see also Hubbard WDT ¶ 4.7 (the 2012 rate structure provides the “necessary flexibility to accommodate the underlying economics of Amazon's various digital music service offerings.”). See also 3/15/17 Tr. 1176 (Leonard) (notwithstanding changes and growth in the streaming marketplace over current rate period, underlying economic structure of marketplace, which made percent-of-revenue based royalty appropriate, has not changed).

Third, the Services' experts further assert that the multiple pricing structures necessary to satisfy the WTP and the differentiated quality preferences of downstream listeners relate directly to the upstream rate structure to be established in this proceeding. For example, Professor Marx opines that the appropriate upstream rate structure is derived from the characteristics of downstream demand. 3/20/17 Tr. 1967 (Marx) (agreeing that rate structure upstream should be derived from need to exploit willingness to pay of various users downstream via percentage of revenue because downstream listeners have varying willingness to pay that should be exploited for mutual benefit of copyright licensees and licensors). Professor Marx further acknowledged that this upstream:downstream consonance in rate structures represents an application of the concept of “derived demand,” whereby the demand upstream for inputs is dependent upon the demand for the final product downstream. Id. Moreover, Dr. Leonard notes that reliance on the Services to identify segmented demand and develop price discriminatory approaches is appropriate because “the downstream company is going to have a lot more information about . . . the business, about what makes sense.” 4/6/17 Tr. 5238 (Leonard).

Regarding a comparison of revenue growth to streaming growth, Professor Hubbard dismisses as economically “meaningless” Copyright Owners' argument that they have suffered relativ e economic injury under the current rate structure simply because the increase in their revenues from interactive streaming has been proportionately less than the growth in the number of interactive streams, leading mathematically to a lower implicit or effective per stream royalty rate. That is, he notes there is no evidence to rebut this prima facie indication of beneficial price discrimination, i.e., no contrary evidence indicating that, if the Services had sought to increase the price of the services available to these low to zero WTP listeners because of higher royalties, they would have paid the higher price, rather than declined to utilize a royalty-bearing interactive streaming service. See 4/13/17 Tr. 5971–73 (Hubbard); see also Dissent at 52.

The Services also link their price discrimination argument to the fact that the marginal physical cost of streaming is zero to the need for a flexible rate structure such as now exists. In this regard, Professor Hubbard notes that, because “[t]he marginal production cost at issue here is—is zero. . . . it's not clear why it's not better to bring new customers into the market on which royalties would be paid and, of course, zero marginal cost incurred.” 4/13/17 Tr. 5917–18 (Hubbard). See also Marx WDT ¶ 97 (“Setting the price of marginal downstream listening at its marginal cost of zero induces more music consumption and variety than per-song or per-album pricing.”).

Professor Marx makes the same argument as to the salutary nature of price discrimination in this context with regard to Spotify's ad-supported approach. Focusing on the first purpose, Spotify is attracting ad-supported listeners who have a relatively low WTP, whether they have low incomes, (a budget constraint) or low interest in music (low “utility,” in the parlance of economists). These listeners, and the advertising revenue they generate are real and reflect the WTP of a large swath of all interactive listeners. See Marx WRT ¶ 115–16 & Fig. 9 (“While I agree that one aspect of the ad-supported service is to provide an on-ramp to paid services, it also has another important aspect, namely to serve low WTP customers. . . . Copyright Owners' economists err in not calculating the impact of the Copyright Owners' proposal on ad-supported services. Ad-supported services currently make up a majority of subscribers and [REDACTED]% of all streams in the industry.”).

Accordingly, a separate tier for an ad-supported service accounts for the different nature of the downstream listenership, allowing the upstream royalty to be based on that characteristic. This differentiation was essentially acknowledged by Copyright Owners late ( too late, actually) when they proposed in their post-hearing filing that “if the Judges intend to include the Spotify ad-supported service in the rate structure and rate calculations, that they do so by establishing separate rates and terms for the ad-supported service. See COPCOL (Corrected) ¶ 228 & n.34. But the PR II-benchmark already incorporates separate rates for free/ad-supported services!

Another important evidentiary factor buttressing the need for price discriminatory rates and structures was the testimony of the Services' survey expert, Mr. Robert Klein, Chair and co-founder of Applied Marketing Systems, Inc. Mr. Klein surveyed 2,101 people (the Klein Survey) who were listeners to streamed music and found, inter alia, that: (1) the majority of listeners would not pay for a monthly streaming subscription; and (2) for those who do subscribe, their demand was elastic, with increases in subscription prices causing overall greater percentage reductions in quantity demanded, moving customers to free, ad-supported and non-streaming alternatives. See Klein WRT ¶¶ 60–67. By contrast, Copyright Owners did not present any survey testimony. The Determination fully credited the Klein Survey, finding as follows:

It is important to note that Copyright Owners' attacks on the Klein Survey are not levelled by any witnesses, nor contradicted by their own survey expert, because Copyright Owners elected not to proffer such an expert in their direct (or rebuttal) cases. Rather, Copyright Owners elected to make a descriptive argument regarding the elasticity of demand among different segments of the market, as opposed to a survey-based or econometric study of price elasticity.

[Although] Copyright Owners attack the Klein Survey on several fronts[,] [t]he arguments made by Copyright Owners are insufficient . . . to seriously weaken the probative value of the Klein Survey. In the end, the Judges are not persuaded by the Copyright Owners' revenue bundling arguments not to adopt a flexible, revenue-based royalty rate.

Determination at 22–23 & n.53; see also Dissent at 64–67 (including point-by-point rejection of Copyright Owners' non-expert criticisms of Klein Survey).

The Services also note that the existing rate structure has produced generally positive practical consequences in the marketplace. Their joint accounting expert, Professor Mark Zmijewski, testified that the [REDACTED] from the sale of product under (former) Subpart A since 2014 has been [REDACTED] over the same period. Expert Report of Mark E. Zmijewski February 15, 2017 ¶¶ 38, 40 (Zmijewski WRT); 4/12/17 Tr. 5783 (Zmijewski); see also 4/13/17 Tr. 5897 (Hubbard) (“the evidence that I reviewed suggests that the copyright holders have actually benefitted from this structure. . . .”).

More particularly, Professor Zmijewski testified that:

• Total revenues reported by the NMPA for NMPA members from all royalty sources [REDACTED]. Zmijewski WRT ¶ 41.

• This [REDACTED]. Id.

• The [REDACTED]. Id.

• Mechanical royalty revenue for the sale of downloads and physical phonorecords [REDACTED]. Id. ¶ 38.

In sum, the foregoing analysis demonstrates the economic reasonableness and appropriateness of the price discriminatory Phonorecords II rate structure and its negotiated safeguards to address the real possibility of revenue diminution. As discussed below, the record evidence also supports royalty rates within the PR II-based benchmark.

The PR II-based benchmark contain several alternate rates explicitly calculated as a percentage of payments made by interactive streaming services to the record companies for sound recording rights. See Addendum to this Initial Ruling. In the Subpart relating to streaming, the (former) subpart B category, the TCC is 22% for ad-supported services and 21% for portable subscriptions. Id.; see also37 CFR 385.13(b)(2) and (c)(2). These percentage figures correspond to sound recording: musical works royalty ratios of 4.55:1 and 4.76:1, respectively.

With regard to these ratios, Copyright Owners' economic expert witness, Dr. Eisenach, stated: “In my opinion, the evidence . . . indicates that the relative valuation ratios implied by the current Section 115 compulsory license . . . represent an upper bound on the relative market valuations of the sound recording and musical works rights.” Id. ¶ 92 (emphasis added). (As an “upper bound,” these ratios would represent the lower bound on the relative market valuations of the reciprocal percentage of the value musical works rights relative to sound recording rights, again, 22% and 21%. ) Thus, there appears to be consensus between Copyright Owners' witness and the Services (who advocate for applying these rates on the price discriminatory tier of their benchmark) that these rates constitute “relative market valuations” (even if they are not Dr. Eisenach's preferred market valuations within the bounded zone of such values).

Dr. Eisenach's testimony regarding the “bounds” of useful market valuations is noteworthy because his acknowledgement is consonant with judicial precedent. The Judges' setting of reasonable rates often requires them to identify a “zone of reasonableness,” within which they identify appropriate statutory rates. See, e.g., Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board, 684 F.3d 1332, 1340 (D.C. Cir. 2012) (The CRB Judges' rate setting can necessitate the finding of a “zone of reasonableness [because] “[s]tatutory reasonableness is an abstract quality represented by an area rather than a pinpoint.”).

The 21% and 22% TCC rates within section 115 identified by Dr. Eisenach as generating the “lower bound on the relative market valuations” imply certain approximate percent-of-revenue rates, i.e., percent of total service revenue (not percent of sound recording revenue). See Dissent at 91, n.133 (sound recording rates clustered between [REDACTED]% and [REDACTED]% of revenue). For example, if the sound recording royalty rate for interactive streaming is [REDACTED]% of revenue, then the musical works rate would be calculated as 0.21 × [REDACTED], which equals [REDACTED]%, (or as .22 × [REDACTED] which equals [REDACTED]%). At the low end of the range, if the sound recording royalty rate is [REDACTED]%, then, applying these TCC figures, the implied musical work royalty rate would be calculated as [REDACTED]% (.21 × [REDACTED]) or [REDACTED]% (.22 × [REDACTED]).

It is important to emphasize and detail the context of these price discriminatory rates. These capped TCC rates are on the “greater of prong” that is compared with the headline 15.1% revenue rate (phased-in) that the Judges are also adopting in this Initial Ruling. As phased in, the headline rate is greater than all the capped TCC-based rates identified in Dr. Eisenach's testimony, supra, [REDACTED]. For 2019, the phased-in headline percentage rate, 12.3%, is [REDACTED] the [REDACTED]% and [REDACTED]% revenue rates derived if the sound recording rates was [REDACTED]%. For 2018, the phased-in headline percentage rate, 11.4%, is [REDACTED] all the rates derived from the capped TCC rates Dr. Eisenach identified as “market valuations” (albeit the lower bound in his opinion). But that is of no negative consequence for Copyright Owners, because they would get paid on the “greater-of” metric (capped TCC or headline rate) under the Phonorecords II -based rate structure the Judges are adopting (For the portable subscriptions, even though the 80 cents/subscriber “lesser-of” portion of the non-headline prong would apply on that prong if it was lower than the capped TCC rate, the actual rate could not be lower than the phased-in headline rate.)

Dr. Eisenach also examined direct agreements between record companies and interactive streaming services that contain rates for sound recordings and mechanical royalties, respectively. See, e.g., id. ¶¶ at 84–91. In such cases, the ratio of sound-recording to musical-works royalties ranged tightly between [REDACTED] and [REDACTED], closely tracking the regulatory ratios implicit in the section 115 TCC. Id. ¶ 92. (The [REDACTED] ratio equates to a TCC rate of [REDACTED]%, and the [REDACTED] ratio equates to a mechanical rate of [REDACTED]%.). He concluded, as he did with regard to the actual section 115 license rates: “In my opinion, the evidence presented . . . indicates that the relative valuation ratios implied by the . . . negotiations under [the statutory] shadow—ranging from [REDACTED] [[REDACTED]%] to [REDACTED] [[REDACTED]%]—represent an upper bound on the relative market valuations of the sound recording and musical works rights. ” Eisenach WDT ¶ 92 (emphasis added).

Dr. Eisenach also identified several additional useful benchmarks. First, he identified what was coined the “Pandora Opt-Out Agreement” benchmark, which reflected a ratio of [REDACTED] of sound-recordings to musical-works in a comparable benchmark setting. This ratio translates to a TCC percent of [REDACTED]%. With sound recording royalty rates of approximately [REDACTED]% to [REDACTED]%, this TCC reflects an effective percentage of total revenue equal to [REDACTED]% to [REDACTED]%.

Second, Dr. Eisenach identified YouTube agreements with music publishers that relate to the combination of a commercial sound recording and a “static image.” The YouTube agreements contain an explicit royalty of [REDACTED]. That [REDACTED]% royalty is a denominator in the ratio concept utilized by Dr. Eisenach, and the numerator is the [REDACTED] sound recording royalty paid to the record companies. YouTube had agreed to pay [REDACTED]% of its revenues, and had agreed to pay [REDACTED] and other record companies [REDACTED]% of revenues. The [REDACTED] ratio reduces to [REDACTED], implying a TCC ([REDACTED]) of [REDACTED]%. The [REDACTED] ratio reduces to [REDACTED], implying a TCC ([REDACTED]) of [REDACTED]%. See Dissent at 101–102.

These additional rates identified in Dr. Eisenach's testimony further confirm the reasonableness of the non-headline rates within the PR II-based benchmark.

Finally, the Judges look at the effective rates paid by Spotify, the largest interactive streaming service in terms of in terms of the number of subscriber-months and the number of plays. See Marx WRT ¶¶ 37–38 & Figs. 8 & 9. Under the PR II based benchmark, Spotify paid on its subscription service an effective “All-In” royalty rate of [REDACTED]% of its total revenues. See Dissent at 80, 115, 149 (and record citations therein). Spotify paid this effective percent-of-revenue rate [REDACTED]. See id. at 29 (Table).

Turning to Spotify's free/ad-supported offering (and as noted supra), Spotify paid royalties under the PR II Rates at an effective “All-In” royalty rate of [REDACTED]%. Spotify paid this effective percent-of-revenue rate [REDACTED]. See id. When Spotify's two tiers are blended and averaged, the effective percent-of-revenue rate is [REDACTED]% of revenue. See id. at 116. The average rate has salience in this proceeding because Spotify's two tiers are interrelated, in that free/ad-supported listeners constitute a pool of potential converts to the subscription tier under this “freemium” model, even as this offering generates royalties under the PR II-based benchmark.

d. Copyright Owners' Concern Regarding Revenue Diminution Is Insufficient To Reject the PR II-Based Benchmark

Copyright Owners argue that what the Services tout as beneficial price discrimination generates an “incredible” level of revenue diminution, including displacement, resulting in a “major problem” that reduces reportable revenues and thus the royalty base. See, e.g., 3/7/22 Tr. 193 (Copyright Owners' counsel). This argument is based upon documents and evidence that demonstrated the following:

• [REDACTED];
• [REDACTED].
• [REDACTED];
• [REDACTED];
• [REDACTED];
• [REDACTED]
• [REDACTED];
• [REDACTED];
• [REDACTED]; and

• Copyright Owners' expert, Dr. Rysman, testified that interactive services often elect to forgo current profit maximization, e.g., by charging lower prices, in order to build a customer base and greater long-run profitability or value, from selling music and non-music products or services to its customers.

CO Initial Submission at 40–42 (and record citations herein).

The Services' economic experts do not ignore the fact that there can be revenue attribution problems when interactive streaming is combined with other products or services. They acknowledge that, even absent any wrongful intent with regard to the identification and measurement of revenue, attribution of revenue across product/service lines of various services can be difficult and imprecise. See, e.g., 4/5/17 Tr. 5000 (Katz) (problem of measuring revenue “certainly a factor that goes into thinking about reasonableness.”).

However, Professor Katz testified that the existing rate structure agreed to by the parties accommodates these bundling, deferral, and displacement issues via the use of an alternative rate prong that would be triggered if the royalty revenue resulting from the headline rate of 10.5% of streaming revenue fell below the royalty revenue generated by that second prong. Katz WDT ¶¶ 82–83; 3/13/17 Tr. 670 (Katz). Moreover, Professor Katz concluded that, because the marketplace appears to be functioning (in the sense that publishers are earning profits and new and existing interactive streaming services continue to operate despite accounting losses), these revenue-measurement issues are being adequately handled by the alternative rate prong, even if an altered second prong might work better. Id. at 738–39. More generally, Professor Katz further noted that, the existing rates within the PR II-based benchmark were performing well, and even if alternative minima might be preferable, no such alternative rates were in the record. See 4/15/17 Tr. 5056–58 (Katz) (under the PR II-based benchmark “the industry . . . was performing well,” but “if someone had a proposal [with] a specific reason why we should adjust this minimum that's something I would have examined,”). But Copyright Owners did not propose alternative rates or minima within the PR II-based benchmark, but instead urged the Judges to disregard the benchmark writ large. Accordingly, there were no alternative rates or minima in the record.

Professor Katz further noted that the PR II-based benchmark rates were established when “ecosystem” entities such as Yahoo—akin to Amazon, Apple, and Google—were in the marketplace. 4/5/17 Tr. 5055–57 (Katz); see also Determination at 31 (and record citations therein) (noting the presence of Microsoft as well as Yahoo as licensees in the interactive market during the Phonorecords II negotiations).

More broadly, the Services' position regarding the use of the two prongs and their alternate rates to ameliorate the revenue-measurement problems is summed up by Professor Katz as follows:

[T]he primary reason [for the two rate prongs] . . . is because of the measurement issues that can come up when having royalties based on a . . . percentage of revenues because there can be issues about how to appropriately assign revenues to a service. And so I think the minim[a] can play an important role when those—you know, when those measurement problems are severe, you can turn to the minimum instead. . . . [W]hat I have in mind, right, is that what would happen if you could imagine an entrepreneur coming along and saying we want to have a service and have some incredibly low price and not a very good monetization model, where a copyright owner would say—in an effectively competitive market, would say, wait a minute, I don't want to license to you on those terms. It's—I just think the possibility of getting a return is so low, I'm not going to do it, even though you, as an entrepreneur, are willing to try this. I as the copyright owner want some sort of, you know, return on it. And that's what the minimum also helps to do.

3/13/17 Tr. 599 (Katz.); see also 3/20/17 Tr. 1900–01 (Marx) (minima protect against revenue measurement problems); 4/7/17 Tr. 5584 (Marx) (statutory minima play “two roles”— protecting the Copyright Owners from “revenue mismeasurement” by creating the “greater of” prong,” but incorporating per subscriber rate prong in “lesser of” component to protect services from the record companies' use of their market power to engage in “manipulation of the sound recording royalties” on which the TCC prong is calculated).

After considering the record, the Judges determine that the Majority had not found—as Copyright Owners claim—that the activities and strategies by the Services were “incredible” or a “major problem. Rather, the Majority's characterization was measured, stating repeatedly that the Services engaged “ to some extent ” in revenue diminution because they “ might focus on long-term profit maximization to promote their long-term growth strategy, which occurs “even absent wrongful intent.” Determination at 20–21, 36, 90; accord, Dissent at 59. In fact, the Majority specifically stated: “The Judges agree that there is no support for any sweeping inference that cross-selling has diminished the revenue base.” Id. at 21 (emphasis added). The Majority (and the Dissent) thus acknowledged the reasonableness of both sides of this issue, recognizing both the Services' use of price discriminatory approaches that can lower per user or per-stream revenues but grow royalties, market share and revenue, as well as Copyright Owners' concomitant desire to protect themselves from reductions in the royalty revenue base, however limited in extent, that would only serve to diminish royalties.

One way the input supplier can avoid this impact is to refuse to accept a percent of revenue form of payment and move to a fixed per-unit input price. This is what Copyright Owners originally and unsuccessfully sought in this proceeding, subject to a bargaining room approach by which they could switch back to the old approach (or any other approach) through purely market-based negotiations, unbounded by the statutory and regulatory standards of “fairness” and “effective competition.” See Dissent at 60.

The Judges must reconcile the parties' competing considerations. A way by which they are both accommodated is through a pricing structure with alternate rate prongs and floors, below which the royalty revenue cannot fall. This is precisely the bargain struck between Copyright Owners and services in 2008 and 2012, and that has been the rate structure through 2017. And, because the Majority and the Dissent found that revenue diminution occurred only “to an extent,” rather than in the pervasive (sweeping”) manner averred by Copyright Owners, there is no sufficient reason in the record to depart from the bargained-for multi-tiered rate structure in Phonorecords II that allows for price discrimination but tempers its impact on royalties through the use of minima and floors.

e. Copyright Owners' Claim of “Inherent” Economic Value Is Belied by the Record, Including Their Own Arguments

Pre-remand, Copyright Owners approached this rate setting process with an overarching premise: A musical work has an “inherent value” that must be reflected in the royalty rates. As the NMPA's president, Mr. Israelite testified, when asked how “inherent” value is defined:

[W]hoever owns an individual copyright is the one to define it. I think that would be the most appropriate definition of it. What someone is willing to license it for would be that inherent value to that owner . . . That would be market value.

3/29/17 Tr. 3707 (Israelite).

If the market for musical works was as atomistic as the above quote assumes, the songwriter of an individual musical work could indeed set his or her own royalty rate, and refuse to license to any streaming service or other distributor who refused to pay that royalty. But that is not how the licensing market works. Songwriters typically assign their licensing rights to music publishers (to avoid ruinous transaction costs). These music publishers control huge “Must Have” repertoires that are offered under blanket licenses to streaming services. (The musical works market of course is subject to a compulsory license, but this is precisely how the unregulated market works for the licensing of sound recordings by labels to interactive streaming services.) It is acknowledged even by Copyright Owners' own expert witness, Professor Watt that the creation of these large collectives generates market power that necessitates rate regulation. See R. Watt, Copyright and Economic Theory: Friends or Foes at 163, 190 (2000) (quoted in Dissent at 35).

Further, this “inherent” market value notion is antiquated as a matter of economics. Although an individual Copyright Owner can announce his or her “asking” royalty, that is not sufficient to generate a “market” royalty, unless and until a licensee agrees to pay it. In market-based economics. that is to say, the economic consensus that has governed economics since the “marginal revolution” in the mid to late 19th century, value is ascertained through the intersection of supply and demand, with the price established at the margin representing the market value of the good or service bought and sold. If there is no demand for a product, be it a musical work or anything else, it has no economic value. Even though costs have been incurred to produce the product, those costs cannot be recovered (or profit earned) absent a sufficient WTP in the market. And, as noted supra, the product being offered and at issue here is comprised of “second copies” of sound recordings (with embedded musical works), which are costless to reproduce for streaming purposes. Of course, these “second copies” do have actual value when they are in demand, and the royalties that their licensing generates must cover: (1) the first copy (creative) costs; (2) the “opportunity cost” (measured by the next best alternative for royalty earnings if the “second copies” could have been supplied through another distribution channel that paid higher royalties to attract the end-user/consumer at issue); and (3) profits to induce the creation of musical works.

Second, the fact that Copyright Owners originally proposed a per-subscriber alternative rate to their per-play rate itself belies their conviction that some “inherent” economic value exists. When the metric of value switches from “per-play” to “per-subscriber,” the focus of value likewise shifts from an emphasis on producer value to consumer value. That is, if there is truly an “inherent” value for a product or service, that singular value cannot divide into two distinct values with the “greater-of” the two controlling. Such an argument gives away the game, so to speak, demonstrating, perhaps unsurprisingly, that economic arguments (not unlike legal advocacy) are often situational—designed to support maximalist positions and the exercise of market power, however acquired. See also Determination at 28 n.64 (rejecting the “inherent value” argument).

f. PR II-Based Benchmark Not “Too Complex”

Copyright Owners and the Majority complained that the PR II-based benchmark is too complex. See Copyright Owners' PFF ¶ 12 (criticizing complexity of PR II Rates as lacking “transparency”); Determination at 36 (characterizing parties' negotiated, renewed, and expanded rate structure as Rube-Goldberg-esque in complexity and impenetrability.”)

After considering this issue on remand, the Judges disagree. If some songwriters or lyricists have been confused by their royalty statements, their confusion of course should be resolved. However, one of the benefits of a collective is that it possesses the expertise and resources to identify and explain how royalties are computed and distributed. Moreover, this claim of complexity cannot serve as a basis to override the multi-part negotiated benchmark that the parties, through their respective trade associations, negotiated and implemented. As the Dissent stated: “There is no good reason why the rate structure that is consonant with the parties' ten-year history and with the relevant economic model should be sacrificed on the slender argument that “simpler is better than complicated.” Dissent at 88.

Further, section 801(b)(1) does not identify “simplicity” as a statutory goal for the setting of rates, rate structure, and terms. Although there is certainly no need for gratuitous complexity, the price discriminatory structure and the associated levels of rates in the PR II-based benchmark that were eliminated by the Majority (while maintaining all the remaining complexity) were most certainly not gratuitous, but rather designed, after negotiations, to establish a structure that would expand the revenues and royalties to the benefit of Copyright Owners and Services alike, while also protecting Copyright Owners from potential revenue diminution by the Services. Moreover, when the market itself is complex—in that the WTP across consumer groups is heterogeneous and the offerings reflect that fact—it is unsurprising that the regulatory provisions would resemble the complex terms in a commercial agreement negotiated in such a setting. For the Judges to demand simplicity in this context would be to sacrifice the specificity that an effectively competitive market requires. See Dissent at 88 (rejecting the simplicity argument by invoking the advice attributed to Albert Einstein that “[e]verything should be made as simple as possible, but no simpler. ”

g. So-Called Statutory “Shadow” Does Not Diminish Value of the PR II-Based Benchmark Rates

Copyright Owners maintain that the rates in the PR II-based benchmark are infirm because, like any benchmark for which a statutory rate is the default, they are not actual market rates. That is, such a rate is said to exist in the so-called “shadow” of the statutory rate. See Dissent at 70 (and citations therein).

The Judges reject this argument for several reasons. First, the argument is undercut by the explicit language of section 115 of the Copyright Act, which states: “In addition to the objectives set forth in section 801(b)(1), in establishing such rates and terms, the Copyright Royalty Judges may consider rates and terms under voluntary license agreements described in subparagraphs (B) and (C).” 17 U.S.C. 115(c)(3)(D). Subparagraphs (B) and (C), respectively, refer to agreements on “the terms and rates of royalty payments under this section” by “persons entitled to obtain a compulsory license under [17 U.S.C. 115(a)(1)]; and “licenses” covering “digital phonorecord deliveries.” Id. Thus, it is beyond dispute that Congress has authorized the Judges, in their discretion, to consider such agreements as evidence, irrespective of—or perhaps because of—the shadow cast by the compulsory license. Thus, the appropriate question is how much weight the Judges, in their discretion, should afford such benchmarks in any particular proceeding.

There is no basis to find, as Copyright Owners suggest, that statutorily-based or influenced benchmarks, including specifically the PR II-based benchmark in this proceeding, are per se inferior to other benchmarks or alternative economic evidence ( e.g., from models, surveys or experiments) that may be unaffected by the shadow. Those other benchmarks or forms of evidence will also be subject to their own imperfections and incompatibilities with the target market and must be identified and weighed accordingly. Thus, the Judges must not only consider (i) the importance, vel non, of any potential so-called “shadow-based” distortionary effects from a benchmark derived from a regulated statutory benchmark market, but also (ii) how any such purported “shadow” effects compare to any distortions generated by other proffered benchmarks and competing alternative economic evidence, e.g., distortions based on complementary oligopoly power, bargaining constraints and product differentiation in other benchmarks, models, surveys or experiments.

The Services' experts discount the foregoing shadow-based criticism. Moreover, the Services laud a statutorily-influenced benchmark in general, and the specific PR II-based benchmark in particular, because the latter reflects more equal bargaining power between licensors and licensees. In this regard, one of the Services' economic expert witnesses, Professor Katz, points out that rates set voluntarily by the parties in a settlement under the “shadow” provide two important benefits. First, with a statutory rate-setting proceeding as a backstop, large licensors cannot credibly threaten to “hold out” and “walk away” from the negotiations without an agreement, thereby negating their ability to use their “must have” status to obtain rates above effectively competitive levels. Second, when, as here, such negotiations are conducted with all the parties at the figurative table—including here, trade associations—no single party has disproportionate market power in the negotiations. See 3/13/17 Tr. 661 (Katz).

The Judges agree that settlement agreements reached in the statutory shadow are useful. Although imperfect when considered in isolation, in that the statutory proceeding is the default backstop, in context they negate the power of any entity simply to refuse to strike a deal. The negation of that power blunts the complementary oligopoly power of licensors of “Must Have” repertoires (whether musical works or sound recordings), making a benchmark agreement reached in the so-called “shadow” advantageous in establishing an effectively competitive rate. See Web IV, supra, 26,316, 26,330–31 (May 2, 2016) (noting counterbalancing effect of statutory license in establishing effectively competitive rates). Further, when such settlement agreements are industrywide, they tend to eliminate disproportionate market power, See Dissent at 72; Web III,79 FR 23102, 23111 (Apr. 25, 2014), aff'd Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Bd., Case No. 14–1098 (D.C. Cir. Aug. 11, 2015) (relying on two settlement agreements).

Nonetheless, Copyright Owners are correct to note that, hypothetically, some licenses might have otherwise been negotiated at rates higher than the settlement rate that was affected by the so-called shadow. But that is simply the tradeoff that the statutory scheme makes in its identification of settlement rates as evidentiary benchmarks. Such a theoretical problem cannot serve to override the salutary aspects of benchmark settlement agreements. See Web IV, supra at 26,630 (rejecting same argument as speculative and “too untethered from the facts to be predictive or useful in adjusting for the supposed shadow of the existing statutory rate.”).

Lastly, with regard to a benchmark affected by the so-called “shadow,” the Judges find that, with regard to the application of the itemized factors in section 801(b)(1), they have the same duty to independently weigh those factors as they do for all otherwise reasonable rates. Thus, the Judges reject the idea that rates and terms reached through a settlement must be understood to supersede—or can be assumed to embody—the Judges' current thinking as to the application of the statutory elements set forth in section 801(b)(1). The Judges are obliged to conduct the four-factor analysis anew when considering a previously adopted settlement in a subsequent proceeding—and they do so infra. Of course, if on such further analysis, the Judges find that the provisions in an otherwise useful benchmark agreement (including those in a benchmark influenced by the so-called “shadow”) do appropriately reflect the four itemized statutory factors in section 801(b)(1), then the Judges may adopt the provisions of that settlement without a factor-based adjustment.

h. Conclusion Regarding PR II-Based Benchmark

Accordingly, the Judges find the PR II Rates to be a useful benchmark. However, this benchmark is modified by the Judges' substitution of the 15.1% headline percentage rate for the 10.5% headline percentage rate in the benchmark.

D. Precedent Permits Judges To Apply Elements of PR II Rates, Rate Structure and Terms Even if Those Are Not Proffered as Benchmarks

The D.C. Circuit has previously held that the Judges have the authority to adopt elements from the existing rate provisions, if they find that those prevailing provisions better satisfy the statutory requisites than any other proposed structures and rates discernible from the record evidence. Music Choice v. Copyright Royalty Bd., 774 F.3d 1000, 1009 (D.C. Cir. 2014). This authority exists even when no party has proffered those provisions in the form of a benchmark.

In Music Choice (concerning the setting of satellite radio royalty rates under the same section 801(b)(1) standard), the CRB Judges rejected the parties' proffered benchmarks and instead relied on a percent-of-revenue rate (13%) that was neither a benchmark nor even the prior statutory rate, but merely “a component of a prior determination.” Id. at 1009. The licensor-party, SoundExchange, argued, like Copyright Owners here, that this component of a prior rate was “stale,” “outdated,” or “obsolete.” Rejecting this argument as “erroneous,” the D.C. Circuit stated that “the Judges did not consider the 13% rate as a current benchmark,” but rather used it to “bridge the gap” caused by the inadequacies of the parties' rejected benchmarks. Id. In so doing, the D.C. Circuit held that the Judges properly resolved “serious problems” with the licensor's proposal, even as it had “partially credited it” and also “used permissible indicia of reasonableness to help fix the rate.” Id.

Music Choice is highly instructive. Here, on remand, the Judges adopt a modified version of the prior rate structure and rates in Phonorecords II. The fact that it was also proffered as a benchmark, in another modified form by the Services, does not render Music Choice inapposite. Rather, because the Phonorecords II provisions were proffered as benchmark evidence, these provisions were placed squarely into the record, allowing the parties and the Judges to address the relative merits. A fortiori, Music Choice underscores the propriety of the Judges approach in this proceeding. That is, even if the Services had not proffered this approach as a benchmark, Music Choice allows the Phonorecords II approach to serve as a guidepost for establishing the rates and rate structure in this proceeding.

Further, here the Judges are adopting actual elements from the prior rate provisions, rather than, as in Music Choice, a mere “component” used to generate the prior rate. A fortiori yet again, Music Choice allows the Judges to prudently utilize the prior rate and rate structure regulations to synthesize a determination in this proceeding. The analogous nature of Music Choice is also seen in the Judges' use in the present case of the “headline” 15.1% revenue rate proposed by Copyright Owners on remand combined with elements from the PR–II regulatory provisions, including its price discriminatory rates. In Music Choice, the Judges likewise “partially credited” the licensor's proposal, which, as noted supra, the D.C. Circuit affirmed.

Finally, the Judges take note that Music Choice also addressed the Judges' findings regarding the setting of another statutory license, for Preexisting Subscription Services (PSS), by using a rate in a settlement from a prior period. This context is also analogous here, because Copyright Owners object to the use of the Phonorecords II rate structure and rates as the product of a settlement. It is instructive to consider how the arguments of the licensor (SoundExchange) in Music Choice mirror those of Copyright Owners in this proceeding:

• SoundExchange notes that this rate “is the product of settlement negotiations that occurred in SDARS I between Music Choice and SoundExchange.”

• SoundExchange argues that the Judges arbitrarily rejected . . . more recent data points in favor of the “outdated” settlement rate.
• SoundExchange maintains that the Judges conceded that the prevailing rate had limited value, as the settlement rate “was negotiated in the shadow of the statutory licensing system and cannot properly be said to be a market benchmark rate.”
• SoundExchange also argues that simply reciting that “nothing in the record persuades the Judges” that the prevailing rate is unreasonable . . . does not show that [it] is reasonable, or that it is supported by the written record.
• [G]iven the lack of creditable benchmarks in the record, the Judges did not err when they used the prevailing rate as the starting point of their Section 801(b) analysis.
• The Copyright Act contemplates that the Judges would . . . consider “prior determinations” and rates established “under voluntary license agreements.”
• [T]he Judges did not err when relying on the settlement rate. The Judges conceded that the settlement rate does not represent a market rate. . . . But . . . the relevant portion of the Copyright Act does not use the term “market rates,” nor does it require that the term “`reasonable rates” be defined as market rates. . . . The Act authorizes the Judges to consider rates set “under voluntary license agreements.”
• Music Choice complains that it agreed to a higher rate to avoid litigation costs, but has not introduced evidence that the settlement was involuntary or otherwise unreasonable. It was not arbitrary, then, for the Judges to consider the voluntary settlement rate.

Music Choice, 774 F.3d at 1012–15. These aspects of Music Choice are highly instructive, considering the Judges' parallel findings regarding the same and similar arguments as discussed supra regarding prior settlement agreements and the so-called “shadow” of the statutory rates.

In sum, Music Choice provides ample support for the conclusion that, even if the Services had not proffered their PR II-based benchmark, the Judges would have acted well within their authority to give the same weight to the PR II rates and structure as they have in this Initial Ruling.

E. Four Itemized Factors in Section 801(b)(1)

The Judges have considered the application of the four itemized statutory factors A through D, in connection with their application of the 15.1% revenue rate and their partial use of the PR II-based benchmark.

1. Factor A

The Judges have explained supra that price discrimination is a “win-win” for Copyright Owners and the Services. By serving low WTP listeners, it brings in new listeners and subscribers who increase royalty payments as well as revenues. Any licensor would prefer to increase its royalties, rather than “leave money on the table,” and a rate structure that effects such an increase (through the concept of “derived demand”) is appropriate. Moreover, for purposes of applying Factor A, a rate structure that increases royalties, ceteris paribus, would induce more production of musical works, a result that Copyright Owners should desire.

This point appears to raise a question: How could Copyright Owners and their economic experts object to a rate structure that inures to their benefit as well? The answer is: They do not object. They are not economic naifs. As stated supra, they advocate for a rate set under the bargaining room theory, through which rate structures can still be negotiated, but not subject to the “reasonable rate” and itemized factor analysis required by law. In those negotiations, as Dr. Eisenach candidly acknowledged, Copyright Owners would have a different threat point to use in order to obtain better rates and terms. 4/4/17 Tr. 4845–46 (Eisenach).

Second, given a heterogeneous downstream WTP, it would not be more profitable simply to equate “availability” with a higher rate. As noted supra, any product that is priced beyond the WTP of a significant portion of the public is unavailable to that segment. Royalties that are aligned with the varying WTP of different classes of listeners will make downstream price discrimination more affordable to the services, driving new revenue and royalties—precisely as the PR II-based benchmark allows. In this regard, Copyright Owners have taken a cramped and unrealistic view of such incentives. In particular, the Judges disagree with Copyright Owners' expert economic witness, Professor Rysman, who startlingly asserted in response to a hypothetical from the bench that even a $10,000 per month subscription price would increase “availability.” 4/3/17 Tr. 4397 (Rysman).

The Judges find Professor Rysman misapprehends the nature of a price signal. If the price is so high as to eliminate or reduce total revenue to creators, in no way will higher rates simply induce the supply of creative works over time. Indeed, even monopolists do not seek the highest price possible, but rather seek to maximize profits. See E. Mansfield & G. Yohe, Microeconomics at 362–63 (11th ed. 2004) (“Monopolies maximize profits by producing where marginal cost equals marginal revenue.”). Thus, even monopolists, who have the most market power, are constrained in their pricing by the demand curve and the marginal revenue it creates. Simply put, although a higher royalty rate might have an immediate superficial appeal, if the consequence will be lower revenues, the high per-play rate would reveal itself as a form of fool's gold.

In sum, the Judges find that the Factor A objective of “maximizing the availability of creative works” is furthered by an upstream rate structure that contains multiple royalty rates reflective of and derived from downstream variable WTP, because it will facilitate beneficial price discrimination. Such price discrimination allows for access to be afforded “down the demand curve,” making musical works available to more members of the public. However, there is no evidence to suggest that the price discriminatory rates should be changed, in order to address the connection between price discrimination and the objective of Factor (A). Accordingly, the Judges find no basis to adjust either the rate structure or the rates based on Factor (A).

2. Factors B and C

The concepts of “fair income,” “fair return” and recompense for costs and other contributions was considered in connection with the setting of the 15.1% revenue rate. In that context, the Judges analyzed the Shapley Value modeling that was designed to generate “fair” rates that allowed the parties to recover their costs and to share the surplus (over and above costs) in a manner that: (1) prevented the “Must Have” Input Suppliers (the record companies and Copyright Owners) from using the essential aspect of their inputs to engage in hold-up by threatening to withhold their respective repertoires; and (2) allocated surplus shares according to each party's contribution to the surplus (as calculated though the “arrival orderings” in the Shapley model).

The PR II-based benchmark was the product of an industrywide negotiation, with the music publishers represented by the NMPA and the interactive streaming services represented by DiMA, their respective trade associations. As explained in the Dissent, supra, at pp. 137–39, when an industrywide settlement is reached, particularly when the default procedure is a contested rate proceeding before the Judges, it contains the same benefits with regard to the avoidance of the “hold-out” effect and the equalizing of bargaining power as produced by Professor Marx's Shapley value modeling. See 3/13/17 Tr. 577 (Katz) (“I think of the shadow as balancing the bargaining power between the two parties.”); Katz CWRT 136, n.236 (“there are market forces that promote the achievement of the statutory objectives in private agreements, such as the 2012 Settlement, when the parties are equally matched (it was an industry-wide negotiation) and the negotiations are conducted in the shadow of a pending rate-setting proceeding that can be expected to set reasonable rates in the event that the private parties do not reach agreement.”).

Accordingly, this benchmark already incorporates the dynamics of a negotiation between parties with mutually countervailing power (although those dynamics required updating of the headline rate to 15.1% to account for the higher revenues, as undertaken by the Majority's Shapley analysis). See Web V,86 FR 59452, 59456 (Oct. 27, 2021) (“the licensor-side complementary oligopoly power could be ameliorated by the “ countervailing power ” of a licensee”).

Therefore, the Judges do not make any adjustment in their application of the PR II-based benchmark pursuant to Factors B and C.

3. Factor D

As noted supra, the Judges understand that a Factor D adjustment is warranted if the rate the Judges would otherwise establish

directly produces an adverse impact that is substantial, immediate and irreversible in the short-run because there is insufficient time for either [party] to adequately adapt to the changed circumstance 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.

Determination at 87.

There is no record evidence to suggest that the Services' PR II-based benchmark, as utilized by the Judges in this Initial Ruling, would create the requisite “adverse impact” to trigger Factor D. The Services certainly do not assert that their own proffered benchmark would be disruptive. With regard to Copyright Owners, the Judges cannot identify any aspect of the PR II-based benchmark that would cause the type of disruption that can serve as an adjustment under the statutory language of Factor D or the Judges' application of same, as quoted above. The Judges understand Copyright Owners' complaint to be principally that [REDACTED] during the Phonorecords II period, [REDACTED] the number of musical works streamed via sound recordings performed on interactive services. However, that is most certainly not any sort of disruption, let alone a disruption cognizable under section 801(b)(1) and under the Judges' application of that provision.

F. Subpart C Offerings Covered by Foregoing Analysis

The Phonorecords II parties also negotiated several new service types—paid locker services, purchased content locker services, mixed service bundles, music bundles and limited offerings. These service configurations were described in subpart C of 37 CFR 385 under the Phonorecords II regulatory provisions. Parness WDT ¶ 13; Levine WDT ¶¶ 38–39; Israelite WDT ¶¶ 28–30. These negotiations spanned more than a year. See 3/29/17 Tr. 3652–55 (Israelite) (involved protracted bargaining, in which NMPA rejected some categories, while others were accepted and became part of subpart C). Id. at 3654–56. The parties ultimately agreed on a structure for subpart C that resembled the subpart B structure, including a headline percentage of revenue royalty rate and per-subscriber and TCC minima. Parness WDT ¶ 14; see also37 CFR 385.22. As with the bundling negotiations relating to subpart B, the parties negotiated and created a bundled service category under subpart C (with certain adjustments to the definition of “revenue.”) 3/8/17 Tr. 161–64 (Levine); 37 CFR 385.21.

Copyright Owners urge the elimination of the subpart C provisions as essentially obsolete because locker services for “purchased content” (new download purchases) and for “paid” downloads (already owned) have largely disappeared, as listeners transitioned away from ownership models to access models. See 3/8/17 Tr. 159–160 (Levine); 3/16/17 Tr. 1458–1461 (Mirchandani); Mirchandani WDT ¶ 33; 3/22/17 Tr. 2523 (Dorn). Copyright Owners also re-assert the same arguments with respect to subpart C as they have for interactive streaming in subpart B. See CORPFF–JS at p.2.

The Services argue that Copyright Owners do not point to any evidence to show that locker services have completely disappeared, emphasizing that Apple and Amazon continue to offer locker service. Joyce WDT ¶ 5; Mirchandani WDT ¶¶ 16–17; 3/22/17 Tr. 2523–25 (Dorn); Ramaprasad WDT, Table 3. More generally, the Services urge the Judges to use the subpart C rate structure as the benchmark for rates in the forthcoming period for the same reasons as they urge the use of the subpart B rates as an appropriate benchmark. See Mirchandani WDT ¶¶ 58–62.

The Judges find no reason on remand to treat the subpart C offerings differently than the manner in which they are treating the subpart B interactive streaming offerings, for the reasons set forth in the Dissent at 118–119. That means, however, that the various “headline” rates for these subpart C offerings must also adjust to 15.1%, whereas the alternative rates (identified in subpart C as “minima” and “subminima)” rates shall remain unchanged.

IV. Change in Definition of Service Revenue for Bundles

The Judges analyze the definition of “Service Revenue” for bundled offerings in the context of the partial adoption of the PR II-based benchmark. As discussed supra, the Judges have found that the PR II-based benchmark is a useful benchmark, particularly because of its features that incentivize beneficial downstream price discrimination that generates more listeners, revenues, and royalties.

A. Background

In their Initial Determination, the Judges adopted a definition of “Service Revenue” ( i.e., a royalty base) for a “Bundle” that provided, in pertinent part:

Service Revenue shall be the revenue recognized from End Users for the Bundle less the standalone published price for End Users for each of the other component(s) of the Bundle . . .

Initial Determination, Attachment A at 7 (§ 382.2 therein).

After the Judges issued their Initial Determination, Copyright Owners submitted a Motion for Clarification or Correction of Typographical Errors and Certain Regulatory Terms which disclaimed any intent to seek rehearing, but sought “clarification or correction” of certain regulatory terms to conform them to what Copyright Owners claimed to be the apparent intent of the Initial Determination. (Motion for Clarification). Copyright Owners purported to bring their motion under the Judges' general regulations governing motions. See37 CFR 303.3 and 303.4 (formerly codified at 37 CFR 350.3 and 305.4).

The Motion for Clarification argued, among other things, that the definition of Service Revenue as applied to bundled offerings should be reworked. Copyright Owners argued that defining the revenue as the total price of the bundle, minus the standalone published prices for the non-streaming offerings in the bundle, undervalued the revenue created by the streaming offerings. They proposed that “Service Revenue” for bundled offerings be defined as the standalone price of the offering (or comparable offerings).

The Services objected to Copyright Owners' styling of their motion as something other than a motion for rehearing. The Services also objected that Copyright Owners had not previously proposed a definition of “Service Revenue” for bundled offerings, and that their “late-proposed” definition was unsupported by the record.

On October 29, 2018, the Judges issued an Order concluding neither party had met the exceptional standard for granting rehearing motions, stating that the parties had failed to present “even a prima facie case for rehearing under the applicable standard”. Amended Order Granting in Part and Denying in Part Motions for Rehearing (Order on Rehearing) (Jan. 4, 2019).

The Judges explained that they nevertheless found it appropriate to resolve the issues that the parties had raised. Order on Rehearing at 2. The Judges added that, to the extent such resolution could be considered a rehearing under 17 U.S.C. 803(c)(2), the Judges resolved the motions on the papers without oral argument. Id.

Regarding the definition of “Service Revenue” for bundled offerings, the Judges summarized the parties' competing arguments:

Copyright Owners presented evidence that the existing approach led, in some cases, to an inappropriately low revenue base—but did so in service to their argument that the Judges should reject revenue-based royalty structures. They did not present evidence to support a different measure of bundled revenue because their rate proposal was not revenue-based. The Services rely on the fact that the approach to bundled revenue in the extant regulations is derived from the 2012 Settlement. The Judges have, however, declined to rely on the 2012 Settlement as a benchmark, as the basis for the rate structure, or, therefore, as regulatory guidance.

The Services have observed correctly that the evidentiary records in Web IV and SDARS III differ from the record in this proceeding.

Order on Rehearing at 17 (emphasis added).

Despite these arguments, the Judges found that neither party presented evidence adequate to support the approach advocated in post-determination filings, because “the `economic indeterminacy' problem inherent in bundling” remained unresolved.” Id. The Judges stated that the Services were the party in possession of the relevant information, and concluded that the Services bore the burden of providing evidence that might mitigate the “indeterminacy problem” inherent in bundling. Because the Judges concluded that the Services had not met that burden, they ruled that they must adopt an approach to valuing bundled revenue that is in line with what the Copyright Owners proposed. As a result, the Judges discarded the formula in the Initial Determination and ruled, instead, that streaming service providers will use their own standalone price (or comparable) for the music component (not to exceed the value of the entire bundle) when allocating bundled revenue. Id. at 16–18.

Consistent with the Judges' Order on Rehearing, the Judges' replaced the definition of “Service Revenue” for a “Bundle” that they had included in the Initial Determination with a new definition in the Determination. The final definition provided, in pertinent part:

Service Revenue shall be the lesser of the revenue recognized from End Users for the bundle and the aggregate standalone published prices for End Users for each of the component(s) of the bundle that are Licensed Activities . . . [or] if there is no [such] standalone price, then the average standalone . . . price . . . for the most closely comparable product or service . . . or . . . the average of standalone prices for comparables.

Determination, Attachment A at 8.

The Services, Copyright Owners and George Johnson appealed the Judges' Determination to the D.C. Circuit. See Johnson, 969 F.3d 363. The Services challenged both the Judges' legal authority and the substantive soundness of the decision to reformulate the definition of “Service Revenue” for bundled offerings, after the Judges had issued the Initial Determination.

The D.C. Circuit examined several authorities under which the Judges may revisit and amend a determination. It addressed the three ways identified in the statute: “(i) order rehearing `in exceptional cases' in response to a party's motion, 17 U.S.C. 803(c)(2)(A); (ii) correct `technical or clerical errors,' id. § 803(c)(4); and (iii) `modify the terms, but not the rates' of a royalty payment, `in response to unforeseen circumstances that would frustrate the proper implementation of [the] determination.' ” Johnson, 969 F.3d at 390. The D.C. Circuit found that the Judges' reformulation of the definition of “Service Revenue” fit none of those categories.

The D.C. Circuit noted that the Judges were explicit that they did not treat the Motion for Clarification as a motion for rehearing under 17 U.S.C. 803(c)(2). Id. Furthermore, the D.C. Circuit noted the Judges' own findings that the Motion for Clarification did not meet the exceptional standard for granting rehearing motions under section 803(c)(2) and that the Copyright Owners failed to make even a prima facie case under the rehearing standard.

In Johnson, the D.C. Circuit found that the change to the definition of Service Revenue for bundled offerings was not an exercise of the Judges' authority under section 803(c)(4) to “correct any technical or clerical errors in the determination[.]” 17 U.S.C. 803(c)(4). The D.C. Circuit observed the substantive nature of the change to the definition and determined that there was nothing technical or clerical about the amendment. The D.C. Circuit found that the Judges did not even purport to modify the terms in response to unforeseen circumstances that would frustrate the proper implementation of the Initial Determination. The D.C. Circuit observed that the Judges never mentioned section 803(c)(4) or unforeseen circumstances as the basis for revamping the Service Revenue definition.

Beyond the explicit statutory authorities for amendments to determinations, the D.C. Circuit addressed arguments for inherent authority to make sua sponte any appropriate substantive or fundamental changes after the Initial Determination. The D.C. Circuit foreclosed reliance on inherent authority, finding that Congress's decision to limit rehearing to exceptional cases, and to confine other post hoc amendments to cases involving technical or clerical errors, would be a nullity if the Judges also had plenary authority to revise their determinations whenever they thought appropriate. The D.C. Circuit noted that the Judges' decision to amend the definition said nothing of the sort, and prior decisions are silent on that topic.

In sum, the D.C. Circuit found that the Judges failed to explain the legal authority for reformulating the definition of “Service Revenue.” In relevant part, the D.C. Circuit ruled

we must vacate the [ ] Determination's bundled offering Service Revenue definition and remand for the [Judges] . . . either to provide `a fuller explanation of the agency's reasoning at the time of the agency action[,]' or to take `new agency action' accompanied by the appropriate procedures.

Id. at 392 (citing Regents, 140 S.Ct. at 1908).

Because the D.C. Circuit determined that the Judges failed to identify any legal authority for adopting the new Service Revenue definition, it found no occasion to address the Streaming Services' separate argument that the definition was arbitrary, capricious, or unsupported by substantial evidence. Id.

The Services and Copyright Owners agreed that the Judges should resolve the definitional issue based on the existing record, after receiving two rounds of additional briefing from the parties. See Services' Proposal for Remand Proceedings (Dec. 10, 2020) (Services' Proposal) at 5–6, 9–10; Proposal of the Copyright Owners for Conduct and Resolution of the Remand (Public) (Dec. 10, 2020) (Copyright Owners' Proposal) at 4–6. The Judges issued an Order Regarding Proceedings on Remand, which, in part, opened briefing on the issue of the adoption of a revised definition of “service revenue” for bundled offerings between issuing the Initial Determination and the Determination. Order Regarding Proceedings on Remand (Dec. 15, 2020). The Judges received the following relevant briefing.

• CO Initial Submission
• Services' Initial Submission
• CO Reply
• Services' Reply

On December 9, 2021, the Judges requested additional briefing. Dec. 9 Order. The Dec. 9 Order sought additional briefing setting forth the parties' views on whether this proceeding constitutes the type of new agency action addressed by the D.C. Circuit, which would allow adoption of a Service Revenue definition without limitation to the definition expressed in the Initial Determination. Additionally, the Judges requested additional evidence that the parties might offer to support adoption of the Service Revenue definitions expressed in either the Initial Determination or the Determination. In response to the Dec. 9 Order, the Judges received the following relevant briefing.

• CO Additional Submission
• Services' Additional Submission

On February 9, 2022, the Judges solicited further briefing on “Whether the D.C. Circuit's Johnson decision permitting the Judges to engage in new agency action in this remand proceeding allows the Judges to engage in new agency action through a reconsideration of Copyright Owners' February 12, 2018 Motion for Clarification as a Motion for `rehearing' pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1.” Sua Sponte Order Regarding Additional Briefing (Feb. 9 Order). In response to the Feb. 9 Order, the Judges received the following relevant briefing.

• Copyright Owners' Brief Responding to Judges' February 9, 2022 Sua Sponte Order Regarding Additional Briefing on New Agency Action Question, and Replying to Services' New Agency Action Arguments in their Joint Supplemental Brief Addressing the Judges' Working Proposal (in Additional Materials Rebuttal Submission of Copyright Owners at Tab B) (Feb. 24, 2022) (“CO Further Briefing”)

• Services' Joint Response to the Judges' February 9, 2022 Sua Sponte Order Regarding Additional Briefing and Rebuttal Regarding “New Agency Action” (Feb. 24, 2022) (“Services' Further Briefing”)

B. Authority for Modification to the Initial Determination

1. Copyright Owners' Position

Copyright Owners assert that this remand proceeding offers a straightforward path to take new agency action and that the law makes clear that new agency action can consist of issuing a new determination on remand. CO Initial Submission at 71. Copyright Owners maintain that:

[T]the new agency action here is a determination after remand proceedings, the Board is largely free to chart its own procedural course, and the Board has done so in its December 15 Order. The Board is not required to undertake any of the procedural steps set forth in 17 U.S.C. 803(b) in order to take such “new agency action.” See 17 U.S.C. 803(d)(3) (requiring only that on remand further proceedings be taken “in accordance with subsection (a)”); 37 CFR 351.15; Intercollegiate Broad. Sys., Inc., 796 F.3d at 125 (“[N]either the Copyright Act nor the Board's regulations prescribe any particular procedures on remand.”) The Circuit's instruction that the action be “accompanied by the appropriate procedures[,]” Johnson, 969 F.3d at 392, does not dictate what those “appropriate procedures” must be but instead plainly refers to these flexible rules. See also Oceana, Inc., 321 F. Supp. 3d at 136 (explaining that when remanding to an agency, a court generally “may not dictate to the agency the methods, procedures, or time dimension, for its reconsideration”).

CO Initial Submission at 71, FN 33.

Copyright Owners acknowledge the Services' position that the asserted procedural error is an “absence of authority” that can never be cured. Id. at 74 (citing Services' Proposal for Remand Proceedings at 10). They note that the D.C. Circuit did not say the Judges lacked the authority to revisit the service revenue definition for bundles on remand. Nor, they observe, did it say the Judges have no authority to review the record evidence and the parties' arguments and reach the same conclusion or a different conclusion on remand. Copyright Owners opine that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. But instead, they note, the D.C. Circuit remanded and said the Judges could take new agency action precisely to cure the asserted procedural defect. Copyright Owners assert that the remand allowed the parties to present the record evidence and their arguments so that the Judges can address the definition “afresh” in the remand determination. Id. at 74.

Copyright Owners argue that 17 U.S.C. 803(d)(3) states only that proceedings on remand must be in accordance with 17 U.S.C. 803(a). They contend that remand proceedings need not be confined to procedures the Services claim are too late in the game for the Judges to follow. The Copyright Owners point to the D.C. Circuit's ruling in Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Bd., that “neither the Copyright Act nor the Board's regulations prescribe any particular procedures on remand.” 796 F.3d 111, 125 (D.C. Cir. 2015) (citing 17 U.S.C. 803(a), (d)(3)). Accordingly, they argue, the Judges can reaffirm the adopted bundled service revenue definition following their review of the parties' submissions without regard to section 803(c)(2) or 803(c)(4). CO Reply at 65–66.

Copyright Owners further argue that the Judges may properly justify the changed definition under section 803(c) as a fuller explanation of the agency's reasoning at the time it was made. They urge that the Judges could explain that, especially in light of the evidence of how the Services misused the prior definition to make service revenue completely disappear, carrying over the prior bundle service revenue from Phonorecords II into the Initial Determination was unintended and inadvertent. CO Reply at 69. Copyright Owners also assert that the Judges could explain that Copyright Owners had, in their Motion for Clarification, identified an “exceptional case” under section 803(c)(2) because the prior definition failed to comport with Judges' precedent and economic principles, and was unsupported by evidence. In addition,

Copyright Owners note that the Judges reheard the evidence and legal arguments as presented in the parties' briefs on the issue and, as a result, may choose to adopt the revised definition. Copyright Owners maintain that for the Judges to do so would not be impermissible post-hoc reasoning, because the D.C. Circuit remanded precisely because the Judges did not provide any reason in the Determination for revising the bundle revenue definition. CO Reply at 69–71.

2. Services' Position

The Services assert that the D.C. Circuit found only “three ways in which the Board can revise Initial Determinations” and that the Judges had failed to establish that the change to the service revenue definition fit any of those three categories. Services' Initial Submission at 64–65 (citing Johnson at 390).

According to the Services the first way the Judges may revise an Initial Determination is to “order rehearing `in exceptional cases' in response to a party's motion, 17 U.S.C. 803(c)(2)(A).” Services' Initial Submission at 65 (citing Johnson at 390). The Services argue that the D.C. Circuit held in Johnson that the Judges' “material revision of the `Service Revenue' definition for bundled offerings does not fall within the Board's rehearing authority under section 803(c)(2)(A)” because “the Board itself . . . was explicit that it `did not treat the [Copyright Owners'] motion[ ]' . . . `as [a] motion[ ] for rehearing under 17 U.S.C. 803(c)(2).' ” The D.C. Circuit also noted that “as the Board found, the Copyright Owners' motion did `not meet [the] exceptional standard for granting rehearing motions' under section 803(c)(2).” Id. (citing Johnson at 390). The Services assert that the Judges were not able to make “a volte-face” and justify on appeal their revision to the definition as an exercise of rehearing authority. As the D.C. Circuit held, agency action must be justified by “reasons invoked by the agency at the time it took the challenged action,” and post-hoc rationalizations are insufficient. Id. (citing Johnson at 390).

The Services add their view that the Judges cannot revisit the decision to deny rehearing without engaging in impermissible post-hoc reasoning. They note that the Supreme Court has explained that, while an agency may “elaborate later” on its “initial explanation” of the reason (or reasons) for its action, it “may not provide new ones.” Services' Initial Submission at 66 (citing e.g., Regents, 140 S. Ct. at 1908). The Services offer that the Judges, having stated that they did not consider the Copyright Owners' motion to revise the definition to be a motion for rehearing, cannot now conclude that the motion qualified as one for rehearing and that the Judges in fact engaged in rehearing. Id.

The Services add that under section 803(c)(2)(A), the Judges can only use their rehearing authority “ `in exceptional cases' in response to a party's motion.” Id. (citing Johnson at 390). The Services argue that the Motion for Clarification cannot be found to have satisfied that standard. The Copyright Owners did not argue that their motion satisfied the “exceptional cases” standard before the Judges or the D.C. Circuit, and have therefore waived that argument. Id.

According to the Services, the second way the Judges may revise an Initial Determination, viz. action to correct a technical or clerical error under section 803(c)(4), cannot be used now to justify any modification of the Service Revenue definition in the Initial Determination. The Services note that the D.C. Circuit held specifically that the Judges' change to the Service Revenue definition could not be construed as correcting a technical or clerical error because it involved a substantive rewrite of the Service revenue definition. Id. at 67 (citing Johnson at 391).

The Services aver that the third way the Judges may revise the terms in an Initial Determination is in response to unforeseen circumstances that would frustrate the proper implementation of the determination. Id. at 67. The Services note that the D.C. Circuit held in Johnson that this authority did not justify the Judges' change to the Service Revenue definition because the Judges did not invoke this authority and “the need to ground the original definition in the record” could not credibly be described as “an unforeseen circumstance.” Id. (citing Johnson at 391).

The Services also note that the D.C. Circuit rejected the argument that the Judges have “inherent authority” to make changes to the Initial Determination. The D.C. Circuit explained that the specific restrictions Congress placed on the Judges' authority in section 803 “would be a nullity if the Board also had plenary authority to revise its determinations whenever it thought appropriate.” Id. (citing Johnson at 391–92). The Services add that even if the Judges offered a new source of authority capable of justifying substantive changes to the Service Revenue definition now, the Judges would be unable to rely on this “uninvoked authority” without engaging in impermissible post-hoc reasoning. Id.

The Services counter Copyright Owners' position that the Judges need not respond to the error the D.C. Circuit identified with this aspect of the Determination and that the Judges' “new agency action” may consist of issuing a new determination on remand. The Services argue that failure to address the legal and factual issues on which the D.C. Circuit remanded would violate the D.C. Circuit's order and would result in a second remand. The Services surmise that the issue of authority to make the changes to the Initial Determination are particularly important in this context, where the D.C. Circuit recognized that the Copyright Act places limits on the Judges' authority to alter an initial determination by defining conditions for rehearing and the types of changes that are permitted absent a rehearing. In this regard, the Services maintain that the Judges cannot do on remand what they lacked authority to do in the first instance. The Services assert that the Judges must resolve the legal question whether there is authority to alter the revenue definition in the Initial Determination. They urge that the remanded issue is not what the substance of the service revenue definition should be as a matter of first impression, but instead is whether the Judges have properly exercised authority to alter the Initial Determination's definition. Services Reply at 52–54.

The Services assert that the Judges have two paths available to them: (1) to provide a “fuller explanation” of the prior conclusion that the Judges had legal authority to revise the Service Revenue definition in the Initial Determination or (2) answer that threshold question through new agency action. The Services maintain that, if they pursue the “fuller explanation” path, the Judges are limited to elaborating on what they said previously, and that they cannot add new reasons they did not initially provide. With regard to what may constitute new agency action, the Services assert that path gives the Judges freedom to consider new reasons that the Copyright Act provided the Judges with the authority to make this change to the Initial Determination. The Services argue, however, that undertaking a new agency action does not, as Copyright Owners claim, obviate the need for the Judges to identify proper legal authority before substantively changing the Initial Determination, such authorities being limited to the authority of section 803(c)(4) or the rehearing authority of section 803(c)(2). Id. at 54–55.

The Services address Copyright Owners' position that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. The Services urge that the D.C. Circuit could not reverse because the Department of Justice raised for the first time on appeal new justifications for the Judges' decision to change the Initial Determination. Instead, the Services maintain, the D.C. Circuit had to remand and give the Judges the opportunity to address the Department of Justice's new justifications in the first instance, as the D.C. Circuit could not rule them out given the posture of the appeal. Id. at 56.

In the Services' Additional Submission, they concede that this remand proceeding is new agency action and that the Judges have provided the parties with sufficient procedural opportunities to present any new evidence and raise any additional arguments regarding the question the D.C. Circuit remanded. Services' Additional Submission at 38. But the Services still insist that the Judges may not alter the Service Revenue definition without first identifying legal authority in the Copyright Act for modifying the Initial Determination. In the Services' view the new agency action avenue provided by the D.C. Circuit merely offers a singular path beyond the Judges' ability to offer a “fuller explanation” of their previous reasoning for revisiting the definition in the Rehearing Order. According to the Services' argument, the new agency action provided for in this remand only offers the additional opportunity to offer new reasons supporting any legal authority for altering the Initial Determination's Service Revenue definition, beyond those that were raised in the appeal. Services' Additional Submission at 38–42

C. Reconsideration of Motion for Clarification as Motion for “Rehearing”

1. Copyright Owners' Position

Copyright Owners argue that the Judges have the authority to engage in new agency action in this remand proceeding through a reconsideration of the Motion for Clarification as a motion for rehearing, pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1. Copyright Owners urge, however, that proceeding in that fashion would add an entirely unnecessary and complicating step. They again suggest that there is no need to reconsider or recharacterize the Motion for Clarification as a motion for rehearing because the remand itself affords the opportunity for the Judges to take new agency action, which, as in a rehearing, permits them to reconsider evidence and arguments, but, unlike a rehearing, is not limited by the constraints of section 803(c)(2). CO Further Briefing, Tab B at 7–8.

Copyright Owners posit that if the Judges engage in new agency action to reconsider the Motion for Clarification as a motion for rehearing under 803(c), and to decide that motion based on all of the evidence in the record supporting the adopted bundle revenue definition and showing the prior bundle revenue definition to be unsupported and unreasonable, they may properly do so. They assert that the while they did not make a request for rehearing on the face of the Motion for Clarification, that is not the same as a finding that the standard could not have been met. The Judges may consider whether, based on the evidence in the record, the rehearing standard has been satisfied on this remand. In Copyright Owners' view, the Judges could conclude, revisiting on remand the question of whether the rehearing standard has now been met, that Copyright Owners have satisfied the “exceptional case” standard for granting rehearing motions under section 803(c)(2). Copyright Owners note that if the Judges do engage in new agency action that reconsiders the Motion for Clarification as a motion for rehearing, the Judges should fully explain their reasoning. Id. Tab B at 8–10.

2. Services' Position

The Services assert that the Judges cannot invoke their rehearing authority by construing the Motion for Clarification as a rehearing motion. They maintain that the D.C. Circuit expressly found that the revision of the Service Revenue definition for bundled offerings does not fall within the Judges' rehearing authority under section 803(c)(2)(A). The Services assert that Copyright Owners did not satisfy either prong of section 803(c)(2)(A), which authorizes rehearing only “upon motion of a participant” and “in exceptional cases.” They note that the D.C. Circuit agreed with the Judges' decision not to treat Copyright Owners' motion as one for rehearing and that the D.C. Circuit also agreed with the Judges' further finding that “Copyright Owners' motion did not meet the exceptional standard for granting rehearing motions.” Services' Further Briefing at 7 (citing Johnson at 390).

The Services add their view that the Judges are bound by the D.C. Circuit's conclusions on this issue. They maintain that because the Judges' section 803(c)(2)(A) rehearing authority is among the grounds that Johnson addressed and determined, the Judges cannot rely on that authority on remand. Id. at 8–9. The Services urge that the Judges already correctly concluded that the Motion for Clarification was not a motion for rehearing, and note that Copyright Owners never presented their motion as one for rehearing. The Services add that because Copyright Owners did not challenge that decision on appeal, it is too late for them to do so now. Id. at 9–10.

The Services argue that Copyright Owners' Motion did not make any attempt to satisfy the exceptional cases standard set out in 17 U.S.C. 803(c)(2)(A). They argue that Copyright Owners did not purport to identify any new evidence, new legal authority, or even a substantive error in the Judges' reasoning in the Initial Determination, but instead the motion asserted that the Judges' inclusion of the definition of service revenue in the Initial Determination was supposedly inadvertent. The Services add that Copyright Owners did not identify any specific evidence in the Phonorecords III record or any aspect of the Initial Determination that suggested the inclusion of this definition was a mistake. Id. at 10.

The Services point out that Copyright Owners' motion did not comply with the procedural requirements for a motion for rehearing. They then urge that the Judges cannot invoke their section 803(c)(2)(A) authority by rewriting a participant's motion to say it is seeking rehearing when that participant specifically and unambiguously disclaimed any intent to seek rehearing. Id. at 11.

The Services note that the Judges previous conclusion that even if the Motion for Clarification had requested rehearing, that motion would not and does not meet that exceptional standard for granting rehearing and failed to make even a prima facie case for rehearing. The Services observe that the Judges apply a strict standard to rehearing motions to prevent parties from using the rehearing process to seek a second bite at the apple by advancing theories and arguments that could have been advanced earlier during the proceeding. Id. at 12. The Services reiterate their view that Copyright Owners' motion did not point to any evidence in the Phonorecords III record at all, and, that the only evidence in the Phonorecords III record concerning bundles supports the longstanding definition of Service Revenue which has been effective in encouraging the Services to offer bundles that benefit Copyright Owners by growing the market for music streaming services. Id. at 14.

The Services finally assert that this is not an extraordinary case where a party has identified an error that, if left uncorrected, would result in manifest injustice. Id. at 15–16. The Services conclude by urging that given this procedural history and the unchanged state of the record since the initial hearing, any claim that Copyright Owners have somehow now satisfied the exceptional case standard would be clear error. Id. at 17.

D. Record Evidence Regarding Definition of Service Revenue

1. Copyright Owners' Position

Copyright Owners assert that the prior bundle revenue definition (published in the Initial Determination) failed to address the “ `economic indeterminacy' problem inherent in bundling” appropriately and in a way consistent with Judges' precedent. CO Initial Submission at 75 (citing Order on Rehearing at 16–18). Copyright Owners proceeded to cite several portions of testimony from the Services' economic experts who acknowledged this problem. Id. They then point to hearing testimony in which Copyright Owners repeatedly raised the “economic indeterminacy” problem and demonstrated what they characterized as the absurd results to which the prior definition had led. Id. at 76. They point out that under the prior definition, service revenue for bundled subscriptions started with revenues recognized from the bundle ( i.e., the price paid by the subscriber) and subtracted “the standalone published price” for all non-music components of the bundle. [REDACTED]. Id.

Copyright Owners point out that the Judges already found with respect to other licenses that such an approach is not only fundamentally unfair, but “absurd.” Id. (citing 81 FR 26316, 26382 (May 2, 2016) (webcaster licenses)); see also83 FR 65210, 65264 (Dec. 19, 2018) (SDARS licenses) (rejecting proposed deductions by service for bundle revenues because of the “acknowledged `economic indeterminacy' problem inherent in bundling”). The Copyright Owners concur with the Judges' correct conclusion that the same reasoning applies to Phonorecords III. Id. at 76–77 (citing Order on Rehearing at 18) (“the `economic indeterminacy' problem inherent in bundling is common to all three proceedings.”). The Copyright Owners offer that Spotify conceded to this flaw in the definition in the Initial Determination, but offered an alternative that contained the same loophole. Id. at 77–78.

Copyright Owners point out that the proponent of a term bears the burden of proof as to adoption. The Judges made clear that the licensee who wishes to offer bundles 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 the licensors. Id. at 79 (citing SDARS III Determination, 83 FR 65210, 65264) (“bundling [is] undertaken to increase [the Services'] revenues and it would be reasonable to assume that [the Services have] information relevant to the economic allocation of the bundled revenue.”). The Copyright Owners contend they presented unrebutted evidence showing the unreasonableness of the Services' proposed definition while the Services offered no evidence to support their definition. Id. at 78, 79 (citing Order on Rehearing at 18). Copyright Owners maintain that no Service offered evidence concerning the separate values of the constituent parts of the bundles, or any other evidence concerning the economic allocation of bundled revenue, let alone the reasonableness of the definition in the Initial Determination. Id. at 80. Copyright Owners assert that in the absence of evidence to support the proposed definition, the Judges may adopt or fashion a definition of service revenue for bundled offerings that comports with the record evidence, which is precisely what the Judges did and can, through new agency action, do again. Id. at 81.

Copyright Owners dispute the Services' assertion that there is support for the Phonorecords II approach to bundles in the record of this proceeding. Instead, Copyright Owners argue, the Services' purported evidence at most supports the benefits of the practice or strategy of bundling. They maintain that the strategy of bundling covered music services with other products or services has nothing to do with whether the Services should be free to reduce the revenue allocable to music to zero. They offer that the definition in the Initial Determination has nothing to do with such benefits, and that those benefits may be equally served by a definition that ensures value is apportioned to the music component in the bundle. CO Reply at 73–76.

2. Services' Position

The Services argue that the evidence in the existing written record addressing bundles shows both that this definition is supported by the Phonorecords II benchmark and that it has proven, industry-wide benefits. Services' Initial Submission at 68. They offer that the Copyright Owners did not propose an alternative definition of service revenue until after the Judges issued the Initial Determination and that any definition they propose now would fail the basic requirement that the Judges must adopt rules “on the basis of a written record.” Id. (citing 17 U.S.C. 803(a)(1) and 803(c)(3)).

Addressing the merits of the definition contained in the Initial Determination, the Services argue that it best serves the goals of the Copyright Act; that as a bright-line, easily administered rule, it continues the broad industry agreement from Phonorecords II. The Services contend the prior definition increases output and incentivizes beneficial price discrimination to reach listeners who would otherwise not pay for music. They argue that the record evidence confirms that the prior treatment of bundles enabled experimentation and variation in the distribution of music with long-term benefits for all parties. They state that Copyright Owners' argument that Services [REDACTED] also demonstrates the broad benefits of the definition of Service Revenue in Phonorecords II because the record showed that arrangement enabled funneling of many of listeners into full-priced, full-catalog services—such treatment of bundles enabled the flexibility and price discrimination that yielded beneficial growth of the royalty pool. The Services allege that Copyright Owners also ignore the extensive royalties that were generated. They add that with the per-subscriber minimum guarantees that the Copyright Owners will still be paid a fair royalty. The Services then cite several portions of testimony from various Services' economic experts who point out the realization of an expanded royalty pool, which the Services offer as proving a functioning marketplace. Id. at 68–74.

The Services then assert that no other definition of service revenue for bundles that has been before the Judges combines both the administrative simplicity of the Initial Determination's definition and the broad price discrimination benefits of promoting discounted bundles. They maintain that while neither the Services nor Copyright Owners submitted evidence specifically addressing the way that customers, Services, or Copyright Owners might value the component parts of bundles, such subjective valuations are unnecessary for the Judges to find ample support for the Phonorecords II approach to bundles in the record. Id. at 75–76.

The Services also argue that while the Judges' decision in SDARS III did involve valuation of the music and non-music components of a bundle, the resolution in SDARS III is inapposite because, here, the rate structure has a way of ensuring that Copyright Owners are fairly compensated for bundles: the statutory minimum payment. Services Reply at 62.

E. Analysis and Conclusions Regarding Definition

1. Remand Proceeding as New Agency Action

Having considered the entirety of the record of this proceeding, a majority of the Judges (Definition Majority) conclude that this remand constitutes “new agency action” and meets all of the criteria to qualify as new agency action. The Judges thus have the opportunity to consider the issue afresh consistent with their procedural rules regarding remands.

The Definition Majority finds that it is unnecessary to attempt to distinguish new “agency action” from “new agency action.” Neither approach is endorsed clearly by the varied judicial interpretations of a new agency action. See R.J. Krotoszynski, Jr., Administrative Law Discussion Forum: “History Belongs to the Winners”: the Bazelon-Leventhal Debate and the Continuing Relevance of the Process/Substance Dichotomy in Judicial Review of Agency Action, 58 Admin. L. Rev. 995 (Fall 2006). As noted by Judge Bazelon, the D.C. Circuit “believed in process-based review, [but] he argued that it was improper for judges to prescribe specific procedures.” Id. at 1001. Judge Bazelon's remand orders focused on providing “genuine opportunities to participate in a meaningful way” and “genuine dialogue” with interested parties, while leaving the agency “free to decide which specific procedures to undertake.” Id.

Several reported cases point to new action as an alternative to a fuller explanation. But few define “new agency action” other than to say, as did the Johnson court, that the agency must take it “accompanied by the [unspecified] appropriate procedures.” Johnson, 969 F.3d at 392. Parties to the original action, already familiar with the issue and the factual and legal background, recognized that the D.C. Circuit identified the adoption of a modified definition in the Determination as one of three issues on remand. In repeated rounds of remand submissions, both the Services and the Copyright Owners included the definition issue. The Judges were not satisfied with the parties' lack of focus on the issue, however, and ordered expressly further briefing on the new agency action issue and sub-issues relating to the adoption of a definition of Service Revenue as it relates to bundled service offerings. See (Dec. 9 Order) at 4; Sua Sponte Order Regarding Additional Briefing (Feb. 9, 2022).

New agency action is not synonymous with justification, or confirmation, of the prior action. New agency action is a procedural mechanism for reconsideration of the record, reopening the record for additional evidence and argument, and adoption of a conclusion based on the expanded record. In this instance, the presentations, written and oral, of participants on remand, together with a re-examination of the original record, support reversion to the definition originally announced in the Initial Determination. Ultimately, given repeated opportunities for legal analysis on the issue, both sides agreed that the remand proceeding itself, with ample notice and multiple opportunities for input was sufficient to constitute new agency action. See CO Further Briefing at 3, 7.

The Services argued, however, that notwithstanding this appropriate new agency action, the Judges remained without authority to adopt the revised definition as a term governing the royalty rates determined in this proceeding. Their arguments regarding procedures undertaken in the Determination are superseded by the Judges' conduct of extensive remand proceedings. The gravamen of the Administrative Procedure Act is transparency in agency rulemaking. Agencies must publish notice of their intentions, provide opportunities for interested parties to comment and object, and finalize regulations only after reconciling objections with the policies and purposes of proposed regulations. The adjudication of this remand proceeding was conducted openly. Interested parties had ample opportunity to object, to comment, and to brief legal and factual issues relating to the Judges' approach to promulgating an appropriate definition of bundled service revenue.

The present analytic approach merely takes the position that the Judges engaged in new agency action by conducting a fully open and broadly explored remand proceeding. Unlike a rehearing or exercise of continuing jurisdiction, this remand proceeding is not limited by the constraints of sections 803(c)(2) or 803(c)(4). Contrary to the Services' assertion, the Judges address the issue on which the D.C. Circuit remanded, the need to exercise authority within the lines drawn by the authorizing statute. This remand proceeding does not, therefore, violate the D.C. Circuit's order.

The Johnson opinion clearly states the two paths by which the Judges may address the issues presented to them on remand; they may either (1) provide “a fuller explanation of the agency's reasoning at the time of the agency action[,]” or (2) to take “new agency action” accompanied by the appropriate procedures. Johnson, 369 F.3d at 392. The Judges chose to pursue the second option: this new agency action. The Judges reiterate: the Services concede that, through this proceeding the Judges have provided the participants with adequate procedural opportunities to present any new evidence on the proper Service Revenue definition for bundles. The Judges also acknowledge, but disagree with, the Services' position that that they must return to the issues as they were presented after issuance of the Initial Determination, regardless of the admittedly complete and valid remand procedure, which constitutes new agency action.

The Judges (the majority on this issue) determine that any confining action on remand to the provisions of sections 803(c)(2)(A) or 803(c)(4) would misconstrue the clear expression of the “new agency action” alternative presented by the D.C. Circuit, as well as chapter 8 of title 17. As the Copyright Owners correctly observed, in a remand proceeding, the Judges are not required to undertake any of the procedural steps set forth in section 803(b) nor are the Judges compelled to consider or be limited by sections 803(c)(2)(A) or 803(c)(4). The statute only requires that the Judges' remand proceedings are in accordance with section 803(a).

The D.C. Circuit observed that the Judges have “considerable freedom to determine [their] own procedures.” SoundExchange v. CRB, 904 F.3d 41 at 61. The D.C. Circuit also cautions that such flexibility must be exercised within the lines drawn by the authorizing statute. Here, the Judges operate within the lines drawn with respect to remand proceedings set forth in chapter 8 of title 17.

2. “Fuller Explanation” of Modification to Initial Determination

Case law regarding development of a “fuller explanation” of an agency's action emphasizes that the agency cannot adopt post hoc reasoning on the same record. See, e.g., SEC v. Chenery Corp., 332 U.S. 194, 201 (1947) (after remand, agency bound to “deal with the problem afresh . . . .”). Certainly, adopting a post hoc argument of appellate counsel, just because it offers a rationale for the agency's original action is impermissible. On the other hand, if the record in the initial proceeding is sufficiently robust to support a reinterpretation or additional reasoning, the agency may justify its initial action with that “fuller explanation” without considering any new evidence. See, Fisher v. Pension Benefit Guar. Corp., 468 F.Supp.3d 7, 20 (D.C.D.C. 2020), aff'd Fisher v. Pension Benefit Guar. Corp., 994 R.3d 664 (D.C. Cir. 2021), rehearing en banc denied, Fisher v. Pension Ben. Guar. Corp., 2021 U.S. App. LEXIS 18793 (D.C. Cir., June 23, 2021) (requirement of new evidence a “novel proposition of law” without precedent). On remand, an agency may elaborate on its prior reasoning, but it may not provide new reasons for the original decision. Fisher, 994 F.3d at 669. If the Judges had chosen in this remand to rest on their Determination regarding the service revenue definition, they might have done so only if they could elaborate on the existing record. In the alternative, the Judges issue a new decision after new agency action. Id.

The Judges, having engaged in new agency action to settle on the definition of service revenue for bundled offerings, do not find a need to address the statutory avenues or the confines that are provided for rehearing or continuing jurisdiction, nor do the Judges pursue the propriety of reconsideration of the Motion for Clarification as a motion for rehearing.

3. Substantive Analysis of Dueling Definitions of Bundled Revenue

The fundamental difference between the impact of the two alternative definitions is simply stated:

Under the Initial Determination: downstream bundling and its price discriminatory effect would be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle;

Under the Determination: downstream bundling and its price discriminatory effect would not be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle.

To explain this difference, the Judges find it helpful to describe (as in the Determination and Dissent) how bundling facilitates price discrimination and how lower royalties for bundled streaming services incentivize such bundling.

Price discrimination occurs when a seller offers different units of output at different prices. See, e.g., H. Varian, Intermediate Economics at 462 (8th ed. 2010). The benefit to the seller arises from attempting to “charge each customer the maximum price that the customer is willing to pay for each unit bought.” R. Pindyck & D. Rubinfeld, Microeconomics at 401 (8th ed. 2013). For all goods, and intellectual property goods such as copyrights in particular, the social benefit is that price discrimination more closely matches the quantity sold with the competitive quantity as the seller or licensor better aligns the price with the WTP of different categories of buyers or licensees. See W. Fisher, Reconstructing the Fair Use Doctrine, 101 Harv. L. Rev. 1659, 1701 (1988).

A seller can engage in price discrimination in several ways. One form is known as “second-degree price discrimination,” by which buyers self-sort the packages and quantities they purchase. ¹⁷¹ See W. Adams & J. Yellen, Commodity Bundling and the Burden of Mo nopoly, 90 Q. J. Econ. 470, 476 (1976) (the profitability of bundling “stem[s] from its ability to sort customers into groups with different reservation price [WTP] characteristics.”). Bundling, i.e., the “practice of selling two or more products as a package,” Pindyck & Rubinfeld, supra at 419, is thus a type of second-degree price discrimination. See A. Boik & H. Takahashi, Fighting Bundles: The Effects of Competition on Second Degree Price Competition, 12 a.m. Econ. J. 156, 157 (2020).

The applicability of these basic economic principles was understood and explained by the parties' experts at the hearing. See, e.g., 3/15/17 Tr. 1224–25 (Leonard) (Google's economic expert testifying that price discrimination through bundling is “very, very common . . . even by pretty competitively positioned firms . . . to sort out customers into willingness-to-pay groups.”); 3/30/17 Tr. 3983 (Gans) (Copyright Owners' economic expert acknowledging that bundling is a form of price discrimination); see also Dissent at 69 (same).

How does this downstream (retail level) benefit of price discrimination impact the setting of upstream royalty rates? As the Majority explained (in summarizing the Services' expert testimony) the linkage is explained by the economic concept of “derived demand”:

[M]ultiple pricing structures necessary to satisfy the WTP and the differentiated quality preferences of downstream listeners relate directly to the upstream rate structure to be established in this proceeding. Professor Marx opines that the appropriate upstream rate structure is derived from the characteristics of downstream demand. 3/20/17 Tr. 1967 (Marx) (rate structure upstream should be derived from need to exploit WTP of users downstream via a percentage of revenue). This upstream to downstream consonance in rate structures represents an application of the concept of “derived demand,” whereby the demand upstream for inputs is dependent upon the demand for the final product downstream. Id.; see P. Krugman & R. Wells, Microeconomics at 511 (2d ed. 2009) (“[D]emand in a factor market is . . . derived demand. . . [t]hat is, demand for the factor is derived from the [downstream] firm's output choice”).

Determination at 19; accord Dissent at 32 (noting that “the upstream demand of the interactive streaming services for musical works (and the sound recordings in which they are embodied)—known as “factors” of production or “inputs”—is derived from the downstream demand of listeners to and users of the interactive streaming services . . . This interdependency causes upstream demand to be characterized as “derived demand.”).

In the present proceeding, the PR II-based benchmark embodies the parties' negotiated definition of Bundled Revenue for purposes of calculating royalties on bundled interactive offerings. This is definition in the Initial Determination. Copyright Owners' preferred definition for Bundled Revenue—the Determination's definition—would not only ignore this agreed-upon definition, but would also de-link the royalty rate from the WTP of purchasers of bundles. The Judges recognize that Copyright Owners have expressed concern the Services could use such bundling in order to diminish revenue otherwise payable on a higher royalty tier. However, the Majority noted that the evidence indicated such diminishment only occurred “in some cases.” Clarification Order at 17. Thus, the Judges find that eliminating the incentive for price discrimination via bundling would be a disproportionate response and inconsistent with the broad price discriminatory PR II-based benchmark they find useful in this proceeding.

Expert testimony in this regard is “substantial evidence” on which the Judges can rely. For example, the D.C. Circuit also relied in Johnson on the testimony of the same witness, Spotify's economic expert witness, Professor Marx, who explained how a downstream “lower willingness (or ability) to pay” among some cohorts of consumers supports definitional terms, for student and family subscribers, that lower royalty rates in order to further “economic efficiency” in a manner that “still allows more monetization of that provision of that service.” Johnson at 392–93. Broadening her lens, Professor Marx also explained that this price discriminatory approach is appropriate “across all types of services and subscribers,” as in “[t]he current law [and in the PR II-based benchmark]” which “accommodates . . . ad-supported services . . . and ` bundled services' through different rate provisions.” Marx WRT ¶ 41 (emphasis added). See also 3/21/17 2182–83 (Hubbard) (Amazon's expert witness testifying that “Prime Music, which is bundled with an Amazon Prime service . . . sort[s] out customers' willingness to pay, with an idea of trying to maximize the number of customers,” and agreeing that this approach constitutes “sorting by way of bundling.”) (emphasis added). Further, Professor Hubbard opined that, given the revenue attribution “measurement problem” associated with bundled products, the “Phonorecords II” approach “with the different categories and the minima . . . address this sort of problem [in] a very good way.” 3/15/17 Tr. 1221 (Hubbard).

As in the case of family and student price discrimination, the beneficial effect of such differential pricing was supported by industry witnesses as well as expert witnesses. See, e.g., Mirchandani WDT ¶ 71 (Amazon executive citing the Phonorecords II-based benchmark provisions regarding bundling that “allowed Amazon to bundle Prime Music with Amazon Prime, enabling Amazon to bring a limited catalog of music [REDACTED]”). In sum, the same type of witness testimony that the D.C. Circuit found sufficient to support price discriminatory student and family plans also supports the use of the price discriminatory bundled definition contained in the Initial Determination.

Given the overall benefits from price discrimination, at first blush it is curious that Copyright Owners would risk “leaving money on the table” by removing the royalty-based incentive for price discrimination via bundling. The Judges have identified this problem earlier in this Initial Ruling, in connection with the broader issue of the overall beneficial price discriminatory structure of the PR II-based benchmark. As the Judges noted in that general price discrimination context, Copyright Owners' own expert economic witnesses acknowledged that they would not irrationally “leave money on the table.” In fact, Copyright owners' aim, according to that testimony, is to create an unregulated space—per the Bargaining Room theory—and to use their complementary oligopoly power to negotiate price discriminatory rates (in bundles or otherwise), which would free them from the section 801(b)(1) requirements of reasonableness and fairness.

The Judges further find that their prior ruling on this issue in SDARS III is distinguishable. There, a proffered bundled revenue definition eliminated the payment of any royalty at all. Copyright Owners quite correctly describe that result as “absurd,” but that is not the result here. Rather, in the present case, the parties' negotiated an approach that the Judges adopted in the Initial Determination requiring royalties to be paid on interactive services bundled with other products or services.

Even more distinguishable is Copyright Owners' assertion that Web IV provides support for their preferred definition of service revenue. The argument is immediately suspect, because Web IV involved per-play royalty rates—not percent-of-revenue rates, making the definition of revenue wholly inapposite. Further, the discussion of the price of an “ice cream cone” in Web IV —on which Copyright Owners rely—had nothing to do with bundling or isolating the WTP for different products or services. Rather, there the Judges criticized a bizarre argument made by a licensee (who had a quantity discount for plays steered in its direction), that was tantamount to arguing that if a vendor sells one ice cream cone for $1.06 but a buyer could buy two for $1.06, that the market price of an ice cream cone is thus only $.06. This argument was indeed fallacious, because the price of an ice cream cone would be the average of the total cost for the two cones, i.e., $.53/cone. Here, the issue is how to address the WTP of different classes of buyers with heterogeneous WTP, not the pricing of a discount for all purchasers buying the same quantity. The parties utilized the Bundled Revenue definition from the PR II-based benchmark (and in the Initial Determination) to address the indeterminacy inherent in the variable WTP among purchasers of the bundles, by setting floors and minima, rather than attempt to sort out the WTP of individual (or individual blocs) of subscribers.

For the foregoing reasons, the Judges find that the definition in the Initial Determination (unlike the definition in the Determination) is consistent with the Judges' other substantive rulings herein. That is, just as the Majority abandoned its Bundled Revenue definition in its Initial Determination because it refused to credit the PR II-based benchmark (even as “guidance”), the Judges here do partially rely on the PR II-based benchmark, and thus find that it supports the Bundled Revenue definition contained in the Initial Determination.

4. Application of Four Itemized Statutory Factors

As the forgoing analysis explains, bundling is a form of price discrimination. Accordingly, the Judges' explanation of how price discriminatory rates in the PR II-based benchmark interrelate with the Factor (A) through (D) objectives in section 801(b)(1) are equally applicable here. Accordingly, the Judges adopt by reference their discussion of those four factors set forth supra in connection with the PR II-based benchmark, and find that there is no basis pursuant to those four factors to adjust the PR II-based benchmark definition of Bundled Revenue.

V. Conclusion

On the basis of the foregoing analyses, and in consideration of the entirety of the record, the Judges make the following determination relating to the issues on remand from the D.C. Circuit.

As noted at the outset, the headline rate for all offerings throughout the Phonorecords III period shall be as follows:

In all other respects, the rates and rate structure of the PR II-based benchmark shall be effective as the rates and structure throughout the Phonorecords III period.

The definition of Service Revenue for bundled offerings throughout the Phonorecords III period shall be the definition contained in the Initial Determination.

VI. Order

In light of the foregoing analyses and conclusions, the Judges hereby order that the participants in this remand proceeding prepare and submit regulatory provisions consistent with this ruling. The participants shall file agreed regulatory language within ten days of the date of this ruling.

The Judges further order that if the participants cannot agree on a joint submission, the Judges will accept separate submissions respectively from (1) Copyright Owners and (2) Services, jointly. In absence of an agreed submission, the participants shall file separate submissions not later than 15 days after the date of this ruling.

The Judges further order that parties shall not file, and the Judges shall not consider, briefing or legal argument beyond necessary explanatory notes to the proposed language, section by section, not to exceed 250 words per proposed section. The Judges specifically admonish the parties that they shall not use these submissions as a basis to object to this Initial Ruling, either explicitly or implicitly by proposing regulatory provisions inconsistent with this Initial Ruling.

The Judges further order that, within 30 days of the date of this Initial Ruling and the attendant dissenting documents, the parties shall file an agreed redacted version of this Initial Ruling, and the dissents, for public viewing.

After the Judges have reviewed the parties' regulatory submissions, the Judges shall adopt and format the necessary regulatory language format terms relevant to this ruling and issue a restricted Initial Determination after Remand, which shall embody their determination of rates and terms. The parties will have an opportunity to suggest redactions from the Initial Determination after Remand before it is issued as a public version.

The parties shall not file any motions seeking rehearing or reconsideration of this Initial Ruling. Subsequent to the Judges' issuance of their Initial Determination after Remand as identified in the immediately preceding paragraph, any party may file a Motion for Rehearing within 15 days of the issuance of said Initial Determination after Remand.

After ruling on any and all Motions for Rehearing as identified in the immediately preceding paragraph, the Judges shall issue a Final Determination after Remand.

So ordered.

B. Order 43 on Phonorecords III Regulatory Provisions (Public Version With Federal Register Naming and Formatting Conventions)

Introduction

The present Order concerns a single issue in dispute among the parties regarding regulatory language implementing the Judges' Initial Ruling and Order after Remand (“Initial Ruling”) entered in this proceeding. Subsequent to filing dueling submissions ( see footnote 2 infra), the parties filed a Joint Submission, informing the Judges that they had “agree[d] on all of the regulatory language” except for certain rate percentages contained in Table 2 of the proposed § 385.21. Joint Submission . . . Regarding Regulatory Provisions Following Initial Ruling and Order (after Remand) at 1 (Nov. 30, 2022) (“Joint Submission”) (eCRB no. 27337).

The Regulatory Language in Dispute

The dispute between the parties is whether the Judges should adopt in the Phonorecords III regulations: (1) the several “Total Content Cost” (“TCC”) rates set forth in the Phonorecords II-based benchmark; or (2) the single 26.2% TCC rate discussed in the Initial Ruling. This dispute relates to nine offerings made by interactive streaming services, as detailed below:

The Issue

At a high level, the remaining regulatory issue is the following:

Whether a 26.2% TCC rate identified in the hearing record, and discussed both on appeal and on remand by the D.C. Circuit, should substitute for TCC rates in the Phonorecords III period, or whether these uncapped TCC rates should be set at the specific levels ranging between 20.65% and 22% set forth in the Phonorecords II-based benchmark adopted by the Judges in the Initial Ruling.

To frame, address, and rule on this issue, in this Order the Judges place the parties' dispute in the context of the prior rulings by the D.C. Circuit and the Judges in connection with this proceeding.

Background

On January 5, 2016, the Judges initiated proceedings to determine the appropriate mechanical license royalty rates and terms for the January 1, 2018 to December 31, 2022 period. See Notice Announcing Commencement of Proceedings in Phonorecords III, 81 FR 255 (Jan. 5, 2016). After the parties filed their written and rebuttal testimonies and engaged in discovery, they participated in a five-week evidentiary hearing presided over by the Judges. See Determination of Royalty Rates and Terms for Making and Distributing Phonorecords, 84 FR 1918, 1920, 1923–1925 (Feb. 5, 2019).

In the Majority Opinion, the Judges adopted a “greater-of” royalty rate structure for the mechanical license, which contained a TCC rate applicable to all categories of offerings. See84 FR 1963; see also Johnson v. Copyright Royalty Board, 969 F.3d 363, 372 (D.C. Cir. 2020) (summarizing the Majority Opinion). More particularly, the Majority adopted the following rates and rate structure:

Majority Opinion at 1918, 1960.

The Services appealed. Among their arguments were the assertions—pertinent to this Order—that the Majority: (i) violated the Services' procedural right to fair notice by choosing a structure that was not advanced by any party; (ii) acted arbitrarily and capriciously by simultaneously combining a TCC prong (phased-in to 26.2% of TCC) with an increase in the percentages on the revenue prong (phased-in to 15.1%); and (iii) failed to reasonably explain its rejection of the Phonorecords II settlement as a benchmark. Johnson, supra, at 376, 380–81.

Copyright Owners argued in opposition that: (i) the Services' procedural rights were not violated because “every component” of the Majority's approach was contained in the hearing record; (ii) the Majority's rate and rate structure rulings were well-reasoned, factually supported and, therefore, not arbitrary and capricious; and (iii) sufficient reasons existed in the record to support the Majority's rejection of the Phonorecords II-based benchmark. Johnson, supra, at 382–383; 387.

The D.C. Circuit vacated and remanded. More particularly, Johnson holds as follows:

1. The Majority Determination “failed to provide adequate notice of the drastically modified rate structure [they] ultimately adopted,” which was beyond “a reasonable range of contemplated outcomes” in “the parties' pre-hearing proposals, the arguments made at the evidentiary hearing, and the preexisting rate structures.” Johnson at 381–82. Accordingly, as to this issue, “[i]f the [Judges] wish[ ] to pursue [their] novel rate structure, [they] will need to reopen the evidentiary record.” Id. at 383.

2. The appellate issue of whether the Majority's adoption of the (phased-in) 26.2% TCC royalty rate was “arbitrary and capricious” could not be addressed—given the absence of “adequate notice” cited in point (1) above. Id.

3. The Majority's derivation, calculation and application of the royalty rate of 15.1% on the revenue prong was proper. The D.C. Circuit explained that, as to this issue, the Majority had engaged in the “type of line-drawing and reasoned weighing of the evidence [that] falls squarely within the [Judges'] wheelhouse as an expert administrative agency.” Johnson at 386. More particularly, the D.C. Circuit approved of the Judges' reliance on “substantial evidence” in the form of expert testimony to set the 15.1% service revenue rate. Johnson, at 384–85 (emphasis added). See also id. at 388 (finding “substantial evidence” for the Judges' finding that an increase in the mechanical royalty rate was necessary to address a “marked decline in mechanical royalty income. . . .”).

4. The Majority's rejection of the Phonorecords II-based benchmark is remanded because the D.C. Circuit “cannot discern the basis on which the [Judges] rejected the Phonorecords II rates as a benchmark in [their] analysis, that issue is remanded to the [Judges] for a reasoned analysis.” Johnson at 387.

On remand, the Judges adopted procedures that mainly followed the parties' requests. More particularly, the Judges followed the D.C. Circuit's directive and reopened the evidentiary record to receive evidence and testimony relating to the TCC issues. See Order Regarding Proceedings on Remand at 2 (Dec. 15, 2020). The post-remand supplementary record added: (1) rate evidence for the 33-months from January 2018 through September 2020, when the parties operated under the Majority's new (but subsequently vacated) regulations including the TCC rates; and (2) new testimony from economic expert witnesses on behalf of Copyright Owners and the Services. See Initial Ruling, passim. However, none of the post-remand evidence submitted and relied upon by the parties specifically addressed as a separate issue the rates for the nine offerings that are the subject of the present Order.

On July 1, 2022, the Judges issued their Initial Ruling —applying Johnson and considering the entire record developed pre-remand and post-remand. In their Initial Ruling, the Judges made several findings that bear upon the issue at hand, viz., whether to adopt in the Phonorecords III regulations the 26.2% TCC rate or the TCC rates (ranging from 20.65% to 22%) from the Phonorecords II-based benchmark. In particular, in the Initial Ruling, the Judges stated the following:

1. The Phonorecords II-based benchmark incorporates price discriminatory features for product differentiation as between: (a) subscription and ad-supported services; (b) portable and non-portable services; and (c) unbundled and bundled services. See Initial Ruling at 67–68 (noting the salutary price discriminatory nature of the Phonorecords II-based benchmark).

2. The Phonorecords II-based benchmark “reflect[s] a rate structure with an adequate degree of competition, because there was a balance of bargaining power [“countervailing power”] between the two negotiating industrywide trade associations, offsetting the complementary oligopoly effects in place when a “Must Have” licensor bargains separately with each licensee.” Initial Ruling at 69.

3. Based upon the available record evidence, the Judges find . . . the Services' Phonorecords II-based benchmark . . . “more than sufficient to satisfy the legal requisites for application, as well as a practical benchmark, when used in conjunction with the 15.1% headline revenue rate advocated by Copyright Owners.” Initial Ruling at 59.

4. “Substantial evidence demonstrates that the Phonorecords II-based benchmark rates, other than the headline rate, are not `too low.' ” Initial Ruling at 73.

5. A Copyright Owner expert witness opined that “the evidence . . . indicates that the relative valuation ratios implied by the current Section 115 compulsory license [ i.e., the Phonorecords II-based benchmark] implies a “ lower bound on the relative market valuations of the reciprocal percentage of the value musical works rights relative to sound recording rights [ i.e., TCC rates] [of] 22% and 21%.” Initial Ruling at 78 (emphasis therein).

6. The royalty rates and terms within subpart C of the Phonorecords II-based benchmark— which include the rates and term for the offerings at issue in this Order — are expressly “covered by [the] foregoing analysis.” Initial Ruling at 93. In rejecting all of Copyright Owners' arguments for different treatment of Phonorecords II-based benchmark rates in Subpart C therein, the Judges declined to adopt Copyright Owners' “re-assert[ion] [of] the same arguments with respect to subpart C” that Copyright Owners advanced in opposing the Phonorecords II-based benchmark “for interactive streaming in subpart B.” See Initial Ruling at 93–94 (“The Judges find no reason on remand to treat the subpart C offerings differently than the manner in which they are treating the subpart B interactive streaming offerings . . . . That means, however, that the various “headline” rates for these subpart C offerings must also adjust to 15.1%, 131 whereas the alternative rates (identified in subpart C as “minima” and “subminima”) rates shall remain unchanged. ”) (emphasis added).

7. The D.C. Circuit had affirmed that: (a) the “headline” percentage royalty rate (not a TCC rate) of 10.5% was too low; and (b) that the Majority had not improperly exercised its authority when it increased that revenue royalty rate to 15.1% (as phased-in over the five-year rate term). Accordingly, on remand, the Judges maintained the 15.1% (phased-in) percentage royalty rate. See, e.g., Initial Ruling at 4, 17.

8. The D.C. Circuit affirmed the Majority's derivation and calculation of the 26.2% TCC rate for use as an input in calculating the 15.1% (phased-in) service revenue percentage royalty rate. However, Johnson vacated and remanded the Majority's application and inclusion of the 26.2% TCC rate. Initial Ruling at 19–20.

For these reasons, the Judges decided in the Interim Ruling that: (1) the overall Phonorecords II rates comprise a “useful benchmark,” when the 15.1% headline percentage rate replaces the 10.5% headline percentage rate for the offerings in Subparts B and C of the Phonorecords II-based benchmark; and (2) “[t]he (phased-in) 26.2% rate [is] unreasonable.” Initial Ruling at 50 n.77; 88; and 93–94.

Procedures Following the Post-Remand Initial Ruling

In the Initial Ruling, the Judges directed the parties to attempt to submit jointly agreed-upon regulatory provisions implementing the Initial Ruling, for the Judges to consider. The Judges further ruled that, if the parties could not agree on all the regulatory language, they should make separate submissions regarding regulatory provisions in dispute. See Initial Ruling at 114.

The parties agreed to many regulatory provisions but disagreed as to several such provisions. Accordingly, they filed separate submissions and respective replies, regarding the regulatory provisions. Services' Joint Submission of Regulatory Provisions (July 18, 2022); Copyright Owners' Submission of Regulatory Provisions to Implement the Initial Ruling (July 18, 2022); Services' Joint Response to Copyright Owners' Submission of Regulatory Provisions (Aug. 5, 2022); Copyright Owners' Response to Judges' July 27, 2022 Order Soliciting Responses Regarding Regulatory Provisions (Aug. 5, 2022).

The Judges considered those submissions and entered an order addressing the disputed regulatory provisions. See Corrected Order regarding Regulatory Provisions following Initial Ruling and Order (After Remand) (Nov. 10, 2022) (“November 10th Order”).

In the November 10th Order, the Judges directed the parties once more to file a joint submission “of regulatory provisions that embody the rulings set forth in Johnson, the Initial Ruling and this [November 10th] Order, and any aspects of the [Majority] Determination (pre-remand) that the parties understand to remain effective after the foregoing rulings.” November 10th Order at 31.

On November 30, 2022, the parties made the Joint Submission (as also identified at the outset of the present Order), in which they provided joint regulatory language no longer in dispute that applied the binding rulings of the Judges and the D.C. Circuit. However, as also noted above, the parties identified the single issue in dispute that relates to the nine service offerings described supra.

The Parties' Respective Arguments in Their November 30th Joint Submission

Copyright Owners' Arguments

According to Copyright Owners, the Initial Ruling “appears to plainly acknowledge that, in light of Johnson, the derivation and calculation of the (phased-in) 26.2% TCC rate percentage cannot be changed.” Joint Submission at 6. More particularly, Copyright Owners aver that, according to the Judges' Initial Ruling, “the D.C. Circuit affirmed the Majority's derivation and calculation of the 26.[2]% . . . TCC rate” and further that “both rate prongs”—the service revenue rate and the TCC rate—were “derived from the same analyses.” Initial Ruling at 19; Joint Submission at 6–7 (quoting Initial Ruling at 19 (emphasis removed)). Further to this point, Copyright Owners rely on the Judges' additional language in the Initial Ruling that the pre-remand Final Determination's “ derivation and calculation of the TCC rate [ i.e., the 26.2% rate] . . . is not subject to further consideration on remand by the Judges.” Joint Submission at 7 (quoting Initial Ruling at 20 (emphasis in Initial Ruling)).

According to Copyright Owners, the foregoing points are consistent with the limited scope of the remand, which “was not opened for new evidence concerning TCC rate percentages.” Joint Submission at 7 (citations omitted). Accordingly, Copyright Owners emphasize that “there is no evidence in the record after remand to support changing the (phased-in) 26.2% TCC rate percentage.” Joint Submission at 7. Copyright Owners—characterizing the former Phonorecords II TCC rates now at issue as newly derived and calculated—maintain that these “new” TCC rate percentages therefore are “foreclosed” by the Initial Ruling and post-remand orders cited above. Joint Submission at 7–8.

Copyright Owners also assert that the TCC rate at issue here—“was not appealed by the Services or challenged during the remand, nor called into question by the Circuit in Johnson. ” Joint Submission at 8 (emphasis removed). The absence of an appeal as to this issue, according to Copyright Owners, means that the only TCC rate supported by Johnson is the 26.2% TCC rate. Joint Submission at 8.

The Services' Arguments

According to the Services, the Judges should adopt in the regulations the TCC percentage rates—ranging from 20.65% to 22%—because those rates are contained in the Phonorecords II-based benchmark adopted by the Judges and thus essentially have been “expressly set out by the Judges” in two prior decisions. Joint Submission at 2 (citing Initial Ruling at 2; November 10th Order at 6 n.13). In light of these prior Orders, the Services characterize Copyright Owners' position as the new argument, improperly seeking regulatory provisions that “reflect the 26.2% rate previously imposed by the [M]ajority in the now-vacated pre-remand Final Determination.” Id.

More pointedly, the Services argue that the Judges' Initial Ruling already expressly considered and rejected application of the 26.2% TCC rate. Id. (citations omitted). Further, the Services maintain that it is because the Judges rejected the 26.2% TCC rate in the Initial Ruling that the Judges had no need to “substantively address the topic of TCC rates” in their November 10th Order. Id. at 4.

The Services further maintain that “Johnson does not compel the Judges to simply reinstate their original pre-remand TCC rates.” Id. To this point, the Services rely on the Judges' post-remand finding that, although the error made by the Majority in adopting the 26.2% TCC rate in the pre-appeal Phonorecords III Determination was procedural, the “consequence . . . was substantive.” Id. (emphasis herein).

For the above reasons, the Services maintain that the Judges could not possibly be required on remand to adopt an express 26.2% in any portion of the Phonorecords III regulations.

Turning from their argument that the 26.2% TCC rate was rejected by the Judges, the Services focus on the Judges' finding in the post-remand Initial Ruling that the “ Phonorecords II benchmark . . . is the `better of the benchmarks proposed by the parties . . . one that satisfies the requirements of 17 U.S.C. 801(b)(1) in all respects,' ” Joint Submission at 5 (quoting Initial Ruling at 2). Because the Phonorecords II benchmark includes the TCC rates now at issue—ranging from 20.65% to 22%—the Services maintain that those rates should properly be included in the Phonorecords III regulations. Id.

The Judges' Analysis and Ruling

Having considered the parties' submissions, the Initial Ruling and all other pertinent material, the Judges rule that the 26.2% TCC rate cannot and shall not be applied in the regulatory provisions now at issue. Rather, the Judges rule that the TCC rates set forth in the Phonorecords II-based benchmark shall be applied in the nine regulatory provisions now at issue, because they are consistent with and give effect to the Judges' Initial Ruling. The more particular bases for this ruling are set forth below.

Most fundamentally, the Judges note at the outset that in the Initial Ruling they expressly did not apply the 26.2% TCC rate in any manner other than as an input—using that TCC rate only as the D.C. Circuit directed—to calculate the 15.1% of service-revenue royalty rate. See, e.g., Initial Ruling at 41 (“[A] careful reading of the remand testimony by Copyright Owners' economists, Professors Watt and Spulber, reveals that neither of them actually testifies that there is sufficient theoretical and empirical evidence to support the . . . 26.2% TCC rate. . . .”) (emphasis in original). See also id. at 40–41 n.69 (contrasting the improper application of the 26.2% TCC as a separate statutory rate from the use of the 26.2% TCC rate as input from a “bargaining model” solely to increase the service revenue rate to 15.1%.).

In this regard, the Initial Ruling has relied upon the clear distinction made in Johnson between the 15.1% service revenue rate and the 26.2% TCC rate. Compare Johnson, supra, at 385 (affirming the Majority's application of the “revenue rate of 15.1%” as “the type of line-drawing and reasoned weighing of the evidence falls squarely within the[ir] wheelhouse as an expert administrative agency”) with id. at 382–83 (vacating the Majority's decision for “significantly hiking the TCC rate to 26.2% from approximately 17% to 22%” without allowing the Services an opportunity to address the issue—an error that was even “worse” than the elimination of caps on certain other TCC offerings.).

Further, the offerings now at issue were contained in the Phonorecords II-based benchmark, and the Judges' application of that benchmark in the Initial Ruling is unambiguous: Other than the new and increased headline rate of 15.1%, “the rates and rate structure of the Phonorecords II -based benchmark proposed by the Services . . .) shall constitute the rates and rate structure for the Phonorecords III period.” Initial Ruling at 2. Accordingly, with regard to the single remaining issue, pertaining to the nine offerings listed supra, the regulatory provisions proposed by the Services in the Joint Submission are fully consistent with the Initial Ruling.

By contrast, Copyright Owners' proposed language introduces a change in the Phonorecords II-based benchmark rates that was never the subject of an evidentiary proceeding pre-or post-remand, whether through live or written testimony. But perhaps more importantly, as a matter of substance, Copyright Owners' proposed regulatory provisions are inconsistent with the language and a key purpose of the Initial Ruling, which is to adopt the Phonorecords II-based benchmark rates, the basis of which were generated consensually by the parties, through negotiations between industrywide trade associations, which prevented unwarranted and disproportionate complementary oligopoly market power from affecting the royalty rates. See Initial Ruling at 69–70.

The Judges also reject Copyright Owners' argument that by maintaining the 20.65% through 22% TCC rates in the Phonorecords II-based benchmark they would be violating their prior rulings regarding the scope of the remand. Citing to the Judges' Order Regarding Proceedings on Remand at 1 (eCRB no. 23390) (“Remand Order”), Copyright Owners state in their Joint Submission that that the remand “was not opened for new evidence concerning TCC rate percentages.” Joint Submission at 7. But the decision to re-open the existing, and robust, evidentiary record only as to rate structure, did not limit the scope of the remand itself, nor consideration of evidence from the underlying proceeding.

Moreover, the Judges find no language in either the Remand Order or the Remand Scheduling Order, and no other basis, that would support Copyright Owners' characterization of the 20.65% through 22% TCC rates in the Phonorecords III-based benchmark as new evidence, given that they were expressly included in that benchmark which had been proffered at the hearing prior to the remand.

Further, the present issue of whether the regulatory provisions implementing the Initial Ruling should apply the Phonorecords II-based benchmark TCC rates or the 26.2% TCC rate is not a dispute regarding the derivation or calculation of a new TCC rate. The Phonorecords II-based benchmark rates are self-evidently not new rates, because they existed in that prior benchmark. Moreover, the present dispute relates to whether the language and reasoning in the Initial Ruling are consistent with maintaining the rates contained in the Phonorecords II-based benchmark for the nine offerings at issue, or whether the Initial Ruling calls for abandoning those benchmark rates and replacing them with the 26.2% TCC rate proffered by Copyright Owners. As explained supra, the 26.2% TCC rate was properly utilized by the Majority as an input (combined with other evidence) in order to calculate the 15.1% service revenue royalty rate. The record reflects no other context in which the 26.2% TCC rate can be utilized, let alone must be utilized. Indeed, as explained supra, the record reflects the Judges' rejection of the 26.2% TCC rate as a stand-alone statutory royalty rate.

The Judges also reject Copyright Owners' argument that the Services somehow waived their argument for maintaining the 20.65% through 22% TCC Phonorecords II-based benchmark rates. More particularly, Copyright Owners incorrectly assert that these rates were “not appealed by the Services. . . .” Joint Submission at 8. Rather, the D.C. Circuit stated unambiguously: “[T]he Streaming Services object to the [Judges'] . . . rejection of the Phonorecords II . . . settlement[ ] as [a] rate benchmark[ ].” Johnson, 969 F.3d at 384; see also id. at 386 (“The Streaming Services argue . . . that the [Judges] arbitrarily rejected . . . [a] potential rate benchmark[ ] . . . the Phonorecords II settlement—without adequate explanation.”).

Moreover, the D.C. Circuit repeatedly noted that it was vacating and remanding the Majority's Determination with regard to, inter alia, the Majority's improper decision to reject the Phonorecords II-based benchmark writ large, i.e., without qualification by the appellate panel that some parts of that proffered benchmark might have been correctly rejected. See Johnson, 969 F.3d at 367, 376, 381, 387. Obviously, virtually all the elements of the Phonorecords II-based benchmark—including the offerings now at issue—were appealed, and not waived, foregone or forfeited by the Services.

Likewise, Copyright Owners are wrong in their claim that the Services had never “challenged” these rate issues “during the remand.” Joint Submission at 8. Rather, the Services argued on remand for the Phonorecords II-based benchmark to be applied comprehensively, without itemizing every element of that proffered benchmark. See Services' Joint Opening Brief (post-remand) at 19–44 (Apr. 1, 2021) (detailing why “the Services' proposal based on the Phonorecords II settlement is reasonable . . . .”); see also Services' . . . Submission of Regulatory Provisions at 2 (July 18, 2022) (“Services' July 18th Submission”) (“[T]he Services have faithfully implemented the task at hand—to use the rates and rate structure of the “Phonorecords II-based benchmark” proposed by the Services during the remand proceeding . . . .”).

Finally, the Judges find and conclude that their ruling in this Order sets forth reasonable rates satisfying the four objectives in the then-applicable (but now superseded) statutory rate standard contained in 17 U.S.C. 801(b)(1). First, with regard to Factor (A), the Judges recognize and follow the D.C. Circuit's ruling that the Majority's decision to increase in the “headline” service revenue royalty rate by 44% from 10.5% to 15.1% was supported by substantial evidence. Johnson at 387–88.

Further with regard to Factor (A), the Judges understand their analysis and reasoning in the Initial Ruling—applying the Phonorecords II-based benchmark and thus rejecting the 26.2% TCC rate—to be applicable to the present dispute regarding the adoption of regulations to implement the Initial Ruling. Accordingly, the Judges adopt by reference herein their analysis and reasoning set forth at pages 90–91 of the Initial Ruling. For those reasons, the Judges decide, as they did in the Initial Ruling, that there is no basis for yet a further increase in the royalty rate based on Factor (A), finding “no evidence to suggest that the price discriminatory rates should be changed, in order to address the connection between price discrimination and the objective of Factor (A).” Id. at 91.

Next, in considering Factors (B) and (C), the Judges' Initial Ruling adopts the Majority's reasoning that the 15.1% service revenue royalty rate provided a “ fair allocation of revenue between copyright owners and services” and it would be “substantively unwarranted to engage in any new consideration on remand of the impact, if any, of Factors (B) and (C) on the otherwise reasonable 15.1% revenue rate.” Id. at 15–16.

In their Joint Submission, the parties have presented no arguments specifically addressing how Factors (B) or (C) might support their proposed TCC rates now at issue. Examining the record, the Judges find and conclude that maintaining the Phonorecords II-based rates ranging from 20.65% to 22% embodies the fairness associated with rates negotiated between industrywide trade associations wielding relatively comparable bargaining power, as discussed supra and in the Initial Ruling. This notion of fairness is embodied in the determination of the reasonable rate and, as can be the case, when one of the four itemized statutory objectives of section 801(b)(1) is bound-up and appropriately addressed within the broader context of setting a reasonable rate, no further adjustment is necessary through an invocation of an itemized statutory factor. See Determination of Royalty Rates and Terms for Making and Distributing Phonorecords (Phonorecords III) 84 FR 1918, 1955, 2015 (Feb. 5, 2019) (Majority and Dissenting Opinions agreeing that “to the extent market factors may implicitly address any (or all) of the four itemized factors, the reasonable, market-based rates may remain unadjusted.”).

Finally, the Judges see no reason to alter their adoption of the Phonorecords II-based benchmark rates for the nine offerings at issue in this Order based upon the final listed statutory objective, Factor (D). In the Joint Submission, Copyright Owners did not make an express argument relating to this factor (nor did the Services). Independently considering the potential application of Factor (D), the Judges find no evidence that the continuation of the Phonorecords II-based benchmark rates for the offerings at issue in this Order would cause any disruption that Factor (D) is intended to address. Further, as noted supra, the Judges have phased-in an increase in the headline service revenue royalty rate from 10.5% to 15.1%—a 44% increase—rendering unreasonable any argument that the present decision to maintain the Phonorecords II-based TCC rates is “disruptive” to Copyright Owners under the statutory Factor (D) standard.

Moreover, the Judges reassert their point in the Initial Ruling that there is no need to independently consider any potential disruption under the Factor (D) standard because the Judges have already found an application of that rate to be unreasonable. See Initial Ruling at 50 n.77. Further, the D.C. Circuit was aware of the existence of the 20.65% to 22% TCC rates in the Phonorecords II-based benchmark for these nine offerings now at issue, and not only declined to affirm the Majority's increase in those rates to 26.2%—a significant increase of 19% to 27% —but also condemned that increase. See Johnson at 383 (“Worse still . . .the [Judges] also raised the total content cost [TCC] rate to 26.2%. . . .That rate previously fell between approximately 17% and 22%”). Nothing in the record suggest that the Judges can or should utilize the narrow statutory “disruption” standard in Factor (D) of section 801(b)(1) as a basis to override the position of the D.C. Circuit or the Judges' analysis in the Initial Ruling as to the inapplicability of the proffered 26.2% royalty rate.

Order

For the foregoing reasons, the Judges shall adopt in the regulatory provisions the several “Total Content Cost” (“TCC”) rates set forth in the Phonorecords II-based benchmark as proposed by the Services.

Within two days of the date of issuance of this Restricted Order, the parties shall file an agreed proposed redacted version for public viewing.

Issue Date: April 26, 2023.

David P. Shaw

Chief Copyright Royalty Judge

C. Dissent in Part as to Section IV of the Initial Ruling and Order After Remand by Judge David R. Strickler (Redacted Version With Federal Register Naming and Formatting Conventions)

I. The Contours of This Partial Dissent

I respectfully Dissent from Section IV of the Initial Ruling and Order after Remand (Initial Ruling). As explained herein, I conclude that the D.C. Circuit's rulings in Johnson preclude the Judges from engaging in “new `agency action.' ” See Johnson v. Copyright Royalty Board, 969 F.3d 363, 386 (D.C. Cir. 2020). Accordingly, I cannot join with the present Majority in its determination that this remand proceeding constitutes “new `agency action' ” consistent with Johnson. That argument is circular and renders useless the D.C. Circuit's careful analysis of the procedures that are and are not available to the Judges after they have issued their Initial Determination.

As further explained herein, the argument is circular because it begins with the D.C. Circuit's ruling that the Determination was improper because it invented a new procedure to change the Bundled Revenue definition that was in the Initial Determination, only to circle back to where it started by creating—through the D.C. Circuit's own remand no less—a further and extra-statutory “new `agency action'”.

The Majority also renders Johnson useless, by adopting a process by which—after the D.C. Circuit has remanded an issue because the Judges lacked procedural authority to rule—the procedural error is essentially honored in the breach, because the remand neuters the effect of the D.C. Circuit's ruling.

I join with the Majority though on its substantive decision to re-adopt the definition of Bundled Revenue set forth in the Initial Determination. As explained infra, I too find that it is clearly preferable to the definition that was swapped into the (Final) Determination. But as explained herein, I reconcile the procedural and substantive points differently. I apply what I believe to be the proper understanding of the D.C. Circuit's ruling—finding, contrary to the Majority, no avenue for “new `agency action'” post-remand. Rather, the Judges must revert to the original—and substantively appropriate—definition of Bundled Revenue in the Initial Determination.

To explicate the bases of this Dissent, my opinion as to this issue is set forth below.

II. Introduction

The Majority and I analyze the definition of “Service Revenue” from “Bundled Offerings” (henceforth “Bundled Revenue” definition) in the context of our partial adoption of the PR II-based benchmark. As discussed supra, the Remand Majority found that the PR II-based benchmark is a useful benchmark, particularly because of its features that incentivize beneficial downstream price discrimination and generate more listeners, revenues, and royalties. As explained below, the Bundled Revenue definition—itself an element within the PR II-based benchmark—also embodies such price discriminatory incentives. Thus, the Judges' analysis of the PR II-based benchmark and the Bundled Revenue definition are connected.

In the Determination, the earlier Majority likewise found the issues relating to the PR II-based benchmark to be bound-up with the question of the appropriate Bundled Revenue definition. But because that earlier Majority rejected the PR II-based benchmark, it likewise rejected the Bundled Revenue definition contained in the Initial Determination. The definition in the Determination thus eliminated the royalty-based incentive to engage in price discrimination via bundling.

In the interregnum between the Initial Determination and the (Final) Determination, the Judges considered Copyright Owners' post-hearing motion which sought, inter alia, to strike the Bundled Revenue definition in the Initial Determination. The Majority agreed with Copyright Owners that the definition in the Initial Determination should be replaced. An important rationale—highly relevant in the present context—was as follows: “The Judges have . . . declined to rely on the 2012 . . . benchmark . . . as the basis for the rate structure, or, therefore, as regulatory guidance.” Amended Order Granting in Part and Denying in Part Motions for Rehearing at 17 (Jan. 4, 2019) (Clarification Order).

Unlike in the Determination, in this Initial Remand Ruling the Judges do rely on the PR II-based benchmark in part because of its price discriminatory aspects. More particularly, because the bundling of interactive services also constitutes a form of price discrimination, the Judges find the PR II-based benchmark definition of Bundled Revenue set forth in the Initial Determination to be substantively reasonable and otherwise consistent with the four itemized factors in section 801(b)(1).

As a procedural matter though, I can neither: (1) offer any further or fuller explanation for why the Majority made this change in the Bundled Revenue definition nor (2) identify any “new `agency action'” that would permit this definitional switch. And contrary to present Majority on remand, I also cannot identify a “new `agency action'” that the Judges can now take to return to the definition in the Initial Determination. But, as explained infra, the Judges need not identify such action, because the absence of a justification for the definitional switch requires the Judges to revert back to the definition in the Initial Determination.

As a substantive matter though, the Judges unanimously agree to replace the post-hearing definition of Bundled Revenue in the Determination and reinstate the definition set forth in the Initial Determination.

III. Background

In this remand proceeding, the parties propose two starkly different definitions of Bundled Revenue. Each has a dramatically different impact on the use of the royalty structure and levels to incentivize price discrimination in the downstream market.

The Services argue in favor of the language contained in the Initial Determination, i.e., in their PR II-based benchmark, which defines Bundled Revenue, in pertinent part, as

the revenue recognized from End Users [ i.e., consumers] for the Bundle less the standalone published price for End users for each of the other component(s) of the Bundle . . . .

Initial Determination, Attachment A at 7 (§ 382.2 therein).

By contrast, Copyright Owners support the Majority's substituted language contained in the Determination, which defines Bundled Revenue, in pertinent part, as

the lesser of the revenue recognized from End Users [ i.e., consumers] for the bundle and the aggregate standalone published prices for End Users for each of the component(s) of the bundle that are License Activities . . . .

Determination, Attachment A at 8 (§ 382.2 therein).

In Johnson, the D.C. Circuit succinctly summarized these conflicting definitions as follows:

In its Initial Determination, the [Judges] directed that the revenue from streaming services that are included in bundled offerings would generally be measured by the value remaining after subtracting the prices attributable to the other products in the bundle.

When the Copyright Owners objected to the substance of that definition in their motion for “clarification,” the Board adopted an entirely new definition of Service Revenue for bundled offerings. . . . This new definition generally measured the value of the streaming component of a bundle as the standalone price of the streaming component.

Johnson at 389.

In the Clarification Order, the Judges succinctly summarized the parties' respective positions. Id. at 17. They noted that Copyright Owners had presented evidence that the PR II-based benchmark definition contained in the Initial Determination “led in some cases to an inappropriately low revenue base,” although the Judges “agree that there is no support for any sweeping inference that cross-selling has diminished the revenue base.” Id. at 17, 21 (emphasis added). The Judges further noted the Services' assertion that the Bundled Revenue definition in the Initial Determination is consistent with the Judges' “endorsement of the classic price discrimination enabled by bundling strategies.” Id.

The Majority resolved this issue in the Clarification Order in favor of Copyright Owners. Specifically, the Majority found that, because of the “indeterminacy problem” inherent in bundling, “the Services—not the Copyright Owners—. . . are in a position to provide evidence of how they price bundles and value the component parts thereof.” Id. at 17–18. However, according to the Majority, although the Services “bore the burden of providing evidence concerning the proper economic allocation of bundled revenue,” they “failed to do so,” and “[b]y default . . . the Judges must adopt an approach to valuing bundled revenue that is in line with what the Copyright Owners have proposed.” Id. at 18.

IV. The Rulings in Johnson Regarding the Bundled Revenue Definition

The Services appealed the Majority's abandonment of the Bundled Revenue definition in their Initial Determination. Their appeal “challenge[s] both the legal authority and the substantive soundness” of this switch.

First, the Services argued that the Majority failed to identify and explain the procedural basis for making the switch after the hearing had concluded. Second, the Services argued that, substantively, the replacement definition in the Determination “was arbitrary, capricious, or unsupported by substantial evidence.” Johnson at 389, 392.

The D.C. Circuit agreed with the Services regarding the procedural issue and therefore vacated and remanded that aspect of the Bundled Revenue definitional switch. In light of its procedural ruling, the D.C. Circuit explicitly declined to rule on the Services' substantive argument relating to the definitional switch. Id. at 392. (“Because the Board failed to explain the legal authority for its late-breaking rewrite, we vacate and remand that aspect of the decision [and] we have no occasion to address the Streaming Services' separate argument that the definition was arbitrary, capricious, or unsupported by substantial evidence.”).

The D.C. Circuit's rulings in Johnson pertaining to this Bundled Revenue Definition were clearly articulated. The D.C. Circuit found that the Majority “failed to explain under what authority” it made a material change to the definition “so late in the game.” Johnson at 389, 392. The D.C. Circuit noted that the Judges expressly declined to treat the Clarification Motion as a motion for rehearing; consequently, the motion did not request and the Judges did not reconsider either evidence or legal argument. Id. at 390. Although appellate counsel offered rationales, the D.C. Circuit rejected counsel's post hoc reasoning. Id. and 391–92. Ultimately, the D.C. Circuit remanded the adopted regulation “either to provide `a fuller explanation of the [Judges'] reasoning at the time of the agency action[,]' or to take `new agency action' accompanied by the appropriate procedures.” Id. at 392, citing Dep't of Homeland Sec. v. Regents of Univ. of Cal. 140 S.Ct. 1891, 1908 (2020).

To be precise, I take note of the following specific rulings in Johnson:

1. “The problem is that the [Majority] has completely failed to explain under what authority it was able to materially rework that definition so late in the game.” Id. at 389.

2. “The [Majority] did not treat Copyright Owners' motion to have the definition changed as a motion for rehearing . . . [because] Copyright Owners' motion did not request a literal rehearing of evidence or legal argument.” Id. at 390 (cleaned up).

3. “The [Majority] nowhere in its order or the [ ] Determination explains the source of its power to make `fundamental' changes under the authorizing statute . . . .” Id. at 392. [same as #1]

4. “[I]t should go without saying that we may not sustain the Board's action based on its attorney's theorizing at oral argument . . . vacillating gestures to uninvoked authority will not do.” Id. at 391–92 (the D.C. Circuit alluding to its rejection of arguments also made only by appellate counsel in support of the Majority's rejection of the PR II-based benchmark earlier in the decision).

“We must vacate the [ ] Determination's bundled offering Service Revenue definition and remand for the [CRB Judges] either to provide `fuller explanation of the agency's reasoning at the time of the agency action[,]' or to take `new agency action' accompanied by the appropriate procedures.” Id. at 392.

V. Remand Procedure Regarding Bundled Revenue Definition

Post-remand, the Judges stated their understanding, as well as the parties' understanding, of the issue on remand with respect to the Bundled Revenue definition:

The Services and Copyright Owners agree that the proceedings on remand should be limited to three issues: * * * [3] the adoption of a revised definition of “service revenue” for bundled offerings between issuing their Initial Determination and [their] Determination.

Order Regarding Proceedings on Remand at 1 (Dec. 15, 2020) (Remand Order).

The parties proposed, and the Judges agreed, that the record would not be re-opened with regard to the Bundled Revenue definitional issue. Rather, the Remand Order permitted the parties only to provide further briefing on this matter. Id. Specifically, the Judges subsequently permitted each party to file simultaneous Initial Remand Submissions and simultaneous Reply Remand Submissions. See Order Adopting Schedule for Proceedings on Remand (Dec. 20, 2020). Thereafter, seeking further analysis on the question of “new agency action,” the Judges solicited, and received, further briefing on this issue. See Notice and Sua Sponte Order Directing the Parties to Provide Additional Materials (Dec. 9, 2021) (Feb. 9, 2021); Sua Sponte Order Regarding Additional Briefing (Feb. 9, 2021).

VI. The Parties' Submissions Regarding Bundled Revenue Definition

In their respective briefing, Copyright Owners and the Services made arguments relating to: (1) the procedural issue, i.e., the Judges' authority, vel non, to switch to a new Bundled Revenue definition in the Determination; and (2) the substantive issue, i.e., the relative merits of the two conflicting Bundled Revenue definitions. See Initial Remand Submission of Copyright Owners at 7–10 (Apr. 1, 2021) (CO Initial Submission); Services' Joint Opening Brief (in Services' Joint Written Direct Remand Submission at Tab D) at 64–76 (Apr. 1, 2021) (Services' Initial Submission); Copyright Owners' Reply Brief on Remand (in Reply Remand Submission of Copyright Owners, Vol. 1) at 64–88 (CO Reply); Services' Joint Reply Brief at 52–63 (Services' Reply).

A. The Procedural Issue

1. Copyright Owners' Arguments

Copyright Owners assert first that the Judges can preserve their post-hearing switch of the Bundled Revenue definition by sidestepping the D.C. Circuit's holding and rationale in Johnson. That is, Copyright Owners maintain that this remand proceeding itself constitutes the necessary form of “new `agency action'” that Johnson invites, while also liberating the Judges from the consequences of the procedural infirmities identified by the D.C. Circuit. More particularly, Copyright Owners argue:

[T]he new agency action here is a determination after remand proceedings[.] [T]he [Judges are] largely free to chart [their] own procedural course, and [they] ha[ve] done so in [their] [Remand] Order. The [Judges are] not required to undertake any of the procedural steps set forth in 17 U.S.C. 803(b) in order to take such “new agency action.” See 17 U.S.C. 803(d)(3) (requiring only that on remand further proceedings be taken “in accordance with subsection (a)”); 37 CFR 351.15; Intercollegiate Broad. Sys., Inc., 796 F.3d at 125 (“[N]either the Copyright Act nor the [Judge's] regulations prescribe any particular procedures on remand.”) The Circuit's instruction that the action be “accompanied by the appropriate procedures[,]” Johnson, 969 F.3d at 392, does not dictate what those “appropriate procedures” must be but instead plainly refers to these flexible rules. See also Oceana, Inc., 321 F. Supp. 3d at 136 (explaining that when remanding to an agency, a court generally “may not dictate to the agency the methods, procedures, or time dimension, for its reconsideration”).

CO Initial Submission at 71 n.33.

Copyright Owners reject the Services' position that the asserted procedural error is an “absence of authority” that can never be cured. Id. at 74 (citing Services' Proposal for Remand Proceedings at 10). They note that the D.C. Circuit did not say the Judges lacked the authority to revisit the service revenue definition from bundles on remand. Nor, they observe, did it say the Judges have no authority to review the record evidence and the parties' arguments and reach the same conclusion or a different conclusion on remand.

Copyright Owners further opine that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. But instead, they note, the D.C. Circuit remanded and said the Judges could take “new agency action” precisely to cure the asserted procedural defect. Copyright Owners assert that the remand allowed the parties to present the record evidence and their arguments so that the Judges can address the definition “afresh” in the remand determination. Id. at 74.

Further, Copyright Owners argue that 17 U.S.C. 803(d)(3) states only that proceedings on remand must be in accordance with 17 U.S.C. 803(a). They contend that remand proceedings need not be confined to procedures the Services claim are too late in the game for the Judges to follow, again relying on the holding in Intercollegiate Broad. Sys., supra, that “neither the Copyright Act nor the Board's regulations prescribe any particular procedures on remand.” Id. at 125. Accordingly, they argue, the Judges can reaffirm the adopted bundled service revenue definition following their review of the parties' submissions without invoking section 803(c)(2) or 803(c)(4) that were ruled inapplicable in Johnson. CO Reply at 65–66.

Also, Copyright Owners argue that the Judges may properly justify the changed definition under section 803(c) as a fuller explanation of the agency's reasoning at the time it was made. They urge that the Judges could explain that, especially in light of the evidence of how (in Copyright Owners' characterization) the Services misused the prior definition to make service revenue completely disappear, the Judges carry-over of the prior Bundled Revenue definition from Phonorecords II into the Initial Determination was unintended and inadvertent. CO Reply at 69.

Copyright Owners also assert that, on remand, the Judges could explain that Copyright Owners had, in their Motion for Clarification, identified an “exceptional case” under section 803(c)(2) because the prior definition failed to comport with Judges' precedent and economic principles, and was unsupported by evidence. In addition, the Judges reheard the evidence and legal arguments as presented in the parties' briefs on the issue and, as a result, chose to adopt the revised definition. Copyright Owners maintain that for the Judges to do so would not be impermissible post-hoc reasoning. They note that the D.C. Circuit remanded precisely because the Judges did not provide any reason in the Determination for revising the Bundled Revenue definition. Copyright Owners note that it was the Services, not Copyright Owners, who appealed the Judges' modification of the bundled service revenue definition; thus, Copyright Owners cannot be penalized for not making every possible argument for affirmance. CO Reply at 70.

Further, and again notwithstanding the holding in Johnson, Copyright Owners argue that the Judges have the authority to engage in new agency action in this remand proceeding through a recasting of the Motion for Clarification as a motion for rehearing, pursuant to 17 U.S.C. 803(c)(2)(A) and 37 CFR 353.1. In this regard, Copyright Owners dismiss the point, raised by the D.C. Circuit, that their Motion for Clarification could not be recast as a motion for rehearing because Copyright Owners had explicitly disavowed that their motion sought rehearing under the statute, and that the Judges agreed. Rather, Copyright Owners maintain that the foregoing is not the same as a finding that the standard could not have been met. In Copyright Owners' view, the Judges could revisit on remand the question of whether the rehearing standard has now been met, and find that Copyright Owners have satisfied the “exceptional case” standard for granting rehearing motions under section 803(c)(2). Copyright Owners add that if the Judges do engage in new agency action that reconsiders the Motion for Clarification as a motion for rehearing, the Judges should fully explain their reasoning. Id. at 8–10.

However, Copyright Owners urge that proceeding in that fashion would add an entirely unnecessary and complicating step. They again suggest that there is no need to reconsider or recharacterize the Motion for Clarification as a motion for rehearing because the remand itself affords the opportunity for the Judges to take new agency action, which, as in a rehearing, permits them to reconsider evidence and arguments, but, unlike a rehearing, is not limited by the constraints of section 803(c)(2). See Copyright Owners' . . . Additional Briefing on New Agency Action . . . Question, etc., Tab B at 7–8 (Feb. 24, 2021).

2. The Services' Arguments

The Services' arguments are based on the reasoning of the D.C. Circuit in Johnson. Specifically, they assert that the D.C. Circuit found only “three ways in which the [Judges] can revise Initial Determinations” via “new agency action,” and the Judges failed to establish that the change to the service revenue definition fit any of those three categories. Services' Initial Submission at 64–65 (citing Johnson at 390).

According to the Services, the first statutory way the Judges may revise an Initial Determination is to “order rehearing `in exceptional cases' in response to a party's motion, 17 U.S.C. 803(c)(2)(A).” Services' Initial Submission at 65 (citing Johnson at 390). The Services argue that the D.C. Circuit held in Johnson that the Judges' “material revision of the `[Bundled] Revenue' definition . . . does not fall within the [Judges'] rehearing authority under section 803(c)(2)(A)” because “the [Judges] [themselves] . . . w[ere] explicit that [they] `did not treat the [Copyright Owners'] motion[ ]' . . . `as [a] motion[ ] for rehearing under 17 U.S.C. 803(c)(2).'” Id. The D.C. Circuit also noted that “as the [Judges] found, . . . Copyright Owners' motion did `not meet [the] exceptional standard for granting rehearing motions' under section 803(c)(2).” Id. (citing Johnson at 390). The Services assert, quoting Johnson once more, that the Judges were not able to make “a volte-face” and justify on appeal their revision to the definition as an exercise of rehearing authority. As the D.C. Circuit held, agency action must be justified by “reasons invoked by the agency at the time it took the challenged action,” and post-hoc rationalizations are insufficient. Id. (citing Johnson at 390).

The Services add their view that the Judges cannot revisit the decision to deny rehearing without engaging in impermissible post-hoc reasoning. They note the Supreme Court has explained that, while an agency may “elaborate later” on its “initial explanation” of the reason (or reasons) for its action, it “may not provide new ones.” Services' Initial Submission at 66, citing e.g., Regents at 1908. The Services offer that the Judges, having stated that they did not consider the Copyright Owners' motion to revise the definition to be a motion for rehearing, cannot now conclude that the motion qualified as one for rehearing and that the Judges in fact engaged in rehearing. Id.

The Services next argue, relatedly, that the Judges cannot simply recast the Services Motion for Clarification as a rehearing motion in an attempt to satisfy the rehearing standard. In this regard, they maintain that Copyright Owners did not argue before the Judges or the D.C. Circuit that their Motion for Clarification satisfied the “exceptional cases” standard, and have therefore waived that argument. Id.

The Services assert that the second statutory way the Judges may revise an Initial Determination, viz. taking “new agency action” to correct a technical or clerical error under section 803(c)(4), cannot be used to justify the modification of the Bundled Revenue definition in the Initial Determination. The Services note that the D.C. Circuit held specifically that the Judges' change in the Bundled Revenue definition could not be construed as correcting a technical or clerical error because it involved a substantive rewrite of the Service revenue definition. Id. at 67 (citing Johnson at 391).

The Services argue that the third and final statutory justification for the Judges to engage in “new agency action” is to revise the terms in an Initial Determination is in response to “unforeseen circumstances” that would frustrate the proper implementation of the determination. Id. at 67. The Services note that the D.C. Circuit held in Johnson that this authority did not justify the Judges' change to the Bundled Revenue definition because the Judges did not invoke this authority and “the need to ground the original definition in the record” could not credibly be described as “an unforeseen circumstance.” Id. (citing Johnson at 391).

The Services also note that the D.C. Circuit rejected the argument that the Judges have an “inherent authority”—unmentioned in the statute—to make changes to the Initial Determination. The D.C. Circuit explained that the specific restrictions Congress placed on the [Judges'] authority in section 803 “would be a nullity if [they] also had plenary authority to revise [their] determinations whenever [they] thought appropriate.” Id. (citing Johnson at 391–92). The Services add that even if the Judges offered a new source of authority capable of justifying substantive changes to the [Bundled] Revenue definition now, the Judges would be unable to rely on this “uninvoked authority” without engaging in impermissible post-hoc reasoning. Id.

The Services also reject Copyright Owners' position that the Judges may sidestep the D.C. Circuit's ruling by issuing a new determination on remand and simply arguing that any ruling after remand qualifies as new agency action pursuant to Johnson. The Services argue that failure to address the legal and factual issues on which the court remanded would violate the D.C. Circuit's decision and would result in yet another remand. The Services emphasize that the issue of authority to make the changes to the Initial Determination are especially important in this context, because the D.C. Circuit recognized that the Copyright Act places limits on the Judges' authority to alter an initial determination by defining conditions for rehearing and the types of changes that are permitted absent a rehearing. In this regard, the Services maintain that the Judges cannot do on remand what they lacked authority to do in the first instance. The Services assert that the Judges must resolve the legal question of whether authority exists to alter the revenue definition in the Initial Determination. Services' Reply at 52–54.

The Services also take note of the alternative path available to the Judges: to provide a “fuller explanation” of the prior conclusion that the Judges had legal authority to revise the Service Revenue definition. The Services maintain that if the Judges pursue the “fuller explanation” path, the Judges are limited to elaborating on what they said previously, and that they cannot add new reasons they did not initially provide. Id. at 54–55; see also Services' Joint Rebuttal Brief Addressing the Judges' Working Proposal at 38–42 (Feb. 24, 2022) (“Services' Additional Submission”).

The Services address Copyright Owners' position that if the only possible outcome were for the Judges to reinstate a definition that lacked any explanation or evidentiary support solely because it was present in the Initial Determination, then the D.C. Circuit would not have remanded the issue but would have simply reversed and reinstated the Initial Determination definition. The Services urge that the D.C. Circuit could not reverse because the CRB's appellate counsel had raised—for the first time on appeal—new justifications for the Judges' decision to change the Initial Determination. Instead, the Services maintain, the D.C. Circuit had to remand and give the Judges the opportunity to address appellate counsel's new justifications in the first instance, as the D.C. Circuit could not rule them out given the posture of the appeal. Services' Reply at 56.

VII. Analysis and Decision

A. The Procedural Issue: Is There “New Agency Action” Available to the Judges?

Having considered the parties' arguments, I conclude that the rulings in Johnson, which clearly rejected all of the Majority's procedural arguments seeking to justify their switch in the Bundled Revenue definition, foreclose any avenue for procedurally justifying this definitional switch. More particularly, I conclude that none of the procedural avenues proffered by Copyright Owners would constitute “new `agency action'” consonant with the holdings in Johnson. Further, I cannot identify any other procedural device ( i.e., an extra-statutory form of agency action) that would permit the switched definition in a manner consistent with Johnson. In addition, I cannot identify any further or fuller explanation that might support the Majority's procedural reasoning for swapping out the Bundled Revenue definition in the Initial Determination and substituting the definition in the Determination.

In reaching this conclusion, I take note of the following specific language in Johnson:

Section 803 identifies three ways in which the Board can revise Initial Determinations. It can (i) order rehearing “in exceptional cases” in response to a party's motion; (ii) correct “technical or clerical errors,”; and (iii) “modify the terms, but not the rates” of a royalty payment, “in response to unforeseen circumstances that would frustrate the proper implementation of [the] determination.”

Johnson at 390 (citations omitted). After identifying these three alternatives, the D.C. Circuit concluded that the CRB Judges “rollout of an entirely new manner for calculating the streaming service revenue from bundled offerings fit none of those categories.” Id.

First, I consider whether in the present case they can engage in “new `agency action'” pursuant to 17 U.S.C. 803(c)(2)(A) by recasting Copyright Owners' Motion for Clarification as a Motion for Rehearing. I conclude that this avenue has been unambiguously cut-off by Johnson and, indeed (as noted in Johnson), by the Judges' own prior ruling:

The [CRB Judges'] material revision of the Bundled Revenue definition . . . does not fall within [their] rehearing authority under Section 803(c)(2)(A). We have that on no less an authority than the [CRB Judges themselves], [who were] explicit that [they] “did not treat the Copyright Owners' motion” to have the definition changed “as a motion] for rehearing under 17 U.S.C. 803(c)(2).” That is because the Copyright Owners' motion did not “request[ ] a literal rehearing of evidence or legal argument.”

Nor could they have because, as the [CRB Judges] found, the Copyright Owners' motion did “not meet [the] exceptional standard for granting rehearing motions” under Section 803(c)(2). . . . [The CRB Judges] explain[ed] that . . . Copyright Owners “failed to make even a prima facie case for rehearing under the [rehearing] standard”.

Johnson, 369 F.3d at 390.

Further cutting off this “rehearing” approach, Johnson also expressly holds that it is a “forceful” principle that the D.C. Circuit “cannot sustain action on grounds that the agency itself specifically disavowed. Id. Moreover, in this Initial Remand Ruling I echo the Majority's ruling in the Clarification Order that Copyright Owners had failed to present “even a prima facie case for rehearing under the applicable standard”. Clarification Order at 2.

Next, I consider whether the Judges can engage in “new `agency action'” by recharacterizing their switch of the Bundled Revenue definition as an attempted correction of “technical or clerical errors,” pursuant to their “continuing jurisdiction” under section 803(c)(4). Once again, they cannot, and the D.C. Circuit has effectively explained why this is so:

The [Judges] do[ ] not even try to squeeze [their] substantive rewrite of the Service Revenue definition into that [§ 803(c)(4)] category. Quite the opposite, the [Judges] admit[ ] that the new definition “represent[s] a departure” from the definition in the Initial Determination, and was a substantive swap designed to “mitigate” the alleged “problem” of the original definition leaving the interactive streaming service providers free to “obscure royalty-based streaming revenue by offering product bundles that include music service offerings with other goods and services[.]” . . . To that same point, the order itself labels the initial and new definitions “diametrically-opposed approaches to valuing bundled revenues.” . . . . Nothing technical or clerical about that.

Johnson at 391.

On remand, I am unable to ascertain any basis for describing or justifying the changed Bundled Revenue definition as a technical or clerical correction. Thus, I conclude that the Judges cannot engage in “new `agency action'” pursuant to this section.

Next, I consider whether the Judges can engage in “new `agency action'”— by trying to squeeze the square peg of their definitional swap into the round hole that is the “unforeseen circumstances” clause in section 803(c)(4). That provision permits the Judges to exercise “continuing jurisdiction” if necessary to modify a regulatory term in a determination in response to “unforeseen circumstances,” if the absence of modification would frustrate the proper implementation of the determination. Once again, Johnson shuts the door:

Come oral argument, the [Judges] attempted to explain that “the unforeseen circumstances would be that [they] initially adopted a definition that was not supported by the record, and that was in fact substantively unreasonable and would frustrate the proper implementation of their determination.” . . . It is hard to see how the need to ground the original definition in the record was an unforeseen circumstance. That is Administrative Law 101. See also17 U.S.C. 803(c)(3) (“A determination of the [Judges] shall be supported by the written record.”).

Johnson at 391 (cleaned up). I agree. The present panel of Judges is bound by the D.C. Circuit's ruling that the overlooking of the need to ground in the factual record the Bundled Revenue definition in the Initial Determination cannot constitute an “unforeseen circumstance.” Accordingly, I am unable to ascertain any basis for describing or justifying the changed Bundled Revenue definition as an “unforeseen circumstance” that would justify their invocation of “continuing jurisdiction.”

I further consider the argument (made by the Judges' appellate counsel and by Copyright Owners) that the Judges have the “inherent authority sua sponte to make any `appropriate' substantive . . . or ‘fundamental’ changes after the Initial Determination . . . that [they] believe[ ] serve `the interests of enhancing the clarity and administrability of the regulatory terms accompanying the [ ] Determination.'” Johnson at 391. The D.C. Circuit made short work of this argument as well, stating that, although the CRB Judges have “considerable freedom” with regard to determining their own procedures

that flexibility must be exercised within the lines drawn by the authorizing statute. Congress's decision to limit rehearing to “exceptional cases,” and to confine other post hoc amendments to cases involving “technical or clerical errors,” would be a nullity if the [Judges] also had plenary authority to revise [their] determinations whenever [they] thought appropriate. The [Judges] nowhere in [their] order or the [ ] Determination explain[ ] the source of [their] power to make “fundamental” changes under the authorizing statute . . . any time [they] deem such changes “appropriate” . . . even after the Initial Determination.

Johnson at 392.

As with regard to the proffered rationales discussed supra, I cannot identify any authority that would allow the Judges to declare for themselves in the present factual and legal context an “inherent” authority to override the Copyright Act and declare their right to engage in “new `agency action.'”

Finally, I consider Copyright Owners' suggestion that the remand itself by the D.C. Circuit permits the Judges, pursuant to the Copyright Act, to engage in any procedure necessary to support their switch in the Bundled Revenue definition. The present Majority essentially adopts this procedural approach. However, I reject that argument as meritless.

The argument begins with a correct premise but seriously veers off course. Copyright Owners correctly note (and the Services do not disagree) that this remand proceeding constitutes “new `agency action.'” Copyright Owners then maintain that, because the Copyright Act does not provide for procedures that govern remand proceedings, the Judges are statutorily unconstrained with regard to the procedures they may adopt. This premise, although perhaps correct in other contexts, is most definitely incorrect in this specific context, given the clear holding in Johnson.

Here, the D.C. Circuit has been unequivocal in identifying the statutory limitations that precluded the Judges from switching out the Bundled Revenue definition in their Initial Determination and replacing it with a different definition in the Determination that was, to use the Majority's phrase, “diametrically opposed” to the prior definition, in that it would eliminate the royalty-based incentive to price discriminate via bundling. But Copyright Owners assert that the remand itself clothes the Judges with the procedural authority to make the very switch that Johnson forbids! I do not understand the D.C. Circuit to have admonished the Majority for its failure to respect the boundaries of its jurisdiction, only to provide them, via remand, with a back-door through which they may circle-back and exceed those very boundaries.

A reading of section 803(a), upon which Copyright Owners rely, provides a further demonstration of the error in this argument. This subsection lists the authorities whose pronouncements the Judges must “act in accordance with,” including, quite unsurprisingly, “the decisions of the court of appeals under this chapter.” 17 U.S.C. 803(a). In the instant case, the D.C. Circuit has unambiguously held that the Judges lacked the statutory authority to make the definitional switch at issue. For the Judges to construe that clear ruling as an implicit invitation to create new extra-statutory remand procedures that contradict the D.C. Circuit's rationale for the remand would be inexplicable and would render useless the procedural ruling in Johnson.

In sum, I cannot and do not understand that the D.C. Circuit intended in Johnson simply to write a meaningless procedural opinion that the Judges could not merely ignore, but use to cleanse the very procedural error the D.C. Circuit had condemned.

Accordingly, the Bundled Revenue definition in the Initial Determination should be reinstated. As explained in the portion of the Initial Remand Ruling in which I join, this reinstatement is harmonious with the entirety of the Judges' findings and conclusions regarding the other remanded issues.

B. The Substantive Issue: The Dueling Definitions of Bundled Revenue

1. Introduction: The Issue as Framed in the Clarification Order

Regarding the definition of “Service Revenue” from bundled offerings, the Judges summarized the parties' competing arguments:

Copyright Owners presented evidence that the existing approach led, in some cases, to an inappropriately low revenue base—but did so in service to their argument that the Judges should reject revenue-based royalty structures. They did not present evidence to support a different measure of bundled revenue because their rate proposal was not revenue-based.

The Services rely on the fact that the approach to bundled revenue in the extant regulations is derived from the 2012 Settlement. The Judges have, however, declined to rely on the 2012 Settlement as a benchmark, as the basis for the rate structure, or, therefore, as regulatory guidance. The Services have observed correctly that the evidentiary records in Web IV and SDARS III differ from the record in this proceeding.

Clarification Order at 17 (emphasis added).

Despite these arguments, the Judges found that neither party presented evidence adequate to support the approach advocated in post-determination filings, because the “economic indeterminacy problem inherent in bundling” remained unresolved. Id. The Judges stated that the Services were the party in possession of the relevant information, and concluded that the Services bore the burden of providing evidence that might mitigate the “indeterminacy problem” inherent in bundling. Because the Judges concluded that the Services had not met that burden, they ruled that they must adopt an approach to valuing bundled revenue that is in line with what the Copyright Owners proposed. As a result, the Judges discarded the formula in the Initial Determination and ruled, instead, that streaming service providers will use their own standalone price (or comparable) for the music component (not to exceed the value of the entire bundle) when allocating bundled revenue. Id. at 16–18.

On remand, the parties have made the following arguments regarding the substance of the Bundled Revenue definition:

2. Copyright Owners

According to Copyright Owners, the prior Bundled Revenue definition in the Initial Determination failed to address the “ ‘economic indeterminacy’ problem inherent in bundling” appropriately and in a way consistent with Judges' precedent. CO Initial Submission at 75 (citing Clarification Order at 16–18). Copyright Owners proceeded to cite several portions of testimony from the Services' economic experts who acknowledged this problem. Id. They then point to hearing testimony in which Copyright Owners repeatedly raised the “economic indeterminacy” problem and demonstrated what they characterized as the absurd results to which the prior definition had led. Id. at 76. They pointed out that under the Initial Determination, the first step in computing Bundled Revenue was to identify revenues recognized from the entire bundle ( i.e., the price paid by the subscriber). The second step was to subtract “the standalone published price” for all non-music components of the bundle. According to Copyright Owners, [REDACTED]. Id. at 76, 83.

Copyright Owners point out that the Judges already found with respect to other licenses that such an approach is not only fundamentally unfair, but “absurd.” Id. (citing Web IV,81 FR 26316, 26382 (May 2, 2016) (webcaster licenses); see also SDARS III,83 FR 65210, 65264 (Dec. 19, 2018) (SDARS licenses) (rejecting proposed deductions by service from bundle revenues because of the “acknowledged ‘economic indeterminacy’ problem inherent in bundling”). Copyright Owners concur with the Judges' conclusion that the same reasoning applies to Phonorecords III. Id. at 76–77 (citing Clarification Order at 18 (“the ‘economic indeterminacy’ problem inherent in bundling is common to all three proceedings.”)). Copyright Owners offer that Spotify conceded to this flaw in the definition in the Initial Determination, but offered an alternative that contained the same loophole. Id. at 77–78.

Copyright Owners also point out that the proponent of a term bears the burden of proof as to adoption. The Judges made clear that the licensee who wishes to offer bundles 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 the licensors. Id. at 79 (citing SDARS III,83 FR 65264 (“bundling [is] undertaken to increase [the Services'] revenues and it would be reasonable to assume that [the Services have] information relevant to the economic allocation of the bundled revenue.”)). Copyright Owners contend they presented unrebutted evidence showing the unreasonableness of the Services' proposed definition while the Services offered no evidence to support their definition. Id. at 78, 79 (citing Clarification Order at 18). Copyright Owners maintain that no Service offered evidence concerning the separate values of the constituent parts of the bundles, or any other evidence concerning the economic allocation of bundled revenue, let alone the reasonableness of the definition in the Initial Determination. Id. at 80. Copyright Owners assert that in the absence of evidence to support the proposed definition, the Judges may adopt or fashion a definition of service revenue for bundled offerings that comports with the record evidence, which is precisely what the Judges did and, through new agency action, do again. Id. at 81.

They further argue that the hearing record and the Judges' precedent and reasoning further explain the unreasonableness of the prior definition and support the adopted bundle revenue definition. Id. at 82. Copyright Owners offer that in contrast to the Services' evidentiary failure, they have provided sufficient evidence showing the unreasonableness of the Services' proposed definition. They maintain that the definition adopted by the Judges in the Determination was consistent with the statutory factors and the evidence in the proceeding showing how the prior definition had been manipulated and “led, in some cases, to an inappropriately low revenue base.” Id. at 83 (citing Clarification Order at 17–18).

Copyright Owners dispute the Services' assertion that there is support for the Phonorecords II approach to bundles in the record of this proceeding. Instead, Copyright Owners argue, the Services' purported evidence at most supports the benefits of the practice or strategy of bundling. They maintain that the strategy of bundling covered music services with other products or services has nothing to do with [REDACTED]. They offer that the definition in the Initial Determination has nothing to do with such benefits, and that those benefits may be equally served by a definition that ensures value is apportioned to the music component in the bundle. CO Reply at 73–76.

3. The Services

The Services argue that the evidence in the existing written record addressing bundles shows both that this definition is supported by the Phonorecords II benchmark and that it has proven industry-wide benefits. Services' Initial Submission at 68. They emphasize that Copyright Owners did not propose an alternative definition of service revenue until after the Judges issued the Initial Determination and that any definition they propose now would fail the basic requirement that the Judges must adopt rules “on the basis of a written record.” Id. (citing 17 U.S.C. 803(a)(1) and 803(c)(3)).

Addressing the merits of the definition contained in the Initial Determination, the Services argue that it best serves the goals of the Copyright Act; that as a bright-line, easily administered rule, it continues the broad industry agreement from Phonorecords II, which “was negotiated voluntarily between the Services and . . . Copyright Owners—strong evidence that its terms are mutually beneficial.” Services' Initial Submission at 69.

The Services contend the prior negotiated definition increases output and incentivizes beneficial price discrimination to reach casual and passive listeners who would otherwise not pay for music and thus would not generate revenue from which royalties could be paid. With regard to [REDACTED]. Id. at 71 (and record citations therein).

They further state that the definition of Bundled Revenue in Phonorecords II also enabled funneling of many of listeners into full-priced, full-catalog services. The Services allege that Copyright Owners also ignore the extensive royalties that were generated. They add that, for casual/passive listeners and those who may be funneled to subscription services, the per-subscriber minimum guarantees that the Copyright Owners will still be paid a fair royalty. The Services then cite several portions of testimony from various Services' economic experts who point out the realization of an expanded royalty pool, which the Services offer as proving a functioning marketplace. Id. at 68–74.

The Services maintain that while neither the Services nor Copyright Owners submitted evidence specifically addressing the way that customers, Services, or Copyright Owners might value the component parts of bundles, such subjective valuations are unnecessary—given that the parties' negotiated handling of the bundling issues provides the Judges with ample support for the PR II-based benchmark definition in the Initial Determination. See id. at 75–76.

The Services also argue that while the Judges' decision in SDARS III did involve valuation of the music and non-music components of a bundle, the resolution in SDARS III is inapposite because, here, the rate structure has a way of ensuring that Copyright Owners are fairly compensated from bundles: the statutory minimum payment. Services' Reply at 62.

C. Analysis and Decision

The fundamental difference between the impact of the two alternative definitions is simply stated:

Under the Initial Determination: downstream bundling and its price discriminatory effect would be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle;

Under the (Final)Determination: downstream bundling and its price discriminatory effect would not be incentivized by a royalty structure that reflects the lower WTP of consumers who subscribe by paying for a Bundle.

To explain this difference, the Judges find it helpful to describe (as in the Determination and Dissent) how bundling facilitates price discrimination and how lower royalties for bundled streaming services incentivize such bundling.

Price discrimination occurs when a seller offers different units of output at different prices. See, e.g., H. Varian, Intermediate Economics at 462 (8th ed. 2010). The benefit to the seller arises from attempting to “charge each customer the maximum price that the customer is willing to pay for each unit bought.” R. Pindyck & D. Rubinfeld, Microeconomics at 401 (8th ed. 2013). For all goods, and intellectual property goods such as copyrights in particular, the social benefit is that price discrimination more closely matches the quantity sold with the competitive quantity as the seller or licensor better aligns the price with the WTP of different categories of buyers or licensees. See W. Fisher, Reconstructing the Fair Use Doctrine, 101 Harv. L. Rev. 1659, 1701 (1988).

A seller can engage in price discrimination in several ways. One form is known as “second-degree price discrimination,” by which buyers self-sort the packages and quantities they purchase. See W. Adams & J. Yellen, Commodity Bundling and the Burden of Monopoly, 90 Q. J. Econ. 470, 476 (1976) (the profitability of bundling “stem[s] from its ability to sort customers into groups with different reservation price [WTP] characteristics.”). Bundling, i.e., the “practice of selling two or more products as a package,” Pindyck & Rubinfeld, supra at 419, is thus a type of second-degree price discrimination. See A. Boik & H. Takahashi, Fighting Bundles: The Effects of Competition on Second Degree Price Competition, 12 a.m. Econ. J. 156, 157 (2020).

The applicability of these basic economic principles was understood and explained by the parties' experts at the hearing. See, e.g., 3/15/17 Tr. 1224–25 (Leonard) (Google's economic expert testifying that price discrimination through bundling is “very, very common . . . even by pretty competitively positioned firms . . . to sort out customers into willingness-to-pay groups.”); 3/30/17 Tr. 3983 (Gans) (Copyright Owners' economic expert acknowledging that bundling is a form of price discrimination); see also Dissent at 69 (same).

How does this downstream (retail level) benefit of price discrimination impact the setting of upstream royalty rates? As the Majority explained (in summarizing the Services' expert testimony) the linkage is explained by the economic concept of “derived demand”:

[M]ultiple pricing structures necessary to satisfy the WTP and the differentiated quality preferences of downstream listeners relate directly to the upstream rate structure to be established in this proceeding. Professor Marx opines that the appropriate upstream rate structure is derived from the characteristics of downstream demand. 3/20/17 Tr. 1967 (Marx) (rate structure upstream should be derived from need to exploit WTP of users downstream via a percentage of revenue). This upstream to downstream consonance in rate structures represents an application of the concept of “derived demand,” whereby the demand upstream for inputs is dependent upon the demand for the final product downstream. Id.; see P. Krugman & R. Wells, Microeconomics at 511 (2d ed. 2009) (“[D]emand in a factor market is . . . derived demand. . . [t]hat is, demand for the factor is derived from the [downstream] firm's output choice”).

Determination at 19; accord Dissent at 32 (noting that “the upstream demand of the interactive streaming services for musical works (and the sound recordings in which they are embodied)—known as `factors' of production or ‘inputs’—is derived from the downstream demand of listeners to and users of the interactive streaming services . . . This interdependency causes upstream demand to be characterized as “derived demand.”).

In the present proceeding, the PR II-based benchmark embodies the parties' negotiated definition of Bundled Revenue for purposes of calculating royalties on bundled interactive offerings. This is the definition in the Initial Determination. Copyright Owners' preferred definition for Bundled Revenue—the Determination's definition—would not only ignore this agreed-upon definition, but would also de-link the royalty rate from the WTP of purchasers of bundles. The Judges recognize that Copyright Owners have expressed concern the Services could use such bundling in order to diminish revenue otherwise payable on a higher royalty tier. However, the Majority noted that the evidence indicated such diminishment only occurred “in some cases” and that such practices were not “sweeping.” Clarification Order at 17, 21. Thus, the Judges find that eliminating the incentive for price discrimination via bundling would be a disproportionate response and inconsistent with the broad price discriminatory PR II-based benchmark they find useful in this proceeding.

Expert testimony in this regard is “substantial evidence” on which the Judges can rely. For example, the D.C. Circuit also relied in Johnson on the testimony of the same witness, Spotify's economic expert witness, Professor Marx, to affirm the inclusion of the price discriminatory structure for student and family plans. Johnson, 969 F.3d at 392–94. Professor Marx explained how a downstream “lower willingness (or ability) to pay” among some cohorts of consumers supports definitional terms, for student and family subscribers, that lower royalty rates in order to further “economic efficiency” in a manner that “still allows more monetization of that provision of that service.” Johnson at 392–93. Broadening her lens, Professor Marx also explained that this price discriminatory approach is appropriate “across all types of services and subscribers,” as in “[t]he current law [and in the PR II-based benchmark]” which “accommodates . . . ad-supported services . . . and ` bundled services ' through different rate provisions.” Marx WRT ¶ 41 (emphasis added). See also 3/21/17 2182–83 (Hubbard) (Amazon's expert witness testifying that “Prime Music, which is bundled with an Amazon Prime service . . . sort[s] out customers' willingness to pay, with an idea of trying to maximize the number of customers,” and agreeing that this approach constitutes “sorting by way of bundling.”) (emphasis added). Further, Professor Hubbard opined that, given the revenue attribution “measurement problem” associated with bundled products, the “Phonorecords II” approach “with the different categories and the minima . . . address this sort of problem [in] a very good way.” 3/15/17 Tr. 1221 (Hubbard).

As in the case of family and student price discrimination, the beneficial effect of such differential pricing was supported by industry witnesses as well as expert witnesses. See, e.g., Mirchandani WDT ¶ 71 (Amazon executive citing the Phonorecords II-based benchmark provisions regarding bundling that “allowed Amazon to bundle Prime Music with Amazon Prime, enabling Amazon to bring a limited catalog of music [REDACTED]”). In sum, the same type of witness testimony that the D.C. Circuit found sufficient to support price discriminatory student and family plans also supports the use of the price discriminatory bundled definition contained in the Initial Determination.

Given the overall benefits from price discrimination, at first blush it is curious that Copyright Owners would risk “leaving money on the table” by seeking to remove the royalty-based incentive for price discrimination via bundling. The Judges have identified this problem earlier in this Initial Remand Ruling, in connection with the broader issue of the overall beneficial price discriminatory structure of the PR-based benchmark. As the Judges noted in that general price discrimination context, Copyright Owners' own expert economic witnesses acknowledged that they would not irrationally leave money on the table. In fact, Copyright owners' aim, according to that testimony, is to create an unregulated space—per the Bargaining Room theory—and to use their complementary oligopoly power to negotiate price discriminatory rates (in bundles or otherwise), which would free them from the section 801(b)(1) requirements of reasonableness and fairness.

The Judges further find that their prior ruling on this issue in SDARS III is distinguishable. There, a proffered bundled revenue definition eliminated the payment of any royalty at all. Copyright Owners quite correctly describe that result as “absurd,” but that is not the result here. Rather, in the present case, the parties' negotiated an approach that the Judges adopted in the Initial Determination requiring royalties to be paid on interactive services bundled with other products or services.

Even more distinguishable is Copyright Owners' assertion that Web IV provides support for their preferred definition of service revenue. The argument is immediately suspect, because Web IV involved per-play royalty rates—not percent-of-revenue rates, making the definition of revenue wholly inapposite. Further, the discussion of the price of an “ice cream cone” in Web IV —on which Copyright Owners rely—had nothing to do with bundling or isolating the WTP for different products or services. Rather, there the Judges criticized a bizarre argument made by a licensee (who had a quantity discount for plays steered in its direction), that was tantamount to arguing that if a vendor sells one ice cream cone for $1.06 but a buyer could buy two for $1.06, that the market price of an ice cream cone is thus only $.06. This argument was indeed fallacious, because the price of an ice cream cone would be reasonably identified as the average of the total cost for the two cones, i.e., $.53/cone, and never as $.06 per cone.

Here, the issue, is how to address the WTP of different classes of buyers with heterogeneous WTP, not the pricing of a quantity discount. The parties addressed this issue by utilizing the Bundled Revenue definition contained in the PR II-based benchmark (and in the Initial Determination) to address the indeterminacy inherent in the variable WTP among purchasers of the bundles, by setting floors and minima, rather than attempt to sort out the WTP of individual (or individual blocs) of subscribers. The “ice cream cone” issue in Web IV is wholly unrelated, and the SDARS III situation, as explained supra, is also distinguishable.

For the foregoing reasons, I find that—even if the Judges had a procedural mechanism by which to support the switch in the Bundled Revenue definition—I would decline to utilize it in this Initial Remand Ruling, because the definition in the Initial Determination (unlike the definition in the Determination) is consistent with the Judges' other substantive rulings herein. That is, just as the Majority abandoned its Bundled Revenue definition in its Initial Determination because it refused to credit the PR II-based benchmark (even as “guidance”), the Judges here do partially rely on the PR II-based benchmark, and thus find that it supports the Bundled Revenue definition contained in the Initial Determination.

VIII. Application of the Four Itemized Statutory Factors

As the forgoing analysis explains, bundling is a form of price discrimination. Accordingly, the Judges' explanation of how price discriminatory rates in the PR II-based benchmark interrelate with the Factor A through D objectives in section 801(b)(1) are equally applicable here. Accordingly, the Judges incorporate by reference here their discussion of those four factors set forth supra in connection with the PR II-based benchmark, and find that there is no basis pursuant to those four factors to adjust the PR II-based benchmark definition of Bundled Revenue.

IX. Conclusion

This Dissent in part is issued as a RESTRICTED document. Within 30 days of the date of issuance, the participants shall file a version of this Dissent with agreed redactions to permit viewing by the public.

Issue Date: July 2, 2022.

David R. Strickler,

Copyright Royalty Judge

D. Dissent in Part Re Benchmark (Redacted Version With Federal Register Naming and Formatting Conventions)

The Copyright Royalty Judges (Judges) sit as a panel in all determination proceedings. See17 U.S.C. 803(a)(2). A majority of two Judges is sufficient to issue a determination. See17 U.S.C. 803(a)(3). If any Judge dissents from the majority determination, that dissenting Judge may issue a dissenting opinion and file it with the majority's determination. Id. The Judges accept this same standard with regard to their issuance of the present Initial Ruling and Order after Remand (Initial Ruling).

The undersigned Judge, author of this dissent in part (Benchmark Dissent) respectfully dissents from the Initial Ruling of the majority (Remand Majority) on the issue of adopting as a benchmark for current rates and terms the rates and terms adopted after a settlement by the parties to the preceding phonorecords proceeding. It should be noted that the Remand Majority adopts the rate structure from Phonorecords II, but retains the headline percent-of-revenue rate adopted in the Determination.

I. Areas of Concurrence

A. Background Statements

The Benchmark Dissent adopts the statements regarding the background and procedural posture of this remand proceeding. See Initial Ruling at 1–2.

B. Percent of Revenue Rate

The Benchmark Dissent agrees with the Remand Majority's retention of the headline percent-of-revenue rate and its phase-in over the period at issue.

C. Definition of Service Revenue for Bundled Offerings

For the reasons articulated in the Initial Ruling and the reasoning of the judge dissenting from that portion of the Initial Ruling, the definition of Service Revenue for bundled offerings contained in the Initial Determination must be adopted. See Initial Determination (Jan. 27, 2018). Adoption of the Phonorecords II (PR II) rate structure requires that the original definition pertain.

II. Area of Dissent

The first function of the Judges is “to make determinations . . . of reasonable terms and rates of royalty payments. . . .” 17 U.S.C. 801(b)(1). Under the statute in effect during the captioned proceeding, the rates shall be calculated to achieve four statutory objectives. Id. The terms of payment of the rates, however, are not subject to any particular statutory restrictions or guidelines. See, e.g., Live365 v. Copyright Royalty Bd., 698 F. Supp. 2d 25, 29–30 (D.D.C. 2010) (“In performing their duties, the [Judges have] broad discretion to . . . impose regulations governing the rates and terms of copyright royalties. . . .”).

In general, in promulgating regulations the Judges aim to effect efficient and effective payment of royalty license fees. Regulations relating to license royalty rates describe the rates the Judges determine to be reasonable, whether presented by agreement of the affected parties or after adjudication. The regulations include, where necessary, methods of calculation of the payable royalties. The regulations also include such provisions as recordkeeping requirements, late fee assessments, and audit authority. As the Remand Majority points out, simplicity and clarity were not among the statutory factors applicable to determining royalty rates in the captioned underlying proceeding. Simplicity and clarity should, however, be paramount among the Judges' considerations in governing rate payment procedures.

In recent proceedings, the Judges have emphasized that the statute requires that they set both rates and terms. At the end of a different royalty rate proceeding, having been confronted with competing proposed regulations, or even with largely agreed regulatory terms, upon which the parties had proffered no evidence, the Judges cautioned counsel in this proceeding:

Please be reminded that the Judges have an obligation to set both rates and terms. . . . In any proceeding, just because a regulation is in the current Code of Federal Regulations does not mean that the Judges are adopting that term. . . . The Judges cannot determine rates or terms without an evidentiary record. . . . The Judges cannot adopt any terms of royalty administration unless the parties present evidence to support their proposed terms.

Tr. 03/08/2017 (Barnett, J.) While chapter 8 of the Copyright Act encourages settlement, the Judges are not mandated to adopt parties' settlements if they find they face opposition that discounts reasonableness or if the proposed regulations are contrary to law. See, e.g., Determination of Royalty Rates and Terms . . . (Phonorecords IV),87 FR 18342, 18347, 18349 (Mar. 30, 2022).

In the proceeding underlying the Determination, the parties proffered a variety of proposed regulations. Copyright Owners contended that the extant rate structure “should be modified and simplified.” Copyright Owners' Amended Proposed Rates and Terms (5/17/2017) at 2. Copyright Owners argued that the ten different rate categories should be “no longer applicable” as Copyright Owners proposed application of the same rates and rate structure to “all interactive streams and/or limited downloads [except bundles], regardless of the business model employed.” Id. at 3. Copyright Owners' rate proposal hinged on a per-unit calculation across the board: the greater of a per-play amount or a per end user amount.

Amazon proposed retaining the PR II rate structure. See Proposed Findings . . . of Amazon (May 13, 2017) ¶ AM–F–25. Amazon argued that the PR II rate structure “enabled Amazon to develop a varied assortment of services. . . .” Id. Amazon contended that the different royalty rates permit price discrimination by the Services. Id. ¶¶ AM–F–47, 49. Amazon conflates price discrimination with provision of heterogenous musical tastes and preferences. Id. ¶ AM–F–48. Amazon's proposal mimicked the regulations adopted by agreement in the immediately prior proceeding.

Apple proposed a per-play rate calculation, which would render the PR II rates and rate structure obsolete. Notwithstanding the different structure, however, Apple offered valid criticisms of the PR II rate structure. Apple termed the PR II rate structure “problematic.” See Apple Inc.'s Findings . . . and Conclusions . . . (May 11, 2017) at 30. Apple argued that the PR II rate structure was “overly complex, economically unsound, and unpredictable.” Id. ¶ APL–F65. Apple acknowledged that these shortcomings resulted in “a loss of trust and overall dissatisfaction with interactive streaming among songwriters. . . .” Id.

Apple noted that across the ten rate categories in the PR II rates, “there are roughly 79 different calculations that can be made.” Id. ¶ APL–F67. Apple argued that the PR II rate structure was “not transparent or easy to understand” for copyright owners and created “uncertainty for services, who may find it difficult to predict which prong . . . will kick in in any given month.” Id. ¶¶ 68–69. Apple opined that, rather than encouraging new business models, the PR II rate structure “tends to stifle innovation around new pricing or distribution models, as services are incentivized to create businesses that fit into the ten pre-defined `boxes.' ” Id. ¶ 70. Apple further argued that the PR II rates were economically unsound because they are based on revenue, which is unrelated to demand for a given copyright owner's song. Id. ¶ 71.

Google's proposal, from which the Majority derived the uncapped TCC rate prong of the Determination, contended that the “fragmented service categories are unnecessary under [its] proposal. . . .” Google, Inc.'s Proposed Findings . . . and Conclusions. . . (May 11, 2017) ¶ GPFF58. Google acknowledged questions regarding the complexity of the PR II rate structure. Google, therefore proposed a rate structure that would both streamline the regulations and protect Copyright Owners' concerns regarding Services' revenue deferment and displacement. Id. ¶ GPFF57.

In the captioned underlying proceeding, the Judges heard little evidence offered in resounding support or vehement objection to the regulations the parties proffered. No party argued or supported the proposition that the PR II rate structure was the only way, or even the best way, to achieve license fee payment.

In this remand proceeding, no party argued against the all-in approach to rate calculation. The parties disagreed regarding retention of “mechanical floors” for configurations for which the Services must pay mechanical royalties both to Copyright Owners in this proceeding under section 115 and to Performing Rights Organizations (PROs) according to the determinations of the “Rate Court.” The parties disagreed over imposition of a cap on the TCC prong in the greater-of percent-of-revenue calculation. They also disagreed over retention or elimination of the per subscriber sub-minima that were featured in the PR II rates.

The Remand Majority cites with approval the remand parties' criticism of the simplified rate structure in the Determination, viz., that it is “virtually as complex as” the PR II rate structure. See Services' Joint Opening Brief (Apr. 1, 2021) at 39. This characterization is a bit of hyperbole. The rate structure in the Determination is an all-in rate with “mechanical floors” where those are warranted. Except for the fundamentally different configurations included in subpart B, it does not set out separate calculations for different delivery configurations. On remand, the Remand Majority chooses to reinstate the PR II rate structure in its entirety, with all of its 79 permutations, changing only the headline percent-of-revenue rate and adding a cap on the TCC rate prong (which is an element of the structure itself). The Benchmark Dissent does not dispute the necessity and propriety of the increased headline percent-of-revenue rate or the cap on the TCC rate prong. Indeed, as noted in the Remand Majority, the D.C. Circuit endorsed the rate increase as well-reasoned and determined well within the Judges' discretion. The D.C. Circuit also found fault with “yoking” the TCC rate alternative to sound recording royalty rates, not subject to the Judges' control, without reins. The basis of this Benchmark Dissent is simply that the regulatory scheme is not efficient, transparent, or mandated by credible evidence; nor is the structure necessary to achieve the purposes of reasonableness and equity.

A. Acceptance of Phonorecords II Settlement as a Proper Benchmark

This is a Dissent in Part. The undersigned Judge does not disagree with the headline rate being retained at 15.1% or with the imposition of a TCC cap, for the reasons elucidated by the Remand Majority. Nonetheless, the Benchmark Dissent continues to disagree with adoption of the entirety of the rate structure adopted by Phonorecords II. As noted above, the Judges solicited evidence to support adoption of regulatory language to effect payment of the rates they established. Copyright Owners, Google, and Apple submitted rate proposals that greatly simplified the rate structure. Their rate structure regulation proposals were crafted to support their varying approaches to rate calculations not adopted by the Judges. Their criticisms of the PR II rate structure are valid, nonetheless, and support the Benchmark Dissent's analysis.

In the underlying proceeding, the Majority declined to label the rate structure and resulting rates incorporated in the regulations promulgated after the Phonorecords II proceeding (rates and rate structure) as a benchmark, or starting point, for determination of new rates and terms in that proceeding. In the Determination in the extant proceeding, the Majority alluded to reasons they found the PR II rates to be inadequate to serve current circumstances. The D.C. Circuit noted that appellate counsel offered further explanation on appeal for the rejection of the PR II rates and rate structure as a benchmark. See Johnson v. Copyright Royalty Board, 969 F.3d 363, 387 (D.C. Cir. 2020). Nevertheless, the D.C. Circuit faulted the Majority for not providing adequate explanation of their rejection of a PR II-based benchmark in the first instance. See id. Indeed, the D.C. Circuit found the Majority's reasoning on the issue in the Determination to be “muddled.” Id. at 386–87.

Copyright Owners argue that the D.C. Circuit's remand for further explanation did not equate to finding error in the Judges' rejection of the PR II-based benchmark. See Initial Remand Submission of Copyright Owners (Apr. 1, 2021) 1, 10 (CO Initial Submission). Notably, the Services did not address the question of a finding of error, but proposed on remand a rate structure substantially similar to that in PR II and offered a benchmark analysis therefor. See Services' Joint Opening Brief (in Services' Joint Written Direct Remand Submission at Tab D) (Apr. 1, 2021) at 19 (Services' Initial Submission).

While the Copyright Owners' parsing of Johnson might be technically correct, the Benchmark Dissent nonetheless accepts the wisdom of revisiting the analysis of the PR II rates and rate structure, focusing on the intricacies of the structure that ultimately come into play in determining the amount of royalty payable. The Benchmark Dissent disagrees that the record in this case demands adoption of the PR II rate structure as a suitable benchmark. The Benchmark Dissent hereby provides a full analysis of this issue, which includes a fuller explanation of the conclusions in the Determination and supports and justifies rejection of the Phonorecords II rate structure.

B. Attributes of a Useful Benchmark

As repeated by the parties in the initial proceeding and in their remand submissions, for an exemplar to serve as a useful benchmark, it must be compared to the target market. The hallmarks of a useful benchmark are: (1) unity of products, (2) unity of sellers, and (3) unity of buyers. In addition, (4) economic circumstances and market conditions can influence the value of a benchmark. See Services' Initial Submission at 20 (citing Determination of Royalt[ies] for Transmission of Sound Recordings. . .,83 FR 65210, 65214 (Dec. 19, 2018) ( SDARS III).

In the Remand Majority opinion, the Judges argue that the PR II rate structure meets “most of the requisites for a useful benchmark.” See Initial Ruling, section III. C. 3. Assuredly, in the real world one is unlikely to find a perfect benchmark; consequently, the Judges in these proceedings look to the best available benchmark(s) and make adjustments to compensate for their shortcomings when compared to the attributes and circumstances of the target rates. The Benchmark Dissent is not so sanguine about one's ability to reconcile the PR II rate structure with current market circumstances pertaining to music streaming (including participants and volumes of sales) almost a decade after the parties agreed to that structure. Because of the recognized gulf in market conditions between Phonorecords II and this Phonorecords III proceeding, the Benchmark Dissent rejects attempts to fit that square peg into the current round hole.

1. Unity of Products—the Same Rights

The PR II rates regulated “sales” of the same licensing rights as those at issue in the current underlying proceeding, viz., the statutory license to utilize musical works embodied in the sound recordings that are the lifeblood of the music streaming services. This factor was not and is not in controversy. In this respect, the Judges could look to the PR II rates as a benchmark.

2. Unity of Sellers—Rightsholders

The songwriter or songwriters own the copyright for musical works, that is, the musical notes and lyrics. In general, songwriters sell or license their works to publishers who fix the works to a physical medium, for example, piano rolls or sheet music. Music publishers also market the musical works licenses to record companies for their sound recordings. In today's market, publishers and songwriters exist in a symbiotic relationship. Without new works, the publishers have no new product to market. To ensure a flow of new product, publishers often subsidize songwriters by providing working space or monetary advances on future sales of licensed work, or publishers might purchase outright the songwriters' copyrights. Whether the rightsholder is a writer, composer, or publisher, the rights are the same, those derived from 17 U.S.C. 106 and limited by 17 U.S.C. 115. See17 U.S.C. 106(1), (3) (exclusive rights); sec. 115 (compulsory licensing). The sellers' interests are aligned.

3. Unity of Buyers—Streaming Services

The Services argue unity of rights and sellers between the time of the PR II rates and the current proceeding. With respect to buyers, the Services allege that the current buyers are “the same or similar. . . .” Services' Initial Submission at 20. The Services argue that the PR II rates involved “either the same type of buyers or the very same buyers as this proceeding.” Id. The license delimits the users it binds. It is axiomatic that current licensees are “of the same type” as licensees in 2012. Describing participants as “similar to those currently in the market” or “of the same type” as current participants is sufficiently imprecise to call into question the unity of buyers required to give great weight to a potential benchmark.

The Services allege that “[m]ost of the participants in Phonorecords III were either directly involved in the Phonorecords II settlement or operated in the market at the time of the settlement.” Id. “Most of the participants” does not reveal which participants were active in Phonorecords II or the reasons for their participation. Amazon began an MP3 digital music service in 2004; it launched steaming in mid-2014. See Written Direct Testimony of Jeffrey Eisenach (Nov. 3, 2016) (Eisenach WDT) ¶ 51. Tab. 2. Apple launched its streaming service in 2019. During the Phonorecords II negotiations, Apple's primary interest was digital downloads from the iTunes store. According to one of its witnesses, Google was, at the time of the Phonorecords II negotiations, “planning to launch a store, a locker, and a subscription service.” Google's participation in the Phonorecords II negotiations was “primarily designed to make sure that our interests were met in—for our forthcoming music service.” 3/8/17 Tr. 157:2–158:2 (Zahavah Levine).

Although the Services argue that the buyers in the current market are the same as, or similar to, buyers at the time of adoption of the PR II rates, the Services then and now advocate differing rate calculations for each music delivery configuration. Indeed, between 2008 and 2012, the delivery configurations multiplied and the parties negotiated different rate structures for those multiple configurations. Acknowledging participation by a service with one configuration—or a plan to launch one configuration—is insufficient to establish a unity of buyers for purposes of rate setting. Almost a decade after the effectuation of the 2012 rates, with new businesses tacking music streaming onto their digital ecosystems, the development of new and different delivery configurations continues to evolve. Nonetheless, the Services would have the Judges adopt a rate structure that specifies current delivery configurations but excludes some current innovations and cannot encompass the next innovations, whatever form they might take.

The Benchmark Dissent acknowledges that buyers of the musical works for which licenses are at issue in this proceeding are of the “same type” as the Phonorecords II buyers. In some instances, they are the same participants. In the current landscape, however, the interests of those buyers are vastly different. The extent to which Apple, Amazon, and Google, were involved in Phonorecords II negotiations bears no resemblance to the interests of those services and their current service configurations. Without greater unity of buyers, the Benchmark Dissent must discount the viability of the PR II rates or rate structure as a useful benchmark in this proceeding.

4. Economic and Market Conditions

The Services argue that the music streaming industry in 2018 was essentially unchanged from 2008 or 2012. See Services' Initial Submission at 20–21. The evidence in this proceeding compels a contrary conclusion. In 2008, musical works distribution consisted primarily of sound recordings reproduced in physical formats (vinyl and CDs) and digital downloads. See Eisenacht WRT ¶ 33 (Feb. 13, 2017). The record reflects that in 2008, of record labels' revenues 96% were derived from sales of physical and digitally downloaded sound recordings; 2.5% from interactive streaming. By 2012, at the inception of the rates that were re-adopted as the PR II rates, musical works sales were beginning to shift from physical media to digital forms. In 2012, 8.1% of record label revenues were attributable to interactive streaming. Id. By 2015, evidence available in this proceeding showed that record labels' revenues from digital downloads approximately equaled revenues from streaming and digital sales were more than double the sales of physical configurations, such as vinyl and CDs. Id. ¶¶ 44–45 and accompanying tables.

Spotify, the dominant pure play streaming service in the U.S., did not enter the U.S. market until mid-2011. See CO Initial Submission at 20–21 (Apr. 1, 2021) and evidence cited therein. Spotify did not participate in the negotiations leading up to the adoption of the 2012 musical works royalty rates. See Eisenacht WRT ¶ 35, n.38. In fact, the record contains evidence that music streaming was not a major factor in setting mechanical license rates in 2008 or 2012. See CO Initial Submission at 19–21, and evidence cited therein. As more and larger streaming services entered the market, music consumption changed in character. Music consumption in the 2018 market had changed character completely from an ownership model to an access model. See Determination at 6.

Further, three of the Services participating in the current proceeding are not pure play streaming services but are multidimensional marketing firms for whom music streaming is only one small facet of the business. From the perspective of those current licensees, the music streaming license is relatively insignificant to their overall financial health. The Judges must, therefore, value the license objectively to assure the conglomerate licensees do not manipulate their revenues so as to reduce music streaming rights below what is fair and reasonable to the rightsholders.

The Services further advocate use of the PR II rates and rate structure as a benchmark because they assert that the multifaceted rate structure is reflective of the Services' own price discriminatory services. The Majority noted the Services' price discrimination as a way to optimally monetize segments of the market with a lesser willingness to pay. Greater accommodation of users less willing to pay results in more streaming and more revenue for the Services at minimal to no marginal cost. A rate determined as a percentage of a service's revenue allows that price discrimination to continue, resulting in additional royalties. The Benchmark Dissent contends, however, that the Judges need not adopt a rate structure with ten different service categories to allow the Services to continue their price discriminatory downstream sales. The payable royalties are a percent of revenue. If the Services receive relatively less revenue by marketing a family plan, for instance, that reduced revenue is the basis for the royalty calculation. Nothing in a simplified rate structure would inhibit price discriminatory service plans. The PR II rates' multi-category structure might encompass the price discrimination the Services employ, but that does not make it a mandatory benchmark for current rates, especially if the target rate structure permits the same flexibility.

C. Adoption of PR II Rates and Rate Structure in Direct Licenses

The Services assert that the PR II rates and rate structure have been adopted in negotiated direct licenses they have signed with rightsholders rendering those rates and that rate structure a valuable benchmark. The Services' witnesses analyzed direct licenses and concluded that the rates closely matched the rates in the PR II regulations. [REDACTED]. Analysis of direct licenses executed belie the Services' assertion that the PR II rates structure is embraced by rightsholders.

D. Additional Shortcomings of PR II Rates as a Benchmark

The D.C. Circuit dismissed the Majority's argument on appeal that (1) the PR II rates were too low and (2) the PR II rates were outdated. The D.C. Circuit noted that these two reasons might support the Majority's conclusions, but they could not be asserted in the first instance on appeal. See Johnson at 386.

1. Rates Too Low

The D.C. Circuit found that the Judges' finding that the PR II rates were too low was not fully articulated until the matter was on appeal. As a result, the D.C. Circuit could not evaluate that reason as support for the final rates. Indirectly, however, the D.C. Circuit nonetheless accepted that underlying reason for the rate changes when it approved the higher rates themselves. See Johnson at 384–86. The adopted rates were soundly grounded in the record evidence. See id. By implication, acceptance of increased rates means the PR II rates were too low to be continued. With or without the “too low” rationale, the final adopted rates prove the point.

2. Rate Structure Outdated

In the Determination, the Majority cited several factors that implied the inadequacy of the PR II rates and rate structure as a compelling benchmark for Phonorecords III. As discussed above, the music streaming industry in 2018 was completely transformed from 2008 or 2012. Both the buyers and the economic market conditions were markedly changed. Referring to the PR II rates as “outdated” encompasses both a temporal element and a structural component.

a. Significance of the Passage of Time

Music streaming in the earlier rate setting periods was in its infancy. Listeners had not yet fully embraced the subscribed access model for music consumption. By 2018, listeners could choose from “a diverse array of streaming offerings.” See WDT of Rishi Mirchandani ¶ 63. Such industry shifts alone could render the PR II rates “outdated.”

b. Clarity and Simplicity

Another salient factor the Majority addressed is the rate structure itself. To understand the PR II rate structure, one needed ten separate full-page flow chart diagrams, each featuring three formulae for calculating greater-of and lesser-of rate components. See Trial Ex. 846. The rates for some consumption configurations included a per-subscriber “mechanical floor” as a failsafe against overreaching by PROs, should the Rate Court increase their rates to an extent that all of the section 115 all-in percent of revenue royalty be consumed by the PROs. See, e.g., [FORMER] 37 CFR 385.13(a)(1) (Standalone non-portable subscription—streaming only [$.15 per subscriber]); [FORMER] 385.13(a)(2) (Standalone portable subscription—mixed use [$.50 per subscriber]) (2018). Other consumption configurations included “minima;” that is a lesser-of calculation comparing a percent of sound recording license costs (TCC) and a per subscriber amount. See, e.g., [FORMER] 37 CFR 385.13(b) (2018). Further, rate calculations differed depending upon, for example, whether the listener streamed on a portable device or a non-portable device; or whether the listener purchased access to the music alone from a pure-play streaming service or as part of a bundled offering, such as “free” streaming for a limited period included in the purchase price of the streaming device.

The rationale for these convoluted rate calculation differences is unknown. They were the product of confidential negotiations among the parties involved in the music streaming business in the first decade of the 21st century. One side of the negotiating table sought reconsideration of those rates. The current licensees are not the same as those who negotiated the 2012 rollover of the 2008 rate scheme. Music streaming business models have witnessed significant growth and change. Meanwhile, the business models employed by songwriters and publishers remain largely unchanged—and not realizing a proportionate capture of the stream of dollars realized by the Services' monetization of ever-more consumption configurations. The marginal cost to the Services of additional streams, regardless of the business configuration or the user's reception device, is zero. The Services, therefore, are in a position to capture increased revenue without an increase in cost of goods sold.

In the end, a sound recording embodying a licensed musical work is being delivered to an end “user”: one song; one listener. The calculation of what royalty the songwriter is entitled to should not rest on the medium of transmission or the location of the listening. See WDS Steve Bogard ¶ 34 (“Streaming music anytime, any place, on any device is the way today's music fans want to enjoy their music. Notwithstanding that the inherent value of a song is the same whether the consumer chooses to buy an album, permanently download an album or a single, or stream music on demand. . . .”). The incremental difference in value to the listener of hearing a song in the car as opposed to through earbuds during a workout is not likely measurable. Certainly, no participant in this proceeding presented any evidence of the relative value of a song to a listener depending on the delivery configuration.

In the interest of making government more transparent and accessible to interested citizens, less is more. Opaque systems and formulae are or should be, in a word, outdated. The fact of settlement does not cure or even address the unnecessary complication of paying a royalty for the use of a statutory license under the PR II rates structure. More importantly, owners of the copyrights being licensed should be able to comprehend, calculate, and verify the sources and amounts of their royalty payments.

3. Not Business Model Neutral

The Services contend that the PR II rate structure is preferable as it is business model neutral. Nothing in the record supports that assertion. In fact, Apple argued that the PR II rate structure stifled innovation as streaming services sought to fit any new business into a business model already defined as one of the ten identified models in the Phonorecords II regulations. The statute does not require that rate structures be business model neutral. The reasonableness requirement demands, however, that the Judges find and adopt reasons for differentiation in rates based on business models.

4. No Evidence of Settling Parties' Subjective Intent

Copyright Owners participating in the current proceeding argued that the Judges should consider the subjective intent of the parties in agreeing to “roll over” the 2008 rates and rate structure into the PR II regulations. The Services countered that subjective intent is irrelevant, as the product of those negotiations serves as objective evidence of the parties' intents. On this question, the Services are correct. The negotiated rates show, objectively, that the negotiating parties agreed to a certain rate structure. The D.C. Circuit criticized the Majority for not including in the Final Determination an explanation of why the subjective intent of the parties to the settlement was a “prerequisite” to adoption of that settlement as a benchmark. See Johnson at 387. The Judges need not, however, accept that objective evidence uncritically.

Negotiating parties' subjective state of mind can serve as convincing evidence of the economic circumstances and the state of the market at the time of the negotiations. While ascertaining the parties' subjective intent in reaching the settlement is not a “prerequisite” to examination of the terms as a benchmark, the Benchmark Dissent finds subjective intent informative and useful as one factor in weighing the value of the settlement as a benchmark.

E. Statutory Factors

The Services argued to the D.C. Circuit that the Majority's rejection of the PR II rates and rate structure was erroneous because the Majority failed to evaluate that structure and those rates under the statutory factors delineated in 17 U.S.C. 801(b)(1). Evaluation under section 801(b)(1) is required by the statute applicable to this proceeding. Nothing in section 801(b)(1) compels the Judges to evaluate compliance with the statutory factors of every proposed potential rate or rate structure. Neither are the Judges required to evaluate every potential benchmark or past rate structure under section 801(b)(1). The Judges are obliged to evaluate any rate structure they intend to adopt against the requirements of section 801(b)(1). If the Judges' promulgated rate structure meets the section 801(b)(1) standard, then the promulgated rate structure can be adopted. Whether other possible proposals might also meet the section 801(b)(1) standard is not at issue in a proceeding.

1. Maximize the Availability of Creative Works to the Public

The Services argue that the PR II rates and rate structure support and contribute to the maximization of musical works. As evidence, they cite the growth of music streaming overall, the profitability of all segments of the music industry. It is beyond question that music consumption has grown exponentially since the co-incident introduction of portable devices and streaming services. Growth continues as those devices and services become increasingly easy to actuate in vehicles.

No participant alleged, however, that music industry success is caused by or even correlated to the PR II rate structure. Coincidence is not probative evidence.

2. Assure Fair Return to Copyright Owner and Fair Income to the Licensee

The Services argued they were receiving a fair income and copyright owners were receiving a fair return under the PR II regulations. Although the Services argued that overall music royalties absorbed an inordinate portion of their revenues, none expressly laid that lack of available revenue at the door of mechanical royalties. Amazon's witness, Dr. Glenn Hubbard described a growing increase in streaming industry revenues and forecasts of continuing growth. See WRT of Glenn Hubbard (Feb. 15, 2017) ¶ 2.23–24 (Hubbard WRT). Dr. Hubbard deconstructed Amazon's increased revenues and concluded that the growth in streaming services' revenue resulted in increased royalty payments to music publishers and other rights holders. Id. ¶ 3.10. When royalty rates are calculated on a percent-of-revenue, the royalty payments increase when revenues increase.

The difficulty with this tautological argument is that revenue growth as between services and rightsholders has not been proportional. And, as Copyright Owners have argued, the rate at which the services share with mechanical rightsholders is the issue in this proceeding. The Judges are not called upon to set annual royalty payment dollar amounts; rather they are mandated to set the rates that drive those dollar amounts. And to adopt regulations that most closely effectuate actual payment to rightsholders, minimizing revenue deferral and other such loopholes. For all of the reasons provided in the Determination and in this Benchmark Dissent, the PR II-based rates and the controlling rate structure do not balance the section 115 fair income-fair return scale appropriately and reasonably.

3. Weigh Relative Roles of Licensors and Licensees in Making the Works Available to the Public

No participant presented evidence to elucidate specifically the relative roles of the parties relating to musical works. Economic evidence assumed that the marginal cost of streaming more music is minimal. This does not discount the services' sunk costs, such as the original technological or capital investments. With respect to the contributions of the copyright owners, the contribution is clear. It all begins with a song. Without new music, the Services could continue by streaming unregulated works, new arrangements or covers of existing works, and non-music content. Whether they would continue to enjoy the growth they have enjoyed over the last decade is unknown. The PR II rates might be a contributing factor to both stability and growth of the industry, but based on the totality of the evidence, the Dissent concludes that with regard to musical works, the relative role of the creator of the musical works, and to a lesser extent, the music publisher, is undervalued.

4. Minimize Disruption

The language for the fourth statutory factor requires the Judges to establish a rate structure in such a way as “[t]o minimize any disruptive impact on the structure of the industries involved and on generally prevailing industry practices.” [FORMER] 17 U.S.C. 801(b)(1)(D). The Services argue that the change in rate structure determined by the Majority in this proceeding is massively, and potentially fatally, disruptive to music streaming services.

Ironically, the music industry has been in a constant state of disruption since the introduction of digital music. From peer-to-peer sharing, to purchased permanent downloads, to interactive and non-interactive streaming, the history of modern music consumption has been a model of disruption. Entry into the streaming market by multifaceted digital ecosystem providers is just the latest significant change in music delivery to consumers. Innovation in music delivery is constant.

Allegedly to minimize disruption, the Services advocated retention of the PR II rates and rate structure. While every aspect of the music industry is experiencing explosive growth, maintenance of the inadequate rates for mechanical licenses is unfathomable. Some change, phased in over time, might be uncomfortable for the licensees, but failure to change rates to acknowledge the music delivery revolution is not an option. With such a dynamic history and uncertain future, a change in mechanical license rates is not just inevitable, but mandatory.

Indeed, the Benchmark Dissent's approach in this proceeding advances the notion that streamed music is streamed music. This is certainly true from the viewpoint of the songwriters and publishers, and of music consumers. Rather than introduce separate rate structures for each new delivery technology or streaming business model, the Judges need to establish a rate that will fairly compensate Copyright Owners for the use of their works and permit a fair return to licensees, regardless of what next technological disruption they might choose to introduce to the industry. In the captioned proceeding, the Majority declined to label the rate structure and resulting rates incorporated in the regulations promulgated after the Phonorecords II proceeding as a benchmark, or starting point, for determination of new rates and terms in this proceeding.

In the Determination, the Majority alluded to reasons they found the PR II rates to be inadequate to serve current circumstances. Nevertheless, the D.C. Circuit faulted the Majority for not providing adequate explanation of their rejection of the PR II benchmark in the first instance. See Johnson at 386–87. Indeed, the D.C. Circuit found the Majority's reasoning on the issue in the Determination to be “muddled.” Id.

F. Rate Structure

For all of the reasons outlined above, the Remand Majority's acceptance and adoption of the Phonorecords II rate structure results in a rate structure in this proceeding that suffers from the same deficits the Benchmark Dissent believes to be inherent in that rate structure. Changing the headline rate and capping the TCC rate prong do not cure the ills of the rate structure itself. True, the PR II-based rates permit price discrimination, which increases revenue, and therefore royalties, in absolute terms. Reinstatement of minima in the TCC prong introduces a failsafe to runaway TCC-based rates. The mechanical floors adopted in the Determination continue, protecting mechanical license rightsholders from runaway performance royalties.

The Benchmark Dissent maintains that all these goals could be met equally well with a streamlined, transparent, fair, and reasonable rate structure, as several of the participants in this proceeding advocated.

III. Conclusion

This Dissent in part is issued as a RESTRICTED document. Within 30 days of the date of issuance, the participants shall file a version of this Dissent with agreed redactions to permit viewing by the public.

Issue Date: July 1, 2022.

Suzanne M. Barnett

Chief Copyright Royalty Judge

List of Subjects in 37 CFR Part 385

• Copyright
• Phonorecords
• Recordings

For the reasons set forth in the preamble, the Copyright Royalty Judges amend 37 CFR part 385 as follows.

PART 385—RATES AND TERMS FOR USE OF NONDRAMATIC MUSICAL WORKS IN THE MAKING AND DISTRIBUTING OF PHYSICAL AND DIGITAL PHONORECORDS

1. The authority citation for part 385 continues to read as follows:

Authority: 17 U.S.C. 115, 801(b)(1), 804(b)(4).

2. Add appendix A to read as follows:

Appendix A to Part 385—Part 385 Applicable to the Period January 1, 2018, through December 31, 2022, as clarified on August 10, 2023

Note:

Cross-references to part 385 in this appendix are to those provisions as contained within this appendix.

PART 385—RATES AND TERMS FOR USE OF MUSICAL WORKS UNDER COMPULSORY LICENSE FOR MAKING AND DISTRIBUTING PHYSICAL AND DIGITAL PHONORECORDS

Subpart A—Regulations of General Application

Subpart B—Physical Phonorecord Deliveries, Permanent Downloads, Ringtones, and Music Bundles

Subpart C—Eligible Interactive Streaming, Eligible Limited Downloads, Limited Offerings, Mixed Service Bundles, Bundled Subscription Offerings, Locker Services, and Other Delivery Configurations

Subpart D—Promotional Offerings, Free Trial Offerings and Certain Purchased Content Locker Services

Dated: July 3, 2023.

David P. Shaw,

Chief Copyright Royalty Judge

David R. Strickler,

Copyright Royalty Judge

Steve Ruwe,

Copyright Royalty Judge

Approved by:

Carla D. Hayden,

Librarian of Congress.