Determination of Reasonable Rates and Terms for the Digital Performance of Sound Recordings and Ephemeral Recordings

AGENCY:

Copyright Office, Library of Congress.

ACTION:

Final rule and order.

SUMMARY:

The Librarian of Congress, upon recommendation of the Register of Copyrights, is announcing the determination of the reasonable rates and terms for two compulsory licenses, permitting certain digital performances of sound recordings and the making of ephemeral recordings.

EFFECTIVE DATE:

July 8, 2002.

ADDRESSES:

The full text of the public version of the Copyright Arbitration Royalty Panel's report to the Librarian of Congress is available for inspection and copying during normal working hours in the Office of the General Counsel, James Madison Memorial Building, Room LM-403, First and Independence Avenue, SE., Washington, DC 20540. The report is also posted on the Copyright Office website at http://www.copyright.gov/carp/webcasting_rates.html.

FOR FURTHER INFORMATION CONTACT:

David O. Carson, General Counsel, or Tanya Sandros, Senior Attorney, Copyright Arbitration Royalty Panel (CARP), P.O. Box 70977, Southwest Station, Washington, DC 20024. Telephone (202) 707-8380. Telefax: (202) 707-8366.

SUPPLEMENTARY INFORMATION:

Table of Contents

I. Background

II. The CARP Proceeding to Set Reasonable Rates and Terms

A. The Parties

B. The position of the parties at the commencement of the proceeding

1. Rates proposed by Copyright Owners

2. Rates proposed by Services

C. The Panel's determination of reasonable rates and a minimum fee

III. The Librarian's Scope of Review of the Panel's Report

IV. The CARP Report: Review and Recommendation of the Register of Copyrights

A. Establishing Appropriate Rates

1. The “Willing Buyer/Willing Seller Standard”

2. Hypothetical Marketplace/Actual Marketplace

3. Benchmarks for setting market rates: voluntary agreements vs. musical works fees

a. Fees paid for use of musical works

b. Voluntary agreements

4. Alternative methodology: Percentage-of-revenue

5. The Yahoo! rates—evidence of a unitary marketplace value

6. Are rates based on the Yahoo! agreement indicative of marketplace rates?

7. Should a different rate be established for commercial broadcasters streaming their own AM/FM programming?

8. Methodology for calculating the statutory rates for the webcasting license

a. Calculation of the unitary rate

b. The 150-mile exemption

9. Rates for other webcasting services and programming

a. Business to business webcasting services

b. Listener-influenced services

c. Other types of transmissions

10. Rates for transmissions made by non-CPB, noncommercial stations

11. Consideration of request for diminished rates and long song surcharge

12. Methodology for estimating the number of performances

13. Discount for Promotion and Security

14. Ephemeral recordings for services operating under the section 114 license

15. Minimum fees

16. Ephemeral recordings for business establishment services (“BES”)

a. Rates for use of the statutory license

b. Minimum fee

17. Effective period for proposed rates

B. Terms

1. Disputed terms

a. Definitions

b. Designated Agent for Unaffiliated Copyright Owners

c. Gross proceeds

2. Terms Not Disputed by the Parties

a. Limitation of Liability

b. Deductions from Royalties for Designated Agent's Costs

c. Ephemeral Recording

d. Definition of “Listener''

e. Timing of Payment by Receiving Agent to Designated Agent

f. Allocation of Royalties among Designated Agents and Among Copyright Owners and Performers

g. Choice of Designated Agent by Performers

h. Performer's Right to Audit

i. Effective date

V. Conclusion

VI. The Order of the Librarian of Congress

I. Background

In 1995, Congress enacted the Digital Performance Right in Sound Recordings Act (“DPRA”), Public Law 104-39, which created an exclusive right for copyright owners of sound recordings, subject to certain limitations, to perform publicly their sound recordings by means of certain digital audio transmissions. Among the limitations on the performance right was the creation of a new compulsory license for nonexempt, noninteractive, digital subscription transmissions. 17 U.S.C. 114(f).

The scope of this license was expanded in 1998 upon passage of the Digital Millennium Copyright Act of 1998 (“DMCA” or “Act”), Public Law 105-304, in order to allow a nonexempt eligible nonsubscription transmission (the “webcasting license”) and a nonexempt transmission by a preexisting satellite digital audio radio service to perform publicly a sound recording in accordance with the terms and rates of the statutory license. 17 U.S.C. 114(a). In addition to expanding the section 114 license, the DMCA also created a new statutory license for the making of an “ephemeral recording” of a sound recording by certain transmitting organizations (the “ephemeral recording license”). 17 U.S.C. 112(e). The new statutory license allows entities that transmit performances of sound recordings to business establishments, pursuant to the limitations set forth in section 114(d)(1)(C)(iv), to make an ephemeral recording of a sound recording for purposes of a later transmission. The new license also provides a means by which a transmitting entity with a statutory license under section 114(f) can make more than the one phonorecord permitted under the exemption set forth in section 112(a). 7 U.S.C. 112(e).

The statutory scheme for establishing reasonable terms and rates is the same for both of the new licenses. The terms and rates for the two new statutory licenses may be determined by voluntary agreement among the affected parties, or if necessary, through compulsory arbitration conducted pursuant to Chapter 8 of the Copyright Act.

In this case, interested parties were unable to negotiate an industry-wide agreement. Therefore, a Copyright Arbitration Royalty Panel (“CARP”) was convened to consider proposals from interested parties and, based upon the written record created during this process, to recommend rates and terms for both the webcasting license and the ephemeral recording license.

II. The CARP Proceeding to Set Reasonable Rates and Terms

These proceedings began on November 27, 1998, when the Copyright Office announced a six-month voluntary negotiation period to set rates and terms for the webcasting license and the ephemeral recording license for the first license period covering October 28, 1998-December 31, 2000. 63 FR 6555 (November 27, 1998). During this period, the parties negotiated a number of private agreements in the marketplace, but no industry-wide agreement was reached. Consequently, in accordance with the procedural requirements, the Recording Industry Association of America, Inc. (“RIAA”) petitioned the Copyright Office on July 23, 1999, to commence a CARP proceeding to set the rates and terms for these licenses. The Office responded by setting a schedule for the CARP proceeding. See 64 FR 52107 (Sept. 27, 1999).

However, the schedule proved unworkable for the parties. RIAA filed a motion with the Copyright Office on November 23, 1999, requesting a postponement of the date for filing direct cases. It argued that the Office should provide more time for the parties to prepare their cases in light of the complexity of the issues and the record number of new participants. The Office granted this request and held a meeting to clarify the procedural aspects of the proceeding, especially for the new participants, and to discuss a new schedule for the arbitration phase of the process. Order in Docket No. 99-6 CARP DTRA (dated December 22, 1999). In the meantime, the Office commenced the six-month negotiation period for the second license period, covering January 1, 2001-December 31, 2002. 66 FR 2194 (January 13, 2000). Ultimately, the Copyright Office consolidated these two proceedings into a single proceeding in which one CARP would set rates and terms for the two license periods for both the webcasting license and the ephemeral recording license. See Order in Docket Nos. 99-6 CARP DTRA and 2000-3 CARP DTRA 2 (December 4, 2000). The 180-day period for the consolidated proceeding began on July 30, 2001, and on February 20, 2002, the panel submitted its report (the “CARP Report” or “Report”), in which it proposed rates and terms to the Copyright Office. It is the decision of this Panel that is the basis for the Librarian's decision today.

A. The Parties

The parties to this proceeding are: (i) The Webcasters, namely, BET.com, Comedy Central, Echo Networks, Inc., Listen.com, Live365.com, MTVi Group, LLC, Myplay, Inc., NetRadio Corporation, Radio Active Media Partners, Inc.; RadioWave.com, Inc., Spinner Networks Inc. and XACT Radio Network LLC; (ii) the FCC-licensed radio Broadcasters, namely, Susquehanna Radio Corporation, Clear Channel Communications Inc., Entercom Communications Corporation, Infinity Broadcasting Corporation, and National Religious Broadcasters Music License Committee (collectively “the Broadcasters”); (iii) the Business Establishment Services, namely, DMX/AEI Music Inc. (also referred to as “Background Music Services”); (iv) American Federation of Television and Radio Artists (“AFTRA”); (v) American Federation of Musicians of the United States and Canada (“AFM”) ; (vi) Association For Independent Music (“AFIM”) ; and (vii) Recording Industry Association of America, Inc. (“RIAA”). Music Choice, a Business Establishment Service, was initially a party to this proceeding, but on March 26, 2001, it filed a motion to withdraw from the proceeding. Its motion was unopposed and, on May 9, 2001, its motion to withdraw was granted.

B. The Position of the Parties at the Commencement of the Proceeding

1. Rates Proposed by Copyright Owners

RIAA proposed rates derived from an analysis of 26 voluntarily negotiated agreements between itself and individual webcasters. RIAA claims that these agreements “involve the same buyer, the same seller, the same right, the same copyrighted works, the same time period and the same medium as those in the marketplace that the CARP must replicate.” CARP Report at 26, citing RIAA PFFCL (Introduction at 8). Based upon these agreements, RIAA proposed the following rates for DMCA compliant webcasting services:

(i) For basic “business to consumer” (B2C) webcasting services:

0.4c for each transmission of a sound recording to a single listener, or 15% of the service's gross revenues.

(ii) For “business to business” (B2B) webcasting services, where transmissions are made as part of a service that is syndicated to third-party websites:

0.5c for each transmission of a sound recording to a single listener

(iii) For “listener-influenced” webcasting services:

0.6c for each transmission of a sound recording to a single listener

(iv) Minimum fee (subject to certain qualifications): $5,000 per webcasting service

(v) Ephemeral license fee:

10% of each service's performance royalty fee payable under (i), (ii), or (iii).

For the section 112 license applicable to the business establishment services, the copyright owners proposed a rate set at 10% of gross revenues with a minimum fee of $50,000 a year.

2. Rates Proposed by Services

Webcasters proposed per-performance and per-hour sound recording performance fees, based upon an economic model, that considered the aggregate fees paid to the three performance rights organizations (ASCAP, BMI, and SESAC) that license the public performances of musical works for radio programs that are broadcast over-the-air by FCC-licensed broadcasters, by 872 radio stations during 2000. From this model, the webcasters derived a per-song and a per-listener hour base rate of 0.02¢ per song and 0.3¢ per hour, respectively. These figures were then adjusted to account for a number of factors, including the promotional value gained by the record companies from the performance of their works. This adjustment resulted in a fee proposal of 0.014¢ per performance or 0.21¢ per hour.

At the end of the proceeding, Webcasters suggested in their proposed findings of fact and conclusions of law an alternative method for calculating royalty fees, namely, a percentage-of-revenue fee structure. Specifically, Webcasters proposed a fee of 3% of a webcaster's gross revenues for all services. The alternative proposal was made with the understanding that the service would be able to elect either option.

Webcasters proposed no additional fee for the making of ephemeral recordings and a minimum fee of $250 per annum for each service operating under the section 114 license.

The Business Establishment Services who need only an ephemeral recording license proposed a flat rate of $10,000 per year for each company.

C. The Panel's Determination of Reasonable Rates and a Minimum Fees

In this proceeding, the Panel had to establish rates and terms of payment for digital transmissions of sound recordings made by noninteractive, nonsubscription services and rates for the making of ephemeral phonorecords made pursuant to the section 112(e) license; either to facilitate those transmissions made or by business establishments which are otherwise exempt from the digital performance right.

The proposed rates are set forth in Appendix A of the CARP Report, which is posted on the Copyright Office website at: http://www.copyright.gov/carp/webcasting_rates_a.pdf.

The proposed terms of payment may be found in Appendix B of the CARP Report, which is posted on the Copyright Office website at: http://www.copyright.gov/carp/webcasting_rates_b.pdf.

III. The Librarian's Scope of Review of the Panel's Report

The Copyright Royalty Tribunal Reform Act of 1993 (the Reform Act), Pub. L. No. 103-198, 107 Stat. 2304, created a unique system of review of a CARP's determination. Typically, an arbitrator's decision is not reviewable, but the Reform Act created two layers of review that result in final orders: one by the Librarian of Congress (Librarian) and a second by the United States Court of Appeals for the District of Columbia Circuit. Section 802(f) of title 17 directs the Librarian on the recommendation of the Register of Copyrights either to accept the decision of the CARP, or to reject it. If the Librarian rejects it, he must substitute his own determination “after full examination of the record created in the arbitration proceeding.” 17 U.S.C. 802(f). If the Librarian accepts it, then the determination of the CARP becomes the determination of the Librarian. In either case, through issuance of the Librarian's Order, it is his decision that will be subject to review by the Court of Appeals. 17 U.S.C. 802(g).

The review process has been thoroughly discussed in prior recommendations of the Register of Copyrights (Register) concerning rate adjustments and royalty distribution proceedings. See, e.g., Distribution of 1990, 1991, and 1992 Cable Royalties, 61 FR 55653 (1996); Rate Adjustment for the Satellite Carrier Compulsory License, 62 FR 55742 (October 28, 1997). Nevertheless, the discussion merits repetition because of its importance in reviewing each CARP decision.

Section 802(f) of the Copyright Act directs that the Librarian shall adopt the report of the CARP, “unless the Librarian finds that the determination is arbitrary or contrary to the applicable provisions of this title.” Neither the Reform Act nor its legislative history indicates what is meant specifically by “arbitrary,” but there is no reason to conclude that the use of the term is any different from the “arbitrary” standard described in the Administrative Procedure Act (APA), 5 U.S.C. 706(2)(A).

Review of the case law applying the APA “arbitrary” standard reveals six factors or circumstances under which a court is likely to find that an agency acted arbitrarily. An agency action is generally considered to be arbitrary when:

1. It relies on factors that Congress did not intend it to consider;

2. It fails to consider entirely an important aspect of the problem that it was solving;

3. It offers an explanation for its decision that runs counter to the evidence presented before it;

4. It issues a decision that is so implausible that it cannot be explained as a product of agency expertise or a difference of viewpoint;

5. It fails to examine the data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made; and

6. Its action entails the unexplained discrimination or disparate treatment of similarly situated parties.

Motor Vehicle Mfrs. Ass'n. State Farm Mutual Auto. Insurance Co., 463 U.S. 29 (1983); Celcom Communications Corp. v. FCC, 789 F.2d 67 (D.C. Cir. 1986); Airmark Corp. v. FAA, 758 F.2d 685 (D.C. Cir. 1985).

In reviewing the CARP's decision, the Librarian has been guided by these principles and the prior decisions of the District of Columbia Circuit in which the court applied the “arbitrary and capricious” standard of 5 U.S.C. 706(2)(A) to the determinations of the former Copyright Royalty Tribunal (hereinafter “CRT or Tribunal”). See, e.g, National Cable Tele. Ass'n v. CRT, 724 F.2d 176 (D.C. Cir. 1983) (applying the Administrative Procedure Act's standard authorizing courts to set aside agency action found to be arbitrary, capricious, and abuse of discretion, or otherwise in accordance with law.”); see also, Recording Industry Ass'n of America v. CRT, 662 F.2d 1, 7-9 (D.C. Cir. 1981); Amusement and Music Operators Ass'n v. CRT, 676 F.2d 1144, 1149-52 (7th Cir.), cert denied, 459 U.S. 907 (1982); National Ass'n of Broadcasters v. CRT, 675 F.2d 367, 375 n. 8 (D.C. Cir. 1982).

Review of judicial decisions regarding Tribunal actions reveals a consistent theme; while the Tribunal was granted a relatively wide “zone of reasonableness,” it was required to articulate clearly the rationale for its award of royalties to each claimant. See National Ass'n of Broadcasters v. CRT, 772 F.2d 922 (D.C. Cir. 1985), cert. denied, 475 U.S. 1035 (1986) (NAB v. CRT); Christian Broadcasting Network v. CRT, 720 F.2d 1295 (D.C. Cir. 1983) (Christian Broadcasting v. CRT); National Cable Television Ass'n v. CRT, 689 F.2d 1077 (D.C. Cir. 1982) (NCTA v. CRT); Recording Indus. Ass'n of America v. CRT, 662 F.2d 1 (D.C. Cir. 1981) (RIAA v. CRT). As the D.C. Circuit succinctly noted:

We wish to emphasize * * * that precisely because of the technical and discretionary nature of the Tribunal's work, we must especially insist that it weigh all the relevant considerations and that it set out its conclusions in a form that permits us to determine whether it has exercised its responsibilities lawfully. * * *

Christian Broadcasting v. CRT, 720 F.2d at 1319 (D.C. Cir. 1983), quoting NCTA v. CRT, 689 F.2d at 1091 (D.C. Cir. 1982).

Because the Librarian is reviewing the CARP decision under the same “arbitrary” standard used by the courts to review the Tribunal, he must be presented by the CARP with a rational analysis of its decision, setting forth specific findings of fact and conclusions of law. This requirement of every CARP report is confirmed by the legislative history of the Reform Act which notes that a “clear report setting forth the panel's reasoning and findings will greatly assist the Librarian of Congress.” H.R. Rep. No. 103-286, at 13 (1993). This goal cannot be reached by “attempt[ing] to distinguish apparently inconsistent awards with simple, undifferentiated allusions to a 10,000 page record.” Christian Broadcasting v. CRT, 720 F.2d at 1319.

It is the task of the Register to review the report and make her recommendation to the Librarian as to whether it is arbitrary or contrary to the provisions of the Copyright Act and, if so, whether, and in what manner, the Librarian should substitute his own determination. 17 U.S.C. 802(f).

IV. The CARP Report: Review and Recommendation of the Register of Copyrights

The law gives the Register the responsibility to review the CARP report and make recommendations to the Librarian whether to adopt or reject the Panel's determination. In doing so, she reviews the Panel's report, the parties' post-panel submissions, and the record evidence.

After carefully considering the Panel's report and the record in this proceeding, the Register has concluded that the rates proposed by the Panel for use of the webcasting license do not reflect the rates that a willing buyer and willing seller would agree upon in the marketplace. Therefore, the Register has made a recommendation that the Librarian reject the proposed rates ($0.14 per performance for Internet-only transmissions and $0.07 per performance for radio retransmissions) for the section 114 license and substitute his own determination (0.07c per performance for both types of transmissions), based upon the Panel's analysis of the hypothetical marketplace, and its reliance upon contractual agreements negotiated in the marketplace.

These changes necessitate an adjustment to the proposed rates for non-CPB, noncommercial broadcasters for Internet-only transmissions as well. The adjusted rate for archived programming subsequently transmitted over the Internet, substituted programming and up to two side channels is 0.02¢, reflecting a downward adjustment from the 0.05¢ rate proposed by the Panel. The new rate for all other transmissions made by non-CPB, noncommercial broadcasters is 0.07¢ per performance per listener. Using this methodology, the Register recommends that the Librarian also reject the Panel's determination of a rate for the making of ephemeral recordings by those Licensees operating under the webcasting license. Because the Panel had made an earlier determination not to consider 25 of the 26 contracts submitted by RIAA for the purpose of setting a rate for the webcasting license, it was arbitrary for the Panel to use these same rejected licenses to set the Ephemeral License Fee. See section IV.13 herein for discussion. Consequently, the Register proposes a downward adjustment—from 9% of the performance royalties paid to 8.8%—to the Ephemeral License Fee to remove the effect of the discarded licenses.

In determining the Ephemeral License Fee for Business Establishment Services operating under an exemption to the digital performance right, the CARP considered separate licenses negotiated in the marketplace between individual record companies and these services. Its reliance on these agreements as an adequate benchmark for purposes of setting the rate for the section 112 license was well-founded and supported by the record. Therefore, the Register recommends adopting the Panel's proposal of setting the Ephemeral License Fee for Business Establishment Services at 10% of the service's gross proceeds. However, the Register cannot support the Panel's recommendation to set the minimum fee applicable to these services for its use of the ephemeral license at $500 when clear evidence exists in the contractual agreements to establish a much higher range of values for setting the minimum fee. Consequently, the Register evaluated the contracts and proposed a minimum fee consistent with the record evidence. The result is a minimum fee of $10,000 per license pro rated on a monthly basis.

Section 802(f) states that “[i]f the Librarian rejects the determination of the arbitration panel, the Librarian shall, before the end of that 90-day period, and after full examination of the record created in the arbitration proceeding, issue an order setting the royalty fee or distribution of fees, as the case may be.” During that 90-day period, the Register reviewed the Panel's report and made a recommendation to the Librarian to accept in part and reject in part the Panel's report, for the reasons cited herein. The Librarian accepted this recommendation and, on May 21, 2002, he issued an order rejecting the Panel's determination proposing rates and terms for the webcasting license and the ephemeral recording license. See Order, Docket No. 2000-9 CARP DTRA 1&2 (dated May 21, 2002).

The full review of the Register and her corresponding recommendations are presented herein. Within the limited scope of the Librarian's review of this proceeding, “the Librarian will not second guess a CARP's balance and consideration of the evidence, unless its decision runs completely counter to the evidence presented to it.” Rate Adjustment for the Satellite Carrier Compulsory License, 62 FR 55757 (1997), citing 61 FR 55663 (October 28, 1996) (Distribution of 1990, 1991 and 1992 Cable Royalties). Accordingly, the Register accepts the Panel's weighing of the evidence and will not question findings and conclusions which proceed directly from the arbitrators' consideration of factual evidence. The Register, however, may reject a finding of the Panel where it is clear that its determination is not supported by the evidence in the record.

A. Establishing Appropriate Rates

1. The “Willing Buyer/Willing Seller Standard”

Sections 112(e)(4) and 114(f)(2)(B), of title 17 of the U.S.C., provide that “the copyright arbitration royalty panel shall establish rates and terms that most clearly represent the rates and terms that would have been negotiated in the marketplace between a willing buyer and a willing seller,” and enumerate two factors that the panel shall consider in making its decisions: (1) The effect of the use of the sound recordings on the sale of phonorecords, and (2) the relative contributions made by both industries in bringing these works to the public. In applying this standard, the Panel determined that it was to consider the enumerated factors along with all other relevant factors identified by the parties, but that it was not to accord the listed factors special consideration. Report at 21; see also Final Rule and Order, Rate Adjustment for the Satellite Carrier Compulsory License, Docket No. 96-3 CARP SRA, 62 FR 55742, 55746 (October 28, 1997).

Nevertheless, when the Panel considered the record evidence offered to establish a marketplace rate, it paid close attention to the two factors set forth in the statute. In analyzing the first factor, which focuses on the interplay between webcasting and sales of phonorecords, the panel found that the evidence offered during the proceeding was insufficient to demonstrate whether webcasting promoted or displaced sales of sound recordings. RIAA's evidence to demonstrate that performances of their sound recordings over the Internet displace record sales consisted of unsupported opinion testimony and consequently, the Panel afforded it no weight. Report at 33. Similarly, the Panel rejected the Webcasters' contention that webcasting promoted sales, affording little weight to its empirical studies. It concluded that the Sounddata survey was not useful for purposes of this proceeding because it focused on the promotional value of traditional radio broadcasts and not the promotional value of webcasting. Id. Likewise, the Panel rejected a study by Professor Michael Mazis because the response rates in the survey study fell below generally acceptable standards. All in all, the evidence on either side was not persuasive. Consequently, the Panel concluded that, for the time period under consideration, “the net impact of Internet webcasting on record sales [was] indeterminate.” Id. at 34.

Broadcasters, however, disagree with the Panel's conclusions. They argue that the Panel should have made an adjustment for the promotional value of the transmissions, noting that the statute singled out this factor for consideration when setting the rates. Broadcasters Petition at 38. They further contend that the record demonstrates that “the promotional value of radio play should be far and away the most significant factor in determining the fair market value of broadcasters simulcast rates.” Id. at 39-40. But all the evidence cited in the record references the interrelationship between radio stations and record companies in the analog world. As noted above, the Panel considered the evidence but did not find it persuasive.

Where the Panel makes a decision based upon its weighing of the evidence, the Register will not disturb its findings and conclusions that proceed directly from the Panel's consideration of the factual evidence. Thus, the Register accepts the Panel's conclusion that performances of sound recordings over the Internet did not significantly stimulate record sales. More importantly, though, the Panel correctly found that promotional value is a factor to be considered in determining rates under the willing buyer/willing seller model, and does not constitute an additional standard or policy consideration to be used after rates are set to adjust a base rate upwards or downwards. Report at 21. Therefore, the effect of any promotional value attributable to a radio retransmission would already be reflected in the rates for these transmissions reached through arms-length negotiations in the marketplace.

As for the second factor, the Panel found that both copyright owners and licensees made significant creative, technological and financial contributions. It concluded, however, that it was not necessary to gauge with specificity the value of these contributions in the case where actual agreements voluntarily negotiated in the marketplace existed, since such considerations, including any significant promotional value of the transmissions, would already have been factored into the agreed upon price. Id. at 35-36. This is not a contested finding.

It is also important at the outset of this review to distinguish the willing buyer/willing seller standard to be used in this proceeding from the standard that applies when setting rates for subscription services that operated under the section 114 license. They are not the same. Section 114(f)(1)(B), governing subscription services, requires a CARP to consider the objectives set forth in section 801(b)(1), as well as rates and terms for comparable types of digital audio transmission services established through voluntary negotiations. See Final Rule and Order, 63 FR 25394, 25399 (May 8, 1998). This standard for setting rates for the subscription services is policy-driven, whereas the standard for setting rates for nonsubscription services set forth in section 114(f)(2)(B) is strictly fair market value—willing buyer/willing seller. Thus, any argument that the two rates should be equal as a matter of law is without merit. See, e.g., Webcasters Petition at 4 (comparing rates set for preexisting subscription services under the policy driven standard with the proposed marketplace rates for nonsubscription services and inferring that the rates should be similar).

2. Hypothetical Marketplace/Actual Marketplace

To set rates based on a willing buyer/willing seller standard, the CARP first had to define the relevant marketplace in which such rates would be set. It determined, and the parties agreed, that the rates should be those that a willing buyer and willing seller would have agreed upon in a hypothetical marketplace that was not constrained by a compulsory license. The CARP then had to define the parameters of the marketplace: the buyers, the sellers, and the product.

In this configuration of the marketplace, the willing buyers are the services which may operate under the webcasting license (DMCA-compliant services), the willing sellers are record companies, and the product consists of a blanket license from each record company which allows use of that company's complete repertoire of sound recordings. Report at 24. Because of the diversity among the buyers and the sellers, the CARP noted that one would expect “a range of negotiated rates,” and so interpreted the statutory standard as “the rates to which, absent special circumstances, most willing buyers and willing sellers would agree” in a competitive marketplace. Id. at 25.

The Services take issue with the Panel's analysis of the hypothetical marketplace. They argue that the willing sellers should be considered as a group of hypothetical “competing collectives each offering access to the range of sound recordings required by the Services,” and not, as the Panel contends, viewed as individual record companies. Broadcasters Petition at 9; Webcasters Petition at 9-10. It is hard to see, however, how competition would be stimulated in a marketplace where every seller offers the exact same product and where more likely than not, the sellers would act in concert to extract monopolistic prices. Possibly sellers would choose to undercut each other, but at some point the price would stabilize. In any event, the Services failed to explain how such collectives would operate in a competitive marketplace. Consequently, the Register rejects the Webcasters' challenge to the Panel's definition on this point and adopts the Panel's characterization of the relevant marketplace, recognizing that for purposes of this proceeding, the major record companies are represented by a single entity, the RIAA.

Turning next to the actual marketplace in which RIAA negotiated agreements with individual services, the Services voice a number of objections to the Panel's decision to rely on the 26 voluntary agreements offered into evidence by RIAA. Specifically, the Services object to the use of the voluntary agreements because they fail to exhibit a range of negotiated rates among diverse buyers and sellers. Broadcasters Petition at 10; Webcasters Petition at 10. They also question the validity of relying on agreements negotiated during the early stages of a newly emerging industry, noting the Panel's admonition to approach such agreements with caution. Report at 47. The reason for the warning was Dr. Jaffe's stated concern that such licenses “may not reflect fully educated assessments of the nascent businesses” long-term prospects.”

The Services also argue that the existence of the antitrust exemption in the statutory license gave RIAA an unfair bargaining advantage over the Services because RIAA represented the five major record companies who together owned most of the works. They contend that RIAA used its superior market power to negotiate supra-competitive prices with Services who could not match either RIAA's power in the marketplace or its sophistication in negotiating contracts. Moreover, they utterly reject the Panel's determination that RIAA's perceived market power was tempered by the existence of the statutory license, which, for purposes of negotiating a fair rate for use of sound recordings, leveled the playing field. Webcasters Petition at 12.

Not surprisingly, RIAA agrees with the Panel on this issue. It maintains that the statutory license offers the Services two clear advantages which more than offset any perceived advantage the RIAA may have had in negotiating a voluntary agreement. First, the license eliminates the usual transaction costs associated with negotiating separate licenses with each of the copyright owners. Second, services may avoid litigation costs associated with setting the rates for a statutory license provided they choose not to participate in the CARP process. RIAA reply at 12.

In essence, both sides articulate valid positions which are supported by the record. RIAA is clearly an established market force with extensive resources and sophistication. In fact, the Panel found that when RIAA negotiated with less sophisticated buyers who could not wait for the outcome of this proceeding, the rates were above-market value, and therefore, not considered by this CARP. Report at 54-56. Nevertheless, it would make no sense for RIAA to take any other position in a marketplace negotiation. Sellers expect to make a profit and will extract from the market what they can, just as buyers will do everything in their power to get the product at the lowest possible price. These are the fundamental principles guiding marketplace negotiations.

Such negotiations, however, were few. For the most part, webcasters chose not to enter into negotiations for voluntary agreements, knowing that they could continue to operate and wait for the CARP to establish a rate. Such actions on the part of the users clearly impeded serious negotiations in the marketplace and support the CARP's observation that the statutory license had a countervailing effect on the negotiation process and limited the ability of RIAA to exert undue marketplace power. See Tr. 9075-77, 9490-94 (Marks) (explaining the difficulties of bringing webcasters to the negotiating table due to the statutory license). Thus, the CARP could only consider negotiated rates for the rights covered by the statutory license that were contained in an agreement between RIAA and a Service with comparable resources and market power.

The only agreement that met these criteria was the Yahoo! agreement. The Panel found that both parties to that agreement entered into negotiations in good faith and on equal footing. Moreover, RIAA's negotiating advantage disappeared. RIAA could not extract super-competitive rates because Yahoo! brought comparable resources, sophistication, and market power to the negotiating table.

Moreover, Yahoo! could have continued to operate under the license and wait for the outcome of this proceeding. Yet, Yahoo!, unlike most of the other Services, did not take this course of action. It wanted a negotiated agreement so that it could fully develop its business model based on certainty as to the costs of the use of the sound recordings. Consequently, it had every incentive to negotiate a rate that reflected its perception of the value of the digital performance right in light of its needs and position in the marketplace. Had RIAA insisted upon a super competitive rate, Yahoo! could have walked away and waited for the CARP to set the rates. RIAA Reply at 13. Thus, it was not arbitrary for the Panel to consider the negotiated agreement between Yahoo! and RIAA. It met all the criteria identified by the CARP (discussed above) that characterized the hypothetical marketplace: Yahoo! was a DMCA-compliant Service; RIAA represented the interests of five independent record companies, and the license granted the same rights as those offered under the webcasting and the ephemeral recording licenses.

The Webcasters make one final argument concerning use of licenses negotiated in the marketplace. They fault the Panel for its reliance on a contract for which there was no prior marketplace precedent for setting a rate. Webcasters Petition at 15. Yet, that alone cannot be a reason to reject consideration of agreements negotiated in the marketplace, albeit at an early stage in the development of the industry. At some point, rates must be set. Such rates then become the baseline for future market negotiations. RIAA recognized an opportunity to participate in this initial phase and moved forward to negotiate contracts with users with the intention of using these contracts to indicate what a willing buyer would pay in the marketplace. However, that was easier said than done. As discussed above, most Webcasters chose not to enter into marketplace agreements, preferring to wait for the outcome of the CARP proceeding in the hope of getting a low rate. Clearly, such resistance to enter into good faith negotiations made it difficult for the copyright owners to gauge the market accurately and find out just what a willing buyer would be willing to pay for the right to transmit a sound recording over the Internet.

3. Benchmarks for Setting Market Rates: Voluntary Agreements vs. Musical Works Fees

The parties offer two very different methods for setting the webcasting rates. RIAA argued that the best evidence of the value of the digital performance right is the actual rates individual services agreed to pay for the right to transmit sound recordings over the Internet. In support of its position, it offered into evidence 26 separate agreements it had negotiated in the marketplace prior to the initiation of the CARP proceeding. The Services take a different approach. They dispute the validity of the contracts as a bases for marketplace rates and offer in their place a theoretical model (the “Jaffe model”) predicated on the fees commercial broadcasters pay to use musical works in their over-the-air AM/FM broadcast programs.

The Jaffe model builds on the premise that in the hypothetical marketplace, copyright owners would license their digital performance rights and ephemeral recording rights at a rate no higher than the rates music publishers currently charge over-the-air radio broadcasters for the right to publicly perform their musical works. Report at 28, citing Webcasters PFFCL ¶¶ 276-78; Jaffe W.D.T. 16-19. To find the rate copyright owners would charge under this model, Webcasters calculated a per performance and a per hour rate by using the aggregate fees that 872 over-the-air radio stations paid in 2000 to the performing rights organizations BMI, ASCAP, and SESAC. It combined the fee data with data on listening audiences obtained from Arbitron to generate an average fee paid by an over-the-air broadcaster per “listening hour.” From this value, Webcasters calculated a per performance fee by dividing the “listener hour” fee by the average number of songs played per hour by music-intensive format stations. Id. These calculations yielded a per song fee of 0.02¢ or, in the alternative, a per listener hour fee of 0.22¢. For purposes of webcasting, these values were adjusted upward to reflect the fact that, on average, webcasters play 15 songs per hour, as compared to the 11 per-hour played on over-the-air radio. The webcaster per hour rate works out to be 0.3 instead of 0.2¢ per hour.

After carefully considering both approaches, the Panel chose to focus on the RIAA agreements. In rejecting Dr. Jaffe's theoretical model, the panel cited three reasons for its conclusion. First, the Panel expressed strong concern regarding the construct of the model, including: 1. The difficulty in identifying all the factors that must be considered in setting a price, and 2. The inherent error associated with predicating a prediction on a “string of assumptions,” especially where the level of confidence in many of the assumptions is not high. Second, the Panel was wary of analogizing the market for the performance of musical works with the market for the performance of sound recordings, finding instead that the two marketplaces are distinct based upon the difference in cost and demand characteristics. And finally, the Panel determined that the Jaffe model was basically unreliable. It could not be used to predict accurately the amount of royalty fees owed to the performing rights societies by a particular radio station. It came to this conclusion after using the model to predict the royalty fees owed by a particular station and comparing that figure to the amount the radio station actually paid. For some radio stations, the model severely underestimated the amount owed to the performing rights societies, thus, drawing into serious question the reliability of the model. Report at 42.

a. Fees paid for use of musical works. The Broadcasters and the Webcasters fault the Panel for disregarding the fees paid for musical works as a viable benchmark. Webcasters Petition at 15, 47. They maintain that Dr. Jaffe's analysis proves that the value of the performance of the sound recording is no higher than the value of the performance of the musical work. Webcasters argue that the fees for musical works constitute a valid benchmark because these rates are the result of transactions between willing buyers and willing sellers over a long period of time, in a marketplace that shares economic characteristics with the marketplace for sound recordings. Webcasters Petition at 48. The Broadcasters agree. They maintain that even under the willing buyer/willing seller standard, “the over-the-air musical works license experience * * * has resulted in fees ‘to which most willing buyers and willing sellers [have] agree[d]’ and constitute ‘comparable agreements negotiated over a longer period, which ha[ve] withstood ‘the test of time.’ ” Broadcasters Petition at 45-46, citing Report at 25, 47.

Broadcasters and Webcasters also object to the Panel's characterization of its proposed benchmark as merely a theoretical model. Webcasters Petition at 51. They maintain that Dr. Jaffe's model was much more than a theoretical model because it used actual data from the musical works marketplace to calculate an analogous rate for use of sound recordings in the digital marketplace. Consequently, these Services contend that the Panel gave inadequate consideration to their proposed benchmark and rejected the model out of hand because it was purported to be only a theoretical model based upon a number of untested assumptions. Broadcasters Petition at 18-19; Webcasters Petition at 18-20, 52.

Finally, the Services argue that the statute does not compel the Panel to consider only negotiated agreements. They also contend, that the reliance on the fees paid for use of the musical works in a prior CARP proceeding to establish rates for subscription services operating under the same license required the panel to give more consideration to the musical works benchmark. Broadcaster's Petition at 1-2; Webcasters Petition at 1-2, 15, 17, 47. Webcasters find support for this last argument in an Order of the Copyright Office issued in this proceeding, dated July 18, 2001.

In that order, the Office acknowledged that in 1998 it had adopted the rates paid for musical works fees as a relevant benchmark for setting rates for subscription services. It stated, however, that the evidence in that case did not support a conclusion that the value of the sound recording exceeded the value of the musical work. Moreover, and directly to the point, the Register's recommendation in the earlier proceeding concurred with the earlier Panel's determination that the musical works benchmark is NOT determinative of the marketplace value of the performance right in sound recordings. The relevant passage states: “The question, however, is whether this reference point (the musical works benchmark) is determinative of the marketplace value of the performance in sound recordings; and, as the Panel determined, the answer is no.” 63 FR 25394, 25404 (May 8, 1998).

The July 18 Order went on to note that in the subscription service proceeding, “[h]ad there been record evidence to support the opposite conclusion, [namely, that the value of sound recordings exceeds the value of musical works], the outcome might have been different.” This statement was an invitation to the parties to provide whatever evidence they could adduce in this proceeding to establish the value of the sound recording. It was not to be read as an absolute determination, that the value of the sound recording in a marketplace unconstrained by a compulsory license is less than the value of the underlying musical work. Instead, the Order stated that “the musical work fees benchmark identified in a previous rate adjustment proceeding as the upper limit on the value of the performance of a sound recording may or may not be adopted as the outer boundary of the “zone of reasonableness” in this proceeding. This is a factual determination to be made by the CARP based upon its analysis of the record evidence in this proceeding.”

It is also important to note that in the prior proceeding, the only reason the Register and the Librarian focused on the musical works benchmark was because it was the only evidence that remained probative after an analysis of the Panel's decision. Each of the other benchmarks possessed at least one fatal deficiency and, consequently, each was rejected as a reliable indicator of the value of the performance of a sound recording by a subscription service. Of equal importance is the fact that the musical works benchmark had never been fully developed in the record, nor had any party relied on it to any great extent in making its case to that Panel. Consequently, it was not arbitrary for the Panel to reject the Services' invitation to anchor its decision for setting rates for nonsubscription services on the prior decision setting rates for preexisting subscription services.

Moreover, the Panel is not required to justify why the rates it ultimately recommended here are greater than the rates preexisting subscription services pay for use of the musical works. That is merely the result of the analysis of the written record before this Panel, and its decision flows naturally from its reliance upon contractual agreements negotiated in the relevant marketplace for the right at issue. This difference in the rates is also attributable to the different standards that govern each rate setting proceeding. As discussed previously in section IV.1, the standard for setting rates for subscription services is policy based and not dependent upon market rates. Consequently, it is more likely that the rates set under the different standards will vary markedly, especially when rates are being set for a new right in a nascent industry.

Nevertheless, the Register agrees with the Services on a number of theoretical points. Certainly, the Panel could have utilized Dr. Jaffe's model in making its decision, either alone or in conjunction with the voluntary agreements, provided that it considered the model's deficiencies, and made appropriate adjustments for the fact that the model required reliance on a string of assumptions to perform the conversion of a rate for the public performance of a musical work in an analog environment, into a comparable rate for the public performance of a sound recording in a digital format. See AMOA v. CRT, 676 F2d 1144 (7th Cir. 1982). But the fact remains that it was not required by law to do so. The Panel was free to choose any of the benchmarks offered into the record or to rely on each of them to the degree they aided the Panel in reaching its decision. See, e.g., Use of Certain Copyrighted Works in Connection with Noncommercial Broadcasting, 43 FR 25068-69 (CRT found voluntary license between BMI, Inc., and the public broadcasters, Public Broadcasting System and National Public Radio, of no assistance in setting rates for use of ASCAP repertoire).

The Register also rejects the Services' contentions that the Panel failed to consider fully Dr. Jaffe's model. See Webcasters Petition at 20, 52. The Panel did consider Jaffe's model and concluded that it need not consider alternative benchmarks that are at best analogous when it had actual evidence of marketplace value of the performance of the sound recordings in the record. Report at 42. It also rejected the offer to utilize the model because the underlying assumptions were in many instances questionable. For example, the Panel did not accept the assumptions that a percentage of revenue model could be converted accurately to a per performance metric, or that the buyers and sellers in the two marketplaces are analogous.

Broadcasters assert that they had established that the value of the musical work is higher than the comparable right for sound recording based on the fees paid for use of these works in movies and television programs. Broadcasters Petition at 24. In addition, they offered a study of the fees paid for these rights in twelve foreign countries where the Services claim these rights are valued more or less equally. Id. at 24, 49. Because the Panel failed to analyze this information, the Services argue, the Panel's rejection of the musical benchmark was arbitrary.

RIAA responds that the information offered on the fees paid for the public performance of sound recordings fails to establish that in these countries sound recordings are valued according to a “willing buyer/willing seller” standard. RIAA Reply at 20, fn 36. In fact, many of the countries surveyed evidently use an “equitable remuneration” standard, which courts have held not to be equivalent to a fair market value. Because it is not possible to ascertain whether any of the rates offered in the survey of foreign countries represented a fair market rate, or that the rights in these countries are equivalent to the rights under U.S. law, the Panel was not arbitrary in its decision to disregard this evidence. The Register also concludes that the Panel's decision not to consider master use and synchronization licenses for use of musical works and sound recordings in motion pictures and television was not arbitrary. At best, these licenses offered potential benchmarks for evaluating the digital performance right for sound recordings, and they may well have been useful had not actual evidence of marketplace value of the sound recordings existed. In any event, they did not represent better evidence than the voluntary agreements negotiated in the marketplace for the sound recording digital performance right.

b. Voluntary agreements. On the other hand, the Panel articulated two affirmative reasons for its focus on the negotiated agreements. First, the statute invites the CARP to consider rates and terms negotiated in the marketplace. Second, the Panel accepted the premise that the existence of actual marketplace agreements pertaining to the same rights for comparable services offers the best evidence of the going rate. Report at 43, citing Jaffe Tr. at 6618.

But in choosing this approach, the Panel did not accept the 26 voluntary agreements at face value. It evaluated the relative bargaining power of the buyers and sellers, scrutinized the negotiating strategy of the parties, considered the timing of the agreements, discounted any agreement that was not implemented, eliminated those where the Service paid little or no royalties or the Service went out of business, and evaluated the effect of a Service's immediate need for the license on the negotiated rate. See Report at 45-59. Ultimately, it gave little weight to 25 of the 26 agreements for these reasons and because the record demonstrated that the rates in these licenses reflect above-marketplace rates due to the superior bargaining position of RIAA or the licensee's immediate need for a license due to unique circumstances. At best, the Panel concluded that the rates included in these agreements establish an upper limit on the price of the digital performance right, and where included, the right to make ephemeral copies. Report at 59.

RIAA objects to the Panel's decision to reject 25 of the 26 agreements on the grounds that the Panel's criticisms were overbroad. RIAA Petition at 34. Specifically, it claims that the Panel mischaracterized its agreement with www.com/OnAir (“OnAir”), arguing that this Licensee paid substantial royalties and its decision to enter into the agreement was not motivated by special circumstances as the CARP claimed. Id. at 31. This observation, however, is not sufficient to overcome the Panel's conclusion in regard to this agreement, especially in light of the testimony of RIAA's own expert witness, Dr. Nagle, who testified the Panel should give no consideration to any agreement with a licensee who cannot survive in the marketplace. Report at 24. Had OnAir continued to operate in the marketplace and renew its license with RIAA, the Panel might have given it more serious consideration. But again, it was not required to do so, especially when the Panel found more probative evidence in the record upon which to rely.

Likewise, RIAA objected to the Panel's decision not to give any weight to the MusicMusicMusic (“MMM”) agreement, arguing in this case that the Panel assumed MMM had renewed its agreement in 2001 for the same reasons that led it to accept a higher than market value rate in 1999. RIAA Petition at 32. Webcasters respond that RIAA misrepresents the facts of the renewal. They maintain that MMM renewed the agreement in 2001 based on “many of the same motivating factors” that led to the initial agreement, including its concerns about its long-term relationship with RIAA in other areas. Webcasters Reply at 29. Because the evidence supports a rationale for MMM to accept a higher than marketplace rate, it was not arbitrary for the Panel to decide not to adopt it as an adequate benchmark. The Panel need not rely on the MMM agreement when it had another agreement negotiated in the marketplace that did not suffer from the same perceived shortcomings.

Specifically, the Panel gave significant weight to the one remaining agreement negotiated—the RIAA-Yahoo! agreement—and used it as a starting point for setting the rates for the webcasting license and the ephemeral recordings license. The Panel found this agreement particularly reliable and probative because: (1) Yahoo! was a successful and sophisticated business which, to date, had made well over half of all DMCA-compliant performances; (2) it had comparable resources and bargaining power to those RIAA brought to the table; and (3) the agreement provided for different rates for different types of transmissions. See Report at 64-67; 70. While the first two reasons offer strong support for the Panel's decision to rely upon the Yahoo! agreement, the third reason is questionable in the context of the Yahoo! agreement because the different rates do not actually represent the parties' understanding of the value of the performance right for these types of transmissions. See discussion infra, section IV.5.

Webcasters, however, argue that the Panel's reliance on the Yahoo! agreement was fatal because it selected a single term out of a multifaceted contract. Webcasters at 22-23. Specifically, they maintain that the webcasting rate did not reflect merely the value of the sound recording, but an abundance of trade-offs that met the needs of RIAA and Yahoo!. Id. at 24. Webcasters make this argument because, in a prior CARP proceeding, the Register had refused to adopt a complicated partnership agreement that purportedly included a rate for the digital performance right as a benchmark for setting the statutory rate. See, Rate Setting Proceeding for Subscription Services, 63 FR 25394 (May 8, 1998). Specifically, the Register concluded that “it was arbitrary for the Panel to rely on a single provision extracted from a complex agreement where the evidence demonstrates that the [rate] provision would not exist but for the entire agreement.” Id. at 25402.

The two agreements, however, are not analogous. The primary purpose of the Yahoo! agreement was to set a rate for use of sound recordings over the Internet. Thus, the noted trade-offs in this agreement were all directly tied to considerations relating to the value of the performance right, and did not affect its validity as a benchmark. Such was not the case with the subscription services agreement offered into evidence in the prior proceeding, where the performance right component was merely “one of eleven interdependent co-equal agreements which together constituted the partnership agreement between [Digital Cable Radio Associates (“DCR”)] and the record companies.” Id.

Along these same lines, the Services challenge the Panel's dependence upon a single contract negotiated between a single seller (RIAA) and a single buyer (Yahoo!), especially in light of the Panel's construct of the hypothetical marketplace. Broadcasters Petition at 14; Live365 Petition at 5; Webcasters Petition at 9, 14. These parties argue that under 17 U.S.C. 114(f)(2)(B), the Panel had discretion to consider negotiated agreements only when the agreements were for comparable types of services in comparable circumstances. Webcasters, including Live365, maintain that Yahoo! had a unique position among webcasters and argue that it was manifestly arbitrary for the Panel to set rates based solely on the rates paid by this one webcaster which by its own admissions was not similarly situated with other webcasters. Live365 Petition at 11; Webcasters Petition at 27. Specifically, they contend that Yahoo! had little concern about getting a reasonable rate for Internet-only transmissions so long as the rate for RR transmissions was favorable and it could continue to grow in this arena. Webcasters note that Yahoo!'s main business was the retransmission of radio re-broadcasts, and that over 90% of all transmissions made by Yahoo! fall within this category. Id. at 28. Consequently, Webcasters maintain that the rates set for Internet-only transmissions in the Yahoo! agreement cannot be fairly applicable to Webcasters at large. Id. at 29.

Broadcasters have other complaints with the Panel's approach. First, they object to the use of the Yahoo! contract to set rates for broadcasters when the buyer in that case was not a broadcaster but a third-party aggregator—a completely different type of business. Second, they fault the Panel for its failure to follow its own dictate to proceed cautiously when viewing contracts negotiated in a nascent industry for newly created rights. Broadcaster Petition at 14. Similarly, Webcasters fault the Panel for relying exclusively on the Yahoo! agreement because it offers only a single, uniform rate for each type of transmission, in contrast to the “range of rates,” involving “diverse buyers and sellers,” that the Panel identified as the hallmark of a willing buyer/willing seller marketplace.” Webcasters Petition at 14. Webcasters also contend that the Yahoo! agreement should not have been considered because it, like the Lomasoft-RIAA agreement, had not been renewed. Webcasters Petition at 41.

Moreover, Live365 questions the Panel's reliance on the Yahoo! contract when it had rejected use of a second similar agreement between MusicMatch (“MM”) and RIAA because MM had accepted higher than marketplace rates for nearly identical reasons to those that account for the inflation in the Yahoo! rates. MM had wished to settle litigation with RIAA and it received a benefit from the inclusion of a Most Favored Nations (MFN) clause in the contract. Yet, in spite of the similarities, the Panel relied on the Yahoo! agreement and disregarded the second one. Such disparate treatment of similarly situated services is arguably arbitrary. Live365 Petition at 13. A closer examination of the agreements, however, reveals a significant difference between the two contracts which allowed the Panel to disregard the MM agreement for further consideration. Most importantly, the MM agreement contained a MFN clause that [material redacted subject to a protective order]. The Panel reasoned that this provision undermined the usefulness of the agreement to establish a marketplace rate because [material redacted subject to a protective order]. Report at 56-57. Such was not the case with the Yahoo! agreement since the MFN clause only allowed Yahoo! to receive a partial benefit commensurate with [material redacted subject to a protective order]. Report at 62.

The Register concurs and agrees with the Panel's observation that it would be unsound to establish a rate for the statutory license using a rate that itself is subject to change based on the outcome of this proceeding.

The Register also finds the other arguments by the parties unavailing. In spite of their objections, the Services' own expert, Dr. Jaffe, agreed in principle with the Panel's approach. In his testimony, he acknowledged that voluntary agreements between a willing buyer and a willing seller would constitute the best evidence of reasonable marketplace value if such agreements were between parties comparable to those using the webcasting license. Tr. 6618 (Jaffe). The Services' argument, of course, is that the Yahoo! agreement is not a comparable agreement for purposes of setting rates for all webcasters, and this appears to be a valid point. Yahoo!'s business model is somewhat unique. Unlike webcasters that create their own programming, Yahoo! merely offers programming by AM/FM radio stations and other webcasters.

Nevertheless, RIAA offers record evidence that contradicts the Webcasters' assertion that Yahoo! is not a comparable service for purposes of this proceeding, noting that many webcasters affirmatively stated that Yahoo! is a competitor. Moreover, RIAA asserts that the number of the performances made by Yahoo! on its Internet-only channels is roughly equivalent to the number of performances made by the other webcasters in this proceeding and, therefore, Yahoo!'s interest in getting a reasonable rate for its Internet-only stations should be comparable to those of the Webcasters in this proceeding. RIAA reply at 33-34.

Because Yahoo! is engaged in both types of transmissions, it is reasonable to accept this agreement as a basis for setting rates for both types of transmissions. Yahoo! has developed a significant business presence in the marketplace for Internet-only transmissions and understands the marketing and business of Internet-only webcasters. Consequently, allegations that Yahoo! has only a de minimis interest in the webcasting field and is thus less interested in getting a reasonable rate for the right to make digital transmissions are without merit. The question, however, is whether each rate in the Yahoo! agreement reflects the actual value of the particular transmission or whether one must consider both rates in concert to understand the valuation process. For a more detailed discussion on this point, see section IV.5 infra.

4. Alternative Methodology: Percentage-of-Revenue

The Panel also carefully considered and rejected a percentage-of-revenue model for assessing fees and determined that a per performance metric was preferable to a percentage-of-revenue model. A key reason for rejecting the percentage-of-revenue approach was the Panel's determination that a per performance fee is directly tied to the right being licensed. The Panel also found that it was difficult to establish the proper percentage because business models varied widely in the industry, such that some services made extensive music offerings while others made minimal use of the sound recordings. Report at 37. The final reason and perhaps the most critical one for rejecting this model was the fact that many webcasters generate little revenue under their current business models. As the Panel noted, copyright owners should not be “forced to allow extensive use of their property with little or no compensation.” Id, citing H.R. Rep. 105-796, at 85-86. Thus, it seemed illogical to set a rate for the statutory license on a percentage-of-revenue basis when in fact a large proportion of the services admit they generate very little revenue, and, therefore, would generate meager royalties even for substantial uses of copyrighted works. Moreover, it is highly unlikely that a willing seller, who negotiates an agreement in the marketplace, would agree to a payment model which itself could not provide adequate compensation for the use of its sound recordings.

Nevertheless, Webcasters and Live365 assert that the Panel acted arbitrarily when it failed to provide a revenue-based royalty option. Webcasters at 54. They maintain that both sides advocated adoption of a percentage-of-revenue option, see RIAA PFFCL, Appendix C; Webcasters PFFCL ¶¶ 283-296, and that it was arbitrary for the Panel to refuse to adopt this approach. See Live365 Petition at 10; see also pg. 11, fn 6. Webcasters also assert that they had made clear that in the event the Panel rejected Jaffe's model, a revenue-based alternative license proposal would be necessary to avoid putting certain webcasters out of business. Webcasters Petition at 56, 60. Moreover, Webcasters reject the Panel's conclusion that the Services' revenue-based fee proposal was untimely. Id.. at 57-60. They maintain that under § 251.43(d) they were allowed to revise their claim or their requested rate “at any time during the proceeding up to the filing of the proposed findings of fact and conclusions of law,” and that the Panel had no authority to alter this provision by order under § 251.50.

In reply, RIAA notes that the Webcasters cite no evidence for their assertion that they reasonably believed the Panel would offer a percentage-of-revenue option and counters their timeliness argument by setting forth the timeline regarding the parties's submissions concerning the rates. RIAA Reply at 62. Evidently at the request of the Webcasters, the Panel issued an order setting November 2 as the deadline for submitting revised or new rate proposals, so that parties were fully aware of each other's position and could style their findings of fact and conclusions of law accordingly. Consequently, the Panel found that the Services' later submission including a proposed rate based on percentage-of-revenue in their PFFCL was untimely. Report at 31, citing Order of November 3, 2001.

After considering the arguments now advanced by the Services concerning the Panel's authority to require final submissions on rates prior to the filing of the PFFCLs, the Register finds that the Panel acted in a lawful manner and within its authority. As RIAA points out in its reply, the Panel has authority pursuant to 37 CFR 251.42 to waive or suspend any procedural rule in this proceeding, including the time by which parties must make final submissions regarding proposed rates. What the Panel cannot do is engage in a rulemaking proceeding to amend, supplement, or supersede any of the rules and regulations governing the CARP procedures. See 37 CFR 251.7. Moreover, the language in § 251.43 is somewhat ambiguous as to when a party can make its final rate proposal, lending itself to two interpretations. For this reason alone, it was prudent for the Panel to issue an order clarifying the application of the rule for purposes of this proceeding. In fact, Webcasters had asked for this ruling and cannot be heard at the end of the process to argue against a ruling that they sought and to which they never objected. Consequently, the Panel was not arbitrary when it found the Webcasters' request for a percentage-of-revenue fee structure untimely.

Moreover, the Panel was not arbitrary for failing to adopt a percentage-of-revenues model merely because some parties voiced an expectation that the Panel would offer such a model as an alternative means of payment. This complaint of unmet expectations is not a substantive argument for finding the Panel's decision arbitrary and, consequently, it will not be considered further.

On the other hand, Live365 does make a substantive argument concerning the Panel's decision not to adopt a percentage-of-revenue model. It notes that the current marketplace uses two types of rate structures, a revenue based model and a performance rate structure, and that the revenue based model is better for start-up and smaller webcasters. Live365 Petition at 8. In fact, Live365 points out that many of the agreements that RIAA negotiated with webcasters incorporated this model. Moreover, Live365 maintains that it was arbitrary for the Panel to propose rates that “had the effect of rendering sound recordings substantially more valuable than musical works, even though the CARP acknowledged that it was rendering no opinion on this issue.” Live365 Petition at 5, 14-15. In its opinion, this result was arbitrary based upon Yahoo!'s stated perception that the value of the performance right for the musical work is comparable to the value of the performance right for the sound recording. Finally, Live365 argues that rates based upon mere perception, as those negotiated in the Yahoo! agreement, are by their very nature arbitrary and should be disregarded. Id. at 15.

RIAA refutes the Services' claim that the Panel was arbitrary because it failed to offer a percentage-of-revenue model. It argues that the record supports the Panel's conclusion that a percentage-of-revenue model would have been difficult to implement because Services use sound recordings to different degrees—a position taken by the Webcasters' own witness. Specifically, Jaffe questioned the appropriateness of using a percentage-of-revenue model where those percentages were based on the economics driving over-the-air broadcasts. RIAA Reply Petition at 52, citing Tr. 6487, 6488, 12582 (Jaffe). Jaffe also acknowledged that it was difficult to assess what the revenue base should be for such a model given the variation of the business models utilized by the webcasters. RIAA also notes that section 114(f)(2)(B) requires the Panel to consider the quantity and nature of the use of the sound recording and argues that a per performance metric automatically accounts for the amount of use by the various services. RIAA Reply at 59.

RIAA also argues that a basic percentage-of-revenue fee structure would frustrate the purpose of the law because it would deny copyright owners fair compensation for use of their works in those situations where a service generates little or no revenue. Certainly, the record contains evidence that a number of webcasters do not expect or intend to earn revenues from their webcasts, see Report at 37; see, e.g., Live365 Petition at 7, maintaining that their use is designed primarily to maintain their over-the-air audience. Because certain Services take this approach, when RIAA did consider using a percentage-of-revenue model, it included a substantial minimum fee proposal in conjunction with the percentage of fee proposal to address the problems associated with low revenue generating businesses. Specifically, the RIAA proposal required that a Service pay either 15% of revenues or $5,000 per $100,000 of a webcasters' operating costs, whichever is greater. RIAA Reply at 61. In this way, RIAA sought to avoid the anomaly of allowing a business unfettered use of the sound recordings without reasonable compensation to the copyright owners. Id. at 54, 61. This formulation, however, would not have given the webcasters the relief they seek through the adoption of a rate based on a percentage-of-revenues. In fact, under RIAA's percentage-of-revenue formulation, many webcasters, including Live365, would have paid more than they will under the Panel's per performance rate structure.

The Register finds that the Panel's decision not to set a percentage-of-revenue fee option was not arbitrary in light of the record evidence. First, it is clear that the Services' primary position was to seek adoption of a fee based upon performances and not a percentage-of-revenue. Indeed, Dr. Jaffe's model proposed a fee model based on listener hours or number of listener songs, and not a rate based upon percentage-of-revenues, because a royalty based upon actual performances would be directly tied to the nature of the right being licensed. Report at 37; Jaffe W.R.T. at 31. Moreover, because they took this position, Services argued for a low minimum rate that would only cover administrative costs and not the value of the performances themselves—an approach the CARP adopted in its Report.

Moreover, the statute does not require the CARP to offer alternative fee structures, and the Services should not have expected the Panel to do so, especially when the Webcasters never advanced a percentage-of-revenues option in their own case. In fact, there is no precedent in the statutory licensing scheme anywhere in the Copyright Act that would support alternative rates for the same right. Clearly, it cannot be arbitrary for the Panel to choose not to deviate from the longstanding practice of establishing only one rate schedule for a license.

5. The Yahoo! Rates—Evidence of a Unitary Marketplace Value

The starting point for setting the rates for the webcasting license is the Yahoo! agreement. In that agreement, rates were set for two different time periods. For the initial time period covering the first 1.5 billion performances, Yahoo! agreed to pay one lump sum of $1.25 million. From this information, the Panel calculated a “blended,” per performance rate of 0.083¢. This value represents the actual price that Yahoo! paid for each of the first 1.5 billion transmissions without regard to which type of service made the transmission. For the second time period, Yahoo! and RIAA agreed to a differential rate structure. One rate was set for performances in radio retransmissions (RR) (0.05¢ per performance) and another rate was set for transmissions in Internet-only (IO) programming (0.2¢ per performance). These rates were first used in early 2000 and do not apply to the first 1.5 billion performances.

However, the CARP did not accept these differentiated rates at face value. The Panel engaged in a far-ranging inquiry to determine how the parties established the negotiated rates. What it found was that Yahoo! agreed to a higher rate for the IO transmissions in exchange for a lower rate for the RR because this arrangement addressed specific concerns of both parties. In particular, RIAA wished to establish a marketplace precedent for IO transmissions in line with rates it had negotiated in earlier agreements, while Yahoo! sought to negotiate rates which, in the aggregate, yielded a rate it could accept. Consequently, the Panel found the rate for the IO transmissions to be artificially high and, conversely, the rates for the RR to be artificially low. For this reason, it made a downward adjustment to the IO rates and an upward adjustment to the RR rates.

Before making this adjustment, though, the Panel had to consider whether it was reasonable to establish separate rates for the two categories of transmissions. In reaching its decision, the Panel considered two facts, the fact that the Yahoo! agreement provided for two separate rates, and the fact that all parties agreed that performances of sound recordings in over-the-air radio broadcasts promote the sale of records. Report at 74. Based on this finding, the Panel concluded that a willing buyer and a willing seller would agree that the value of the performance right for RR would be considerably lower than for IO transmissions. Moreover, it attributed the existence of the rate differential in the Yahoo! agreement to the promotional value enjoyed by the copyright owners from the performance of the sound recordings by broadcasters in their over-the-air programs, and not to promotional value attributable to transmissions made over the Internet. Report at 74-75. Specifically, the Panel found that, “to the extent that Internet simulcasting of over-the-air broadcasts reaches the same local audience with the same songs and the same DJ support, there is no record basis to conclude that the promotional effect is any less.” Report at 75.

This finding, however, did not prompt the Panel to make any further adjustment for promotional value, finding instead that the differential rates in the Yahoo! agreement already reflect “marketplace assessment of the various promotion and substitution effects, along with a myriad of other factors.” Report at 87. Primary among these factors were the Most Favored Nations (MFN) clause and the cost savings to Yahoo! in avoiding CARP litigation. The Panel reasoned that Yahoo! was willing to accept somewhat inflated royalty rates in exchange for the costs it saved by not participating in the CARP proceeding, and for the MFN clause which had some indeterminate value for Yahoo!.

RIAA disagrees with the Panel's analysis and these findings. As an initial matter, it maintains that there was no record evidence to support a separate rate for commercial broadcasters. RIAA Broadcaster PFOF 24-52. Second, it argues that the Panel adopted a two-tier rate structure for RR and IO transmissions based on the different rates in the Yahoo! agreement, and its mistaken view of the significance of an exemption in the law for a retransmission of a radio station's broadcast transmission within a 150 mile radius of the radio broadcast transmitter in setting the rate for radio retransmissions. See 17 U.S.C. 114(d)(1)(B).

Although RIAA maintains that in its negotiations with Yahoo! it had argued that the value of the radio retransmission should not be based on the location of the original radio broadcast transmitter, it claims that it was nervous about the application of the 150-mile radius exemption to retransmissions made by third-party aggregators, like Yahoo!. Consequently, RIAA maintains that it agreed to a lower rate for radio retransmissions, knowing that its arguments for not exempting these transmissions were weak, and because Yahoo! agreed to pay for each transmission without regard to the exemption. The resulting adjustment for the 150-mile exemption consisted of a reduction to the base rate, 0.2¢, and reflects the fact that about 70% of all radio retransmissions fall within the 150-mile zone. In addition, RIAA agreed to a further reduction to compensate Yahoo! for any “competitive disadvantage” it faced if commercial broadcasters were found to be totally exempt from the digital performance right under a separate exemption.

The Panel, however, did not credit RIAA's explanation and concluded that this concern over the exemptions, especially the 150-mile exemption, had no bearing on Yahoo!'s negotiations. The Panel steadfastly maintained throughout its report that Yahoo!'s only aim in the negotiation process was to achieve a rate that translated into an acceptable overall level of payment, and that it did not concern itself with the legal consequences of the 150-mile exemption. Report at 66-67. Thus, the Panel characterized RIAA's arguments in regard to the 150-mile exemption to be nothing more than a “red herring” and without effect in the negotiation process. Id. at 85. Consequently, the Panel found that Yahoo! willingly granted RIAA's request for the “whereas clause,” relating to the transmissions within the 150-mile radius, because it cost Yahoo! nothing. Yahoo!'s perception of the clause, however, did not alter the significance of the “whereas clause” to RIAA, who wanted the provision included in the agreement because it would allow RIAA to argue before this CARP that the 0.05¢ rate for radio retransmissions represents a real rate of 0.2¢, which was discounted to account for the legal uncertainties at the time of the negotiation. Report at 67.

Webcasters had problems with the Panel's analysis, too. It found fault with the Panel's approach to setting rates for webcasting based on the rates in the Yahoo! agreement. Webcasters object to the methodology used by the Panel in calculating the proposed rates, especially the use of an inflated rate as a starting point for setting the rates for IO transmissions. Moreover, they contest the use of any rate for IO transmissions contained in the Yahoo! agreement because Yahoo! had less interest in negotiating a favorable rate for these transmissions, which constituted only 10% of its business. Webcasters Petition at 30-40. Instead, Webcasters argue that Yahoo! agreed to the 0.2¢ rate for IO transmissions only because it obtained a significantly lower rate for its radio retransmissions, and that any number of possible combinations of rates could have been set to achieve Yahoo!'s targeted rate. Because of this, Webcasters argue that the endpoints settled upon in the agreement were patently arbitrary. The Register concurs with the Webcasters' analysis on this point and finds that the Panel's use of the IO rate was arbitrary because of the IO rate, which, in and of itself, did not reflect what the willing buyers and willing sellers had agreed to in the Yahoo! deal.

Another flaw in the Panel's reasoning, according to Webcasters, was its reliance on the 0.083¢ “blended rate” as the lower end of the acceptable range of IO rates. They argue that this rate should not even be considered because it was never negotiated as a performance rate at all. This observation, however, overlooks the fact that Yahoo! actually paid this rate for 1.5 billion performances without regard to the nature of the performances. The fact that the rate was not negotiated as a separate rate for Internet-only transmissions does not diminish its usefulness for purposes of this proceeding. As the Panel asserted throughout this proceeding, it is hard to find better evidence of marketplace value than the price actually paid by a willing buyer in the marketplace.

The question, however, is whether the rates in the Yahoo! agreement represent distinct valuations of Internet-only transmissions and radio retransmissions. Ultimately, the Register concludes that they do not and, therefore, the Panel's reliance on these specific rates for IO transmissions and radio retransmissions as a tool for setting the statutory rates is arbitrary. The fundamental flaw in the Panel's analysis, though, is not its acceptance of the Yahoo! agreement as a starting point. Rather, it is the Panel's determination that the differential rate structure reflects a true distinction in value between Internet-only transmissions and radio retransmissions based upon the promotional value to the record companies and performers due to airplay of their music by local radio stations. The Panel reached this conclusion in spite of the fact that nothing in the record indicates that the parties considered the promotional value of radio retransmissions over the Internet when they negotiated these rates.

RIAA maintains, and the Broadcasters concur, that no evidence exists to support the Panel's determination that Yahoo! and RIAA considered and made adjustments for the promotional value of radio retransmissions. RIAA Reply at 48; Broadcasters Petition at 39. In fact, the Broadcasters argue that it was “ ‘patently’ arbitrary for the Panel to conclude that promotional value was a “likely influence” on Yahoo!'s RR rate when the record evidence showed that neither party had ever suggested anything of the kind.” Broadcasters Petition at 39. The Register agrees and finds that the Panel's reliance on promotional value to justify the price differential for IO transmissions and radio retransmissions was arbitrary. The Panel's speculative conclusion that “this factor was likely considered by RIAA and Yahoo!, and is evidently reflected in the resulting difference between RR and IO negotiated rates,” only serves to undermine the validity of the Panel's final analysis on this point. See Report at 75.

Moreover, the Panel's own earlier findings with regard to the studies offered to show that the Internet has a promotional effect contradicts its later finding concerning the promotional effect derived from radio retransmissions over the Internet. After considering the two studies offered into evidence by the Services, the Panel categorically stated that it “could not conclude with any confidence whether any webcasting service causes a net substitution or net promotion of the sales of phonorecords, or in any way significantly affects the copyright owners' revenue streams.” Report at 33-34. It noted that “the Soundata survey presented by Mr. Fine evinced a net promotional effect of radio broadcasts, but said little about the net promotional effect of the Internet—and nothing about the net promotional effect of webcasting.” Id. at 33. It went on to say that “for the time period this CARP is addressing, the net impact of Internet webcasting on record sales is indeterminate. Id. at 34. These observations do not support a conclusion that radio retransmissions have a greater impact than IO transmissions on record sales or that either form of transmission has any impact on record sales.

However, the CARP did conclude that “to the extent promotional value influences the rates that willing buyers and willing sellers would agree to, it will be reflected in the agreements that result from those negotiations.” Id. But therein lies the problem. As discussed above, RIAA and Yahoo! did not consider promotional value when negotiating the Yahoo! agreement, therefore, its effect cannot be reflected in the IO and RR rates set forth in the Yahoo! agreement.

However, rejection of the CARP's conclusion on this point does not nullify the usefulness of the Yahoo! agreement. The Register accepts the Panel's determination that the Yahoo! agreement yields valuable information about the marketplace rate for transmissions of sound recordings over the Internet, and is a suitable benchmark for setting rates for all the reasons discussed in section IV.3, supra. Moreover, a careful review of the record support's the Panel's further finding that in effect, the real agreement between Yahoo! and RIAA was for a single, unitary rate for the digital performance of a sound recording and not the two separate rates set forth in the agreement—rates, which the Panel found were artificially high (for IO transmissions) and low (for RR).

The Register accepts the CARP's conclusion that the differential rate structure was developed to effectuate particular objectives of the parties, distinct and apart from establishing an actual valuation of the performances. Specifically, the Panel found that RIAA obtained an artificially high IO rate in an attempt to protect its targeted valuation of IO transmissions for use in this proceeding and Yahoo! received an “effective rate” it could accept. Because the record evidence supports this finding, Report at 65, referring to Tr. 11256-57; 11281 (Mandelbrot); Panel Rebuttal Hearing Exhibit 1 at 4; Tr. 11279-81, 11395-96 (Mandelbrot); Tr. 10237-38 (Marks), it was not arbitrary for the Panel to reach this conclusion. Report at 64-65 (noting that “Yahoo!'s primary concern, as characterized by its negotiator, was to negotiate a license agreement under which it would pay ‘the lowest amount possible’, that “Yahoo! was willing to accept a higher IO rate in exchange for a lower RR rate in order to achieve the lowest overall effective rate for all its transmissions” (emphasis added), and that Yahoo! was pleased to achieve the lowest possible overall rate.”); (noting that “the bottom line” combined rate was of paramount importance to Yahoo!). Report at 74. Moreover, Yahoo! maintains that it would not have paid the 0.2 cent rate for the IO transmissions but for the rate it received for radio retransmissions because the two rates, when considered together, yielded an acceptable “effective rate” for all transmissions. The testimony of David Mandelbrot, the Yahoo! representative, is particularly informative on this point.

Question: When you entered into the agreement with the RIAA, just looking at the 0.2 cents per performance rate for Internet-only broadcasting, you didn't consider that an unfair rate, did you?

Answer: Mandelbrot: We considered it a higher rate than we would have paid if we were just negotiating an Internet-only rate. I would say we did not consider it an unfair rate in the totality of the entire agreement, which was that we were getting the 0.05 cent rate for the radio retransmissions.

Mandelbrot Tr. at 11347-11348. This statement supports a finding that Yahoo!, the willing buyer in this case, did not accept the stated IO rate as an accurate reflection of what it would be willing to pay for the right to make those transmissions.

There is also scant evidence to indicate that Yahoo! gave any serious consideration to the effect of the 150-mile exemption for certain radio retransmissions when negotiating the IO and RR rates. Mandelbrot maintained that the exemptions were of little significance to Yahoo!, since it was “looking to use whatever [it] could to get as low a rate as possible.” Id. at 11381; see also 11331 (Mandelbrot admits using the ambiguities in the law, even though they thought the arguments in their favor were weak, solely for the purpose of getting “an effective rate that we could live with”). Again it is clear that Yahoo!'s focus was the negotiation of a rate at the lowest possible level that would allow it to conduct business without concerns about copyright violations.

Where such determinations are based on the testimony and evidence found in the record, the Register and the Librarian must accept the Panel's weighing of the evidence and its determination regarding the credibility of a witness. Likewise, the Register and the Librarian may not question findings and conclusions that proceed directly from the arbitrators' consideration of factual evidence in the record. In this instance, the Panel credited Mandelbrot's testimony and his characterization of the negotiation process, specifically concluding that his testimony was credible, and that Yahoo! understood the argument based on the 150-mile exemption had no significant impact on the rates ultimately negotiated. Report at 67. Consequently, we must accept the Panel's assessment on this point, which leads to the conclusion that the “effective rate” achieved through the unique rate structure represents the value these parties placed on the performance of a sound recording, without regard to origin of or the entity making the transmission.

Based upon a modification to the Panel's approach for calculating rates for making transmissions of sound recordings under statutory license that accepts as much of the Panel's reasoning as possible, the base rate for each performance is 0.07¢ (rounded to the nearest hundredth). The methodology for calculating this rate is presented and discussed in full in section IV.8.

6. Are Rates Based on the Yahoo! Agreement Indicative of Marketplace Rates?

Many webcasters, including Live365, maintain that the proposed rates derived from the Yahoo! rates do not reflect what a willing buyer would pay in the marketplace for the right to make these transmissions. Live365 maintains that the Panel incorrectly analyzed the evidence in the record. First, it notes that the Panel itself found that many of the rates in the voluntary agreements were prohibitively high, including a revenue-based royalty set at 15% of a webcaster's gross revenue. Live 365 Petition at 16. It then argues that it was arbitrary for the Panel to make this finding and then propose rates that exceed the rates it deemed to be excessive, and more than the market could bear. Id. To make its point, Live365 uses the Panel's per performance rate and calculates how much certain services would pay for the digital performance right and translates that amount into a percentage of revenue metric. In each of the cited examples, the amount to be paid based on the proposed per performance rate (as expressed as a percentage of revenues) is considerably higher than that that would be required under any of the percentage-of-revenue models proposed by any party at any time. For example, under the Panel's proposed rates, one service would purportedly pay 21% of its gross revenue, a figure which is considerably higher than the 15% of gross revenues contained in many of the voluntary agreements ultimately rejected by the Panel. Based on this observation, Live365 contends that the Panel's proposal runs counter to the evidence and, therefore, it is arbitrary. Id. at 18.

Moreover, Live365 argues that the Panel failed to account for relevant market factors, including how much a webcaster can pay. Id. at 19. Webcasters voice similar concerns, arguing that the adoption of a per performance rate will cause ruin to many webcasters who to date have yet to generate a viable income stream. Webcasters Petition at 60. In place of this structure, webcasters assert that a percentage-of-revenue model must be adopted in order to address the economic situation facing small, independent webcasters. They maintain that those Services that entered into voluntary agreements based on a percentage-of-revenue will remain in business while those operating under the statutory license with its per performance royalties will not. Webcasters Petition at 62-63. In the eyes of the Webcasters, such a result reflects unexplained disparate treatment of similarly situated parties, and requires an adjustment to eliminate this unjust and arbitrary result. Webcasters also argue that the Panel failed to articulate a rational basis for failing to offer an alternative rate structure based on percentage-of-revenue.

In addition, Live365 argues, as do the Broadcasters, that Yahoo! is a substantially different type of business from small start-up webcasters who would be unwilling to pay the same rates as Yahoo! for the use of sound recordings. Thus, it contends that the Yahoo! rates do not reflect what these buyers would be willing to pay in the marketplace. The implication is that these businesses have expended significant monies on start-up costs, including software, infrastructure development, and bandwidth, and having not yet established substantial revenue streams would be unable or unwilling to pay the same rates. Live365 Petition at 7, 11. Moreover, Live365 argues that the rates set by the Panel thwart Congressional intent “by making Internet performances of sound recordings economically unviable for many webcasters.” Live365 Petition at 21.

RIAA takes exception with the Webcasters and Live365 on these issues. It analyzes how much certain webcasters and Live365 pay, as a percentage-of-revenue, for sales and marketing cost, personnel cost and bandwidth. The results show that a company's costs for these services can amount to more than 100 times the amount of a company's revenue, whereas the projected costs of the royalties for transmitting sound recordings for the same time period are no more than 2 times the amount of a company's revenue. RIAA Reply at 57. In all cases, these costs reflect the start up nature of the industry, and not the ultimate make or break point of the business. Thus, a proposed fee that results in royalty payments above the current revenue stream for a webcaster is not atypical or unexpected. Certainly, if that were the measure of the value of these services, then the costs for employment, hardware, and marketing—so essential to establishing and maintaining the business—must also be viewed as excessive and above the fair market value for each of these services. Clearly, that is not the case, nor can one rationally conclude that it should be the case.

Moreover, RIAA notes that the courts have historically upheld rates set by the CRT, even when users have argued that the rates would cause the business to cease certain operations. Where the intent of Congress is to set a rate at fair market value, as in this proceeding, the Panel is not required to consider potential failure of those businesses that cannot compete in the marketplace. See National Cable Television Ass'n. v. CRT, 724 F.2d 176 (D.C. Cir. 1983) (holding that rates set at fair market value were proper even though cable operators argued that the rates were prohibitively high and would cause them to cease transmission of the distant signals at issue.).

The law requires only that the Panel set rates that would have been negotiated in the marketplace between a willing buyer and a willing seller. It is silent on what effect these rates should have on particular individual services who wish to operate under the license. Thus, the Panel had no obligation to consider the financial health of any particular service when it proposed the rates. It only needed to assure itself that the benchmarks it adopted were indicative of marketplace rates.

7. Should a Different Rate be Established for Commercial Broadcasters Streaming Their Own AM/FM Programming?

Although RIAA had argued that the rate for commercial broadcasters should be the same as the rate for Internet-only webcasters, the Panel did not agree. It did agree, however, that the rate for commercial broadcasters should be the same as the rate adopted for radio retransmissions and that these rates should be based on the Yahoo! agreement.

It noted that the Yahoo! agreement established rates for retransmissions of the same types of radio station signals as those directly streamed by commercial broadcasters. Consequently, it put the burden of proof on the broadcasters to present evidence to distinguish between the direct transmission of their programs over the Internet and the retransmission of the same programming made by a third-party. Broadcasters were unable to offer any compelling evidence on this point. Thus, in the end, the Panel was unable to distinguish between commercial broadcasters and radio retransmisions, stating that “the record was utterly devoid of evidence implying a higher rate [for commercial broadcasters] and insufficient [evidence] to warrant a lower rate.” Report at 84-85. (emphasis in the original).

Nevertheless, Broadcasters are troubled by the Panel's use of the Yahoo! agreement to set rates for broadcasters for two main reasons. First, they argue that Yahoo! represents a substantially different type of business. Second, they maintain that the Panel must make affirmative findings that the businesses are comparable before applying the same rates to both Services. Broadcasters Petition at 26-27.

Indeed, Yahoo! offers a plethora of services, making available hundreds of radio stations, local television stations, video networks, concerts, CD listening programs, Internet-only music channels and educational and entertainment video programs. Id. at 28. Nevertheless, an examination of the record clearly shows that both business models are fundamentally comparable in at least one all-important way: they simulcast AM/FM programs over the Internet to anyone anywhere in the world who chooses to listen. Even accepting the fact that Broadcasters say their fundamental business is to provide programming to their local audiences, the potential for reaching a wider audience cannot be denied. Given that the record indicates that 70% of Yahoo!'s radio retransmissions are to listeners within 150 miles of the originating radio station's transmitter, Yahoo!'s business with respect to radio retransmissions seems to be very similar. Moreover, the fact that Yahoo! offers many additional services is not relevant to this proceeding because the Yahoo! agreement only addressed the rates Yahoo! paid for streaming sound recordings over the Internet. Had the contract been tied to other services offered by Yahoo!, it might well have been inappropriate to use this contract in this context. That is not the case and so it was not arbitrary for the Panel to rely on the Yahoo! contract to set the rate for broadcasters who stream their own programming over the Internet.

Commercial broadcasters then take another approach and argue that they never would have agreed to the rates that Yahoo! paid because their purposes for streaming differ from Yahoo!'s purposes. Commercial broadcasters assert that they began streaming in order to have a presence “in the online world, to maintain the local radio brand, and as a convenience to their regular over-the-air listeners.” Broadcasters Petition at 29. They then note that many commercial broadcasters have already ceased streaming because of an increase in costs. They cite this fact as evidence of their assertion that they would only be willing to pay a significantly lower rate than a third-party aggregator like Yahoo! See Broadcasters Petition at 31, fn 25 (offering examples of decisions made by radio stations to cease their streaming operations because of bandwidth fees and dispute over royalty fees between AFTRA and the advertising agencies). They also cite the testimony of David Mandelbrot, who testified that Yahoo! feared broadcasters would be unwilling to absorb the rates Yahoo! negotiated for streaming AM/FM programming. Id. at 32. Based upon this evidence, the Broadcasters and Live365 conclude that the Panel acted in an arbitrary manner in setting the rates that will put many services out of business. Live365 Petition at 15, 18.

However, the Panel did consider the differences between the two business models, speculating that it was entirely possible that the cost to stream AM/FM programming would be lower for broadcasters than for third-party aggregators like Yahoo! Id. at 84-85. Had Broadcasters made that argument or similar ones to show that Yahoo! received greater value from its streaming activities, the Panel may well have set a lower rate for Broadcasters who stream their own programming. Id. at 85. But as the Panel observed, it cannot make adjustments based on mere speculation. So when the Panel found no record evidence to distinguish these services, it had no reason to offer a separate rate for commercial broadcasters who stream their own AM/FM signal over the Internet. Id. at 84.

Moreover, RIAA points out that Yahoo! never even tried to pass along the costs of the transmissions to the radio stations. Thus, no determination could be made as to whether the broadcasters would have accepted the rate and paid it, or rejected it out of hand. RIAA Reply at 45. RIAA's observation is persuasive, as is the Panel's general observation that the record did not contain any evidence to support a different rate for commercial broadcasters. Thus, the Panel's decision not to set a different rate for commercial broadcasters was not arbitrary.

For these reasons, the Register accepts the Panel's decision not to differentiate between simulcasts made by commercial broadcasters and simulcasts of the same programming made by a third-party aggregator. Accordingly, the rate for commercial broadcasters streaming their over-the-air radio programs on the Internet is the unitary rate gleaned from the Yahoo! agreement.

8. Methodology for Calculating the Statutory Rates for the Webcasting License

a. Calculation of the unitary rate. In section IV.5, the Register rejected the Panel's determination that the Yahoo! agreement provided a basis for establishing different rates for Internet-only transmissions and radio retransmissions. Instead, a determination was made that the Yahoo! agreement justified only a single rate applicable to all transmissions, without regard to the source of the transmission. To calculate this unitary rate, it is necessary to determine what Yahoo! paid for the initial 1.5 billion performances, based on the lump sum payment, and what it expected to pay for transmissions after that time.

The first calculation was actually done by the Panel based upon Yahoo!'s agreement to pay RIAA $1.25 million for the first 1.5 billion transmissions made by Yahoo!. It divided the amount paid by the number of performances ($1.25 million/1.5 billion performances) to get a “blended” rate of 0.083¢ per performance. Report at 63. To determine the “effective rate” for the second period, a calculation must be made to account for the differential IO and RR rates, 0.2¢ and 0.05¢, respectively, set forth in the agreement and the relative proportion of Internet-only transmissions to radio retransmissions. This is a simple arithmetic calculation and one that Yahoo! had already performed in order to gauge the actual costs of the performances under the differentiated rate structure. This calculation yielded an “effective” or “blended” rate of 0.065¢ per performance based upon Yahoo!'s expectation that 90% of its transmissions would continue to be radio retransmissions with the remaining 10% being Internet-only transmissions [((9 × 0.05¢) + (1 × 0.2¢))/10]. Report at 63, citing Tr. 11279, 11292 (Mandelbrot), Panel Rebuttal Hearing Exhibit 1 at 7.

Now the question is how to reconcile these values to determine the unitary rate. Although an argument can be made for adopting either value, it makes more sense to use both values and take the average of the two. In this way, the final unitary rate captures the actual value of the performances made in the initial period (for which Yahoo! paid a lump sum for the first 1.5 billion performances) and the projected value of the transmissions at the agreed upon rates for the remainder of the license period; and it falls within the range of acknowledged values of these transmissions. Courts have long acknowledged that rate setting is not an exact science, and all that is necessary is that the rates lie within a “zone of reasonableness.” See National Cable Television Assoc. Inc. v. CRT, 724 F.2d 176, 182 (D.C. Cir. 1983) (“Ratemaking generally “is an intensely practical affair. The Tribunal's work particularly, in both ratemaking and royalty distributions, necessarily involves estimates and approximations. There has never been any pretense that the CRT's rulings rest on precise mathematical calculations; it suffices that they lie within a “zone of reasonableness”'). Thus, the record here supports a “zone of reasonableness” between 0.083¢ and 0.065¢.

Accordingly, the Register recommends that the rate for making an eligible nonsubscription transmission of a sound recording over the Internet under section 114 be set at 0.07 cents per performance, per listener, the midpoint of the “zone of reasonableness.”

Determination of this rate, however, is not necessarily the end of the rate-setting process. Webcasters had argued for a downward adjustment to the rates proposed by the Panel to compensate for litigation cost savings and added value due to MFN clause. Such arguments apply with equal force to the unitary rate proposed by the Register. Webcasters Petition at 42-43. The Webcasters' argument is well taken and, based on the record evidence, it is reasonable to assume that the rates in the Yahoo! agreement are slightly higher to account for these two factors. See Report at 68-69. However, there is a problem in making an adjustment to the proposed rate where the record contains no information quantifying the added value of the factors that purportedly resulted in inflated rates. See Report at 29 (discussing lack of record evidence quantifying value of any factor, other than promotional value, that allegedly influenced the negotiated rates). The potential (but apparently unquantifiable) added value attributable to these 2 factors might present a problem if the Register were proposing a rate at the high end of the 0.065¢-0.083¢ range, but because the Register is recommending a rate in middle of the “zone of reasonableness,” it is safe to conclude that the recommended rate falls into that zone of reasonableness even taking these factors into account.

Similarly, Broadcasters argued for a downward adjustment of the simulcast rate to account for the promotional value associated with over-the-air broadcasts. Broadcasters Petition at 41. The record, however, does not support this suggestion. Indeed, the Panel did acknowledge that over-the-air radio retransmissions had promotional value, but it concluded that “the net impact of Internet webcasting on record sales is indeterminate.” Report at 34. This is not to say that webcasting, including simulcasting of over-the-air radio programming, has no promotional value. It only means that the record companies gain similar benefits from both types of transmissions. Consequently, no adjustment is necessary.

b. The 150-mile exemption. Under section 114(d)(1)(B)(I), any retransmission of a nonsubscription broadcast transmission is exempt, as a matter of law, from the digital performance right, provided that “the radio station's broadcast transmission is not willfully or repeatedly retransmitted more than a radius of 150 miles from the site of the radio broadcast transmitter.” During the course of the negotiations between RIAA and Yahoo!, there was a great deal of uncertainty regarding this provision and whether it applied to transmissions made over the Internet. See discussion above, section IV.a.5.

As noted above (section IV.a.5.), in its Petition, RIAA argued that during the course of the negotiations between RIAA and Yahoo!, there was a great deal of uncertainly regarding this provision and whether it applied to transmissions made over the Internet. RIAA argued that because of this uncertainty, it had been willing to agree to a lower radio retransmission rate. In fact, RIAA pointed out that its chief negotiator had advised its negotiating committee that RIAA's arguments against application of the 150-mile exemption to a retransmitter such as Yahoo! “are not particularly strong.” RIAA Petition at 20.

Confronted with the assertions made in RIAA's petition which indicated that RIAA itself had had considerable doubts on the subject at the time of the negotiations, the Register felt compelled to determine whether radio retransmissions over the Internet to recipients within 150 miles of the radio transmitter are, in fact, eligible for the section 114(d)(1)(B) exemption. The Register issued an order on June 5, 2002, asking the parties to brief two legal questions concerning the 150-mile exemption. The first question asked whether a retransmission over the Internet of a radio station's broadcast transmission to a recipient located within 150 miles of the site of the radio broadcast transmitter is an exempt transmission pursuant to 17 U.S.C. 114(d)(1)(B). The second question then queried whether the exemption would still apply to radio retransmissions made within the 150-mile radius by a Licensee, in the case where that same service is simultaneously retransmitting the radio station's broadcast transmission of one or more recipients, located more than 150 miles from the site of the radio broadcaster's transmitter.

Section 114 could be read as allowing a Licensee to take advantage of the exemption for all Internet retransmissions of a radio broadcast to recipients within a 150 mile radius of that radio station's transmitter. The statutory language, however, does not make clear whether that same Licensee would retain the benefit of the exemption for those transmissions if additional retransmissions of the radio broadcast signal were also made “willfully” or “repeatedly” outside the 150-mile radius.

A critical piece in the analysis is the meaning of the word “retransmission.” Each retransmission of a radio signal over the Internet may be viewed as a discrete, point-to-point transaction to be considered on its own merit without reference to further retransmissions made by the Licensee. Alternatively, the reference to “willful and repeated” may require consideration of each retransmission, together with all other retransmissions, made by the Licensee to multiple listeners over a period of time, both inside and outside the 150-mile radius.

Having considered the parties' responses, the statutory language and its relationship to section 112, the Register now concludes that the exemption is not applicable to radio retransmissions made over the Internet. While Copyright Owners and Performers offer many arguments in support of their position that radio retransmissions within 150 miles of the radio station's transmitter are not exempt, and while Broadcasters offer many arguments to the contrary, the critical piece of the analysis—and the argument that the Register finds persuasive—is found in the text of section 112(e). This section provides a statutory license for making ephemeral recordings only to “a transmitting organization entitled to transmit to the public a performance of a sound recording under the limitation on exclusive rights specified by section 114(d)(1)(C)(iv) or under a statutory license in accordance with section 114(f).” 17 U.S.C. 112(e)(1).

The statutory license for ephemeral recordings in section 112(e) was enacted as part of the same section of the DMCA—section 104—that expanded the section 114 statutory license to include webcasting. The purpose of this ephemeral recording statutory license was to enable business establishment services and services using the new section 114 statutory license for webcasting to make the ephemeral recordings they need to make in order to facilitate their licensed transmissions, and in recognition of the fact that the exemption in section 112(a) permitting the making of a single ephemeral recording might not be adequate. See H.R. Rep. 105-796, at 89-90.

Congress expressly provided in the DMCA amendments that business establishment services operating under the section 114(d)(1)(C)(iv) exemption are eligible for the section 112(e) statutory license for ephemeral recordings in order to facilitate Internet transmissions by business transmission services. Congress's failure to do the same for services operating under the section 114(d)(1)(B) exemption demonstrates that Congress did not contemplate that that exemption would be available to services making retransmissions via the Internet.

Moreover, if section 114(d)(1)(B) were interpreted as providing an exemption for a radio retransmission over the Internet, when that retransmission is to a recipient located within 150 miles of the radio station's transmitter, the Licensee could not make ephemeral recordings to facilitate such an exempt retransmission. This interpretation would put the Licensee in the illogical position of having a right to retransmit the radio signal, but no means of accomplishing the retransmission without negotiating private licenses to make ephemeral recordings to facilitate the exempt transmissions. At the same time, the Licensee could operate under a statutory license for making the ephemeral recordings to facilitate its non-exempt transmissions beyond the 150-mile radius made pursuant to the section 114(f) statutory license. As RIAA points out in its response to the June 5 Order: “Such a result is inconsistent with one of the purposes of the DMCA statutory licenses to create efficient licensing mechanisms for copyright owners and webcasters,” citing H.R. Rep. 105-796, at 79-80 (1998). Consequently, the better interpretation of the section 114(d)(1)(B) exemption is to consider all retransmissions of a License in the aggregate, which logically means that no Internet retransmissions are exempt under section 114(d)(1)(B).

Based on the interplay between sections 112 and 114, the better interpretation of the law is that the exemption does not apply to radio retransmissions made over the Internet.

9. Rates for Other Webcasting Services and Programming

a. Business to business webcasting services. Some Services provide specialized Internet radio-like stations to businesses rather than directly to consumers. These business-to-business webcasting services (B2B) are in many respects analogous to business establishment music services and can provide programming customized to the demographics of the customers of a particular business. Report at 78. For this reason, RIAA had proposed setting a higher rate for business to business webcasting services than for business to consumer (B2C) services. The Panel, however, rejected this suggestion, finding that the evidence did not support a higher rate for B2B services. It found that most of the agreements for such services had rates near or below the predominant rate set for standard Internet-only transmissions. Report at 79. Thus, the Panel concluded that it had “found insufficient evidence to support a separate rate for syndicator services”, and set the rate accordingly at 0.14¢ per performance, just as it had for Internet-only performances. Id.

RIAA argues for a premium rate for these Services, because they syndicate their programming through third-party non-entertainment websites. RIAA maintains that these transmissions are outside the scope of the webcasting license, and consequently, services should pay a premium when they make transmissions through non-entertainment websites. RIAA Petition at 50-52. In response, Webcasters argue that the “value of the performance does not change merely because of the technology of the webcaster or the fact that the sound recording is heard when it is accessed at a third-party website rather than the originating webcaster's website.” Webcasters Reply at 57. Moreover, they maintain that RIAA offered no evidence to demonstrate that these transmissions should be valued at a higher rate. In fact, the record indicates the opposite. Most of the RIAA voluntary agreements which permit the licensee to distribute its webcasts to third-party websites contain no premium for this practice. Id. at 59.

Thus, based on the weight of the evidence, it was not arbitrary for the Panel to conclude that a separate rate should not be set for syndication services. The Panel is responsible for weighing the evidence and so long as the record supports its decision, the Register will not second-guess the Panel's finding of fact. Nevertheless, this determination does not end the inquiry. RIAA correctly cites section 114(j)(6) of the Copyright Act for the proposition that an eligible nonsubscription transmission does not include those made by a service whose primary purpose is to sell, advertise, or promote particular products or services other than sound recordings, live concerts, or other music-related events. Thus, in any given case a determination would have to be made to ascertain whether such transmissions are covered under the statutory license. This proceeding, however, is not the appropriate vehicle for such a fact-specific determination. If a court determines that the transmissions made by a particular business-to-business service fall outside the scope of the webcasting license, then those transmissions are acts of copyright infringement unless the service obtains licenses from the copyright owners. In such cases, an infringement action would be the appropriate course of action, rather than the imposition of a premium rate for such transmissions as suggested by RIAA. No rate—premium or otherwise—can be set for a transmission that does not comply with the terms of the license.

b. Listener-influenced services. There was also much discussion about listener-influenced services that allow the listener some control over the programming through on-line ratings and skip-through features. RIAA's position first and foremost is that these services do not qualify for the webcasting license. However, RIAA also proposed a much higher rate for these services in the event the Panel discerned a need to set a separate rate for these services. Again, the Panel found no record support for setting a separate and higher rate for listener-influenced services. It rejected the agreements between RIAA and non-DMCA compliant services because the rates in those agreements were for rights beyond those granted under the statutory license. Nor could the Panel discern from the record evidence which services would be subject to the basic webcasting rate as distinguished from the rate for listener-influenced services. Consequently, the Panel decided “that so long as a service complies with, and is deemed eligible for the statutory license, it should not pay a separate rate based upon listener influence.” Report at 81.

The Register finds the Panel's analysis to be consistent with the law, and thus accepts the Panel's decision not to set a separate rate for transmissions which might not come within the scope of the license. Again, if transmissions made by a listener-influenced service are determined to be outside the scope of the statutory license, the proper course of action would be for the parties to negotiate a voluntary agreement for these transmissions, or for the copyright owner to file a copyright infringement suit against the service. The Panel has no authority to propose a rate for any transmission which cannot be made lawfully under the statutory license.

c. Other types of transmissions. A broadcaster may stream three different types of programming in addition to a simulcast of its AM/FM radio signal: (1) “Archived” (previously aired) radio programming; (2) “side channels” (Internet-only programming); and (3) “substituted programming” (programming that replaces over-the-air programming that has not been licensed for simulcast over the Internet). The question for the Panel was whether such programming is the same or substantially similar to radio retransmissions or Internet-only programming.

In making its decision, the Panel first considered the definition of a “radio retransmission performance.” It found that the record failed to provide a coherent and workable definition, rejecting both the definition set forth in the Yahoo! agreement and the one that was included in the defunct settlement agreement between RIAA and the commercial broadcasters. Instead, it adopted the definition of the term provided by Congress in the statute which defines the term as “a further transmission of an initial transmission * * * if it is simultaneous with the initial transmission.” See 17 U.S.C. 114(j)(12). Based on this definition, the Panel concluded that a transmission made as part of archived programming, side channels or substituted programming was something other than a radio retransmission and, therefore, not entitled to the lower rate proposed for radio retransmissions. Instead, it agreed with RIAA that the programming was essentially the same as Internet-only programming, and without any record evidence to substantiate a different rate, should be subject to the 0.14¢ IO rate.

Broadcasters do not contest the Panel's determination with respect to side channels, and they recommend that the Librarian provide that the side channel rate be set at the webcaster rate expressly without prejudice to reconsideration in a subsequent CARP proceeding. Broadcasters Petition at 56. They do, however, object to the imposition of the rate for IO transmissions on the performances of sound recordings made during the transmission of an archived program or a substituted program. Id. at 55. Broadcasters' arguments no longer have any relevance under the statutory rate structure proposed by the Register, which proposes a single, unitary rate for all transmission. This fact in conjunction with the Panel's observation that the Yahoo! agreement did not differentiate or even recognize these alternative categories supports a determination that no separate rate should be set for these transmissions.

10. Rates for Transmissions Made by Non-CPB, Noncommercial Stations

National Public Radio (“NPR”) and the National Religious Broadcasters Music License Committee (“NRBMLC”) were the only two representatives of non-commercial stations participating in this proceeding. NPR reached a private settlement with the Copyright Owners during the proceeding and withdrew. In considering what the rate should be for the stations represented by NRBMLC and any other noncommercial station operating under the statutory license, the panel first considered past CARP decisions involving the statutory licenses. It found that a prior CARP had considered and distinguished commercial stations and noncommercial stations on the basis of their financial resources, noting that noncommercial stations depend upon funding from the government, business, and viewers, whereas commercial broadcasters generate a revenue stream through advertising. Report at 89, citing CARP report adopted by Librarian on September 18, 1998, Noncommercial Education Broadcasting Rate Adjustment Proceeding, 63 FR 49823. Moreover, the earlier Panel determined that a rate set for a commercial station is an inappropriate benchmark to use when setting a rate for the same right for noncommercial stations because of these economic differences between these businesses. Specifically, it acknowledged that use of a rate set for a commercial broadcaster would overstate the market value of the performance for a noncommercial station.

Next, the Panel examined RIAA's approach, which focused on the amount the performing rights organizations (“PROs”) were awarded in the 1998 Noncommercial Education Broadcasting Rate Adjustment Proceeding for use of their works by noncommercial stations. It adduced that they received 1/3 the amount of the fees paid by the commercial stations. Based on this precedent, RIAA offered the noncommercial stations a rate that corresponds to 1/3 the rate to be paid by commercial broadcasters. The Panel, finding no other evidence in the record to support a different rate, adopted the RIAA proposal for radio retransmissions, and proposed a rate of 0.02¢ per-performance (one-third of the 0.07¢ per performance rate, rounded to the nearest hundredth of a cent) for these transmissions only. Just as with the commercial broadcasters, the Panel found that archived programming subsequently transmitted over the Internet, transmissions of substituted programming, and transmissions of side channels constitute a transmission more akin to an Internet-only event. Consequently, it proposed a per performance rate for noncommercial broadcasters of 0.05¢ (one-third the rate paid by commercial broadcasters and webcasters for IO transmissions) for each sound recording included in these transmissions. This rate, however, is meant to apply only to the first two side channels—and not to additional side channels—in order to avoid the possibility of a noncommercial broadcaster gaining a competitive advantage over the commercial broadcasters and webcasters who initiate Internet-only programs and do so at a higher cost.

Non-CBP broadcasters argue in their petition to set aside the CARP report, that the Panel failed to set the appropriate rates in two ways. They contend that the Panel ignored the record evidence which clearly established that the noncommercial stations are fundamentally different from commercial broadcasters and webcasters, and less viable economically, thus requiring the Panel to establish a lower rate for these stations. They also dispute, like the Webcasters and the commercial broadcasters, the Panel's decision to reject, as a benchmark, the amount of royalty fees these services pay for the use of the underlying musical works in an analog market under a separate compulsory license. Non-CPB Petition at 4. They then calculate a ratio between what a commercial broadcast station pays for use of the musical works in the analog world and what on average the non-CPB stations pay in the same market, based on an estimation of the number of stations, and the amount of royalties the stations paid for use of musical works in their over-the-air broadcasts. From these calculations, they suggest that a noncommercial broadcaster, on average, pays only 1/34th the amount of royalties that a commercial station pays for use of the same musical works and argue for a rate equal to 1/34th the amount that commercial broadcasters will pay. Alternatively, they request a flat rate of $100 per station, see Non-CPB, Noncommercial Broadcasters Reply Petition at 5, and argue that in no case should the rate exceed 1/3 the rate adopted for commercial broadcasters. Non-CPB, Noncommercial Broadcasters Petition at 9.

NRBMLC also turned to the rates for the statutory noncommercial broadcasting license and argued that the rates for the webcasting license should be based upon the rates currently paid to performing rights organizations for use of the musical works in over-the-air programs under this license. The Panel rejected this proposal on a number of grounds. First, it noted that those rates were the subject of prior settlements which stated that the negotiated rates for the noncommercial license were to have no precedential value for future rate setting proceedings for the noncommercial license. In light of this term, the Panel found the rates for the statutory noncommercial license had no relevance to the current proceeding. Not only were the rates for a totally different right, but they apparently have no precedential value for considering future statutory noncommercial rates for use of the musical works. Report at 90. Second, the panel considered rates proposed by Dr. Murdoch, the expert witness for NPR, who at the request of the Panel made an attempt to identify an appropriate rate for noncommercial stations based on the fees currently paid to the PROs. Although she complied with the request of the Panel, she expressed severe reservations about her own conclusions, citing numerous problems with her own calculations. Report at 91. For these reasons, the Panel rejected Murdoch's proposed rates.

RIAA supports the Panel's decision, noting that the non-CPB, noncommercial broadcasters failed to offer any differential rate for this type of service in its direct case or an expert witness who could support their ultimate request for a $100 flat rate. The only witness who testified on behalf of this group was Joe Davis, who works for a commercial broadcaster, and had only anecdotal information concerning noncommercial stations. Because of his lack of expertise in this area, the Panel did not credit his testimony. Such action on the part of the panel is not arbitrary.

Nor was it arbitrary for the Panel to decide not to rely on the statutory rates set for use of the musical works by noncommercial broadcasters. The arbitrators rejected the non-CPB, commercial broadcasters' request to look to these rates because the agreements, at the insistence of the parties to the agreements, are not even considered precedent for setting future rates for the use of the musical works. If anything, it would be arbitrary to rely on these values as a benchmark for setting rates for a completely different category of works when they had no acknowledged value for readjusting the rates for the works to which they do apply. Had the Panel wished to use these rates, it needed at the very least an opportunity to examine the circumstances surrounding the adoption of the “no precedent” clause. It would have also required record evidence to substantiate such bold assertions on the part of the users as the notion that these rates were set at a rate higher than what would have been negotiated in the marketplace. Non-CPB Broadcasters Reply Petition at 7; RIAA Reply at 11. Because of these infirmities, the Register finds the Panel did not act arbitrarily in rejecting the rates set for the section 118 license as a benchmark.

Thus, in the end, the Panel accepted RIAA's proposal to set the rate for noncommercial broadcasters at one-third the rate established for commercial broadcasters. The Panel also provided a separate rate for archived programming subsequently transmitted over the Internet, substituted programming and up to 2 side channels set at one-third the rate established for Internet-only transmissions. The Panel made this adjustment based on its determination that a noncommercial broadcaster should not be subject to commercial rates when streaming programming consistent with the educational mission of the station, over the Internet. Report at 94. However, the Panel imposed a limitation on the use of this reduced rate for Internet-only transmissions to avoid the possibility that a non-CPB broadcaster could use its unique position to essentially become a commercial webcaster.

The Register accepts the Panel's methodology for setting the rate for noncommercial broadcasters. The rates proposed by the Panel, however, must be adjusted to reflect the Register's recommendation to set a unitary rate for both commercial broadcasters and webcasters. Using the proposed base rate of 0.07¢ and reducing this value by two-thirds, the adjusted rate for non-CPB, noncommercial broadcasters is 0.02¢ (one-third of 0.07¢, the base rate for all transmissions, rounded to the nearest hundredth) per performance, per listener. This rate shall apply to a simultaneous retransmission of the non-CPB, noncommercial over-the-air radio programming, archiving programming subsequently transmitted over the Internet, substituted programming, and up to two side channels. The rate for all other Internet-only transmissions is 0.07¢.

One last disputed issue raised by the non-CPB, noncommercial broadcasters is the imposition of the same $500 minimum fee that the CARP set for all other licensees. They argue that a $500 minimum fee far exceeds any reasonable rate that should be imposed on this category of users in light of the financial considerations that distinguish them from the other services. Non-CPB Broadcasters Reply Petition at 10. In support of this position, the users cite Dr. Murdoch's testimony to illustrate that the Internet license for use of SESAC's repertoire is less than $100. But this is not the total amount that a noncommercial station would pay; it would also have to pay fees to BMI and ASCAP in order to license all the works included in the sound recordings covered by the section 114 license. The minimal amount that a webcaster must pay to cover the combined works administered by the three PROs is $673, more than the proposed minimum rate to operate under the section 114 license. Webcasters PFFCL ¶ 363. In any event, the Panel set the rate at $500 to cover administrative costs to the copyright owners and access to the sound recordings. It was not arbitrary to impose a minimum fee on the Non-CPB, noncommercial broadcasters that merely covers costs for these rudimentary purposes nor can it be deemed excessive in light of what these entities pay the PROs for the public performance of musical works.

11. Consideration of Request for Diminished Rates and Long Song Surcharge

RIAA requested a surcharge for songs longer than five minutes. RIAA PFFCL ¶ 210. Its request was denied because the Panel did not find that such a charge was included in most of the relevant license agreements. Report at 105. RIAA, however, argues that the Panel misread the Yahoo! agreement. RIAA Petition at 42. It notes that Yahoo! could estimate the number of performances it made by multiplying its listening hours by a fixed number of performances and that when it did so, the record companies received compensation for [material redacted subject to a protective order] performances, even though Yahoo! may have only played, for example, 5 12-minute classical recordings in an hour. Id. The Yahoo! agreement, however, does not require that it employ the estimation methodology; it merely states that Yahoo! may make this calculation. Thus, there was no probative evidence that the marketplace valued a classical sound recording, or similar sound recordings of longer than average duration, at a different rate. Consequently, it was not arbitrary for the Panel to reject RIAA's suggestion to impose a “long song” surcharge. In any event, it is highly likely that this concern will be addressed for the time period to which these rates apply, since most services will be using the estimation formula for calculating the number of performances which assumes 15 performances for each aggregate tuning hour. See section IV.11, infra.

On the other side, webcasters asked that there be no royalty fee for songs that are less than thirty seconds long, citing technology problems or the use of song-skip functions. Webcasters Petition at 71. The Panel disagreed and saw no need to make any adjustment. It noted that the use of the blended rate from which it calculated the proposed rates was itself based upon figures which already took into account problem performances that had occurred during the initial period. This adjustment was expressly made for the first 1.5 billion transmissions only. Report at 106-107. The Panel chose not to make a similar adjustment for subsequent performances because the Yahoo! agreement did not provide for such an adjustment.

Likewise, the Panel determined that the use of the skip function provides a benefit to webcasters and it saw no need to penalize copyright owners for the benefit that flowed to the users through a conscious use of a function provided by the service. Moreover, none of the negotiated agreements provided for any reduction in rate for skipped songs. Report at 107. Consequently, the Panel did not provide a lower rate or exemption for truncated performances resulting from use of the skip song function.

The Webcasters object to the Panel's conclusion, maintaining that the Panel failed to adequately explain its decision and consider relevant evidence. See Webccasters Petition at 71. They contend that the Panel should have given more weight to three of the 26 agreements, which provided an exemption for performances less than thirty seconds in duration. Such action, would itself, have been arbitrary. Clearly, the Panel could not rely on these agreements when it had already disregarded them for purposes of establishing the royalty rates.

Moreover, RIAA makes a number of arguments in support of the Panel's decision. First, it notes that the performance of even a portion of a sound recording without a license is an infringement of a copyright owner's rights. As such, there is no a priori reason for making 30-seconds-or-fewer performances exempt from royalty obligations. Second, RIAA cites 17 U.S.C. 114(h)(2)(B) to demonstrate that Congress recognized the value of performances of limited duration and the right to license such performances. Specifically, this section exempts copyright owners licensing public performances of sound recordings from the requirement to make these sound recordings available on no less favorable terms or conditions to all bona fide entities, when they are licensing promotional performances of up to 45 seconds in duration. RIAA Reply at 71-75. These arguments support the Panel's decision not to exempt performances of thirty seconds or less, and as such, its decision is neither arbitrary nor contrary to law.

The Panel did, however, grant the users an exemption for incidental performances, citing the existence of a similar term in the Yahoo! agreement as the basis for its decision. Specifically, the Panel “exclude[d] transmissions or retransmissions that make no more than incidental use of sound recordings, including but not limited to, certain performances of brief musical transitions, brief performances during news, talk and sports programming, commercial jingles, and certain background music.” Report at 108. This is not a disputed provision.

With the agreement of the parties, the Panel also exempted performances of sound recordings made pursuant to a private license agreement. Id.

The Register notes, however, that the Webcasters' concerns regarding the Panel's determination not to grant its request to impose no royalty on songs less than 30 seconds in duration are ameliorated for the current licensing period. Under the proposed terms of payment, a service may estimate the number of performances for purposes of determining the extent of copyright liability on an “Aggregate Tuning Hour” basis, which calculates payment on the basis of 15 performances per hour. This approach alleviates a Licensee's obligation to account for and pay for each performance, including those that are less than 30 seconds in duration.

12. Methodology for Estimating the Number of Performances

Until each service can account for each performance, and is required to do so, there is a need for a methodology that will allow a service to make a reasonable estimate of the number of performances. Accordingly, the Panel proposes the following procedure:

For the period up to the effective date of the rates and terms prescribed herein, and for 30 days thereafter, the statutory licensee may estimate its total number of performances if the actual number is not available. Such estimation shall be based on multiplying the licensee's total number of Aggregate Tuning Hours by 15 performances per hour (1 performance per hour in the case of retransmissions of AM and FM radio stations reasonably classified as news, business, talk or sports stations, and 12 performances per hour in the case of all other AM and FM radio stations).

Report at 110.

The Broadcasters object to the Panel's formulation for estimating the number of performances, arguing that for many program formats, e.g., news, business, talk, or sports stations, the estimate would likely significantly overstate the use of music by these stations. Broadcasters Petition at 57. However, they do not offer an alternative methodology for calculating these performances. Moreover, a mere likelihood of overstating the values in some cases is not enough to undo the Panel's formulation.

Likewise, Webcasters argue that the 30-day cutoff period for using the methodology for estimating the number of performances is arbitrary because there is no record support for this determination. Webcasters Petition at 72. Instead, they propose allowing the Services to employ this methodology through the remainder of the current licensing period, which ends December 31, 2002, since it will be used, in any event, by most Services for purposes of calculating their liability for their past usage of the sound recordings. Id.

What is troubling about this provision is the Panel's determination to require a full accounting of each performance beginning 30 days after the effective date of the order setting the rates and terms. The Report documents that many services are not currently equipped to track or accurately account for each performance, and the Register agrees. In fact, until the issuance of final rules regarding Records of Use, there are no requirements for tracking these performances. Because the Office has yet to establish just how a service will account for its use of the sound recordings, the Register determines that the proposed timeframe for requiring a strict accounting is arbitrary. Instead, the rule shall require that a Service begin accounting for each performance in accordance with the rules and regulations regarding Records of Use 30 days after the effective date of final rules. These rules shall determine what information needs to be calculated to determine which sound recordings have been performed, how many of such performances occurred, and when and how often such information shall be collected by the Services. Meanwhile, interim rules are being promulgated that will, for the immediate future, impose more modest reporting requirements on Services.

In the meantime, for the remainder of the period covered by this proceeding (i.e., through December 31, 2002), Services may estimate the number of performances in accordance with the Panel's formulation. While this is not the perfect solution, it represents a reasonable approximation of the number of performances. And in those cases where a Service believes the formulation overestimates the use of the sound recordings, it has the option of actually counting the number of performances and calculating the royalties accordingly. Certainly, it cannot be seriously argued that a Service would be unduly burdened by undertaking this task. Conversely, if after accounting for each of the performances in the programs which are allowed to use the one performance per hour estimate, the Service finds its programming performs more sound recordings than the approximation, a Service benefits from use of the Panel's methodology.

13. Discount for Promotion and Security

RIAA proposed a 25% discount to any service that includes promotional and security features beyond those required under either the webcasting license or the ephemeral recording license. Because that proposal would exceed the scope of the terms set forth in the law, the Panel declined RIAA's invitation to provide for such discounts within the context of the statutory license. Report at 110. It is clear that the Panel may reject such a proposal, as it did here, because the statutory license does not expressly require that such a rate be established. No party contested the Panel's determination on this issue. Therefore, the Register sees no reason to question the Panel's decision.

14. Ephemeral Recordings for Services Operating Under the Section 114 License

A transmitting organization entitled to make transmissions of sound recordings under the webcasting license may also make a single ephemeral copy of each work to facilitate the transmission under an exemption in the law or it may make multiple copies of these works pursuant to a statutory license. See 17 U.S.C. 112(a) and (e), respectively. In addition to setting rates and terms for the webcasting license, the Panel in this proceeding had the responsibility for setting the rates for the ephemeral recordings. The Office combined these section 112 and section 114 proceedings because the licenses are interrelated and the beneficiaries of the license, just as the users, are in most instances the same for both the webcasting license and the ephemeral recording license. However, there is one group of users of the ephemeral recording license that is exempt from the digital performance right—services which provide transmissions to a business establishment for use by the business establishment within the normal course of its business (“business establishment services”). 17 U.S.C. 114(d)(1)(C)(iv).

During the proceeding, the Services argued that these “ephemeral” copies have no economic value apart from the value of the performance they facilitate. Webcasters Petition at 67; Broadcasters Petition at 50. In support of this position, the Services cite with approval a Copyright Office Report which stated that the Office found no rationale for “the imposition of a royalty obligation under a statutory license to make copies that have no independent economic value, and are made solely to enable another use that is permitted under a separate license.” Report at 98, citing U.S. Copyright Office, DMCA Section 104 Report at 114, fn 434 (August 2001). The Panel also contended that experts on both sides took this view. Webcasters Petition at 66, citing Jaffe W.D.T. 52-54; Tr. at 6556; Tr. at 2632 (Nagle). Had there been nothing more, the Panel might have agreed with the Services and adopted the Office's position. In construing the statute, however, the Panel found that Congress did not share the Copyright Office's view. Instead, the Panel found that Congress required that a rate be set for the making of ephemeral copies in accordance with the willing buyer/willing seller standard. Report at 98-99.

The Panel utilized the same approach in setting rates for the ephemeral recording license as it had in setting the rates for the webcasting license. Report at 104. It first examined the 26 RIAA agreements for evidence that market participants paid a fee to make ephemeral copies and how much they paid. Of the 26 agreements, fifteen did not contain any rate for the ephemeral license and did not purport to convey this right; two used a percentage of overall revenues; eight used a percentage (calculable to 10%) of the performance royalty fees paid; and one paid a flat rate per use of the license for a year (calculable to 8.8% of the performance royalty fees paid). Id. From this, the Panel identified a range of rates between 8.8% and 10% of the performance fees paid. It then chose to place significant weight on the 8.8% value because it was derived from the information in the Yahoo! agreement to which the Panel has given considerable weight throughout this proceeding. Id. However, the Panel did not rely solely on the Yahoo! agreement in this instance, choosing instead to give minimal weight to the eight other agreements that set the ephemeral rate at 10% of the performance rate, and so rounded the 8.8% value up to 9.0%. Id. Both Webcasters and Broadcasters filed Petitions to Modify in which they object to the Panel's approach to setting the ephemeral rate. They argue that the evidence supports their position that the ephemeral copies have no independent economic value apart from the performances they facilitate. In the alternative, they maintain that the value of the ephemeral copies is included in the royalty fee for the performance of the sound recording. Consequently, they contend that the appropriate way to set the ephemeral rate would be to determine the economic value of the ephemeral copies and reduce the performance rate by that amount. Webcasters Petition at 67; Broadcasters Petition at 51.

Moreover, the Services disagree with the Panel's use and analysis of the voluntary agreements for setting this rate. Specifically, they cite the lack of an ephemeral rate in 15 of the 26 agreements, even though it is clear that these recordings are necessary to effectuate a performance, as evidence of RIAA's view that the making of ephemeral copies had only a de minimis value. Broadcasters Petition at 52. For this reason, webcasters and broadcasters argue that RIAA placed little value on these copies and implicitly acknowledged that the value of these recordings is at best de minimis. They then criticize the Panel's methodology, asserting that the calculation of the ephemeral rate based upon the rates derived from the Yahoo! agreement for a per performance model, totally ignored the fact that Yahoo! agreed to pay a flat fee once it began making payments on a per performance basis, without regard to the number of performances. Webcasters Petition at 69; Broadcasters Petition at 53. Finally, Webcasters object to any use of the non-Yahoo! agreements in calculating this rate because the Panel had already found these agreements to be unreliable for purposes of setting the marketplace rates. Similarly, the Broadcasters question the Panel's reliance on eight of the agreements that it had rejected earlier as “unreliable benchmarks.” Id. at 54.

The non-CPB, noncommercial broadcasters adopt the objections to ephemeral recording rate put forth by the commercial broadcasters. Noncommercial Broadcasters Petition at 11.

On the other hand, RIAA supports the Panel's determination in general, noting that the CARP relied primarily on the Yahoo! agreement to calculate the ephemeral rate for webcasters. It maintains, however, that the Panel should have afforded the 25 voluntary agreements more weight and set the rate at 10% of the performance rate in deference to the fact that many RIAA licensees had agreed to a negotiated or effective ephemeral rate of 10%. RIAA Reply at 68. RIAA also challenges the Services' complaints in general, noting that in spite of all the objections to the Panel's determination, the Services fail to offer any evidence regarding an alternative rate.

The Panel's approach in setting the ephemeral rate was not arbitrary. It calculated the rate based on the fees Yahoo! actually paid to RIAA for the right to make ephemeral reproductions. Use of the Yahoo! agreement for this purpose was perfectly logical, and consistent with the general approach taken by the Panel in determining rates for webcasting. What causes concern, however, is the Panel's reliance, even to a small degree, on the ephemeral rates set forth in eight of the 25 voluntary agreements it had previously repudiated. Such action is arbitrary unless the Panel can offer a clear explanation for its actions. It did not do so and, in fact, it stated that its review of the 26 licenses “reveals an inconsistent, rather than a consistent, pattern.” Report at 100. Moreover, the Panel conceded that these agreements “do not represent evidence which establishes RIAA's proposed rate.” Id. at 104. Nevertheless, the Panel granted “very modest effect” to those agreements which have ephemeral rates around 10% to justify its decision to round the 8.8% effective rate up to 9%. Considering those agreements is clearly arbitrary and, consequently, to the extent the Panel gave any weight to any license agreement other than the Yahoo! agreement, it acted in an arbitrary manner. Accordingly, the rate for the ephemeral license for licensees operating under section 114 should be set at 8.8% of the performance rate.

15. Minimum Fees

The Panel established a minimum fee of $500 for each licensee for use of the webcasting license and the ephemeral recording license. These rates are in line with those negotiated by RIAA and the 26 services with which it reached an agreement. The Panel determined that RIAA would not have negotiated a minimum fee that failed to cover at least its administrative costs and the value of access to all the works up to the cost of the minimum fee. Report at 95. The adoption of the $500 minimum, however, is predicated on the adoption of a per performance rate and not a percentage-of-revenues. The Panel implied that had it decided to adopt a percentage-of-revenue model, the minimum fee would have been more substantial because the Panel would have had to consider more carefully the impact of start-up services with little revenue. Report at 95.

Because the minimum rate is calculated to cover at least the administrative costs of the copyright owners in administering the license and access to the sound recordings, the Panel applied the rate to all webcasting services and made it payable as a non-refundable advance against future royalty fees to be paid during that year, due upon the first monthly payment of each year. Moreover, the Panel offered no proration of the fee, making it due in full for any calendar year in which a service operates under the statutory license. Report at 96.

RIAA objects to the low value for the minimum fee set by the Panel because it fails to take into account the broad range of rates established in the licenses RIAA negotiated in the marketplace. Moreover, as a policy matter, RIAA contends that use of the lowest value set forth in a single agreement discourages copyright owners from adopting a low minimum fee in a single instance to accommodate special circumstances for a particular service. RIAA Petition at 44-45. Finally, RIAA faults the Panel for justifying its choice by comparing the $500 minimum fee to the amount that the Services pay the performing rights organizations (PROs) under a blanket license. RIAA rejects this rationale on two fronts. First, the minimum fee does not approximate the amounts that are paid to the PROs, and second, use of the musical works benchmark has been found by the CARP to be an inappropriate measure for establishing fees in this proceeding.

In response, Broadcasters first note that RIAA never disputed the Panel's understanding for the existence of a minimum fee, or claimed that a higher fee is necessary to achieve the stated purposes of the minimum fee. Namely, the minimum fee is meant to cover the costs of incremental licensing, i.e., the cost to the license administrator of adding another license to the system without regard to the number of performances made by the Licensee, see Webcasters PFFCL ¶ 361, and access to the entire repertoire of sound recordings. Broadcasters Reply at 12-13; Webcasters Reply at 52-53. Moreover, they claim that the minimum fee is in line with the fees paid to the performing rights organizations which can serve as a benchmark for the minimum because “they serve the same purposes that the CARP identified in setting the minimum fees for the statutory license at issue.” Broadcasters Reply at 14; Webcasters Reply at 52, 55. The Services, however, do not blindly accept the Panel's proposed fee, arguing first that the record supports a much lower minimum fee. They also strenuously object to RIAA's request for a $5,000 minimum, arguing that such a high minimum would be confiscatory for most users of the license, especially for those radio stations that play little featured music. Broadcasters Reply at 16; Webcasters Reply at 56.

None of these arguments compel the Librarian to reject the proposed $500 minimum. The Panel set a minimum rate to accomplish two purposes, and none of the parties argue that the $500 fee falls outside the “zone of reasonableness” for such rates. If anything, the fee may be viewed as too low, if one takes into account the minimum amounts paid to the performing rights organizations for the blanket license for performing musical works. Together each Service must pay, at the very least, a total of $673 to the three performing rights organizations to cover access to the musical works for use over the Internet and the incremental cost of licensing—the very purposes for which the minimum fee is being set in this proceeding.

Whether to utilize the musical works benchmark was a decision for the Panel and it chose not to do so. This approach was not arbitrary. As it had done throughout this proceeding, the Panel could choose, as it did, to rely on agreements negotiated in the marketplace between willing buyers and willing sellers. Moreover, the Panel could propose any rate consistent with the agreements so long as the proposed rate would cover costs for administering the license and access to the works. For this reason, the Panel examined the agreements offered into evidence by the RIAA and chose the lowest value that RIAA had accepted in a prior agreement. It did so because it assumed that an entity would not agree to a minimum rate that would result in a loss. Had RIAA truly believed that the $500 minimum fee was inadequate to cover at least the administrative costs and the value of access, the Panel reasoned that it would have required a higher fee. This approach is not arbitrary and, consequently, the proposed minimum fee is adopted for the period covered by this proceeding.

16. Ephemeral Recordings for Business Establishment Services (“BES”)

a. Rates for use of the statutory license. Business establishment services are well-established businesses, which have offered their services for many years. Among the established businesses in this group are AEI Music Network, Inc., DMX Music, Inc., Muzak, Inc., PlayNetwork, Inc. and Radio Programming and Management Inc. Two of the old guard, AEI and DMX, and one new service, Music Choice, participated in this proceeding. At an early stage of this proceeding, but after filing a direct case, Music Choice withdrew from the proceeding.

Of the services offered by AEI and DMX only those services that transmit musical programs to their customers via cable or satellite in a digital format are eligible for the ephemeral recording license. The Panel referred to this aspect of the business as the “broadcast model” of the service. Through this process, these services make hundreds of thousands, if not millions, of copies of the sound recordings. The law allows these services to perform sound recordings publicly by means of a digital transmission under an exemption in section 114. However, Congress did not exempt these services from copyright liability when making copies of these works in the normal course of their business. Rather, Congress created a statutory license to cover the making of ephemeral recordings by these services. In its proposed findings of fact and conclusions of law, DMX and AEI proposed a flat fee of $10,000 per year for each company for the making of buffer and cache copies, but argued in the alternative for a zero rate. See DMX/AEI PFFCL ¶ 44. In support of the alternative position, DMX/AEI argued that Congress had only envisioned a minimal rate to compensate the copyright owners for the use of ephemeral copies. It also cited the Copyright Office's Section 104 DMCA Study for the proposition that ephemeral recordings have no independent economic value apart from its use to facilitate transmissions. However, as RIAA points out, these businesses have always paid for such copies. Report at 115-116, citing RIAA Reply to DMX/AEI PFFCL ¶¶ 8-12. RIAA asked that rate be set at 10% of gross revenues with a minimum fee of $50,000 a year and asked the Panel to refrain from setting rates tailored to the needs of specific companies. RIAA made the later request because AEI/DMX asserted that its digital database is already covered by preexisting licenses and therefore, it does not need an ephemeral license in order to make these phonorecords. Consequently, AEI/DMX asked the Panel to set a rate to cover only the cache and buffer copies it needed to facilitate its transmissions and to exclude the value of the database copies when setting the rate for the ephemeral license. In fact, AEI/DMX contends that it was arbitrary for the Panel to set a rate “for all ephemeral copies which may be utilized in the operation of a broadcast service” when it had received evidence for setting a rate only for buffer and cache copies. DMX/AEI Petition at 4. It also maintains that the statute contemplates that the Panel set rates according to the needs and desires of the parties. Id. at 8-10.

RIAA disagreed with this approach, asking the panel to establish a technology-neutral rate to cover the making of all copies that a business establishment service may need to make under the license. It also proposed that the CARP rely on license agreements between the copyright owners and Business Establishment Services when fashioning the appropriate rate and not the 26 voluntary licenses considered when setting the webcasting rates.

As an initial matter, the Panel had first to decide which copies and how many are covered by the ephemeral recording license. This is a necessary step in the process, because the statutory license allows a transmitting organization to make and retain no more than a single phonorecord of a sound recording, except as provided “under the terms and conditions as negotiated or arbitrated under the statutory license.” Section-by-section analysis of the H.R. 2281 as passed by the United States House of Representatives on August 4, 1998, Committee Print, Serial No. 6, 105th Cong., 2d Sess., p. 61.

Thus, the Panel considered and ultimately rejected DMX/AEI's request for a rate that only covered certain types of ephemeral copies. It did so in large part because it determined that Congress had “intended to create blanket licenses which would afford each licensee all the rights necessary to operate such a service,” and noted that in this case, that would include “the right to make any and all ephemeral copies utilized in a broadcast background music service.” Report at 118. This interpretation of the law is consistent with the purpose of the section 112 license.

In creating the ephemeral recording license, Congress sought to provide a way for any licensee or business establishment service to clear all the reproduction rights involved in making digital transmissions of sound recordings under section 114. Congress “intended [this provision] to facilitate efficient transmission technologies, such as the use of phonorecords encoded for optimal performance at different transmission rates or use of different software programs to receive the transmissions.” H.R. Rep. No. 105-796, at 90 (1998). These copies are known as “ephemeral recordings.” “The term “ephemeral recording” is a term of art referring to certain phonorecords made for the purpose of facilitating certain transmissions of sound recordings, the reproduction of which phonorecords is privileged by the provisions of section 112.” Id. Because the purpose of the license is to facilitate a lawful transmission of a sound recording under a statutory license or exemption, it would appear that the license covers not only the first reproduction of the sound recording on a company's server, but also all intermediate copies needed to facilitate the digital transmission of the sound recording.

The mere fact that the license covers different ephemeral recordings that may be catalogued in different ways does not mean that a separate rate must be set for each category. Had the record supported different rates for different categories of ephemeral recordings, or for different types of business establishment services, it is conceivable that the Panel might have chosen to differentiate among these categories or types of businesses by assigning different rates to each one. See also Order (dated July 16, 2001) (advising Panel that it could set different rates for different business models, provided that the record supported such a decision). Whether such an approach would have been arbitrary would depend upon the findings of the Panel in light of the record evidence and, more importantly, upon whether the proposed rates covered the making of all ephemeral copies needed to facilitate the digital transmission of a sound recording under the section 114 business to business exemption.

The section 112 license is without question for the benefit of all services operating under the business to business exemption and not just DMX/AEI. A rate tailored only to meet the specific needs of a single service would by its very nature be arbitrary if the rate failed to cover the entire scope of the license. The fact that DMX/AEI has chosen to license the copies in its database through a private agreement and use the statutory license to cover the remaining ephemeral copies would not relieve the Panel of its responsibility to set rates for all ephemeral copies which fall within the scope of the license, including those copies in a DMCA compliant database. Other business establishment services using a DMCA-compliant database exist and may choose to meet their copyright liability by operating under the statutory license. See RIAA reply at 18; Report at 116. It is without question that such a service may take advantage of the statutory license without participating in a CARP proceeding.

Once these rates are set, a Service can either operate entirely under the statutory license or, alternatively, the Service may choose to make some ephemeral copies under the statutory license and others under a private agreement. These choices, however, have no bearing on the responsibility of the Panel to establish a rate, or a schedule of rates, that would allow a Service to utilize the license to the full extent of the law.

In fashioning the rate, the Panel considered the arguments put forth by the parties and ultimately rejected DMX/AEI's basic premise that Congress had contemplated a de minimis rate to compensate for “leakage” (use of ephemeral copies to make phonorecords for sale) and, its interpretation of what it characterized as the Copyright Office's view that such copies have no independent economic value. This decision was reached after examining the statute and its legislative history and finding nothing that directly supported the “leakage” theory. Moreover, the Panel had already determined that its responsibility was not to give effect to the Copyright Office's view on how the law should change. Instead, it determined that its duty was “to follow the current Congressional mandate set forth in section 112(e)(4) and determine a separate rate for ephemeral copies' based upon the willing buyer/willing seller standard. Report at 98-99. Thus, the Panel rejected AEI/DMX's proposal to set a low rate based upon its finding that these entities have always paid substantial royalties to record companies in exchange for the use of its complete catalogue. Report at 119.

In any case, the starting point for setting the rates for the ephemeral recording license as it applies to business establishment services is the statute. It provides that, as with the rates for the webcasting license, the rates should be those that “most clearly represent the fees that would have been negotiated in the marketplace between a willing buyer and a willing seller.” 17 U.S.C. 112 (e)(4). Thus, the Panel turned to actual agreements that have been negotiated in the marketplace to discover how the market values these rights. As discussed previously, the use of rates negotiated in the marketplace is not arbitrary. It eliminates the need to try to value specific economic, competitive, and programming factors because the parties would have already accounted for these considerations during the negotiation process and their impact would be reflected in the negotiated rates.

Both sides seem to agree with the Panel's approach. RIAA had no complaint with the Panel's use of voluntarily negotiated licenses in setting the ephemeral rates for business establishment services. Moreover, DMX/AEI's own counsel acknowledged that marketplace agreements were appropriate benchmarks for establishing the rates for the rate for the section 112 license and conceded that the agreements relied upon were worthy of consideration. Tr. 9577-78 (Sept. 12, 2001). Nevertheless, DMX/AEI did argue that the proposed rate constitutes an undue financial burden that thwarts Congress' intent to facilitate the adoption of new technologies. DMX/AEI Petition at 11.

The question is which agreements should be considered when setting the rates for the ephemeral reproductions. Having found that the business establishment services offer a completely different type of service from webcasting, the Panel rejected DMX/AEI's invitation to use the ephemeral rates negotiated by the webcasters. Report at 121. Instead, the Panel opted to use the license agreements that had been negotiated between individual record companies and background music services as a benchmark for setting the relevant section 112 rates even though, in some instances, the license conveyed some rights to the licensee beyond the reproduction and distribution of the sound recording. The Panel was not troubled by this observation, however, because it found that in all cases the right to copy and distribute the works was by far the most important right for which the licensee paid royalties. Moreover, it noted that the rates did not fluctuate through the year even when a service altered its method for delivering music. Thus, the Panel used the rates reflected in these licenses to establish a range of rates (10-15% of gross proceeds) for consideration. See Report at 117; see e.g., RIAA Reply to AEI/DMX at 2. From this data, it found that “background music companies and record companies would agree to a royalty of at least 10% of gross proceeds,” and set the rate accordingly. Report at 126.

RIAA agrees with the Panel's approach, and that it was appropriate for the Panel not to consider contracts for ephemerals made in the course of webcasting because these businesses are not comparable with Business Establishment Services. They serve different customers and operate under different economic business models with different delivery methods. For example, Business Establishment Services make reproductions of sound recordings and deliver them via cable or satellite for use by the establishment for the enjoyment of their customers. These differences are further underscored by transactions in the marketplace. RIAA notes that within a single license with one business entity, it negotiated a separate rate for webcasting ephemeral copies and a separate rate for ephemeral copies used by the Business Establishment Service. RIAA reply at 24-25. The fact that RIAA negotiated separate rates for the making of ephemeral recordings for different services supports a finding that the businesses are not comparable. Therefore, it was not arbitrary for the Panel to decline to consider the ephemeral rates set forth in the licenses between the webcasters and the record companies when establishing a rate for Business Establishment Services.

Moreover, an examination of the record evidence clearly shows that the 10% of revenues rate set by the Panel is not an arbitrary figure. RIAA Exhibits 9 DR, 10 DR, 11 DR, 12 DR, 13 DR, 14 DR, 26 DR, 27 DR, 28 DR, 60-A DR, 66 DR-X, Knittel Rebuttal Ex. 22; Knittel W.D.T. 14-15. It represents the low end of the range of rates set forth in the agreements between the major record labels and Business Establishment Services. The fact that two agreements, negotiated during a period of uncertainty whether there was a legal obligation to pay anything for the satellite transmissions they covered, reflect a lower rate does not change the outcome. See Report at 124. As RIAA points out, the rate in one of these agreements was reset at a substantially higher rate once the initial contract with the lower rate expired. RIAA Reply to AEI/DMX at 25, fn 25. Nor is there any reason to reject the Panel's determination, as DMX/AEI contends, because the Panel failed to adjust for the promotional value to the record companies or bring these rates into line with those set for Subscription Services in the previous proceeding. As the Panel stated on several occasions, it is unnecessary to adjust a marketplace-negotiated rate for the promotional value that flows to the record companies because that benefit would already be reflected in the contract price, if it were important to the parties.

Likewise, DMX/AEI's second premise for rejecting the Panel's determination must also be discarded. It argued that the Panel set an arbitrarily high rate for Business Establishment Services when compared to the rate set for Subscription Services in an earlier proceeding. DMX/AEI Petition at 19-20. As discussed in a previous section, see section IV.3, rates set for Subscription Services in a prior proceeding are just not comparable to rates under consideration in this proceeding. Marketplace rates for making reproductions of sound recordings for use by a Business Establishment Service have no established relationship to rates set under a totally different standard for the public performance of sound recordings by Subscription Services. There is no established nexus between the industries, the marketplaces in which they operate, or the rights for which the rates are set. To make any adjustments to the ephemeral rate based on the rate for the digital performance rate adopted for the Subscription Services in a previous proceeding would itself be patently arbitrary.

b. Minimum fee. The statute also requires the Panel to set a minimum fee for use of the license. Using the same licenses, it determined that the minimum fee should be $500 a year based on its observation that most, although not all, willing buyers have not agreed to a fee approaching RIAA's proposed rate of $50,000 a year and that some agreements include no minimum fee at all. Because there is no discernable trend in the licenses, the Panel chose to adopt the same fee it proposed for the webcasting licenses because it is calculated to cover at least the administrative costs of the license.

RIAA argues that a $500 minimum is too low and contradicts the record evidence, citing the existence of significantly higher rates in many of the industry agreements and the lack of any agreement with a minimum as low as $500. RIAA Petition at 46-47. RIAA further contends that the CARP by its own reasoning should set a significantly higher minimum fee where, as here, the ephemeral rate is based on a percentage-of-revenue model. Id. at 49. The Copyright Owners are concerned that a low minimum rate will increase “the risk that a service, especially a new one, will make a large number of ephemeral copies and not generate revenues, effectively giving the service a blanket license for free.” Id. Consequently, the Copyright Owners ask the Librarian to adopt their proposal and set the minimum fee for use of the ephemeral license at rate no lower than $50,000.

DMX/AEI objects to RIAA's request for a higher minimum fee. It maintains that RIAA requested rate is inconsistent with record evidence, which establishes that either DMX/AEI currently pays [material redacted subject to a protective order] in its direct licensing agreements with the major labels for On-Premises services or that it is disproportionately high when compared with the minimum fees paid by other members of the background music service industry. DMX/AEI Reply at 7. Accordingly, AEI/DMX urges the Librarian not to entertain the RIAA's request.

An examination of the relevant agreements reveals that almost all of these agreements have a substantial minimum fee for the making of ephemeral recordings and that all of those minimum fees are considerably greater than the $500 minimum proposed by the CARP. Consequently, the Panel's decision to adopt a $500 minimum fee when no contract considered by the Panel contained a minimum fee as low as $500 is arbitrary. The minimum fees in the agreements before the CARP were by and large significantly higher than the $500 fee proposed by the CARP and should have served as the guiding principle in setting the minimum fee for the Business Establishment Services, especially in light of the Panel's earlier observation that a percentage of revenue fee requires the establishment of a substantial minimum fee to offset the risk that a start-up Service with little revenue could operate without paying adequate royalty fees for use of the license. Moreover, RIAA notes that each contract before the CARP was between a Business Establishment Service and a single record label. It then makes the argument that “[i]f a business establishment service is willing to pay a minimum fee [significantly higher than the minimum fee proposed by the Register] for access to just one label's sound recordings, the value of the blanket license to all copyrighted recordings must be higher.” RIAA Petition at 46. Based on this evidence, the Panel should have set the minimum fee for the section 112 license as it applies to Business Service Establishments at a significantly higher level, and it was arbitrary not to have done so.

The Register notes that minimum fees have been as low as $5,000 and as high as the $50,000 minimum proposed by RIAA. The purposes of the minimum fee, however, are to cover the costs of administration and insure an adequate return to the copyright owners based upon the value of the right with respect to the overall fee for use of the license. For these reasons, the Register proposes a minimum fee of $10,000 per Licensee. The fee is at the low end of the range of negotiated minimum fees and is in line with DMX/AEI's own valuation of the license at $10,000 per year. Admittedly this fee appears high when compared with the minimum fee for the eligible nonsubscription services, but it serves to balance the risk associated with setting a statutory fee based upon a percentage of revenues instead of a fee that would charge a specific fee for each reproduction.

17. Effective Period for Proposed Rates

The rates and terms proposed by the parties were the same for each time period under consideration by the Panel. Consequently, the Panel proposed, and the parties agreed, that the same rates and terms would apply to both periods: (1) October 28, 1998 (the effective date of the DMCA) through December 31, 2000; and (2) January 1, 2001, through December 31, 2002. The Register finds that it was not arbitrary for the Panel to propose the same rates and terms for both periods under consideration.

B. Terms

Sections 112(e)(4) and 114((f)(2)(B) require that the CARP propose and the Librarian adopt terms for administering payment for the two statutory licenses. The Panel stated that, as with rates, the standard for setting these terms is what the willing seller and the willing buyer would have negotiated in the marketplace. The Panel did not interpret the standard to include necessarily setting terms that “represent the optimum alternative from the standpoint of administrative convenience and workability.” It reasoned that such considerations were “not part of the governing standard for the Panel, nor [were they] a matter on which [the Panel] would have either record evidence or institutional expertise.” Consequently, the Panel made no determination pertaining to administrative efficiency, choosing instead to defer to the expertise of the Librarian. Report at 129.

For the most part, the terms proposed by the Panel are those to which all parties to the CARP proceeding have agreed in negotiations. For this reason, the Panel accepted all terms on which the parties agreed, finding that where there was agreement, the terms meet the statutory standard under which these terms must be set. Moreover, the Panel found that there was evidence in the record to support adoption of most of these terms.

The Register is skeptical of the proposition that terms negotiated by parties in the context of a CARP proceeding are necessarily evidence of terms that a willing buyer and a willing seller would have negotiated in the marketplace. Especially when those terms relate to administration of the receipt and distribution of royalties by collectives that are artificial (but necessary) creations of the statutory license process, rather than entities likely to be created in an agreement between a copyright owner and a licensee, the fiction that those terms reflect the reality of the marketplace is difficult to accept.

Not all of the terms recommended by the Panel are terms that the Register would have adopted if her task were to determine the most reasonable terms governing payment of royalties. However, in light of the standard of review, the Register recommends accepting the terms adopted by the Panel except in the relatively few instances where the Panel's decision was either arbitrary or not feasible. See Report at 129 (“we must defer to the expertise of the Librarian the final evaluation of the administrative feasibility of terms which willing buyers and willing sellers would agree to in marketplace negotiations”). The discussion that follows addresses, first, the terms recommended by the Panel that one or more parties have asked the Librarian to reject. Following that discussion, the Register discusses those terms recommended by the Panel that, although they are acceptable to the parties, she proposes to modify or reject, because they are arbitrary or contrary to law.

1. Disputed Terms

The parties were unable to reach a consensus with respect to two issues: (1) The incorporation of specific definitions for the terms, “Affiliated,” “AM/FM streaming,” “Broadcaster,” and “Non-Public;” and (2) the designation of an agent for unaffiliated copyright owners.

a. Definitions. The Panel carefully considered the utility of incorporating the proposed terms for Affiliated,” “AM/FM streaming,” “Broadcaster,” and “Non-Public.” It decided to reject the webcasters” request to adopt the disputed terms and definitions, noting that the terms were not applicable to the rate structure ultimately adopted by the Panel. The Parties have filed no objection on this point and the Register finds no reason to include a definition of these terms in the regulations.

Notwithstanding the Panel's decision as to these terms, it did incorporate other terms that were necessary for the administration of the license. The proposed definitions for these additional terms are based upon submissions from the parties made at the Panel's request. See, Services' Submission of Definitions; Proposed Definitions of the Recording Industry Association of America, Inc. (Feb. 12, 2002). Again, no party has filed an objection to the Panel's decision to propose additional terms the purpose of which is make the regulatory framework clearer and more functional.

b. Designated Agent for Unaffiliated Copyright Owners. Read literally, section 114 appears to require that Services pay the statutory royalties directly to each Copyright Owner. As a practical matter, it would be impractical for a Service to identify, locate and pay each individual Copyright Owner whose works it performed. As a result, in the administration of the predecessor statutory license for noninteractive subscription services, a Collective was appointed to receive and distribute all royalties. The RIAA has served as the Collective for the nonsubscription services.

In this proceeding, the Parties proposed and the CARP agreed to a modification of the single-collective model. Licensees making transmissions of a public performance of a sound recording pursuant to the statutory license in section 114 and/or making ephemeral recordings of these works under the statutory license in section 112(e) would make all payments owed under these licenses to the designated “Receiving Agent.” The Receiving Agent would then make further distribution of the royalty fees to the two Designated Agents who would then distribute the royalty fees among the Copyright Owners and Performers in accordance with the methodology set forth in the regulations.

The CARP accepted the proposal of the parties to designate a single Receiving Agent, SoundExchange, in order to maximize administrative efficiencies for the Copyright Owners and Performers, on the one hand, and Licensees, on the other. SoundExchange is a nonprofit organization formed by RIAA for the purpose of administering the sections 112 and 114 statutory licenses. It has over 280 member companies, affiliated with more than 2,000 record labels accounting for over 90% of the sound recordings lawfully sold in the United States. W.D.T. at 4 (Rosen). SoundExchange is governed by a board comprised of representatives of Copyright Owners and Performers and, under a recent reorganization, the Copyright Owners and artists representatives will have equal control over the SoundExchange Board. AFM/AFTRA PFFCL ¶ 6.

In addition to its role as a Receiving Agent, the CARP accepted the Parties' proposal that both SoundExchange and Royalty Logic, Inc. (“RLI”) serve as Designated Agents. RLI is a for profit subsidiary of Music Reports, Inc. and was created to offer a competitive alternative to SoundExchange. W.D.T. at 2 (Gertz). The purpose of having two designated agents is to provide Copyright Owners with the option of electing to receive their royalty distribution from either SoundExchange or RLI. The Receiving Agent will allocate royalties to the two Designated Agents based on the Copyright Owner's designation.

However, the parties could not agree on which Designated Agent would distribute funds to Copyright Owners who failed to make an election. The Webcasters proposed that RLI be named the agent for unaffiliated Copyright Owners, but Copyright Owners and Performers asked the Panel to designate SoundExchange as the agent for those copyright owners.

After carefully considering the role of the Designated Agent for unaffiliated copyright owners and the record evidence, the Panel made a determination to name SoundExchange as the Designated Agent for those copyright owners who fail to expressly designate either SoundExchange or RLI as their agent to receive and distribute royalties on their behalf. The primary reason for this designation was the preference expressed by the Copyright Owners and the Performers. The Panel reasoned that the Services had no real stake in deciding this issue because their responsibilities and direct interest end with the payment of the royalty fees to the Receiving Agent. Moreover, AFM and AFTRA, which represent artists who are among the beneficiaries of the license, expressed a strong preference for the designation of SoundExchange as the agent in these instances. The Copyright Owners made this choice based on the non-profit status of SoundExchange, its experience with royalty payments, and the fact that SoundExchange has agreed to a reorganization that gives artists substantial control over its operations. The Panel agreed with the reasons articulated by the Copyright Owners and Performers and found that the probable outcome of a marketplace negotiation would have been the selection of SoundExchange.

Broadcasters contest the Panel's decision to designate SoundExchange as the agent for unaffiliated copyright owners. They assert that there is no record evidence to support the Panel's observation that this was the inevitable outcome of marketplace negotiations, in spite of the actual requests made by Copyright Owners who participated in this proceeding. Broadcasters Petition at 59-60.

The Copyright Owners and Performers disagree, and assert that unlike the Licensees whose only concern is whom to pay and when, copyright owners and performers have a vital interest in how their royalty fees are collected and distributed and have expressed a strong preference for SoundExchange as the designated agent. See RIAA Reply at 81; AFM/AFTRA Reply at 2. Certainly, Performers believe that SoundExchange will make fair and equitable distributions and not deduct additional costs beyond those necessary costs incurred to effectuate a distribution. AFM/AFTRA Reply at 2-3 (“SoundExchange is subject to the joint and equal control of copyright owner and performer representatives with an interest in maintaining an efficient operation that will distribute the maximum possible license fees, that SoundExchange is a nonprofit organization so that no copyright owner's or artist's royalty share will be diminished by anything other than necessary distribution costs, and that SoundExchange is experienced and has demonstrated its commitment to identifying, finding and paying performers during its distribution of Section 114 and 112 subscription service statutory license fees.”); see also RIAA Reply at 83.

The CARP's decision to designate SoundExchange as the agent for unaffiliated copyright owners is fully supported by the record evidence and, consequently, it is not arbitrary. First, the fact that Copyright Owners and Performers commend SoundExchange to the Panel is direct evidence of their preference for a non-profit organization that has already invested heavily in a system designed to locate and pay Copyright owners and Performers. It would be arbitrary to ignore their wishes where, in fact, the alternative agent represents primarily broadcasters, television stations, and other Licensees—not Licensors. See AFM/AFTRA PFFCL concerning terms ¶ 13. Second, SoundExchange is a non-profit collective that will deduct only necessary distribution costs. On the other hand, RLI, the entity competing for the agency designation, is a for-profit organization whose acknowledged goal is to make a profit. In fact, RLI has suggested that it needs the designation from the CARP in order to generate enough revenues to make it worthwhile to take on the role of an agent for purposes of making distributions of statutory license royalty fees. See Services Proposed Findings (12/18/01) at ¶ 16. In addition, RLI has been unable to say just how much it expects to deduct as reasonable costs, making it impossible to ascertain whether designation of RLI would be in the best interest of the unaffiliated copyright owners. Third, Performers and Copyright Owners have a direct governance role in the operation of SoundExchange, thereby insuring their interests are not neglected or overshadowed by the interests of the agent. AFM/AFTRA Reply at 4; AFM/AFTRA PFFCL concerning terms ¶ 6. Performers have expressed strong concerns about the designation of an agent who has no mechanism or apparent interest in providing the Copyright Owners and Performers with a means to voice their concerns. See AFM/AFTRA PFFCL concerning terms ¶ 9 (noting that designation of RLI as the agent for unaffiliated copyright owners would have the undesirable effect of forcing these non-members “into an agency relationship with an entity that not only is not governed by Copyright Owners and Performers, but also is not even required to obtain their guidance and input regarding policies, procedures or distribution methodologies.” ).

For all the foregoing reasons, the Register concludes that the CARP was not arbitrary in designating SoundExchange as the agent for unaffiliated copyright owners. Of the four factors considered by the Panel, each weighs in favor of SoundExchange. Of course, any Copyright Owner or Performer can affirmatively choose RLI to act on its behalf as a Designated Agent.

c. Gross proceeds. As discussed earlier, the Panel proposed the adoption of a rate for Business Establishment Services making ephemeral recordings under section 112 at 10% of gross proceeds. The Panel recognized the necessity of also formulating a definition of “gross proceeds” in order to make the rate workable. To meet this need, it opted to incorporate, with minor modifications to accommodate the section 112 license, the definition used in many of the background music agreements even though the definition is less than clear on its face as to what constitutes gross proceeds. The lack of specificity, however, did not trouble the Panel because it expected the parties to adopt the understandings within the industry developed during the normal course of dealings.

RIAA does not share the Panel's view. It objects to the proposed definition of “gross proceeds,” arguing that the provision fails utterly to define the term in any meaningful way. It also contends that it is arbitrary to rely on industry practices to flesh out the industry's understanding of the term when no record evidence exists about these practices. To remedy this situation, RIAA proposes that the Librarian adopt the definition of “gross proceeds” for a Business Establishment Service that is set forth in the agreement between SoundExchange and MusicMusicMusic (“MMM”). RIAA Exhibit No. 60A. RIAA asserts that this is the only record evidence on this point. RIAA petition at 52-54.

DMX/AEI rejects RIAA's suggestion that the Librarian adopt a definition from an agreement with MMM, “an unsophisticated licensee, who by its own admission is unlikely to pay any significant royalties pursuant to the agreement.” DMX/AEI Reply at 3. RIAA's proposed definition of “gross proceeds” would include fees generated by equipment rental, maintenance services, advertising of all kinds, and revenues payable to a licensee from any source in connection with the licensee's background music service. Id. at 5. DMX/AEI argues that such a definition is utterly contrary to the normal practice of using proceeds derived solely from the delivery of copyrighted sound recordings to business establishments.

As a general principle, terms pertaining to a statutory license must be defined with specificity. At first blush, the proposed definition of “gross proceeds” does not appear to meet this standard, merely reciting that a Business Establishment Service must pay a sum equal to ten percent of the licensee's gross proceeds derived from use of the musical programs that are attributable to copyrighted recordings. However, record evidence suggests the definition may be as simple as the CARP's characterization of the term. Barry Knittel, in discussing the promotional funds established for the benefit of the record companies from gross proceeds, stated that the money placed into these accounts comes from the company's gross revenues, and that these revenues are generated from all the billings for music. Tr. 8384 (Knittel). This statement suggests that the determination of what constitutes “gross revenues” is not a mystery and that it is merely the amount the Business Establishment Services receive from their customers for use of the music. This approach, however, does not necessarily appear to capture in-kind payments of goods, free advertising or other similar payments for use of the license. See RIAA Petition at 54.

Consequently, the Register proposes to expand on the CARP's approach and adopt a definition of “gross proceeds” which clarifies that “gross proceeds” shall include all fees and payments from any source, including those made in kind, derived from the use of copyrighted sound recordings to facilitate the transmission of the sound recording pursuant to the section 112 license. See RIAA Exhibit No. 60A DR. (Second Webcasting Performance and Webcasting and Business Establishment Ephemeral Recording License Agreement). The Register finds it necessary to expand upon the proposed definition to avoid any confusion on this point and not as a means to capture additional revenue streams not contemplated by the Panel or by the parties to such agreements. Because the record fails to enumerate the types of revenue that may be received in kind, the Register finds it unwise to include even an illustrative list when there is little evidence of what specific types of revenues should be considered in the calculation of “gross proceeds.” Thus, the definition of “gross proceeds” shall be as follows:

“Gross proceeds” shall mean all fees and payments, including those made in kind, received from any source before, during or after the License term which are derived from the use of copyrighted sound recordings pursuant to 17 U.S.C. 112(e) for the sole purpose of facilitating a transmission to the public a performance of a sound recording under the limitation on the exclusive rights specified in section 114(d)(1)(c)(iv).

2. Terms Not Disputed by the Parties

a. Limitation of Liability. One of the terms proposed by the Parties and adopted by the CARP was that “A Designated Agent shall have no liability for payments made in accordance with this subsection with respect to disputes between or among recipients.” The Parties explained that the purpose of this provision was to “mak[e] clear that so long as a Designated Agent complies with the requirements adopted by the Copyright Office for distributing royalties, then a beneficiary of statutory royalties cannot sue such Designated Agent for payments made in accordance with Copyright Office regulations. Any dispute among recipients should be resolved among themselves.”

The Register understands the desire of SoundExchange and RLI to insulate themselves from liability in cases where Copyright Owners or Performers dispute the Designated Agent's allocation of royalties. The Copyright Office's experience with distribution proceedings for the statutory licenses for which royalties are initially paid to the Copyright Office provides ample evidence that individual copyright owners and performers often believe they are being paid less than their fair share of statutory license royalties, and it is natural for a Designated Agent to wish to avoid having to defend against such claims.

Moreover, as has become apparent in the course of the pending rulemaking proceeding relating to notice and recordkeeping for the use of sound recordings under the statutory licenses, the information that Licensees will be providing to the Designated Agents about which (and how many) sound recordings they have performed will be far from perfect, and the Designated Agents necessarily will have to make difficult judgments in determining how to allocate royalties. If the Designated Agents had comprehensive information identifying each and every performance transmitted by a Licensee, and each and every Copyright Owner and Performer for each performance, in theory they could pay each Copyright Owner and Performer his or her precise share of royalties. In the real world—or at least for the remainder of the period for which this proceeding is setting rates and terms—some Copyright Owners and Performers inevitably will receive less than their precise share of the royalty pool, and others will receive more than their precise share. The Designated Agents should not be held to an impossibly high standard of care.

Unfortunately, neither the CARP nor the Librarian have the power to excuse a Designated Agent (or, for that matter, anyone else) from liability for a breach of a legal obligation. If a Designated Agent has in fact wrongfully withheld or underpaid royalties to a Copyright Owner or Performer, the law may provide a remedy to the Copyright Owner or Performer.

Although the Librarian cannot excuse the Designated Agents from potential liability, he can adopt terms that provide a mechanism that will make claims by disgruntled Copyright Owners or Performers less likely, or at least less viable. The Register therefore recommends that in place of the ultra vires provision excusing the Designated Agents from any liability, the Librarian provide that the Designated Agents must submit to the Copyright Office a detailed description of their methodology for distributing royalty payments to nonmembers. This information will be made available to the public, and any Copyright Owner or Performer who believes the methodology is unfair will have an opportunity to raise an objection with the Designated Agent prior to the distribution, thereby giving the Designated Agent the opportunity to address the problem before the Copyright Owner or Performer has suffered any alleged harm. This provision is modeled on a provision proposed by the parties to the previous CARP proceeding to establish rates and terms for noninteractive subscription services under section 114. See proposed 37 CFR 260.3(e), in Notice of Proposed Rulemaking, Determination of Reasonable Rates and Terms for the Public Performance of Sound Recordings, 66 FR 38226, 38228 (July 23, 2001).

The Register also proposes that the Librarian adopt a term that provides a Designated Agent with an optional mechanism pursuant to which the Designated Agent may request that the Register provide a written opinion stating whether the Agent's methodology for distributing royalty payments to nonmembers meets the requirements of the terms for distribution set forth in the implementing regulations. Although such an opinion by the Register would not be binding on a court evaluating a claim against a Designated Agent, it can be assumed that a court would find the opinion of the Register persuasive.

The Register anticipates that under this scheme, a Designated Agent that acts conscientiously and in good faith in the distribution of royalties will not be found liable to a Copyright Owner or Performer who is dissatisfied with his or her share of the distribution.

b. Deductions from Royalties for Designated Agent's Costs. The parties had proposed, and the CARP agreed, that Designated Agents be permitted to deduct from the royalties paid to Copyright Owners and Performers “reasonable costs incurred in the licensing, collection and distribution of the royalties paid by Licensees * * * and a reasonable charge for administration.” The Register recommends that the provision permitting deductions for costs incurred in licensing be removed from this provision. See § 261.4(i). Although a Designated Agent may happen to engage in licensing activities, licensing per se is not among the responsibilities of a Designated Agent under the terms of the statutory license. The purpose of the Designated Agent is to receive and distribute the statutory royalty fees. There is no justification for permitting a Designated Agent to deduct costs incurred in licensing activity from the statutory royalties, and the CARP's acquiescence in this term was therefore arbitrary.

There was also a suggestion in testimony presented to the CARP that it would be proper for a Designated Agent to deduct from statutory royalties its costs incurred as a participant in a CARP proceeding. Tr. 11891-11893 (Williams). Nothing in § 261.4(i), including the references to “reasonable costs incurred in the collection and distribution of the royalties paid by Licensees,” can properly be construed as permitting a Designated Agent to deduct from the royalty pool any costs of participating in a CARP proceeding. Such activity is beyond the scope of collection and distribution of royalties. Of course, Copyright Owners and Performers may enter into agreements with a Designated Agent permitting such deductions, but a Designated Agent may not make such deductions from royalties due to unaffiliated Copyright Owners and Performers or those who have simply designated a Designated Agent without specifically agreeing to permit such deductions.

c. Ephemeral Recording. The Register recommends that a definition of “Ephemeral Recording” be added to the definitions. This definition incorporates by reference the requirements set forth in section 112(e).

In a related provision, the Register has harmonized the language of §§ 261.3(b) and (c) and makes clear that beneficiaries of the statutory license for ephemeral recordings may make any number of ephemeral recordings so long as they are made for the sole purpose of facilitating the statutory licensees permitted transmissions of performances of sound recordings. The regulatory text proposed by the parties and accepted by the Panel provided that for Business Establishment Services, the section 112 royalty shall be paid “[f]or the making of unlimited numbers of ephemeral recordings in the operation of broadcast services pursuant to the Business Establishment exemption contained in 17 U.S.C. 114(d)(1)(C)(iv),” (emphasis added), but that for webcasters, the section 112 royalty shall be paid “[f]or the making of all ephemeral recordings required to facilitate their internet transmissions.”

A literal reading of section 112(e) might lead to the conclusion that the ephemeral recording statutory license permits only the making of a single ephemeral recording, but the statute qualifies that provision by stating “(unless the terms and conditions of the statutory license allow for more),” and the legislative history makes clear that the terms established by the Librarian in this proceeding may include terms permitting the making of additional ephemeral recordings. H.R.Rep. 105-796, at 89. Therefore, it is appropriate that the terms make clear that statutory licensees may make more than one ephemeral recording to accomplish the purposes of the statutory license.

The reference to “all” ephemeral recordings “required” to facilitate webcasters” transmissions, and the reference to “unlimited” recordings for Business Establishment Services” “operation”, are arguably inconsistent with each other and somewhat ambiguous. To clarify that the scope of the section 112 statutory license is similar for both types of service, and to more accurately reflect the appropriate scope of that license, the Register recommends that the regulatory language provide, in the case of webcasters, “[f]or the making of any number of ephemeral recordings to facilitate the Internet transmission of a sound recording,” and in the case of Business Establishment Services, “[f]or the making of any number of ephemeral recordings in the operation of a service pursuant to the Business Establishment exemption.” (Emphasis added).

d. Definition of “Listener”. The definitions of “Aggregate Tuning Hours” and “Performance” both include references to a “listener” or to “listeners.” It is not clear from the text of these definitions whether each person who is hearing a performance is a “listener” even if all the persons hearing the performance are listening to the same machine or device (e.g., two or more persons listening to a performance rendered on a single computer). Clearly the intent is that all persons listening to a performance on a single machine or device constitute, collectively, a single “listener,” because “listener” is used here to assist in defining what constitutes a single performance. Indeed, it would be difficult to implement an interpretation that counted all individuals in such circumstances as separate “listeners.” Accordingly, the Register recommends including a definition that provides that if more than one person are listening to a transmission made to a single machine or device, those persons collectively constitute a single listener.

e. Timing of Payment by Receiving Agent to Designated Agent. The terms proposed by the Parties and accepted by the CARP included a provision requiring that the Receiving Agent pay a Designated Agent its share of any royalty payments received from a Licensee within 20 days after the day on which the Licensee's payment is due. While the Register recognizes that such a provision would, in principle, be unobjectionable, she concludes that under current conditions it is administratively unfeasible.

As the parties recognized in their commentary on this provision, “The parties do not know either the payment methodology that will be used to calculate royalties or the types of information that will be reported by Licensees. Such determinations cannot be made before the conclusion of this proceeding and the Notice and Recordkeeping Proceeding.” Appendix B, p. B-10. However, they assumed that the Receiving Agent and the Designated Agent could agree on a “reasonable allocation method” even in the absence of any firm data.

The Register is skeptical. It is apparent at this point in the rulemaking on notice and recordkeeping that obtaining accurate reports of Licensees' use of sound recordings will be difficult, particularly during the first few months. Moreover, the initial reports of use will require reporting on less than a monthly basis, making it impossible in many instances for the Receiving Agent to make any determination whatsoever as to a Designated Agent's allocated share during at least the first month or two in which royalties are paid. Reports on past use of sound recordings (i.e., from October 28, 1998, to the present) will present an even more formidable challenge. It is difficult to imagine that 20 days after the Receiving Agent has received the first royalty payments from Licensees, the Receiving Agent and the Designated Agent will have any reliable information from which they can ascertain how the proceeds should be allocated. The Register therefore recommends that the proposed requirement that payment be made within 20 days of the day on which the Licensee's payment is due be replaced by a requirement that the payment be made “as expeditiously as is reasonably possible,” a more flexible term that recognizes the difficulty in establishing a specific deadline. The Register cautions that during the first few months of operation of the system of reporting and or royalty payment, “expeditious” payment under the circumstances may be a matter of many weeks, if not months.

It can reasonably be expected that for future periods governed by future CARPs or negotiated agreements, more stringent requirements of prompt payment will be appropriate. But it must be recognized that in this initial, transitional period, delays will be inevitable.

f. Allocation of Royalties among Designated Agents and Among Copyright Owners and Performers. The terms proposed by the Parties and accepted by the Panel provide that the Receiving Agent allocate royalty payments to Designated Agents “on a reasonable basis to be agreed among the Receiving Agent and the Designated Agents,” and that the Designated Agents distribute royalty payments “on a reasonable basis that values all performances by a Licensee equally.” The Panel accepted these terms, but observed that a “determination of how royalty payments should be apportioned between the Designated Agents cannot be made until the parties know the rate structure adopted by the CARP (in the first instance) and the Librarian of Congress (on review) and the outcome of the Notice and Recordkeeping Proceeding.” Appendix B, at p. B-10. Similarly, the Panel remarked that “The terms do not specifically provide how a Designated Agent should allocate royalties among parties entitled to receive such royalties because such allocation will depend upon the rate structure adopted by the CARP (in the first instance) and by the Librarian of Congress (on review) and may be affected by the types of reporting requirements that are adopted by the Copyright Office in the Notice and Record-keeping Proceeding for eligible nonsubscription transmissions and business establishment services.” Id., p. B-12.

The Register recommends that the provisions for allocation of royalty payments among Designated Agents and for allocation of royalties among parties entitled to receive such royalties be clarified, making explicit the relationship between the notice and recordkeeping regulations and the allocation of royalties. Each of these provisions should provide that the method of allocation shall be based upon the information provided by the Licensee pursuant to the regulations governing records of use of performances.

The Register has some trepidation about the provision in § 261.4(a), proposed by the Parties and recommended by the CARP, that provides that apportionment among Designated Agents “shall be made on a reasonable basis that uses a methodology that values all performances equally and is agreed upon among the Receiving Agent and the Designated Agents.” (Emphasis added). The regulation does not provide what happens in the event that the Receiving Agent and the Designated Agents cannot agree on an allocation methodology. One could recommend a provision that gives the ultimate decisionmaking power to one of the parties or to a third party, but instead, the Register proposes the addition of § 261.4(l), which would simply provide that in the event of a stalemate, “either the Receiving Agent or a Designated Agent may seek the assistance of the Copyright Office in resolving the dispute.”

g. Choice of Designated Agent by Performers. A literal reading of the terms recommended by the Panel would permit a Copyright Owner to select the Designated Agent of its choice, but would require a Performer to accept the Designated Agent selected by the Copyright Owner; and the Panel's report appears to agree with this interpretation. Report at 132. However, the Report does not articulate any reason for the decision to deprive Performers of the same right to choose that is given to Copyright Owners, and the commentary in Appendix B is silent as well.

As the Panel acknowledged, “Copyright owners and performers, on the other hand, have a direct and vital interest in who distributes royalties to them and how that entity operates” Report at 132 (emphasis added). The Register agrees. It was arbitrary to permit Copyright Owners to make an election that Performers are not permitted to make. The Register can conceive of no reason why Performers should not be given the same choice. Accordingly, the Register recommends that § 261.4 be amended to provide that a Copyright Owner or a Performer may make such an election. See § 261.4(c) of the recommended regulatory text.

The Register has also inserted a housekeeping amendment to provide that for administrative convenience, a Copyright Owner's or Performer's designation of a Designated Agent shall not be effective until 30 days have passed.

h. Performers' Right to Audit. The terms proposed by the Parties and accepted by the CARP provided that a Copyright Owner may conduct an audit of a Designated Agent. These provisions also include safeguards to ensure that a Designated Agent is not subjected to more than one audit in a calendar year.

However, the terms do not provide that Performers have a similar right to conduct an audit of a Designated Agent, despite the fact that Performers, like Copyright Owners, depend upon the Designated Agent to make fair and timely royalty payments. The Parties' commentary in Appendix B states that audit rights are limited to Copyright Owners “rather than the entire universe of Copyright Owners and Performers, which could number in the tens of thousands.” Appendix B at p. B-24. The commentary suggests that it would be impracticable for a Designated Agent to be subject to audit from individual Performers. Apart from reproducing the Parties' commentary, the Panel offered no observations on this point.

The Register fails to understand how it would be “impracticable” to permit Performers, who depend on a Designated Agent for their royalty payments, to initiate an audit of the Designated Agent when the Copyright Owners may do so. The Designated Agent is given sufficient protection by virtue of the provision that it can be subject to only a single audit in a calendar year, by the provision that the party requesting the audit must bear the presumably considerable costs of the audit, and by the provision that any audit “shall be binding on all Copyright Owners and Performers.” The Register, therefore, recommends that the audit provisions be amended to permit not only Copyright Owners, but also Performers, to initiate an audit.

i. Effective date. Section 114(f)(4)(C) states that payments in arrears for the performance of sound recordings prior to the setting of a royalty rate are due on a date certain in the month following the month in which the rate is set. The effective date of the rates, however, is not necessarily the date of publication in the Federal Register. The Librarian has often set the effective date of a rate several months after the initial announcement of the decision. See Determination of Reasonable Rates and Terms for Subscription Services, 63 FR 25394 (May 8, 1998) (setting the effective date for the rate for subscription services three weeks after the date of publication of the final order in the Federal Register); Rate Adjustment for the Satellite Carrier Compulsory License, 62 FR 55742 (October 28, 1997) (announcing an effective date of January 1, 1998, set to coincide with the next filing period of the statements of account).

Section 802(g) provides that the effective date of the new rates is “as set forth in the decision.” 17 U.S.C. 802(g). The Register has interpreted the term “decision” to mean the decision of the Librarian, since section 802(g) only refers to the decision of the Librarian. Thus, this provision has been interpreted as providing the Librarian with discretion in setting the effective date. Moreover, the courts have held that an agency normally retains considerable discretion to choose an effective date, where, as here, the statute authorizing agency action fails to specify a timetable for effectiveness of decisions. RIAA v. CRT, 662 F.2d. 1, 14 (D.C. Cir. 1981).

In setting an effective date, the Register has considered the impact of the rate on the Licensees and the administrative burden on the Office in promulgating regulations to insure effective administration of the license. Clearly, there will be a burden on many Licensees who, by law, are required to make full payment of all royalties owed for transmissions made since the effective date of the DMCA, October 28, 1998, on or before the 20th day of the month next succeeding the month in which the royalty rate is set. Moreover, the Copyright Office is in the midst of promulgating rules governing records of use that will be used to make distribution of royalty fees in accordance with the terms of payment.

Consequently, the Register proposes an effective date of September 1, 2002, which will require the Licensees to make full payment of the arrears on October 20, 2002. Payment for the month of September shall be due on or before November 14, 2002, the forty-fifth (45th) day after the end of the month on which the rate becomes effective, in accordance with the term proposed by the parties and adopted by the CARP. Similarly, all subsequent payments shall be due on the 45th after the end of each month for which royalties are owed. This payment schedule provides the Licensees with additional time to make the initial payment and any necessary adjustments in their business operations to meet their copyright obligation.

V. Conclusion

Having fully analyzed the record in this proceeding, the submissions of the parties, the Register of Copyrights recommends that the Librarian adopt the statutory rates for the transmission of a sound recording pursuant to section 114, and the making of ephemeral phonorecords pursuant to section 112(e), as set forth below:

In addition, the Register recommends that the Librarian adopt the terms of payment proposed by the CARP, as modified in the recommendation, and set September 1, 2002, as the effective date for the statutory rates and the terms of payment.

VI. The Order of the Librarian of Congress

Having duly considered the recommendation of the Register of Copyrights regarding the Report of the Copyright Arbitration Royalty Panel in the matter to set rates and terms for Licensees making certain digital performances of sound recordings under section 114(d)(2) and those making ephemeral recordings under section 112(e), the Librarian of Congress fully endorses and adopts her recommendation to accept the Panel's decision in part and reject it in part. For the reasons stated in the Register's recommendation, the Librarian is exercising his authority under 17 U.S.C. 802(f) and is issuing this order, and amending the rules of the Library and the Copyright Office, announcing the new royalty rates and terms of payment for the sections 112 and 114 statutory licenses.

List of Subjects in 37 CFR Part 261

• Copyright
• Digital audio transmissions
• Performance right
• Recordings

Final Regulation

In consideration of the foregoing, part 261 of 37 CFR is added to read to as follows:

PART 261—RATES AND TERMS FOR ELIGIBLE NONSUBSCRIPTION TRANSMISSIONS AND THE MAKING OF EPHEMERAL REPRODUCTIONS

Authority: 17 U.S.C. 112(e), 114, 801(b)(1).