The 44% Raise the D.C. Circuit Sent Back: Johnson v. Copyright Royalty Board

How the D.C. Circuit vacated the Copyright Royalty Board's Phonorecords III streaming mechanical rate hike for inadequate notice and unexplained reasoning, and what the remand meant for songwriters.

In Johnson v. Copyright Royalty Board, No. 19-1028 (D.C. Cir. Aug. 7, 2020), the U.S. Court of Appeals for the District of Columbia Circuit reviewed the most consequential government-set music-royalty rate of the streaming decade — and largely refused to let it stand. The Copyright Royalty Board (CRB), through its three Copyright Royalty Judges, had concluded the Phonorecords III proceeding by raising the statutory mechanical royalty paid by interactive streaming services to songwriters and publishers, restructuring the formula, and removing a longstanding cap. The streaming services appealed. Writing for the panel, Judge Patricia Millett held that the Board had violated basic principles of reasoned administrative decisionmaking, and the court vacated and remanded the determination in significant part.

At a glance

Matter: Johnson v. Copyright Royalty Board, No. 19-1028
Forum: U.S. Court of Appeals for the District of Columbia Circuit (panel opinion by Judge Millett)
Decided: August 7, 2020
Agency proceeding below: Phonorecords III, Copyright Royalty Board, No. 16-CRB-0003-PR (2018–2022)
Subject matter: The Section 115 compulsory mechanical license and the CRB’s percentage-of-revenue rate structure for interactive streaming and limited downloads
Headline change under review: Increase of the “all-in” headline royalty rate from 10.5% toward 15.1% of revenue, phased in over five years, plus removal of the cap on the total content cost (TCC) prong
Disposition: Vacated and remanded in part for inadequate notice and insufficient explanation; affirmed in part

The statutory machinery behind a government-set rate

Section 115 of the Copyright Act creates a compulsory license: once a musical work has been distributed to the public, anyone may reproduce and distribute it — including by interactive stream — without the songwriter’s individual permission, so long as they pay a royalty at the rate the government sets. That rate is not negotiated in the open market. It is fixed by the Copyright Royalty Judges in periodic Phonorecords proceedings, trial-like rate determinations governed by the Administrative Procedure Act.

The Phonorecords III determination, covering 2018 through 2022, was hailed by songwriter advocates as a landmark. The Board adopted a multi-pronged, “greater-of” formula that calculated the royalty pool from a service’s revenue and from the total content cost it paid for sound-recording rights, then compared the results against per-subscriber floors. Most visibly, the Board increased the headline percentage-of-revenue rate from 10.5% to 15.1%, phased in across the five-year term — the figure popularly described as a roughly 44% raise — and it removed the cap that had previously limited the TCC prong. The streaming services — among them Spotify, Amazon, Google, and Pandora — petitioned for review. (Apple Music notably did not join the appeal.)

Why the determination did not survive review

The D.C. Circuit’s opinion is, at bottom, an administrative-law decision rather than a copyright-policy one. The court did not hold that songwriters were entitled to less; it held that the Board had not done the work the APA requires to justify the structure it chose.

Three failures drove the remand. First, inadequate notice. The court found that the Board’s ultimate rate structure — combining the uncapped TCC prong with the substantially higher headline rate — was not within the “zone of reasonably contemplated outcomes” that the parties had been given a fair chance to address. The Board had assembled its final formula from pieces of various proposals, some never squarely presented at the hearing, leaving the services without the opportunity to test the combined structure with evidence. Notice that lets a party anticipate and respond is a bedrock of reasoned rate-setting, and the court found it lacking.

Second, unexplained rejection of a benchmark. The Board had declined to treat a prior settlement — the Subpart B configuration carried forward from earlier proceedings — as a benchmark for market value, but it did so without an adequate, evidence-based explanation for why that settlement evidence was unreliable. An agency may discount benchmark evidence, but it must say why in terms a reviewing court can follow.

Third, an unauthorized post-decision change. After issuing its initial determination, the Board substantively redefined a key term bearing on “Service Revenue” without identifying the authority that permitted a substantive revision outside the narrow lane reserved for technical or clerical corrections. That, too, the court could not sustain.

The throughline is procedural discipline. Even a rate that may be substantively defensible must emerge from a process in which the regulated parties had fair notice and the agency gave reasons a court can review.

What the remand actually preserved — and unsettled

Because the court vacated and remanded rather than reversing outright, the Phonorecords III rates were not erased in fact. The Board had to revisit the structure, give proper notice, and explain its reasoning. After further proceedings on remand, the Board ultimately re-adopted the 15.1% headline rate and uncapped TCC structure — this time with the procedural record the D.C. Circuit had demanded — and those final rates were later published. The practical lesson for the industry was that the songwriters’ increase was vulnerable not because it was too generous, but because of how it was reached.

For licensees, the decision underscored that a percentage-of-revenue rate is only as durable as the administrative record beneath it. For songwriters and publishers, it was a reminder that even a hard-won rate can be delayed for years by a remand, with retroactive recalculation and re-distribution following long after the period it governs.

Open questions

How much process is enough? The court demanded that the final structure fall within a “zone of reasonably contemplated outcomes,” but the boundaries of that zone — how far an agency may recombine party proposals before notice fails — remain fact-bound and contestable in future proceedings.
What weight do settlements carry as benchmarks? The opinion requires reasoned explanation before discounting settlement-derived evidence, yet leaves open how much deference such private deals deserve as proxies for market value.
Does the remand model invite delay? With each Phonorecords cycle subject to APA review, the gap between a rate’s nominal effective date and its final, litigated form may keep widening — a structural tension the decision does not resolve.

Implications

Rate-setting is administrative law first. A statutory royalty rate stands or falls on notice and reasoned explanation, not on whether the number is fair in the abstract.
Notice must cover the structure, not just the inputs. Combining individually disclosed elements into an undisclosed whole can itself be a notice failure.
Benchmarks demand reasons. An agency that rejects settlement or market evidence must explain, on the record, why it is unreliable.
Victories can be provisional. A favorable rate subject to remand may not become final — or payable in recalculated form — for years.
The CRB’s process improved under pressure. The remand forced a fuller record that ultimately supported the same rate, illustrating how appellate review can harden, rather than overturn, an outcome.

Frequently asked questions

What is a mechanical royalty, and why does the CRB set it? A mechanical royalty pays the songwriter and publisher for reproducing and distributing a musical composition — distinct from the public-performance and sound-recording rights. Because Section 115 makes that license compulsory, the rate is fixed by the Copyright Royalty Board rather than negotiated, so the precise formula the Board adopts governs what every interactive streaming service must pay.

Did the D.C. Circuit cut songwriters’ royalties? No. The court vacated and remanded the determination for procedural and explanatory failures under the Administrative Procedure Act; it did not rule that the rate was too high. On remand, the Board re-adopted the same 15.1% headline rate and uncapped structure with a fuller record.

What was the “total content cost” cap that the Board removed? The TCC prong ties part of the royalty calculation to what a service pays for sound-recording rights. Earlier rules capped how high that prong could push the total; Phonorecords III removed that cap, a change the court found the services had not received adequate notice to contest in combination with the higher headline rate.