The Royalty Statement Is Not the Operating Record

What six months of building settlement infrastructure taught me about $4 billion catalogs, 175 percent split claims, and why payouts are not proof.

I learned the central fact of this piece reading the MLC’s 2024 Annual Royalty Recap on a Sunday evening.

The largest mechanical royalty settlement system in the United States processed $1.05 billion in blanket mechanicals in 2024. Its initial match rate, the percentage that found a registered work on the first pass, was 84.1 percent. Six months of reprocessing later, the same year’s match rate was 89.8 percent. I read that paragraph three times.

Before I started building a settlement system, I would have told you 84 percent was a poor match rate. It is not. It is, in context, a remarkably good one. The number that mattered was the other one, the five-point upward drift that happened after the cash had already left the building. That number told me something about how the industry actually works that no amount of vendor pitching had managed to communicate: the first statement is not the final statement. Reprocessing is not the exception. It is the normal case.

If you book the first statement as truth, your books diverge from reality every month. Indefinitely.

Article 1 was about why the agreement leaves no trace at indie scale. This piece is about where that absence shows up later, when money has to move.

What this looks like one floor down

When I describe this to people who do not run royalty operations, they nod politely. When I describe it to people who do, the response is usually some version of: yes, obviously, this is most of what we do.

The translation, for the polite nodders: a statement arrives a month or two later than expected. Titles don’t match, a Spanish-language version of the same recording got reported under a slightly different artist string and didn’t link to the master. One line item shows a meaningful amount of revenue against an asset, but you cannot find the agreement that authorized the version that earned it. Someone on the team spends a week reconstructing commercial truth from PDFs, emails, and a spreadsheet that lives on someone else’s hard drive. The next month’s statement supersedes part of the previous month’s. Nobody re-issues the prior month’s statement to the writer or the co-producer; the correction is just absorbed quietly into the running totals, and if someone asks why the year-to-date number moved, the answer is some version of “reprocessing.”

This is fine, technically. It is also why so many audits in this industry find money.

The MLC’s numbers are useful because the MLC is doing it correctly

I want to be careful here. The MLC publishes more honest operational numbers, more often, than any private label or publisher in the country. They are not the villain of this piece. They are the source of evidence.

Inside the 2024 recap: while running fresh distributions, the MLC continued to reprocess earlier periods. It paid out an additional $46.5 million for the nine 2021 distributions, $41.7 million for the twelve 2022 distributions, and $44.4 million for the twelve 2023 distributions. That is roughly $132.6 million of catch-up payments, sent in 2024, for usage that occurred two to four years earlier.

By the end of 2024, the MLC was still holding approximately $519.8 million in blanket royalties pending distribution: $273.8 million unmatched, $225.2 million unclaimed, and $20.8 million on hold for legal, ownership, termination, overclaim, or payee-review reasons. That is not failure. That is what a serious settlement system looks like when the underlying record is still catching up. Add it to the historical unmatched pool that DSPs transferred to the MLC at inception, roughly $427 million, adjusted to about $397 million after the Copyright Royalty Board finalized the Phonorecords 3 rates in August 2023, and the picture sharpens further. By mid-2025, more than $200 million of the historical pool had been matched and distributed. The other half is still there, in a database of 44 million registered works, getting closer to its rightsholders every month, never asymptoting to one hundred percent.

This is what reconciliation looks like when you do it well. It is iterative. It is updated after the cash leaves the building. It depends on data the rightsholder may not have submitted yet. None of this is a critique of the MLC. The MLC is operating exactly as a serious settlement system should. The critique is of every operating record downstream that treats the first statement as audited fact.

SoundExchange tells the same story on the recording side: $1.05 billion in gross distributions for 2024, up 4.9 percent over 2023, with $909 million of that in statutory royalties under Section 114. The agency crossed $12 billion in cumulative distributions by early 2025. Its audited annual report still explains why royalties can remain undistributed: legal proceedings, missing reports of use, faulty data, or absent proxies. Different rights type. Same shape. The settlement system runs first. The data catches up second. The economic adjustment lags both.

These are among the most transparent collective settlement systems in U.S. music. If the most transparent systems run iterative reconciliation as their normal operating state, every layer above them, labels, admins, distributors, catalog funds, is doing the same thing, but without the disclosure.

What this costs when it hits a public income statement

The clearest dollar number on what unreconciled rights data costs sits in a Peloton 10-Q.

In March 2019, the National Music Publishers’ Association and fourteen of its members, Downtown, Pulse, peermusic, Ultra, Big Deal, Reservoir, Round Hill, TRO Essex, The Royalty Network, and others, sued Peloton over the use of more than a thousand copyrighted compositions without licenses. The complaint sought $150 million in damages. It was amended later to seek roughly $300 million. The case settled February 24, 2020.

Peloton’s 10-Q for the quarter ending March 31, 2020, attributed a $44.9 million increase in general and administrative expense to settlements with Flywheel Sports and Downtown Music Publishing. The broader litigation-and-settlement total for the quarter was $49.3 million. Billboard and The Hollywood Reporter, contemporaneously and citing the timing, traced the bulk of that number to the NMPA settlement.

Two things matter about Peloton for anyone reading this who runs an indie label, an admin firm, or a catalog fund.

The first is the plaintiff list. It is also a roster of the independent publishing layer. Reservoir is now public on Nasdaq. Round Hill was acquired by Concord at a 67 percent premium in late 2023. Several of those companies now sit squarely in the customer segment every serious rights-infrastructure company is trying to serve.

The second is that the case is not really about Peloton. The case is about what happens when a counterparty, any counterparty, any size, cannot tie the song to the agreement to the use to the statement. When that fails, infringement litigation becomes the cleanup mechanism. It is slow, expensive, public, and asymmetric. The publishers got their settlement and a “joint collaboration agreement to optimize Peloton’s music licensing systems and processes.” Peloton got the $49.3 million quarter.

The same chain breaks at the catalog finance layer

Reservoir Media’s 10-K for the fiscal year ended March 31, 2025, filed with the SEC on May 28, 2025, discloses this on its Recorded Music segment:

“Other income, net during Fiscal 2025 consisted of a $104 thousand gain recorded on the disposal of an equity investment (the ‘Investment Gain’) and the Company’s share of proceeds related to underreported usage for acquired music catalogs that pertained to periods prior to the Company’s acquisition of the music catalogs, which totaled $823 thousand (the ‘Recovery Income’).”

Stop and read that twice. A Nasdaq-listed catalog company has booked, in audited segment reporting, $823 thousand of full-year cash from reconciling underreported usage of a catalog for periods before it owned the catalog. The thesis of this article is now a P&L line item with its own three-letter capitalized acronym in a 10-K.

The same filing also unwinds Reservoir’s older “Royalty Dispute”, which Reservoir’s FY2024 10-K describes as having “commenced in 2017, and was settled in October 2023.” A six-year-running dispute that produced a $2.7 million write-off of recoupable legal expenses, plus $620 thousand of interest expense as the carrying cost. It is footnote 16. Of course it is.

Hipgnosis Songs Fund disclosed the same operational shape at the fund layer. Accrued income and receivables of $111.9 million as of March 2022, of which $7.3 million was international PRO lag that the fund itself said “can run up to 24 months.”Round Hill stated the same structural truth in finance language: royalty income carries an inherent reporting lag of three to six months in the U.S., three to twelve months outside it.

This is not poor management. It is the structural shape of music revenue as it shows up in audited financials for sophisticated rights buyers. The Hipgnosis valuation collapse in 2024, Shot Tower Capital cut the portfolio’s fair value to $1.948 billion, 26.3 percent below the prior year’s number, and separately found a 7.6 percent NAV error from double-counted accrued revenue, was the publicly visible failure mode for a problem every catalog fund manages quietly.

If your finance model treats the royalty statement as the operating record, you will be wrong by some amount that grows with portfolio size, geographic footprint, and acquisition pace.

The Hipgnosis arc, finally

If Reservoir is the clean version of the catalog story, Hipgnosis is the messy one. It is also, in the last twenty-six months, the clearest public laboratory experiment in what happens when operating data becomes a financing question.

The condensed version, because it has now happened three times in two years:

· March 2024: Shot Tower Capital, hired by the new Hipgnosis Songs Fund board after the previous board lost a shareholder vote, valued the portfolio at $1.948 billion, 26.3 percent below the previous September 2023 number. A separate finding: the operative NAV had to be cut another 7.6 percent for accrued revenue that had been double-counted. Shot Tower concluded the investment adviser had “failed to perform to music publishing industry standards” on acquisition underwriting, financial statements, and conflicts management. The investment adviser publicly disputed this.

· April 2024: Blackstone wins a bidding war against Apollo-backed Concord, acquires HSF at $1.31 per share, valuing it at $1.572 billion.

· October 2024: The acquired catalog is re-securitized as Lyra Music Assets, $2.36 billion catalog basis, KBRA-rated.

· March 2025: Blackstone combines HSF, Hipgnosis Songs Assets, and Hipgnosis Song Management into a single entity. Rebranded Recognition Music Group. New CEO. “Fundamentally different,” per the press release.

· July 2025: Recognition raises another $372 million via a second ABS, the first music ABS rated publicly by both Fitch and S&P, not just KBRA. The rating agencies were getting more comfortable with the asset class.

· June 2025: Sony Music Publishing buys Hipgnosis Songs Group, the publishing administration subsidiary, ~4,400 songs from the Big Deal Music acquisition. First Sony bite.

· May 11, 2026: Sony Music Publishing, in partnership with Sony Music Group’s new joint venture with Singapore’s sovereign wealth fund GIC, and with participation from Sony Bank, agreed to acquire the entire 45,000-song Recognition catalog from Blackstone, subject to customary closing conditions. Don’t Stop Believin’, Under the Bridge, Single Ladies, Hallelujah, Sweet Dreams, All I Want For Christmas Is You. Reported deal value: approximately $4 billion.

The same operating dataset that produced a 7.6 percent NAV error in March 2024 has now passed through three institutional configurations in twenty-six months, with implied valuations climbing from $1.948 billion to roughly $4 billion. The catalog is the same. The songs are the same. The reconciliation problem is the same. What changes is who is depending on the operating record being defensible, and how much money is riding on it being right.

Sony, GIC, and Sony Bank are the ones now agreeing to underwrite that.

The upstream cause, which is structural

The U.S. Copyright Office, in its 2021 unclaimed-royalties report, said the quiet parts out loud, across two adjacent paragraphs in its data quality section.

The first: labels routinely supply different data about the same sound recordings to different DSPs, and DSPs alter and normalize that data on ingest. The report cites Music Reports’ observation that metadata rows for a single recording can differ in fields and content across services. Same song, same recording, different data, different platforms, different downstream payments.

The second: in a symposium example documented in the same report, disputed ownership splits between parties could add up to 175 percent, forcing the publisher or label receiving the data to hold the money until the parties resolved their shares.

One hundred. Seventy. Five.

When the input is 175 percent claimed and 100 percent payable, money is held. When data about the same recording differs across DSPs, the same underlying recording can be matched, reported and paid against different records depending on which surface carried the use. When the chain from recording to underlying work to ownership share to claimant identity is broken at any link, the iterative reprocessing the MLC publishes about is the correction mechanism. It is not optional. It is structural.

Judge Aleta Trauger of the Middle District of Tennessee, ruling in Eight Mile Style v. Spotify in August 2024, a case Spotify won on equitable estoppel grounds, wrote that Spotify “was in the practice of licensing catalogs without knowing, with any specificity, what was in them.” She also called it “close to unthinkable” that the publisher did not know about the alleged infringement for years. That is not plaintiff allegation; that is a federal court’s description of how the largest music streaming service in the world operated for years. Eight Mile Style appealed to the Sixth Circuit and then voluntarily dropped the appeal in December 2025. The factual finding stayed in the opinion.

Meanwhile, the MLC’s bundling lawsuit against Spotify, dismissed January 29, 2025 and ostensibly settled law, was given new life when the MLC filed an amended complaint on October 1, 2025. Spotify’s own April 2026 6-K filing, on SEC EDGAR, now reads:

“If the MLC were to ultimately be entirely successful in its claim alleging that Spotify’s Premium Service is not a bundle, then the liability in relation to the period March 1, 2024 to March 31, 2026 would be approximately €410 million, plus potential penalties and interest, which we cannot reasonably estimate.”

That number was €94 million when Spotify first disclosed an estimate in November 2024. The MLC has also sought permission to take an interlocutory appeal of the court’s prior ruling that the Premium Service is a bundle. Discovery on the amended complaint is in progress.

A jury is being asked whether a definitional disagreement about what counts as a bundle, and how the audiobook component should be valued for the bundling math, is worth most of half a billion dollars. The classification question is itself a chain-of-evidence question: revenue line categorization is upstream of the royalty calculation, and the parties do not agree on the upstream reconstruction.

Why this matters now, specifically

Music IP is now a rated asset class, financed at institutional scale. Per Skadden, citing GlobalCapital, cumulative music royalty securitization issuance from 2020 to 2024 exceeded $8 billion, with $3.5 billion issued in 2024 alone. The deals are not small. Concord’s $1.76 billion July 2025 ABS, backed by 1.3 million copyrights. Hipgnosis’ $1.47 billion securitization. Carlyle’s $464 million Litmus-backed deal. KKR via Kobalt. HarbourView. Influence Media. Duetti’s $80 million securitization of independent-artist royalties, the first of its kind. Music royalties are being financed like institutional cash-flow assets. The operating record underneath them now has to become finance-grade too.

KBRA, the most prolific rater of music ABS to date, said this directly in 2023: “Music royalty ABS requires a relatively complex infrastructure to manage the underlying collateral of intellectual property (IP). This includes tracking and protecting copyrights, promoting interest and usage, collecting royalties, and distributing those payments to copyright holders. Music IP assets are not homogeneous; the asset valuations and the legal diligence are bespoke, dynamic processes.”

Translation, in rating-agency language: catalog-level variance is visible across publicly rated transactions, and the diligence is bespoke, not formulaic.

The next thing rating agencies notice, they always notice, is whether the underlying operating data actually supports the cash flow projections that were modeled. When that question gets asked seriously, the match rate stops being a back-office detail and becomes an input to valuation, an input to coverage ratios, an input to whether the next deal closes.

What this changes in practice

The part of the stack I’ve been working on is the chain of evidence: agreement → asset version → revenue line → statement line → cryptographic proof that the whole chain reconciles.

Every settlement batch we finalize produces a signed bundle: the statement, the source manifest, the split posture at finalization, an audit log excerpt, and a verification path. The bundle is signed against a public key published at a stable URL. Anyone, a counterparty, an auditor, a rating agency analyst, a fund LP, can independently re-verify any production bundle against that public key with standard tools.

A worked example, with commands, lives at splitgraf.com/trust. The cryptographic specifics matter less than the operational one: when reprocessing happens, and reprocessing always happens, the next bundle carries the new state with the same chain intact. The operating record travels with the statement.

I came into this six months ago thinking the hard part was the cryptography. It is not the hard part. The hard part is much earlier: making sure the thing you are about to sign is the thing that is actually true.

Closing

The industry is better than it used to be. The MLC says so. SoundExchange says so. The catalog operators with the cleanest filings say so.

Better is not reconciled. The royalty statement is still, too often, a payout artifact rather than a defensible operating record. Reprocessing is no longer an exception. Catalog finance is asking diligence questions the operating data was not designed to answer. AI training will ask harder ones. A 45,000-song catalog has now changed hands twice and is in the process of changing hands a third time, all in twenty-six months, at an implied valuation that has more than doubled. Sony, GIC, and Sony Bank are not the kind of institutions that tolerate a 7.6 percent NAV adjustment quietly.

Statements should prove why money moved, not just that it moved.

The operating record needs to travel with the payment.

Arpit writes about music rights infrastructure at SplitGraf. The verification protocol described above can be reproduced against any production settlement bundle at splitgraf.com/trust. Sources for every dollar figure, court ruling, and corporate event in this piece are linked in the footnotes. Get in touch with me arpit@splitgraf.com

Source notes

• The MLC, 2024 Annual Royalty Recap: $1.05B processed; 84.1% initial / 89.8% current match rate; reprocessing payouts of $46.5M (2021), $41.7M (2022), $44.4M (2023); year-end pending of $519.8M ($273.8M unmatched + $225.2M unclaimed + $20.8M on hold). https://pages.themlc.com/2024-annual-royalty-recap

• The MLC, 2024 Annual Report: 44M+ registered works; $3B cumulative milestone (May 2025). https://www.themlc.com/hubfs/The%20MLC%202024%20Annual%20Report.pdf

• The MLC, Historical Royalties page: $427M initial pool, adjusted to $397M after CRB Phono 3 finalization. https://www.themlc.com/historical-unmatched-royalties

• SoundExchange press release, Feb 24, 2025: $12B cumulative distribution milestone; Q4 2024 distributions of $248.6M; full-year 2024 gross distributions of $1.05B, up 4.9% over 2023. https://www.soundexchange.com/news/soundexchange-surpasses-12-billion-cumulative-distribution-milestone/

• SoundExchange 2024 audited fiscal report: $909M statutory royalties in 2024. https://www.soundexchange.com/wp-content/uploads/2025/05/2024-SoundExchange-Fiscal-Report-Audited-Financials.pdf

• Hollywood Reporter, Peloton and NMPA Settle, Dismiss Copyright Infringement Lawsuit (Feb 27, 2020): 1,000+ compositions; 14 NMPA member plaintiffs.

• Peloton Form 10-Q for quarter ending March 31, 2020: $44.9M G&A increase attributed to Flywheel Sports + Downtown Music Publishing settlements; broader $49.3M litigation-and-settlement total.

• Reservoir Media, Inc. Form 10-K, FY2025, filed May 28, 2025. SEC EDGAR: https://www.sec.gov/Archives/edgar/data/1824403/000141057825001379/rsvr-20250331x10k.htm. Verbatim Recovery Income disclosure of $823K full-year. FY2024 10-K: Royalty Dispute “commenced in 2017, and was settled in October 2023”; $2.7M write-off of recoupable legal expenses; $620K interest expense.

• Shot Tower Capital, Hipgnosis Songs Fund, Operational and Valuation Review (March 2024): $1.948B fair value (26.3% below prior September 2023); 7.6% NAV adjustment for double-counted accrued revenue; HSM “failed to perform to music publishing industry standards” finding (publicly disputed by HSM).

• Hipgnosis Songs Fund 2022 results: $111.9M accrued income and receivables; $7.3M international PRO lag “can run up to 24 months.”

• Recognition Music Group corporate history: Blackstone acquisition of HSF April 2024 at $1.31/share = $1.572B; combined entity rebranded March 2025; Lyra Music Assets ABS October 2024 ($2.36B catalog); second ABS $372M July 2025 (first publicly Fitch/S&P-rated music ABS).

• Sony Music Publishing + GIC + Sony Bank agreed acquisition of Recognition Music Group catalog, May 11, 2026. Reuters: https://www.reuters.com/legal/transactional/sony-music-buy-recognition-music-catalog-source-says-deal-4-billion-2026-05-11/. ~45,000 songs. Subject to customary closing conditions.

• U.S. Copyright Office, Unclaimed Royalties: Best Practice Recommendations for the Mechanical Licensing Collective (July 2021): data bifurcation discussion citing Music Reports on metadata row differences across DSPs; symposium example of split claims summing to 175 percent. https://www.copyright.gov/policy/unclaimed-royalties/unclaimed-royalties-final-report.pdf

• Eight Mile Style v. Spotify USA Inc. and Harry Fox Agency, 3:19-CV-00736 (M.D. Tenn., Aug. 15, 2024), Judge Aleta A. Trauger. Bloomberg Law, Dec 18, 2025: “Eminem Publisher Drops Bid to Revive Spotify Copyright Suit.” Sixth Circuit case nos. 24-5849 / 24-5894.

• MLC v. Spotify, 1:24-cv-03809 (S.D.N.Y.): dismissed Jan 29, 2025; amended complaint filed Oct 1, 2025; MLC seeking interlocutory appeal. Spotify Form 6-K, fiscal Q1 2026 (period ended March 31, 2026), filed via SEC EDGAR: https://www.sec.gov/Archives/edgar/data/0001639920/000162828026027951/spot-20260331x6xk.htm. Verbatim disclosure of approximately €410M exposure if MLC fully prevails. €94M baseline from Music Ally coverage of November 2024 Spotify 6-K.

• Skadden, The Resurgence of Music Securitization (July 2025), citing GlobalCapital 2025: cumulative issuance 2020–2024 over $8B; $3.5B in 2024 alone. https://www.skadden.com/-/media/files/publications/2025/07/the_resurgence_of_music_securitization_issuer_and_investor_appeal_in_the_data_driven_era.pdf. Concord’s $1.76B July 2025 ABS via MBW; Carlyle/Litmus $464M April 2025 via MBW.

• KBRA, Music Royalty ABS: The Beat Goes On, Business Wire press release, March 31, 2023. Key Takeaways verbatim. https://www.businesswire.com/news/home/20230331005349/en/KBRA-Releases-Research-Music-Royalty-ABS-The-Beat-Goes-On