Investing in Music Royalties

On March 4, 2024, the board of Hipgnosis Songs Fund published the preliminary valuation it had commissioned from Shot Tower Capital, a Baltimore advisory firm that valued music catalogues for a living. The review put the portfolio at a midpoint of $1.93 billion, a multiple of 15.9 times net royalty income before contingent catalogue bonuses were deducted, and 26.3% below the value the fund itself had reported on September 30, 2023. Adjusted only for that number, the fund’s operative net asset value fell from $1.7392 a share to $1.1657.

Two weeks later it fell again. On March 18 the fund disclosed that it had double-counted accrued revenue in the same calculation, and restated the pro-forma figure to $1.0765. On March 28 the full due-diligence report followed, and its finding, reported by Billboard and the trade press that week, was not that music had stopped earning. It was that the fund had overstated its revenue, its earnings and the size of the stakes it held in the catalogues it had bought. The manager called the report inaccurate and misleading. Three announcements, two restatements of the same net asset value, inside twenty-four days.

Four months later, on July 29, 2024, Blackstone’s acquisition of the fund became effective at $1.31 a share, valuing the equity at $1.584 billion. Shareholders approved it with 99.97% of votes cast, on a 59.21% turnout. The fund had listed in London in July 2018 at 100p, raised £202 million in what Music Business Worldwide called the largest London IPO of that year, and gone on to raise more than £1.05 billion of equity in all. Six years later it was gone — folded into a private Blackstone business that was renamed Recognition Music Group in March 2025 — and the songs — Neil Young, Shakira, Blondie, thousands of others — went on earning exactly as much as they had earned the week before.

Nothing about the music changed. What changed was the price a buyer would pay for a dollar of its annual income. That number, the multiple of net publisher share, is the whole subject of this guide: what it means, what it has been, what it actually buys, and why a stream of royalties that never misses a payment can still be a bad investment if you pay twenty times for it.

Every recorded song you have ever heard is two pieces of property owned, usually, by different people. The first is the composition: the melody and the lyric, the thing a songwriter writes down. The second is the master, sometimes called the sound recording: one specific performance of that composition, fixed to tape or to a hard drive, usually paid for by a record label. When Whitney Houston sang “I Will Always Love You,” Dolly Parton owned the composition and Arista owned the master, and both were paid every time the record played.

This division is the first thing a buyer has to get straight, because the two sides of the song behave differently as assets and trade at different prices. Publishing — the composition side — is the more protected of the two. Much of its income is set by statute or by rate courts rather than negotiated, it survives re-recordings and cover versions, and its term runs from the author’s life plus seventy years for anything written since 1978. Masters are the commercial side. Their streaming income is negotiated between the label and the platform with no statutory floor, they can be displaced by a re-recorded version, and they carry the label’s own deductions before anything reaches the artist.

Why publishing costs more per dollar

That asymmetry shows up directly in the market’s prices. On Shot Tower Capital’s tally of transactions above $20 million of enterprise value, reported by Billboard, publishing catalogues changed hands at an average of 16.1 times net publisher share in 2024, against 16.7× in 2023; recorded-music catalogues fetched 13.0 times net label share in 2024, against 13.8× the year before. Include the iconic transactions — Shot Tower’s term for assets valued above $200 million — and publishing averaged 17.5× in 2024 against 18.4× in 2023, while recorded music’s own peak year was 2022 at 16.3×. About three turns of multiple on a like-for-like basis, year after year, is what the market pays for statutory protection and a longer term.

The largest deals blur the line, because they are sold as bundles. Sony’s June 2024 acquisition of Queen for a reported $1.27 billion, the largest music-rights transaction on record, bought the publishing, the name-and-likeness rights and the recorded music outside North America; its October 2024 purchase of Pink Floyd for a reported $400 million covered recorded music and likeness but explicitly not the songwriting, which the writers kept. Reading any headline number, the first question is which of the two copyrights it bought.

A song earns money in four distinct ways, each with its own payer, its own collection society and its own rate mechanism, and a buyer who cannot name which stream a listing is selling should not bid on it.

Mechanical royalties are paid for the reproduction of a composition — historically the pressing of a record, and in the streaming era the copy a service makes to deliver you the song. In the United States this is a compulsory licence under Section 115 of the Copyright Act: anyone may reproduce a composition provided they pay the statutory rate, which the Copyright Royalty Board sets. Mechanicals for digital delivery — interactive streaming, limited downloads and permanent downloads alike — are collected in the US by the Mechanical Licensing Collective, created by the Music Modernization Act of 2018 and operating the blanket licence since January 1, 2021; it has distributed more than $3 billion since its first payment in April 2021. Physical product sits outside that blanket licence and is still licensed work by work, through the Harry Fox Agency and its peers or directly.

Performance royalties are paid when a composition is publicly performed — on the radio, in a bar, at a stadium, in a stream. These are collected by performing rights organisations, which license their whole repertoire in a blanket and distribute what they collect. ASCAP, the largest US society, collected $1.945 billion in 2025, up $110 million or 6.0%, of which $1.471 billion was domestic (up 5.3%) and $474 million foreign (up 8.2%). It made a record $1.759 billion available for distribution, up $63 million or 3.7%, of which $1.304 billion came from US-licensed and administered performances. Over the decade to 2025 its revenue compounded at 6.7% a year and its distributions at 7.3%. BMI, its main rival, has not published revenue figures since 2022; after converting to for-profit status and selling control to New Mountain Capital in a deal that closed in the first quarter of 2024, it targets distributing 85% of licensing revenue and retaining about 15%, up from roughly 10% as a not-for-profit.

Synchronisation is the licence to marry a song to a moving image: a film, an advertisement, a video game, a television title sequence. There is no statutory rate and no society. Every sync is a bilateral negotiation, both copyrights must consent, and the fee is whatever the parties agree. Sync is the lumpy, high-margin part of a catalogue, and it is also where a single placement can double a song’s trailing twelve months and mislead an unwary buyer for years.

Neighbouring rights and digital performance cover the master side: the payment owed when a recording, rather than a composition, is publicly performed. In the United States SoundExchange collects the statutory digital performance royalty from satellite, internet and cable music radio, splitting it 50% to the rights owner, 45% to the featured artist and 5% to a session-musician fund. It distributed $991.5 million in 2025 on unaudited figures, 5.9% below 2024 — fewer settlements, lower reported SiriusXM revenue and the cumulative effect of a SiriusXM underpayment — and has passed $13 billion since 2003. That dispute is instructive about how fragile collection can be: SoundExchange put the underpayment above $400 million, sued in 2023, and had its complaint dismissed on August 7, 2025 on the ground that the Copyright Act gives it no private right of action; it has appealed to the Second Circuit. The US also grants no terrestrial radio performance right for masters at all, which is why American neighbouring-rights income is smaller than a European catalogue’s.

The scale behind those streams is why the asset class exists. The IFPI’s Global Music Report of March 2026 put global recorded-music revenue at $31.7 billion in 2025, up 6.4% and an eleventh consecutive year of growth, with paid subscription streaming up 8.8% to 52.4% of the total. Spotify alone said it paid rights holders more than $11 billion in 2025, over 10% more than in 2024. Those are the industry’s own figures, published by the industry. Whether a particular song shares in that growth is the question section six answers.

Between the listener’s subscription and the cheque a royalty owner receives sit four or five intermediaries, each taking a percentage, and the number a catalogue is valued on is the one at the bottom of that chain. Getting the vocabulary right is not pedantry; it is the difference between a 16× multiple and a 22× multiple on the same asset.

Start at the top. A streaming service collects subscription and advertising revenue and pays out a share of it — roughly two-thirds at Spotify, on the company’s own account in its January 2026 payout disclosure — into two pools, one for masters and one for compositions. The master pool goes to labels and distributors, who deduct their own share before paying the artist under whatever contract exists. The composition pool splits into mechanicals, routed through the Mechanical Licensing Collective in the US, and performance, routed through the PROs.

Gross, net, and the writer’s half

The composition side then divides again, by convention rather than by law. Songwriter share is the half of publishing income that belongs to the writer personally; publisher share is the other half, which belongs to whoever publishes the song. A publishing catalogue for sale may be one, the other, or both, and the multiple quoted means something different in each case.

Net publisher share, or NPS, is what remains to the publisher after songwriter royalties and other mandatory payments have gone out. It is the industry’s standard denominator: when Shot Tower reports 16.1× for 2024, it means 16.1 dollars of price per dollar of annual NPS. The recorded-music equivalent is net label share, or NLS, the label’s income after artist royalties. Because these are net figures, a catalogue with an unusually generous artist contract will have a lower NPS on the same gross revenue and, at the same multiple, a lower price. Two catalogues with identical Spotify streams can be worth very different sums.

Administration is the last deduction and the easiest to miss. A publisher without its own collection infrastructure pays an administrator — a major, or a specialist like Kobalt or Songtrust — commonly 15% to 25% of what it collects to register the works, chase the societies and account for the money; Songtrust, to take a published schedule, charges 15% on performance income and 20% on everything else. At auction that cut has usually already come off the number in the listing. Not always. Ask.

The single most valuable feature of a US publishing catalogue is that a large share of its income is set by a federal tribunal rather than negotiated by a party with more leverage than you. Three Copyright Royalty Judges, appointed by the Librarian of Congress, set the Section 115 statutory rate in five-year proceedings named Phonorecords I through V, and the outcome of each proceeding reprices every mechanical royalty in the country at once.

Phonorecords IV covers 2023 through 2027 and is the period in force. Its terms came from a settlement between the National Music Publishers’ Association, the Nashville Songwriters Association International and the Digital Media Association, which the Board accepted in a Final Rule published December 30, 2022 and effective January 1, 2023. The headline rate is a percentage of a streaming service’s revenue, and the schedule is public and fixed for all five years: 15.1% in 2023, 15.2% in 2024, 15.25% in 2025, 15.3% in 2026 and 15.35% in 2027.

Write those steps out and they are a tenth of a point, then three of a twentieth. That matters because several trade write-ups of the settlement described the increase as half a percentage point a year, a figure that has since been copied from one summary to the next; it cannot be reconciled with a 15.35% endpoint, and a model built on it credits a US publishing catalogue with statutory income the rule does not provide. Check the Final Rule, not the write-up.

The floor beneath the percentage

The percentage of revenue is not the only test, and on the services where the alternative binds it is not the number that pays. The all-in pool a service owes is the greater of that revenue percentage and a floor built from what the service already pays record labels for the masters — for a stand-alone portable subscription, 26.2% of total content cost — with a per-subscriber mechanical floor underneath both. Over the Phonorecords IV period the total-content-cost and per-subscriber components rose considerably more than the headline rate did, which is why a catalogue’s mechanical income can grow faster than the 15.1-to-15.35 series suggests. For physical product and permanent downloads the 2026 statutory rate is a separate schedule entirely: 13.1 cents per work, or 2.52 cents per minute of playing time, whichever is greater.

What Phonorecords V will decide

The next period, 2028 through 2032, was still being litigated as of September 10, 2026. The Board commenced the Phonorecords V proceeding in January 2026. On June 29, 2026 a group of participants including the major record companies, the NMPA, the NSAI, the Music Artists Coalition and A2IM filed a partial settlement covering the Subpart B configurations — physical phonorecords, permanent downloads, ringtones and music bundles — which would carry the existing structure forward with inflation adjustments.

The Board posted that settlement in the Federal Register on July 10, 2026 with comments due August 10, and it drew objections from songwriter groups. The Songwriters Guild of America and Word Collections argued that the proposal would restart the physical and download rate at Phonorecords IV’s opening 12 cents in 2028 rather than carrying forward from the 13.1 cents of 2026. The streaming mechanical rate for 2028 onward was unresolved as of September 10, 2026; the Judges directed participants unable to settle to file written direct statements by October 5, 2026.

Price a publishing catalogue on a fifteen-year horizon and three separate CRB determinations will land inside your holding period, each able to move a large slice of your income without your doing anything. That cuts both ways, and it is why publishing multiples carry a premium to masters: a rate the government sets is a rate no counterparty can quietly renegotiate against you.

A multiple of net publisher share is a price, expressed as a number of years of current income, and it is the single most misunderstood figure in this market. Buying a catalogue at 16× NPS does not mean you get your money back in sixteen years. It means that in year one you receive one-sixteenth of what you paid, and everything after that depends on what the income does.

Invert it and the fog clears. A multiple of m is an entry yield of 1/m. Ten times trailing income is a 10.0% first-year yield; 13× is 7.7%; 16× is 6.25%; the 19.4× that Shot Tower recorded as the average publishing multiple in 2021 is 5.2%. Those are gross, pre-tax, pre-cost yields on an asset with no liquidity, no daily price and no manager to call.

The multiple is a discount rate in one equation

The arithmetic that connects the two is the growing perpetuity. If a catalogue earns NPS in the current year and its income changes at a constant rate g for ever, and you discount at r, the value is NPS ÷ (r − g), which means the multiple m = 1 ÷ (r − g). Every multiple you see in a press release is therefore a statement about two numbers the buyer has assumed, and you can back one out if you assume the other.

Do that with the market’s own prices and the last five years stop being mysterious. At 19.4× the buyer needs r − g = 5.2%. Assume income grows 3% a year with streaming and the implied discount rate is 8.2%, which in 2021, with a ten-year Treasury under 2%, was a fat spread for an asset a bank would lend against. Assume instead that income declines 3% a year and the implied discount rate is 2.2%, which is worse than a government bond. The same cash flow, the same catalogue, the same price: the entire difference between a good deal and a bad one is the sign in front of g.

That is what repriced in 2022 and 2023. Rates rose, so the r a buyer needed rose with them, and the growth story softened, so the g they would assume came down; both moved the wrong way at once and the multiple, being the reciprocal of their difference, fell hard. Shot Tower’s series shows the result: publishing including iconic transactions from 19.4× in 2021 to 17.5× in 2024, recorded music from a 16.3× peak in 2022 to 13.0×.

The market did not stop functioning at the lower level; it got busier. Citrin Cooperman’s music and entertainment valuation practice, led by Barry Massarsky, told Billboard in April 2026 that it had priced 566 catalogues worth nearly $13 billion in 2025, against 557 catalogues worth $10.7 billion in 2024, and that multiples had held steady for a third consecutive year. Average deal size rose from about $19 million to nearly $23 million. Publishing-only catalogues accounted for 41% of the combined value, mixed publishing-and-masters catalogues about 40%.

Every valuation in this market rests on a forecast of how a song’s income falls away with age, and the shape of that curve is the assumption that decides whether a multiple was cheap or expensive. The industry’s standard method, developed most visibly by Barry Massarsky of Citrin Cooperman, formerly senior economist at ASCAP, is a discounted cash flow built on the last three to four years of performance, mechanical and synchronisation income, projected forward on a decay curve toward what he calls a steady state. New music decays toward that plateau; a standard from decades ago is already sitting on it.

Why old songs are the stable ones

This is the counterintuitive fact that makes the asset class work, and Luminate’s data measures it. In the first half of 2025, tracks released within the previous eighteen months — the industry calls them current — took 24.2% of US on-demand audio streams, and everything older, the catalogue, took 75.8%. On a release-year cut, 43% of US on-demand audio streams came from tracks released in the five years 2021 to 2025, leaving 57% to everything before that. Current music’s total US plays declined in 2025 even as total streams grew.

Catalogue’s share keeps rising partly because the denominator keeps exploding. Luminate counted 5.1 trillion on-demand audio streams globally in 2025, up 9.6% on 2024, spread across 253 million individual tracks, with about 106,000 new ones arriving daily; 120.5 million tracks drew ten streams or fewer all year. A catalogue with an audience competes against an infinite supply of music that has none.

Decay is net of two forces

The number that matters to a buyer is not a song’s decay rate but its net rate — song-level decay minus the growth of the pool it draws from. Global recorded-music revenue rose 6.4% in 2025 and has risen every year since 2015; paid subscriptions grew 8.8%. A catalogue losing 8% of its share of streams a year inside a pool growing 6% a year is losing about 2% of income a year, not 8%. This is why institutional buyers were able to assume flat-to-growing income at 17× and not obviously be fools, and it is why the growth of the pool is the first thing to check when you underwrite one.

Three forces work the other way. Subscription price rises have carried much of the growth behind that 6.4% figure for 2025, and a price rise is a one-off that cannot repeat indefinitely. Per-stream rates fall when consumption grows faster than revenue, because a fixed pool divided by more plays pays less per play; the quoted $0.003 to $0.004 Spotify range is an average of a divided pool, not a rate card, and it varies by territory, by tier and by whether the listener pays. And the flood of new uploads dilutes every existing work’s share by construction.

The dollar-age test

Royalty Exchange reports two numbers on every listing: the LTM multiple, the price against the last twelve months of income, and dollar age, a time-weighted measure of how old the songs generating that income are. The second is the useful one. A $50,000 trailing year from a single track released fourteen months ago is a different asset from a $50,000 trailing year spread across two hundred songs averaging twenty-five years old, and the first is worth a fraction of the second’s multiple. Buy young income at a catalogue multiple and the decay curve takes the difference out of you.

The catalogue market of 2019 to 2022 was a rates trade wearing a music-industry costume, and reading it that way makes the sequence obvious. Streaming had turned a declining industry into a growing one; interest rates were at zero; and a royalty stream from a forty-year-old song looked, to a certain kind of buyer, like a long-duration inflation-linked bond with better stories at dinner. Money arrived accordingly.

What the largest catalogues fetched

The headline transactions are the record everyone remembers. Bruce Springsteen sold to Sony in December 2021 for a figure reported at around $500 million, covering both his recordings and his songwriting. Bob Dylan sold his songwriting to Universal Music Publishing in December 2020; Sony announced the purchase of his recorded catalogue in January 2022. Sony paid a reported $1.27 billion for Queen in June 2024, the largest music-rights transaction on record, taking the publishing, the name-and-likeness rights and the recorded music outside North America; and a reported $400 million for Pink Floyd in October 2024 — recordings and name-and-likeness only, with the songwriting staying with the writers. Two nine-figure headlines, two different bundles of copyright, and the difference is most of what a buyer needs to know about either.

Alongside the strategic buyers came the financial ones. Hipgnosis Songs Fund listed in London in July 2018 and raised £202 million at its IPO, then more than £1.05 billion in total across placings in April, August and October 2019 and July and September 2020. Round Hill Music Royalty Fund followed with a second London vehicle. Blackstone funded a private companion vehicle for Hipgnosis’s manager. Primary Wave, Concord, HarbourView, Litmus, Influence Media and a dozen others raised institutional capital against the same thesis: Brookfield Asset Management committed $1.7 billion to a permanent-capital vehicle alongside Primary Wave in October 2022, in a package the parties put at about $2 billion, and took a minority stake in the publisher itself.

The turn

When the Federal Reserve began raising rates in March 2022, the discount rate in every one of those models moved, and the multiple, being the reciprocal of the gap between that rate and the growth rate, could not stay where it was. Nothing about the songs changed; the cost of the money that bought them did.

Both listed vehicles then failed the market test at once, and in a specific way that is worth understanding: their shares traded below the value their own boards published for the assets. A closed-end fund cannot issue new stock below net asset value without diluting its holders, so a persistent discount ends the fund’s ability to grow, which ends the manager’s fee growth, which eventually ends the fund.

Round Hill went first. Concord agreed in September 2023 to buy the Round Hill Music Royalty Fund for about $468.8 million, or $1.15 a share — a 67.3% premium to the September 7, 2023 close of $0.69. Read that premium backwards and it is the more useful number: to be worth 67% more in a trade sale than in the market, the shares had to have been trading at a little over half of what a strategic buyer would pay. Shareholders approved on October 18, 2023 with about 99% of votes cast in favour, the scheme became effective on November 1, and more than 150,000 songs across 51 catalogues — works recorded by The Beatles, Elvis Presley, James Brown and Billie Holiday among them — left the public market.

The rise and end of Hipgnosis Songs Fund is the best-documented failure in this asset class, and every retail buyer should read it as a checklist rather than as a scandal. The fund did not lose money because songs stopped earning. It lost money because of the gap between what a manager reported and what an independent valuer found, and because a closed-end wrapper punishes that gap without mercy.

The chronology

The fund listed in July 2018 and raised £202 million, and by late 2020 had raised more than £1.05 billion of equity across five further issues. In October 2023 the cracks opened in public. On October 16 the fund disclosed that its independent valuer, Citrin Cooperman, had cut the expected retroactive royalty payments for 2018 to 2022 arising from the previous rate period from $21.7 million to $9.9 million, a reduction of about 54%, and the board withdrew the 1.1p interim dividend it had declared three weeks earlier to stay inside a covenant on its revolving credit facility. The shares fell more than 10% on the news to 66.31p.

Ten days later the shareholders ended it. At the continuation vote on October 26, 2023, 83.2% of votes cast went against continuing the fund and 84.1% against a proposed $440 million sale of catalogues to a sister vehicle; shareholders removed the chairman the same day, and the board was obliged to bring forward proposals for reconstruction, reorganisation or winding up within six months.

In March 2024 the board published Shot Tower Capital’s work in three separate announcements over twenty-four days, and the order of them is the part worth memorising. On March 4 came the valuation: a portfolio midpoint of $1.93 billion, at 15.9× net royalty income before contingent catalogue bonuses, 26.3% below the September 30, 2023 figure, with a range of $1.80–$2.06 billion. Pro-forma operative NAV fell from $1.7392 to $1.1657 a share, 92.08p at the March 1, 2024 exchange rate.

On March 18 the fund announced an amendment to that operative NAV. It had double-counted accrued revenue in the same calculation, and the pro-forma figure was restated to $1.0765 (85.03p), a further 7.6% down. This is the instalment most summaries drop, and it is the one that tells you the most: the error was not in the independent valuer’s work but in the fund’s own accounting of the number it had published a fortnight earlier. On March 28 the full due-diligence report landed, with the substantive findings — that the fund had overstated revenue, earnings and the size of its stakes in catalogues. The manager called the report inaccurate and misleading. The board said it would use free cash flow to pay down debt and would not resume dividends for the foreseeable future.

Hold the three dates apart, because collapsing them into one is how this episode gets told badly. A valuation cut, an accounting restatement and a diligence report are three different kinds of bad news, and only the first of them is about the price of music.

The endgame ran from April to July 2024. Concord opened at $1.16 a share ($1.4 billion), raised to a final $1.25 ($1.5 billion). Blackstone bid $1.31, valuing the equity at $1.584 billion, and won; 99.97% of votes cast approved at the meeting on July 8, on a 59.21% turnout, and the transaction became effective July 29, 2024. The fund continues under Blackstone as a private business: in March 2025 it was merged with the two related Hipgnosis entities and renamed Recognition Music Group, holding more than 45,000 songs and recordings across 145 catalogues.

Read the chart as a ranking of who was right. The manager’s number was $1.74. Adjusted for the independent valuation, and then again for a double-count the fund found in its own arithmetic, it was $1.08. Two informed strategic buyers, bidding against each other with full diligence access, settled at $1.31 — a quarter below the reported NAV and above the independent one. When a self-reported mark and a real bid are that far apart, the bid is the number.

Invest Alternative runs a daily index across the alternative-asset markets we cover, and there is no music-royalty series on it. That absence is not an oversight, and it is the most useful thing our own data can tell you about this asset class.

Our composite, still provisional, stood at 100.271 on September 8, 2026, against a base of 100 on September 2, 2025 — up 5.74% over thirty days and 0.29% over the year, with a high of 108.418 on January 28, 2026 and a low of 90.553 on June 25, 2026. Every category inside it is built from something priced daily: an ETF, an exchange feed, an auction tape, a spot quote.

Music royalties have no such input that isolates the asset. The listed securities that touch this market — Reservoir Media on Nasdaq, the three majors, a handful of streaming platforms — are operating companies whose share prices carry staff, signings, debt and label economics alongside the catalogue, so a series built on them would measure the music business rather than the price of a dollar of song income, which is the thing this guide is about. The Financial Alternatives sleeve carries a Private Credit series at 83.157 on the same date, down 16.79% over the year, proxied on a listed BDC ETF, precisely because listed vehicles produce a price every day.

Until July 2024 there were two daily prices for a diversified pool of song catalogues carried at a published net asset value: Hipgnosis Songs Fund and Round Hill Music Royalty Fund, both on the London Stock Exchange. Both were bought and delisted within nine months of each other, and with them went the only continuous public mark that could be read against a stated NAV for the catalogues themselves. What remains is a market of private transactions valued by a handful of advisory firms and reported in arrears, in ranges. Weigh that plainly before sizing a position: this is an asset whose price is discovered a few hundred times a year, by professionals, in private, and disclosed to you as an annual average.

A non-institutional buyer has three honest routes into royalty income: buy the copyright outright at auction, buy a security a platform has issued against copyrights it owns, or buy a listed company or trust that owns them. They price differently, they fail differently, and the fee that decides the outcome sits in a different place in each one.

Whole-asset auction

Royalty Exchange is the established venue. Sellers list a defined royalty stream — a writer’s share, a producer’s points, a small publishing catalogue — and investors bid in a timed online auction. The company describes more than 27,000 registered investors, over 2,000 completed transactions and more than $170 million of volume. There is no listing charge. The buyer pays a fee of 1% of the final deal amount or $500, whichever is greater, and the seller a 15% marketplace fee on the sale price, set before the listing goes up; an annual membership, priced at $4,997 in the first year and $997 thereafter, waives the buyer fee and cuts the seller’s to 12%. Every listing carries an LTM multiple and a dollar-age figure, and an order-book feature lets buyers post standing interest rather than wait for an auction.

Two cautions, and the first is about the numbers the venue publishes about itself. Clearing multiples are described by the platform, never published as an audited distribution; the ranges quoted by the platform and by third-party reviewers of it do not agree with each other, and neither do the case histories it markets, such as an asset it reports as having returned 17.76% a year from Q3 2001 through Q4 2013. Treat every vendor performance figure as marketing until you have seen the statements behind it.

The second is the shape of the buyer fee, which is regressive by construction. Because it is 1% or $500, whichever is greater, the $500 floor binds on everything below a $50,000 deal — which is most of what a first-time buyer looks at. On a $10,000 winning bid the floor is 5% of the price, turning a 10× entry into 10.5× and costing half a point of yield before anything else happens. Above $50,000 the same fee is 1% and barely registers. The venue is cheapest for the buyers who need it least.

Fractional shares in a platform’s copyrights

Two venues sell fractions rather than the asset. In both, the fee that matters is not the one on the tariff page; it is the one already inside the price you pay.

Jukebox: the spread, not the fee

Jukebox, which operated as JKBX until it rebranded in February 2025 under a new chief executive, sells fractional royalty shares in individual songs through an SEC-qualified offering structure, and describes shares in more than 60,000 songs representing over $6 billion of music royalties. Its offering documents provide for a royalty fee of 1.0% of the gross monies received on the underlying rights. The headline fee is not the whole cost: an independent review of its filings found the issuer had acquired 86 assets for $45.3 million and listed them to investors at $50.3 million, and the roughly $5 million between those figures is paid by the buyer in the price rather than charged as a fee. The appeal is the ticket size; the cost is that you are buying a security issued by a platform rather than the copyright itself, which adds an issuer to the chain between you and the money and leaves you dependent on that issuer for any way out.

SongVest: the sourcing fee

SongVest takes the same fractional approach through Regulation A+ offerings sold as SongShares, with units that have started around $16. Its filings are the reason to read this route carefully rather than to dismiss it: the offering circulars provide for a sourcing fee of up to 25% of the purchase price of the underlying royalty asset (earlier offerings used 16%), on top of an administrative fee of up to 10% of the royalty payments collected. A sourcing fee is charged against your capital before a dollar of income arrives, so it functions as an increase in the multiple you paid, and at that scale it is not a small one.

The listed route: two securities, no fund

Two music-rights securities can be bought in an ordinary brokerage account, and it is worth being exact about what each one is, because “you cannot buy music royalties on an exchange” is repeated widely and is not true as stated.

Reservoir Media (Nasdaq: RSVR) is a music-rights operating company, not a fund. It reported revenue of $175.7 million in the year to March 31, 2026, of which $116.8 million was music publishing, up 9%, on operating income of $38.2 million. Its share price carries the staff, the signings, the debt and the acquisition pipeline alongside the catalogues, so what you own is the music business, not a dollar of song income at a stated price. Mills Music Trust (OTC: MMTRS) is the opposite trade: closer to a pure royalty claim, and about as undiversified as an asset gets. It is a 1964 grantor trust over a single catalogue, with 277,712 units, that passed through $2.35 a unit in 2025 on total distributions of about $653,000 — and that spent the year disputing calculation-method underpayments of $271,140 with the payer, which is a fair illustration of the collection risk section fourteen describes.

What genuinely does not exist is the thing in between: a listed, diversified vehicle that publishes a net asset value for its catalogues. Both of those were acquired in 2023 and 2024. Neither Reservoir nor Mills publishes one, so neither gives you the number that made the closed-end funds useful — an independent mark you could measure the share price against. That is the structural fact about retail access here, and it is why the sections on tax, term and diligence carry more weight in this guide than they would for an asset you could buy in one line and value from a factsheet. Two siblings in this hub sit on the same problem from other angles: Investing in Litigation Finance on a privately negotiated claim to someone else’s cash flow, and Investing in Pre-IPO Shares on paying a price no public market has tested. The wrappers that hold assets like these are the subject of Investing in Listed BDCs and Investing in Interval Funds and Non-Traded BDCs, where the fund carries a mark and the reader can watch what a buyer will actually pay for it.

Long before a retail investor sees a royalty stream, the institutional side has priced it as collateral, and the bond market’s verdict is a better sanity check on catalogue values than any press release. Music royalty asset-backed securities pool a catalogue’s cash flows, tranche them, and sell rated notes against the pool. KBRA, which rates most of them, says it has rated $12.9 billion of music royalty-backed bonds since 2020.

The largest deal on record is Concord’s $1.765 billion securitisation, closed in July 2025, backed by more than 1.3 million copyrights and described by the issuer as the largest and longest-tenured term securitisation of music rights ever done; it refinanced a $1.750 billion 2022 series that had been rated single-A-plus in its senior tranches. HarbourView raised $500 million from KKR in June 2025 through a private securitisation, after $500 million in March 2024. Chord Music brought a $500 million deal against a catalogue valued at $830 million. KBRA expects 2026 issuance to fall about 25%, from more than $3.3 billion in each of the two preceding years to a little over $2.5 billion, which it attributes primarily to consolidation among issuers rather than to deteriorating collateral.

What the bond market is telling you

An agency willing to put a single-A rating on notes backed by song royalties is making a supportive statement about the stability of that cash flow. But note the structure: bondholders take the first, most predictable slice at a fixed coupon, with covenants and a reserve account, and the equity — the catalogue owner, and by extension you — takes the volatile remainder. When a listing offers a yield close to what an investment-grade music bond pays, you are being asked to take equity risk for debt returns. Our hub flagship, Investing in Private Credit, covers how that trade is priced across the wider market.

The US tax treatment of a purchased royalty stream is the least glamorous part of this guide and the part most likely to change your return, because royalties are ordinary income and the top federal rate on ordinary income is 37%. Everything below is the published rule rather than advice, and the entity structure and the income-forecast election in particular need a CPA who has done this before.

The income is ordinary

Royalty receipts are ordinary income, reported on Schedule E for property held for the production of income, and taxed at your marginal rate. Because royalties from investment property are portfolio income rather than passive income, the passive-loss rules generally do not quarantine deductions against them. Royalties are also net investment income, so the 3.8% surtax under Section 1411 applies once modified adjusted gross income passes $200,000 for a single filer or $250,000 on a joint return — thresholds that have not been indexed since they took effect in 2013 and therefore catch more people every year.

Put the pieces together and the rate is easy to state. An investor in the 32% federal bracket living in a state that charges 5% therefore pays about 40.8% of every dollar received, before any deduction; at the top federal rate of 37%, made permanent by the 2025 tax act and reached above $640,600 of taxable income for a single filer in 2026, the same arithmetic gives 45.8%. The worked example below uses the lower of the two, 40.8%, and says so in its chart caption.

Substitute the top-bracket 45.8% into that example and the net return moves by less than a tenth of a percentage point, which is the counterintuitive part and worth understanding before you spend money on structuring. The reason is that the cost-recovery deduction described next shelters the income at your marginal rate, and the recapture at the exit is taxed at the same marginal rate; raise the rate and you enlarge both sides of the trade at once. In this asset the tax rate is close to a second-order variable. The tax character — ordinary all the way through, with no capital-gains rate at the end — is the first-order one, and it does not change with your bracket.

The deduction: cost recovery on the copyright

The offsetting mechanism is depreciation of what you paid. Two regimes exist and the difference is important. Section 197 amortises acquired intangibles straight-line over fifteen years, but by its own terms it does not apply to a direct purchase of a copyright or an interest in one unless the copyright is acquired as part of the purchase of a trade or business. A stand-alone royalty purchase at auction usually is not.

What is available instead is the income forecast method under Section 167(g), which Congress wrote for exactly this kind of property: motion picture films, video tapes, sound recordings, copyrights, books and patents. Each year’s deduction is your cost multiplied by that year’s net income from the property, divided by the total income anticipated from the property through the end of the tenth taxable year after the year it is placed in service. The forecast is revised as reality arrives, and a look-back rule under Section 167(g)(2) charges or refunds interest, computed on Form 8866, where actual income diverges materially from the estimate.

The practical effect is that the deduction is front-loaded, because the early years produce the most income and therefore claim the largest fraction of basis. On the worked example below, the first-year deduction exceeds the first-year income, and the position throws off a paper loss for a decade. That deferral is worth real money — and it reverses at the exit, because the depreciation you took is recaptured as ordinary income when you sell.

At the exit

A royalty interest held more than a year and sold at a gain is a capital asset in the buyer’s hands, taxed at long-term rates up to 20% plus the 3.8% surtax — but only to the extent the gain exceeds the depreciation claimed, because depreciation is recaptured first, at ordinary rates. On a catalogue whose income declines and whose basis has been mostly written off, almost the entire sale proceeds are recapture and the capital-gains rate never applies. A seller of a catalogue they created faces a different and older question, with its own case law, that is beyond this guide.

Numbers make the argument better than adjectives, so here is a complete round trip on a single retail purchase, with every fee, every tax and the arithmetic shown. The assumptions are ours and are stated so you can change them.

The purchase

You win an auction at $10,000 for a royalty interest whose last twelve months of income was $1,000 — a 10.0× LTM multiple, at the cheap end of what retail lots clear at. The buyer fee is 1% or $500, whichever is greater, so at this size it is the $500 floor: your cash out is $10,500 and your effective multiple is 10.5×. You allow $250 a year for bookkeeping and the annual recomputation the income forecast method requires. You assume income declines 3% a year, net of streaming growth. You hold ten years and sell at 9× the then-trailing twelve months, paying the 15% marketplace fee.

The ten years

Income falls from $970 in year one to $737 in year ten, and totals $8,490. Total income anticipated through year eleven, the denominator the income forecast method requires, is $9,205, so year one’s deduction is $10,500 × ($970 ÷ $9,205) = $1,106 — more than the income it offsets. Across ten years you deduct $9,684 of your $10,500 basis against $8,490 of income and $2,500 of costs, which produces a cumulative tax refund of $1,507 at a 40.8% marginal rate rather than a bill. Net cash from the income stream over the decade: $7,497.

The exit

Year ten’s income of $737 at 9× is $6,637 gross, less the 15% marketplace fee, leaving $5,641. Your remaining basis is $816, so the gain is $4,825 — all of it recapture of depreciation already claimed, taxed at 40.8% for $1,969. Net proceeds: $3,673.

What it returns

Total cash back is $11,170 on $10,500 out. Two returns come out of that, and which one you quote decides whether this looks like a bond or like a mistake.

The gross figure is 4.39% a year over the ten years. It counts the purchase price including the buyer fee, the royalties as they arrive, and the sale proceeds after the 15% marketplace fee — and it counts nothing else. It is before the $250 a year of administration and before every dollar of tax. The net figure is 0.92%, the same ten years after the administration cost and after tax in both directions: the refunds the cost-recovery deduction generates while you hold, and the recapture bill when you sell.

Quote the right one for the comparison you are making. A published index return for equities or for an ABS is a gross, pre-cost, pre-tax number, so 4.39% is the like-for-like figure against a bond yield or an index; 0.92% is what actually reaches you. Do not read the 3.5-point gap between them as the tax bill, which is the mistake a “pre-tax versus after-tax” label invites. Split it and 3.0 points are the $250-a-year administration cost and 0.46 of a point is tax. On a position this size the bookkeeping costs six times what the government does, because the cost-recovery deduction shelters almost all of the income while the flat annual fee does not shrink with it.

The 10.0% entry yield the multiple implied ends as 0.92% net, and none of that gap is a criticism of the asset. It is the arithmetic of buying a small, illiquid, ordinary-income claim through a venue whose fees do not scale down with the ticket. Change the assumptions and the answer changes with them, which is the point of the next three lines.

Every cell in that grid is the same pair on the same basis — gross of the administration cost and tax, then net of both — and the exit is held at 9× in all three, so the only thing changing is the decay rate. Three lessons fall out of the arithmetic.

The decay assumption dominates everything. Holding the exit at 9×, the swing between flat income and a 6% annual decline is 4.89 points of net annual return, from 3.38% to −1.51%. That is more than the entry multiple, the fees and the tax combined. Every hour of diligence belongs to the income statements, not to the bid.

The exit multiple is the second variable, and a smaller one than most bidders assume. At the base case’s −3% decay, selling at 7× instead of 9× gives 3.09% gross and −0.11% net; selling at 11× gives 5.53% and 1.85%. That is roughly 1.2 points of gross IRR, or about one point net, for every two turns. Assume you sell below what you paid: your catalogue will be a decade older, and dollar age is precisely what the next buyer prices.

The fee stack is small but permanent. The $500 in and the 15% out came to $1,496 here — 15% of the bid, and a year and a half of income spent on the two transactions that bracket the decade.

Royalty income rarely stops; it gets diverted, reclassified, or turns out never to have been yours. The failures in this market are documentary rather than dramatic, which is precisely why they catch retail buyers.

Chain of title

The right you are buying was created by a contract, often decades ago, between people who are dead. Splits get recorded wrongly at the societies, co-writers get added after the fact, an administrator’s registration conflicts with a society’s, and the money sits unmatched. Shot Tower’s finding on Hipgnosis was partly this: the fund’s stakes in some catalogues were smaller than reported. Before you bid, insist on society statements in the seller’s name for three or four years, and check the registration at the PRO and the Mechanical Licensing Collective yourself.

Termination rights

US copyright law lets an author or their heirs claw back a grant. For grants made on or after January 1, 1978, Section 203 allows termination during a five-year window beginning 35 years after the grant was executed — or, where the grant covers the right of publication, the earlier of 35 years from publication and 40 years from execution — with notice served not less than two and not more than ten years before the effective date. For pre-1978 grants, Section 304(c) allows termination in a five-year window opening 56 years after the copyright was secured.

The right belongs to the author and, if the author has died, to a statutory list of a surviving spouse, children, grandchildren and executors; it cannot be contracted away in advance. This is the most under-priced risk in retail royalty listings, because a catalogue whose window opens inside your holding period may simply stop being yours, and the notice will arrive years before it does.

Platform and issuer risk

The venue is not the asset. Royalty Exchange’s own listed vehicle, Royalty Flow, bought 25% of the Eminem royalties payable to FBT Productions from albums released 1999 to 2013 for $18.75 million and planned a Nasdaq listing through Regulation A-plus; the offering was cancelled in April 2018 after the exchange withdrew its provisional approval, and the purchase was funded internally instead. Neither that nor JKBX’s February 2025 rebrand cost a direct copyright owner anything, which is the point: own the copyright and a platform’s troubles are someone else’s.

Concentration and re-recording

A stream built on one or two songs is a single-security position with no diversification and total dependence on one artist’s continued relevance. On the master side there is an additional hazard with no publishing equivalent: an artist can re-record their own compositions and direct their audience to the new versions, which leaves the original master earning a fraction of what it did. Publishing income survives that manoeuvre; masters do not.

This asset rewards preparation more than selection, because the diligence is where the mispricing lives and there is no index to fall back on. A workable sequence:

1. Decide the size before the asset. A single royalty interest is an illiquid, undiversified, ordinary-income position with a decade-long holding period. Size it at a weight you would be content to write off; above low single digits of a portfolio it is a business, not an investment.
2. Read three months of listings without bidding. Record the LTM multiple, dollar age, stated term and clearing price on every auction. A distribution of your own is worth more than any published range.
3. Convert every listing to an implied discount rate. Use r = 1/m + g, write down the g you assumed, and compare the answer with the ten-year Treasury and an investment-grade music ABS the same week.
4. Demand the statements. Three to four years of society and administrator statements in the seller’s name, with performance, mechanical and sync income shown separately. Re-strike the multiple on the three-year average.
5. Establish the tax position before you bid. Confirm with a CPA whether Section 197 or the Section 167(g) income forecast method applies, and what entity you will hold it in. The question is the timing of the deduction, not your bracket: as the worked example shows, the marginal rate barely moves the answer, while a fifteen-year straight line instead of a front-loaded forecast moves it a great deal.
6. Bid the price your arithmetic gives you, and let auctions go. With no daily price and 15% round-trip friction, the only reliable edge is refusing to pay the multiple everyone else pays.
7. Reconcile every statement against the society portal for two years. Under-collection and mis-registration are the common failure modes, and both are fixable if caught early.

These are the readings that would change our view, each with an as-of date so a reader in a later year can check them against the record of the day rather than against this one.

• The average publishing multiple. 17.5× NPS including iconic transactions in 2024, from a 19.4× peak in 2021, with Shot Tower Capital itself expecting the ex-iconic average to drift from 16.1× in 2024 to about 15.1× by 2028 (Shot Tower Capital via Billboard, April 2025). Above 19× means rates have fallen a long way or buyers are underwriting growth again; below 14× puts institutional pricing where a retail bid at 10× starts to look expensive rather than cheap.
• The recorded-to-publishing spread. 13.0× against 16.1× on a like-for-like basis in 2024, a spread of about three turns. A spread under two turns says masters are being underwritten as though they had a statutory floor, which they do not.
• Phonorecords V. The 2028–2032 streaming mechanical rate was unresolved as of September 10, 2026, with a partial settlement on physical and download configurations posted for comment in the Federal Register on July 10, 2026 and objected to by songwriter groups; written direct statements from participants who cannot settle are due October 5, 2026. A settlement continuing the escalation from 15.35% is the base case in every publishing model; a decision that flattens or cuts it reprices every US catalogue.
• Global revenue growth. $31.7 billion in 2025, up 6.4%, paid subscription up 8.8% (IFPI, March 2026). Growth below about 3% makes the flat-income assumption behind mid-teens multiples untenable, because song-level decay would no longer be offset.
• Catalogue share of consumption. 75.8% of US on-demand audio streams in the first half of 2025 came from tracks older than eighteen months (Luminate). A sustained fall would undermine the entire premise that old songs are the stable ones.
• Music ABS issuance. KBRA expects a little over $2.5 billion in 2026, about 25% below each of the two preceding years, having rated $12.9 billion since 2020. Issuance is the clearest read on whether the leveraged bid is returning or retreating.
• SoundExchange distributions. $991.5 million in 2025, down 5.9% on unaudited figures. A second consecutive fall would say master-side statutory income is shrinking structurally rather than settling a dispute.
• A diversified vehicle that publishes a net asset value. None was listed as of September 10, 2026, the first time since 2018 that has been true. Two music-rights securities are buyable — an operating company, Reservoir Media (Nasdaq: RSVR), and a single-catalogue trust, Mills Music Trust (OTC: MMTRS) — but neither publishes a NAV for its rights, so neither restores the number that mattered. A new closed-end fund would, and its discount to net asset value would immediately become the most informative figure in the market.
• A US terrestrial performance right for masters. There is none, which is the single largest structural discount on American recorded-music catalogues. The American Music Fairness Act, reintroduced in the 119th Congress as S.326, would create one; a Senate Judiciary intellectual-property subcommittee heard it in December 2025 and it had passed neither chamber as of September 10, 2026. Enactment would add an income stream that no current master valuation contains.
• Our tape. The IA Composite stood at 100.271 on September 8, 2026, +0.29% over a year, with no music-royalty series in it. We will add one when a daily-priced instrument exists to build it from.

Every reading above carries the date it was taken. Check each against the source of the day before you act on it, and treat a figure whose date has moved as a new question rather than as a confirmation.