The statement arrives months after the quarter it describes. It runs to forty pages and shows a figure with no visible relationship to how often the song was played. A singer whose track spent the spring on playlists opens the file and finds a balance that would not cover a month of rent, printed beside columns labeled mechanical, performance, distribution fee and unrecouped. She asks her manager what the columns mean, and her manager calls the label, which sends a second document, just as long, answering a slightly different question.
Nothing on those pages is an error. That is what makes the arrangement so hard to argue with.
Any explanation starts with an awkward fact. A single recording carries two separate copyrights, and the money for each travels through different pipes. The composition is the song itself, the melody and words that a publisher administers on the writer's behalf; the sound recording, usually called the master, is the particular performance captured in the studio, and it almost always belongs to whoever paid for the session. One stream pays both. Two payments are collected by different organizations, calculated on different formulas and delivered on different schedules, which is the first reason an artist's statement looks nothing like a listener's intuition about fairness.
Performance royalties are the half that reaches songwriters. Performing rights organizations license the right to play a composition in public, then divide what they collect between a writer's share and a publisher's share; the writer's share is legally the writer's and cannot be signed away. What remains is the publisher's share, and it goes to whoever controls the copyright. For a songwriter who signed a publishing deal at twenty-two with no lawyer in the room, that is frequently not the songwriter. A hit that plays in supermarkets and sports bars for a decade can generate steady income for a company that had nothing to do with writing it.
The split sounds technical. It decides who gets rich.
Mechanical royalties cover the reproduction of a composition. That is why a songwriter once earned per CD pressed and now earns per stream. In the United States, the Music Modernization Act replaced a patchwork of individual licensing negotiations with a blanket license and a nonprofit collective to administer it, and the rates themselves are set by a federal tribunal called the Copyright Royalty Board. Sync licensing, the use of a song in a film, a game or an advertisement, remains the one corner where a publisher and a writer negotiate a price freely, and for many working songwriters it is the difference between a career and a hobby.
Money also thins out at every border it crosses. A publishing administrator takes a cut to register and collect, a sub-publisher in each foreign territory takes another, and the local collecting society deducts its own administrative fee. Each layer is defensible on its own. Stacked, they can consume a meaningful share of the earnings before anyone writes a check, and the writer personally chose few of those intermediaries.
The master side is where the arithmetic turns aggressive.
A streaming service pays the label, and the label pays the artist under a contract signed when the artist had the least bargaining power they will ever have. Royalties are calculated against a defined base, and that base carries deductions: packaging, free goods, promotional copies, a distribution fee. Advances are recoupable, so the artist sees nothing until the label has recovered what it spent on the record. Deals that hand the label a slice of touring and merchandise revenue exist because masters alone rarely repay the investment, and because the label knows the artist needs that income to finance the next album.
An advance is not a payment. It is a loan against a future the lender also controls.
Everyone else on the record sits at a different distance from the money. A producer with bargaining power negotiates points on the master and collects for years, while a producer without it takes a flat fee and vanishes from the ledger. Mixers, engineers and session musicians are almost always work-for-hire, paid once, with no claim on the recording. Union session players do receive small residual payments through a special fund, an exception that illustrates how unusual the arrangement is. A backing vocalist who made the chorus work gets scale and a line in the credits; a guest artist singing the hook on someone else's single takes a flat fee, occasionally a point, and no share of the composition unless the writers agreed to that in advance.
One structural fact shapes every negotiation above. Streaming services generally commit a fixed share of their revenue to rights holders, so the total pool grows with subscriptions and advertising, not with listening. More plays do not create more money; they divide the same money more finely, and a catalog that expands every week mechanically dilutes what any single stream is worth. Artists experience this as an unexplained decline in per-play rates. The explanation is arithmetic. How the pool is divided is a separate fight. Under the pro-rata model the services use, every subscription dollar goes into one pot and is paid out in proportion to total plays, which favors an artist with broad but shallow listenership over a niche act with devoted fans; a user-centric model would divide each subscriber's fee only among the artists that subscriber actually played, paying small acts more and superstars less. That is the reason the platforms have shown no urgency about adopting it.
Then there is the problem of identification. Payments travel on identifiers: a code for the recording, a code for the composition, and a set of writer splits agreed at the time of release. If a split sheet is never filed, or a distributor misspells a name, or a foreign society cannot match a work to its owner, the money is collected and then held. This unmatched pool is the industry's black box, large enough that collecting societies now staff matching teams whose entire job is to work out who a payment belongs to. Small errors compound: a missing code can strand a payment for years, and by the time it is matched, the writer may have left the business. A songwriter cannot be paid for a song the system does not know they wrote.
Every participant is responding rationally to the incentives the structure creates. Labels want advances recouped and catalogs that keep paying, which is why they buy proven songs sooner than they develop unproven writers. Streaming services want the lowest sustainable per-stream cost. Publishers want long control of copyrights, and they acquire them from artists who need cash now. Songwriters hold the weakest hand in the room, and it strengthens only when one of their songs becomes a hit. Older anomalies sit alongside the new ones. American terrestrial radio pays songwriters and publishers for the composition and pays performers and labels nothing, a carve-out that has survived decades of lobbying and turns broadcast airplay into promotion rather than income; digital radio pays both sides, at a statutory rate. The inconsistency is not a design flaw. It is the residue of who had lobbyists in the room when the rules were written.
Changing any of this means changing rules fixed in legislation, consent decrees and standard contracts. Statutory rates could rise; consent decrees governing the largest performing rights organizations could be revised or retired. European regulators have already imposed transparency requirements obliging services and labels to explain how a royalty figure was reached. Requiring writers to register their splits at the moment of release would shrink the black box more than any rate increase, and it is the reform with the least glamour and the most effect. None of it is impossible. The question the industry has not answered is whether its problem is really about how much money moves through the system, or whether it built a set of pipes that carry cash faster than they carry the paperwork identifying whose cash it is.