YouTube says it can’t always figure out who to pay for a song. Should that shape EU copyright legislation -- and the rates it pays to all rightsholders?
A few weeks ago, YouTube CEO Susan Wojcicki wrote, in a Financial Times op-ed, that the streaming service couldn't accurately pay royalties to all the rightsholders of "Despacito." Her point was that a proposed revision of the EU Copyright Directive would expose YouTube to legal liability for hosting music whose rightsholders can't be identified, but this seemed like quite a claim. How complicated could it be to track down the owners of the composition underlying last year's biggest hit?
As it turned out, figuring out which of the song's owners wasn't getting paid, and why, was more complicated than I could have imagined.
For weeks, lobbyists for Google, YouTube's parent company, have been giving presentations to EU policymakers in Brussels making the same point Wojcicki did in more specific terms: Although YouTube knows who wrote and publishes the composition, it can't identify the right entity to pay for 12.5 percent of the rights in some territories. And if the Directive passes in its current form, Wojcicki wrote, "we might have to block videos like this to avoid liability."
By the standards of today's hits, some of which involve a half dozen writers, "Despacito" is relatively simple: It was written by Luis Fonsi and Erika Ender, who are both published by Sony/ATV and each credited with 37.5 percent of the song, and Daddy Yankee, who publishes the remaining 25 percent through his company Los Cangris Inc., which has an administration deal with Sony/ATV outside the U.S. (The remix also credits other writers, including Justin Bieber, but YouTube's issue is with the original.) So I started by calling Sony/ATV.
A Sony/ATV spokesperson told me that the problem doesn't lie with the royalties it collects. But not all publishing royalties flow directly to Sony/ATV, which collects mechanical royalties and some performance royalties in Europe through SOLAR, its joint venture with the collection societies PRS for Music and GEMA. Some royalties flow to songwriters from their U.S. collection societies (such as ASCAP or BMI), which have reciprocal agreements with their foreign counterparts. And YouTube says it doesn't always know what rightsholders have authorized which foreign entities to collect money on their behalf. "We don't know who the agent is for this 12.5 percent," a YouTube spokesperson told me. "In other words, we know the authors, but we do not know the licensing entities that own or administer rights on behalf of the authors."
In Rumsfeldian terms, this is a known unknown -- YouTube knows it can't identify payment data for the entire composition. But the company also faces unknown unknowns, such as user-uploaded songs without any authorship information at all. Right now, the service can stream first and ask questions later because it benefits from safe harbors -- under the Digital Millennium Copyright Act in the U.S. and the Electronic Commerce Directive in the EU -- that allow it to use material uploaded by users as long as it responds to takedown notices. But Article 13 in the new version of the Copyright Directive -- which the branches of the European Union are now negotiating to finalize -- would change that. Depending on how the final legislation looks, YouTube could have the same legal responsibility as online services like Spotify.
That would come as good news to most rightsholders, who believe that the current system of safe harbors gives YouTube a significant negotiating advantage when it comes to licensing content. Although YouTube does a decent job of identifying music and blocking what it's not authorized to stream, the scale at which it operates means that plenty still gets through. The time and expense of filing takedown notices gives YouTube some advantage when it comes to setting prices and conditions for use.
It took a few days of phone calls to figure out who wasn't being paid -- and, even now, it's hard to be sure. (YouTube said it couldn't tell me who owned the portion of the composition it couldn't identify, since it gets that information from rightsholders themselves.) But after a few days I received a phone call from Daddy Yankee's lawyer, Edwin Prado, who told me it appears that Daddy Yankee isn't getting paid on some U.S. YouTube uses for part of the composition that he controls. The problem may come from YouTube not having the correct information in the U.S. -- Los Cangris is a small company run partly by Daddy Yankee's family -- which could have created an issue in other territories as well. Prado is still looking into this.
In Wojcicki's op-ed, and in Brussels, YouTube implies that its problem with "Despacito" shows that publishing rightsholders need to get their act together when it comes to providing accurate data. But Prado didn't see it that way. "That's an excuse," he told me. YouTube knows where to license the rights to "Despacito" because the service approached Los Congris for a promotional license, Prado says. (YouTube confirmed this.) "How does the right hand not speak to the left hand?"
Both publishers and YouTube seem to believe that the service's problems accounting to "Despacito" rightsholders is damning for the other side. The truth is that no one comes out of this looking too good. YouTube is set up to depend on information submitted online, but executives there could have brought the problem to the attention of rightsholders before EU politicians. But the issue isn't as simple as Prado suggests, either: Knowing who wrote a song isn't the same as knowing whom to pay for the rights in Europe. YouTube relies on rightsholders to supply correct publishing data, so "the ability to find rights data is a reflection of the capability of the industry to understand and identify content it licenses," a spokesperson told me. Shouldn't rightsholders just supply accurate information along with the compositions they license?
Sure. But when it comes to popular songs, in most cases YouTube already has the licenses it needs, even if they don't come with all of this information. There's a big difference between now knowing how rights are divided and not having them at all. Sure, it's difficult for YouTube to get all the licenses it needs. But is that an unreasonable price to pay for streaming every music video ever created, plus thousands of clips recorded by iPhones held aloft during concerts? Especially when being able to do so gives YouTube a substantial advantage over companies like Spotify, which can only offer officially released tracks and not those concert clips?
For all of their problems, music publishing companies have a powerful incentive to get their data in order: They need to get paid. Under the current safe harbors, YouTube's incentives aren't nearly as obvious -- as long as it responds to takedown notices, it's in the clear. It has made substantial investments in content-identification technology, presumably both to help rightsholders and to improve its own business. But it has little legal obligation to do so.
Both sides cast this as a black-and-white issue, but the final legislation may offer more nuance. The version of the Copyright Directive passed by European Parliament doesn't offer YouTube much room for error, but the one passed by the European Council is a bit more forgiving -- it requires YouTube to take steps to avoid copyright infringement but also allows some margin for error. (The final legislation will emerge from negotiations that start with these, as well as another version, from the European Commission.) It will be difficult to negotiate a compromise that works -- some rightsholders are already worrying that the EU Council version of the legislation could create a new version of safe harbor that's more limited but also offers more legal clarity.
The issues around accounting for "Despacito" certainly proves that the publishing business needs to be more transparent -- especially online and internationally. But it may not say as much about Article 13 as either side would like.