The Copyright Society of the USA’s Annual Meeting took place over the past couple of days in Toronto. One of the many interesting sessions was one on music licensing, where we from the U.S. got to learn that Canada has just as much of a “sausage factory” as we do when it comes to the legal framework for digital music, just with different ingredients and different factory processes (no puns please about “prah-cess” vs “proe-cess”).

We also got to learn that the Music Modernization Act (MMA), having passed the House in a rare show of bipartisan unanimity, is facing some concerns on its journey through the Senate. Some of these concerns are the same ones I raised back in February about the structure of and incentives around the Mechanical Licensing Agency specified in the bill.

To recap: the MMA that’s currently making its way through the Senate is a rollup of three separate pieces of legislation that would reform statutory music licensing in different ways. The biggest of these — the one that was originally called the Music Modernization Act — purports to fix various problems with reproduction and distribution (so-called mechanical) licensing for musical compositions in interactive streaming services such as Spotify, Apple Music, and YouTube Music.

These services have enormous catalogs of music tracks, and they ingest new tracks at the rate of tens of thousands every day. Currently, the services have to clear the rights on underlying compositions for every track individually by filing so-called Notice of Intentions (NOIs) with the rights holders (unless they have deals in place with the rights holders already). This involves figuring out which compositions underlie music tracks and disbursing royalties to rights holders.

Music services get data feeds from record labels (and indie aggregators), not music publishers or songwriters. Information about the compositions that underlie sound recordings is sometimes inaccurate, incomplete, or contested. This makes the task of matching recordings to compositions complex, grungy, and open to “garbage in, garbage out” errors. Most music services outsource their entire mechanical licensing functions to third party providers such as the Harry Fox Agency (HFA) and Music Reports. Regardless, the risk of data errors lays the services open to potential allegations of copyright liability, and indeed various lawsuits have been filed.

The MMA proposes to fix this problem in two ways. The simpler one is to replace the composition-by-composition licensing requirement with a blanket license, which exists for various other types of digital music licensing (such as performance licenses for noninteractive webcasting, as offered by the likes of iHeartRadio and Pandora). The NOI system was designed when the primary reason for obtaining a mechanical license for a musical composition was to record a cover version of it; the system worked fine for that. But for interactive streaming music services that purport to have “all music” in their catalogs, one-composition-at-a-time licensing is a needless nuisance, and the blanket license is a no-brainer.

It’s the other provision that is now drawing some concern. It establishes a single Mechanical Licensing Agency that would process music composition licensing for all of the interactive streaming services. This would be a nonprofit agency appointed by a process run by the U.S. Copyright Office, overseen by an advisory committee, and subject to a review every five years.

The concern is about the structure for the Mechanical Licensing Agency. It’s not specified in much detail, and its oversight is minimal. It would be a big gain in efficiency — in effect, a single agency would do what each interactive music service (with largely identical music catalogs) has to do separately on its own today. But it would replace a market where several services compete for this business with a nonprofit monopoly.

Different stakeholders I’ve spoken to about this have widely divergent preconceived notions about how the Agency would be run and who would do the actual work. The funding for the Agency would come from fees paid by the music services, which are not the entities that have incentives to ensure accuracy and accountability. Companies like HFA and Music Reports would not necessarily be motivated to contribute the considerable expertise and infrastructure they have built up over years of performing this function.

In short, the MMA is an opportunity to institutionalize accuracy and accountability into a process that has been rife with black boxes and various flavors of educated guessing; but the structure of the Mechanical Licensing Agency risks squandering that opportunity in favor of a scheme where the temptations will be to minimize costs and sweep the data problems under the proverbial rug instead of fixing them. I’ve suggested a couple of ways to improve accuracy and accountability within the legislation’s existing framework.

Stakeholders — and apparently Senators on the Judiciary Committee that is responsible for this bill — are beginning to share this concern. And it’s not just the rights holders who are starting to complain. On the music licensing panel at the CSUSA Annual Meeting, Charlie Sanders of the Songwriters Guild of America raised the issue. But so did Sarah Rosenbaum of Google — who, notably, used to work at Music Reports. As she explained it to me, Google is concerned that when digital services (like the newly launched YouTube Music) are accused of offering material without proper licenses or not paying royalties properly, Google tends to get bad PR no matter whose fault it is.

Whatever the motivations, it’s nice to see that the right concerns are being raised in the right quarters about the MMA. It would be good to improve the prospects that the Mechanical Licensing Agency will bring much-needed accuracy and transparency and help the music industry clean up its various rights data, adding to the many other worthwhile aspects of the bill in its current form.