The Music Modernization Act (MMA) has been getting traction in the United States. It was introduced into the Senate two weeks ago, and it has become part of a larger package of music copyright legislation which has become known as the “Music Bus” bill. More and more organizations have announced their support for the MMA. Many industry discussions are about improving the draft legislation, on the presumption that it will pass.

One such discussion took place recently between officials at the US Patent and Trademark Office — which advises on copyright issues for the Executive branch of government — and myself. I gave them a description of what I thought was wrong with the MMA; they asked me how I’d propose fixing it in the legislation. I gave an off-the-cuff answer; later on I realized that my answer wasn’t very well thought out. So here’s an attempt to improve it.

I described the substance of the MMA a few weeks ago, but briefly (and in relevant part): the MMA streamlines the process by which on-demand music services — Spotify, Google Play, Apple Music, Rhapsody, Tidal, etc. — obtain so-called mechanical licenses to stream musical compositions and pay appropriate royalties.

The current law guarantees a music service the right to do this, as long as it issues some paperwork called Notice of Intentions (NOIs) to composition rights holders for each track in its catalog, or has a licensing deal in place with the rights holder already. For a big service like Spotify, that means having to check licensing for tens of thousands of new tracks every day — a major administrative hassle, not to mention a risk of legal liability if the paperwork isn’t done correctly or in a timely manner. Now that on-demand streaming constitutes the majority of music industry revenue, this has become a very big deal.

Part of the administrative hassle as well as the source of many errors is the need to figure out which musical composition underlies each sound recording and who owns the rights to that musical composition. (Or which compositions, if the track is a mashup.) Music services get information about sound recordings from record labels and other recorded music aggregators, but they often don’t get the information about underlying compositions. So they hire third party agencies such as Harry Fox Agency (HFA) and Music Reports Inc. (MRI) to do that job — or in the case of Google, they buy an agency (RightsFlow) outright.

The MMA promises to make it much easier for music services to get mechanical licenses and pay the royalties on them. First, it provides a blanket license, meaning that the services will no longer have to check licenses and possibly file paperwork for each track they ingest into their catalogs. Second, it calls for the establishment of a single licensing collective that will process the mechanical royalties; this includes matching compositions to the sound recordings played by the music services’ users, finding the rights holders to those compositions, and paying them accordingly. The licensing collective will be a nonprofit organization selected by the U.S. Copyright Office, funded by administrative fees from music services, and subject to review every five years.

My views on the MMA were concerned with the licensing collective. Processing mechanical licenses from on-demand music streaming — which amounts to hundreds of billions of transactions per year, and growing fast — is a very hard job, particularly because of the problem of matching sound recordings to compositions. (In that sense, it’s harder than the job that SoundExchange currently has in processing royalties in sound recordings from digital radio services like Sirius XM, Pandora, iHeartRadio, and AM/FM stations’ Internet streams.) Mechanical license administration is a small market consisting of a handful of for-profit companies that compete with one another and continuously evolve their processes to increase their accuracy and efficiency.

The MMA would replace this competitive market with a single nonprofit entity that would only be subject to lightweight oversight. Any rights holder that’s not satisfied with the results would have no legal recourse other than through the dispute resolution scheme defined in the MMA and Copyright Office’s 5-year review of the licensing agency; rights holders wouldn’t be able to sue music services for unpaid mechanicals, as they are doing now. Furthermore, although the MMA requires the mechanical licensing collective to make a database of musical compositions and their matching recordings available to the public online, its processes for distributing royalties to the appropriate parties would be opaque.

The existing for-profit mechanical licensing agencies would have incentive neither to apply to become the MMA’s licensing agency nor to lend their considerable expertise to help ensure that it’s done right. As a result, the licensing agency under the MMA could be motivated merely to do the minimum job necessary to earn re-certification at the next five-year review.

That’s the problem. A mechanical licensing agency that does a mediocre job is likely to disadvantage long-tail and older songwriters and music publishers. More generally, it will mean lost opportunities for improving accuracy and efficiency through better data and technology, and likely a diminution of the innovative efforts in that regard from places like the Open Music Initiative and various startups.

(Having said that: SoundExchange has stated informally that it is interested in taking on the role. It has an excellent reputation across the industry, and it is gaining experience with mechanical licensing through its acquisition last year of the Canadian mechanical licensing agency CMRRA. SoundExchange was also designated by the Copyright Office and is subject to similar oversight.)

If that’s the problem, then what is the solution?

There’s no doubt that replacing the existing composition-by-composition system with a blanket license is a huge gain in efficiency — as is eliminating the massive redundancy of every music service having to process its own mechanical licenses on more or less the same body of material (even if it’s done through a handful of outside agencies). The question is how to replace the incentives that the single nonprofit agency lacks compared to a competitive market of for-profit entities.

One answer is to hold the mechanical licensing agency accountable by conducting regular independent audits for accuracy and efficiency in royalty processing. It’s possible to require this within the bill’s existing governance framework, with modest added language.

The bill currently establishes an operations advisory committee of at least six people, with equal representation from the musical composition rights holder (songwriter and music publisher) and music service provider constituencies. This committee’s job is “to make recommendations to the board of directors concerning the operations of the mechanical licensing collective, including the efficient investment in and deployment of information technology and data resources.” It also calls for the Register of Copyrights (the head of the Copyright Office) to appoint a best practices working group, with representation from the mechanical licensing collective, copyright owners, digital music services, and performing rights societies (like ASCAP and BMI), to “consider and advise on best practices to minimize the incidence of unidentified and unmatched musical works and facilitate and encourage the exchange of ownership information and prompt access to such information by and among such parties.”

That’s a good starting point, but it falls short of real accountability for the licensing agency. The best practices working group should be specifically tasked with defining metrics and criteria that the licensing agency must meet for accuracy of recording-composition matching and rights holder identification, efficiency of royalty distribution, and so on. It should also recommend and vet entities to audit the mechanical licensing agency for compliance with these criteria annually — entities that should have no conflicts of interest with any of the parties. In addition, the best practices working group’s membership should be expanded to include at least one independent, unaffiliated expert whom the Register (i.e., not one of the interested parties) would appoint.

To ensure independence, both the audits and the independent expert(s) could be funded through the Copyright Office instead of through the administrative fees that music services must pay to fund the licensing agency.

The operations advisory committee should then oversee the annual audits and convey the results to the Register of Copyrights, who should publish them and take them into consideration when determining whether to re-designate the licensing agency or look for an alternative. It will most likely take a few years of experimentation to determine the right set of metrics and success criteria, given that the existing mechanical licensing agencies keep their various techniques and success rates closely guarded secrets.

The Music Modernization Act is a great opportunity to solve a vexing problem in music licensing by making it more efficient and less risky. But the opportunity to improve accuracy and accountability in royalty distribution shouldn’t be passed up. The above is just one idea to help ensure that this doesn’t happen.