How Collaboration Is Stifling Innovation in the Music Streaming Market

The subscription streaming market, dominated by some of the world’s most innovative companies (Spotify, Apple, Amazon, and Google) is surprisingly homogenous: same products, same prices, same catalogs.  At the same time, there are countless other music streaming startups that have “died on the vine,” as Rutgers professor Michael Carrier once wrote.  Why is this?  In a new article, I examine how the major record labels (Universal Music, Sony Music, and Warner Music) have created a market largely devoid of choice by leveraging a contracting framework that was once believed to promote innovation.

Previously, academics touted “braiding,” the interweaving of contract provisions that facilitate information exchange with informal enforcement norms, as a means to promote innovation, especially in a world of heightened uncertainty.  The thinking was that braiding facilitates the development of trust, and as the potential costs of leaving the collaboration rise, the risk of opportunism decreases.  In other words, the theory suggested that relationships can be structured in a manner that is naturally positive for society.  However, that theory is under-explored and assumes away all externalities.  As my article shows, a collaborative approach to stifling innovation can overcome braiding’s pro-innovation bias, with tangible outside effects.  The way that the major labels license streaming services shows how.

The streaming industry demonstrates that braiding, at least when taken out of the bilateral-contract context, enables firms to lawfully accomplish objectives that might otherwise violate antitrust law.  Consider a comparison between the record labels and book publishers.

Spotify was founded to provide music fans with an alternative to piracy. Unlike file-sharing sites, it would pay copyright owners a share of advertising revenue in exchange for a license.  But it needed the catalogs from the major record labels, which together collectively control 75 percent of the recorded music market and nearly all popular music.  Obtaining licenses from  the major record labels to offer music in the United States, however, took two-and-a-half years following Spotify’s European launch.  And, at the labels’ behest, Spotify was forced to develop a subscription offering (rather than offer only a free, ad-supported service), agree to a 55 percent revenue share (backed by large up-front “minimum guarantee” payments), and grant equity in the company on terms highly favorable to the labels.

The book publishers also encountered challenges as their distribution market transitioned to digital.  After Amazon offered popular books well below the publishers’ standard prices, the publishers and Apple agreed on a new retail model that returned to them control over end-user pricing.  Threatening to pull their content from Amazon if it did not agree to the new model, the publishers successfully eliminated the platform’s ability to undercut Apple or any other distributor on price.

Both sets of rights-holders appeared to be fighting for control over innovation and profits in their new downstream markets, but the similarities ended there.  The book publishers and Apple were found guilty of a price-fixing conspiracy.  The labels, however, have adopted a more sophisticated approach to controlling their adjacent distribution market.  They engage in independent, confidential negotiations with the streaming services but “braid” formal information-sharing contract provisions with informal enforcement norms to implement parallel contract terms and privately order the industry.  The result is that each of the popular services offers essentially the same product at the same price, and braiding looks more like a recipe for tacit (lawful) collusion.

The labels’ agreements with the services have been summarized in detail by government officials in the United States and Europe and industry experts.  Each is negotiated individually, but also includes a most-favored-nation provision (MFN) that ensures the label gets the benefit of the highest revenue share any of them can secure.  The contracts also include detailed license restrictions that severely limit what licensees can do with the catalogs. These latter terms effectively require the platforms to seek permission from the labels for any meaningful product change.  Both kinds of parallel provisions implicitly require the services to share with each of the labels any information that sheds light on how the others are acting.  For example, was a better revenue share agreed to?  What product is being approved?

Long-term collaborations (including tacit collusion) are prone to defection, but for informal enforcement norms, such as trust, that raise switching costs.  In the recording industry, trust has grown over years of industry consolidation, coordinated strategy against new technologies (including litigation), and repeat player dynamics on the individual and institutional levels.  Trust is a tool for self-enforcement and erases the need for more explicit coordinated action among the labels, which reap substantial benefits from the ongoing arrangement.

Take money, for instance.  The recording industry collectively earned over $12 billion in 2022.  But the labels are also able to control innovation (whether to products, features, or business models) in the downstream subscription-streaming market.  Incumbents are hamstrung by the terms of their licenses, while startups have struggled to gain traction and investment.  Venture capitalists have shied away from music startups due, in part, to the labels’ high license fees, which are enough to consume a fundraising round, and authority over the company’s product development.  Notably, the areas in which there is meaningful startup activity in online music – user-generated content sites, for example – do not require direct licenses from the record labels (although Universal’s recent dispute with TikTok is likely to shed light on how the record labels are working with those sites).

In my article, I propose solutions that chip away at the braided framework that is enabling the labels’ behavior.  There are existing options in copyright and antitrust law that have already been explored; to those I offer additional ideas that are rooted in tax and transparency.  For example, drawing inspiration from the European practice of imposing levies on large content distribution companies, I propose a penalty tax on the surplus that a record label makes by virtue of the application of an MFN provision.  My goal is to create incentives for defection that encourage the labels to break out of their collaboration.

This post comes to us from Rachel Landy, a visiting assistant professor at Cardozo School of Law. It is based on her recent article, “Downstreaming,” available here.