Rejoinder on “Frontiers: The Interplay of User-Generated Content, Content Industry Revenues, and Platform Regulation: Quasi-Experimental Evidence from YouTube”

1. Introduction

We thank Tushnet (2023) for the thoughtful commentary on Wlömert et al. (2023), and we greatly appreciate a legal expert’s reflections on our results and implications. In this rejoinder, we would like to address three issues against the background of this commentary. First, we examine the relevance of the study’s setting for questions concerning platform regulation. Second, we discuss the economic implications of “safe harbor” provisions for the bargaining power of user-generated content (UGC) platforms vis-à-vis rightsholders. Third, we reflect on the ongoing debate about whether UGC platforms stimulate or cannibalize demand.

The argument that guides the interpretation of the results from the quasi-experiment and the implications that can be drawn from the study is that the current “safe harbor” provisions, which dictate the procedures that platforms must follow to limit their copyright liability, increase their bargaining power vis-à-vis rightsholders. This increased bargaining power is reflected in substantially lower payouts to rightsholders by UGC platforms compared with pure music streaming services such as Spotify or Apple Music.

2. Platform Regulation

Regarding the question how the setting examined by Wlömert et al. (2023) relates to platform regulation, Tushnet (2023) notes that “During the dispute with GEMA, Google used the same mechanism to prevent GEMA-associated music from being available in Germany at all. What the authors are actually testing, then, is the difference between blocking—carried out by Content ID—and monetizing—also carried out by Content ID. That doesn’t involve a safe harbor.”

We agree that the Content ID system¹ was in place in the pre- and posttreatment periods in our study and that the quasi-experiment that we study does not involve a move from a non–safe harbor environment to a safe harbor regime (or vice versa). At the same time, however, we believe that valuable insights can be gained by comparing the situation in which no UCG content was available (our pretreatment period) to the situation in which UGC content is available and artists and labels receive some compensation through Content ID (our posttreatment period). In particular, it seems reasonable to assume that under a very strict regulatory regime (e.g., platforms would have to clear rights for all UGC that is uploaded), less content would be available (similar to our pretreatment period). In contrast, in a very lenient regulatory regime (e.g., where platforms are not responsible for the content that is uploaded), much more content would be available because platforms would have no incentive to restrict the supply of content. The latter bears similarity with our posttreatment period because of the strong increase in supply of UGC content.

Hence, our analysis of how the sharp increase in the supply of UGC, enabled by the agreement between GEMA and YouTube, affects demand in other channels can be used to learn how regulations that affect the supply of UGC in these markets will affect demand in other channels. We believe that this is very important, specifically because the payments differ strongly across platforms and channels, and we will discuss this heterogeneity in payments in the next section.

3. Bargaining Power of UGC Platforms

Tushnet (2023) questions “that legal “safe harbors” are what enable YouTube to negotiate payments for less than Spotify does for music, even though YouTube has a lot of content that doesn’t rely on major labels’ music—or any music at all—and still drives lots of engagement and revenue.”

All available evidence that we are aware of points to a substantial and significant difference in the payments per stream that different providers make to artists and labels, to the extent that, for example, the payments from YouTube Content ID are only a fraction of what other services such as Spotify, Apple Music, or Deezer pay per stream.

Table 1 summarizes data that were collected for the year of 2019 and reports the average revenue per stream (Trichordist 2020), and it indicates that, although Content ID accounts for 51% of all streams, it accounts only for 6.4% of payouts. The International Federation of the Phonographic Industry (IFPI), an industry association that represents the music industry globally, has come to similar conclusions by comparing the revenue that services generate for the industry per user, noting that while Spotify paid record labels US$20 per user in 2017, YouTube is estimated to have returned less than US$1 per music user (IFPI 2018, p. 27). Our discussions with music industry executives also confirmed the validity of the numbers reported in Table 1. Music industry representatives have thus been worried about this mismatch and labeled it the value gap, that is, “the mismatch between the value that some user upload services extract from creative content and the revenue returned to the creators” (https://www.ifpi.org/rightsholders-unite-in-calling-for-an-effective-solution-to-the-value-gap/).

Table 1. Streaming Market Global Shares 2019 (Selected Services)

Why are payouts from Content ID lower than, for example, from Spotify, Apple Music, or Deezer? Of course, we cannot know for sure, but we believe that there is a strong economic argument that safe harbor regulations give platforms bargaining power over music labels and artists. This argument is also made forcefully by Liebowitz (2018), who argues that because operators of UGC platforms can provide their users with access to copyrighted materials without paying copyright owners for those materials, the economic logic predicts that these sites will be unwilling to pay a full market price equivalent to what they would be willing to pay if they had to obtain permission from copyright owners before providing their users with access to that material. This gives UGC platforms a bargaining advantage when they negotiate prices for permission to use copyrighted works on their sites, which UGC platforms want because the permission still has some marginal value to them. Liebowitz (2018) provides additional analyses in support of this argument.

This argument is also reflected in the U.S. Copyright Office’s report in Section 512 of Title 17; for example, on p. 82, the report summarizes the concerns by music industry representatives that draw a connection between the bargaining power of UGC services and Safe Harbor regulations (U.S. Copyright Office 2020). Although many platforms (e.g., YouTube) have evolved their business models since then, for example, by placing stronger focus on paid music subscriptions, the discussion remains relevant in the industry (Blistein 2021).

Even though we obviously cannot rule out the possibility that other factors may lead to the low payouts from Content ID, we believe the economic logic is strong enough to make the arguments outlined above a plausible explanation. Although we have the impression that music is very much in YouTube’s DNA (e.g., most of the most frequently viewed videos are music videos, see https://en.wikipedia.org/wiki/List_of_most-viewed_YouTube_videos) we agree with Tushnet (2023) and note that YouTube has a wide range of video content: not only music as the pure-play music services like Spotify. As a result, YouTube is less dependent on licensing deals with rightsholders, which we expect will further increase the bargaining power of UGC platforms.

It should also be noted that not only UGC platforms incur transaction costs in processing takedown requests, but so do rightsholders who bear the burden of identifying their content and notifying UGC platforms of copyright infringement. Therefore, these transaction costs for rightsholders constitute another important source of bargaining power for UGC platforms.

In summary, we believe that the high share of demand (51% of all streams) versus the low share of revenue generated by this large number of streams (6.4%) reflects this substantial bargaining power of UGC platforms. In other words, we think that it is reasonable to assume that the potential fallback or outside option, provided by safe harbor regulations, enables platforms to negotiate a license that has favorable terms for platforms. Furthermore, YouTube’s large share of the music market justifies a detailed analysis of the economic impact of this particular platform to inform the regulatory debate.

4. UCG’s Heterogeneous Effect on Demand in Other Channels

An important question that follows from the observation that UGC platforms pay less per stream than other services is whether and for what content UGC platforms cannibalize demand in other channels with higher payouts. Regarding this point, Tushnet (2023) suggests “that if the history of collective licensing is any guide, it will be the less-successful creators whose interests are sacrificed, deprioritizing discovery of music and further rewarding those who already are successful.”

We thank Tushnet (2023) for contributing this valuable historical perspective, and our findings agree with this historical assessment. In Wlömert et al. (2023), we quantify how the availability of UGC after the GEMA settlement stimulates demand on other streaming channels for most songs and we show that cannibalization occurs for new releases and hit content. In contrast, most content (i.e., from smaller, less popular artists, older songs) benefits from the availability of UGC content.

However, because of the demand distribution that is heavily skewed toward very successful artists and tracks, the overall industry-wide revenue effect is negative. This finding also implies that a stricter regulatory regime would likely benefit established artists at the expense of less-established artists, in line with the arguments of Lunney (2020). We agree that this tradeoff should be considered by policymakers in their evaluation.

The differentiated results that we provide allow rightsholders to use this information for their individual negotiations with platforms. In this way, we provide transparency in a market that is currently facing a relatively new strong UGC platform: TikTok. Based on our empirical results, we would expect that in particular small and niche artists should ensure in potential negotiations that their content remains visible on TikTok due to likely beneficial spillovers to other channels.²

5. Conclusion

We are grateful to Tushnet (2023) for her detailed and thoughtful comments on our paper (Wlömert et al. 2023) and to Olivier Toubia for inviting us to write this reply that allows us to further explain our perspective on the interpretation of the results of the quasi-experimental study. We hope that the paper as well as this interdisciplinary discussion will stimulate further research in this important area of copyright and platform regulation, that needs to balance the interests of copyrights holders, platforms, and consumers.