Introduction
In October 2015, following a phase II investigation, the European Commission published its decision to approve a JV between three of the largest collective management organisations (CMOs) in Europe, PRSfM (UK), GEMA (Germany) and STIM (Sweden).[1]M.6800-PRSfM/ STIM/ GEMA/ JV.
As in other recent music industry mergers, the Commission investigated whether the combination of the parties’ repertoires would enhance their negotiating power when licensing digital service providers (“DSPs”) that commercialise music, such as iTunes and Spotify.[2]The Commission also separately investigated concerns that the JV might affect copyright administration services provided to the music publishers. The parties ultimately agreed behavioural commitments … Continue reading In three previous mergers in the music industry (Universal/BMG, Sony/Mubadala/EMI Music Publishing, and Universal/EMI) the Commission concluded that the repertoires of different licensors were competing with one another, and so required significant divestments – a standard analysis of horizontal substitutes.[3]M.4404 – Universal/BMG Music Publishing, M.6459 – Sony/Mubadala Development/ EMI Music Publishing, M.6458 – Universal Music Group/EMI Music.
The Commission’s conclusion in the present case, that larger CMO repertoires are not able to command higher royalties, is in marked contrast to those in its recent investigations in this sector. This was attributable to the complementary nature of the Parties’ repertoires, which in turn provided a clear pro-competitive incentive for a JV that would allow them to license those rights more effectively and efficiently.
The Commission’s Assessment
The framework explored by the Commission was drawn from labour market economics, considering negotiations between labour unions and firms. In stylized labour markets, where every worker performs the same task (i.e. workers are substitutable), they have more bargaining power when they negotiate collectively through a union than when they negotiate individually. This is because when workers are substitutable an additional employee adds less value than the previous employee. As a result, the average damage to a company per worker lost increases as more workers threaten to stop working collectively (as part of a union).
However, if the starting assumption of worker substitutability is reversed, and if workers were instead complementary, then the result would also be reversed: the aggregation of workers would result in less bargaining power for the union. To continue the labour markets illustration, different types of workers might be complementary if a firm needed both plumbers and electricians to operate a plant. In that case, there are increasing benefits to hiring an electrician after having hired a plumber (since only by hiring both can the firm operate). For a firm that has hired a plumber and is negotiating with an electrician, this means that the marginal benefit of the electrician is higher than the average value of the two workers together. In this situation a union made up of plumbers and electricians would have less bargaining power than the electrician alone.
Given its concern that the aggregation of the parties’ repertoires could lead to stronger bargaining power and, accordingly, higher royalty rates, the Commission tried to assess whether large repertoires achieved higher royalty rates than smaller repertoires. The Commission used two approaches: looking at contractually agreed royalty rates, and using data on actual royalties paid by DSPs to CMOs, which were calculated as a proportion of the DSPs’ retail revenues related to each CMO’s repertoire.
While the Commission found no indication that larger CMOs had negotiated more favourable terms than those with smaller repertoires, there are a number of challenges that affect the reliability of this type of analysis. In particular, the calculated royalty rate paid by the DSP to different CMOs might reflect differences in the popularity or commercial appeal of different repertoires, as well as the DSPs’ independent retail pricing decisions.[4]Royalty rates are often calculated as a proportion of DSPs’ retail prices, and may be further complicated by minima per stream or per download, or periodic lump sum payments; DSPs might choose to … Continue reading If these effects were material, it would be wrong to simply look for causation from repertoire size to effective royalty rates, when an apparent relationship (or lack thereof) might have been driven by factors far removed from the bargaining power of the CMOs. The Commission acknowledged some of the challenges in this empirical exercise, and may build on the experience in future investigations.
Having found no signs of substitutability through these empirical exercises, the Commission then investigated other features of the market in order to corroborate this finding. The Parties’ repertoires were widely considered to be individually, pre-merger, essential repertoires for DSPs wishing to operate successfully in Europe. Consistent with this, the Commission observed that DSPs typically licensed all repertoires, not merely a selection of repertoires as might be expected if repertoires were substitutes (as in the case of different brands of milk in a supermarket, where it is not necessary to stock every single brand in order to have a credible offering as a supermarket).
However, the implications of the DSPs’ decision to stock all repertoires, while intuitive in this case, might need to be more carefully considered in other contexts. It cannot be presumed that where a supplier chooses to stock all (or even a large selection of) brands this is proof that those brands are complementary. For example, if an online retailer choses to stock 10 brands of USB stick for sale, this would not mean that those brands of USB stick were complementary. Rather, this might reflect the low incremental costs of stocking additional brands online, and while an online retail offering might be more credible by offering 10 brands, there may be many candidate brands that could contribute to that level of credibility. In each case it will be necessary to consider the relevance and balance of these different effects.
The Commission also sought to distinguish the case of CMO repertoires from earlier cases involving the aggregation of music publisher or recorded music company repertoires, highlighting the distinct linguistic and cultural bases of the parties’ repertoires, as well as the non-discrimination objectives of the CMOs.
Conclusion
The Commission’s preparedness to assess any hypothetical theory of harm against the available evidence must be welcomed. In doing so, it acknowledged the challenges in some of the empirical assessments of evidence that were undertaken. If intuitive, high-level empirical exercises are significantly confounded by complicating factors, then it is necessary to properly account for these factors in order to achieve a reliable assessment. The Commission sensibly considered a wide range of factual evidence in its assessment, looking not only at direct evidence of differences in royalty rates, but also the testable predictions of its model, such as licensing behaviour by DSPs and CMOs.
________________________
To make sure you do not miss out on regular updates from the Kluwer Competition Law Blog, please subscribe here.
References
