2017 year's Pros and Cons seminar addressed the issue of Market Definition.
Contributors (from the left): Rikard Jermsten, Ana Sofia Rodrigues, Lars Sørgard, Richard Whish, Sebastian Wismer, Kate Collyer, Amelia Fletcher, Nikodem Szadkowski, Hila Nevo, Marios Iacovides, Lucrezia Busa and Joe Farrell (on the screen).
Summary of the seminar
Richard Whish, Emeritus Professor in King’s College London, welcome all speakers and told that debate on market definition is relevant and timely, as it has passed 20 years since Commission Notice on the definition of relevant market was first released. He recalled the time, when the hypothetical monopolist test was first introduced, many doubts and questions from anti-trust enforcers and lawyers. Over time, the hypothetical monopolist tests have been established and applied by people who are engaged in competition economics or competition laws as an important step in competition analyses. Over time we have realized that defining a market through the hypothetical monopolist test is more about conceptualization of a market. We have also realized that there are formidable problems with applying the SSNIP (small but significant and non-transitory increase in price) test, as markets have become more complex than ever.
Private Labels and Relevant Markets
The first session of this year’s seminar was commenced by Dr Hila Nevo from Tel-Hai Academic College and the University of Haifa, whose main question was how to define the relevant market when private label products were involved. On the one hand, private label products compete with branded products for consumers on the retailer’s shelves, but on the other hand, producers of the branded products and the retailers have a vertical relation, as the branded products producers often produce private label products.
The SSNIP question is whether private label products impose a sufficiently strong competition pressure to induce consumer substitutability. Another question relating to market definition can be which alternative retailers are available for the suppliers, and whether the supplier is able to effectively compete without access to a particular retail outlet.
Dr Nevo pointed out that market definition itself does not determine authorities’ assessment in any mechanistic way, and that one needs to take into account constraints from outside the relevant market, or the degree to which competitive constraints – inside and outside of the market—are more important than others. Market definition is a useful tool to organize our thoughts, but many problems can appear and we have to deal with them case by case.
Nikodem Szadkowski, Deputy Director of Polish Office of Competition and Consumer Protection, commented that market definition becomes more complex when private label products are involved, and that authorities should not apply naïve approach. One example is when branded products and private label products compete in downstream, while there are only few producers who produce both types of the products. If one defines separate markets for the supply of branded products and private label products, taking into account competition between those products downstream, odd results may follow such that very high market shares upstream do not indicate of market power. Therefore it may be helpful to look at suppliers of private label products separately.
He commented also that market definition is still a useful framework, even if market shares may not mean much in a particular setting. Market definition is a framework allowing us to tell coherent stories. There would appear particular problems with complicated markets with vertical integration and differentiated products. But it is a flexible framework that we all are used to. He rounded off with the comment that it is better to make a right decision using clunky market definition, rather than to reach a bad one by dogmatically clinging to “the most rational” definition.
Revisiting the hypothetical monopolist test: the role of common sense market definitions
Amelia Fletcher, professor in University of East Anglia, opened the second session with a comment that “market” is a common sense concept, which also makes economic sense, and that market shares have been serving a relevant proxy for market power. The concern arises because market definition remains highly determinative of case outcomes. In fact, the hypothetical monopolist test is still more often decisive in more complex markets, and in these markets, the test is plagued by major debate and controversy over details of its methodology and application.
Professor Fletcher wrapped up the presentation by two distinctive situations; firstly, market definition and market shares can be used as a preliminary indicator of market power and likely competitive harm. Here, a ‘common sense’ approach to market definition can usefully be employed, albeit cautiously and informed by the hypothetical monopolist test framework. In case of full case analysis and final case decisions, based on detailed empirical evidence and analysis, market power and competitive concerns should be assessed directly, albeit drawing on the hypothetical monopolist test methodology, rather than indirectly via market delineation and following market shares.
Dr Ana Sofia Rodrigues, Chief Economist at Portuguese Competition Authority, took up her comment on Professor Fletcher’s presentation by exemplifying some of the ongoing market definition debates; whether defining the relevant market is a necessary step or direct assessment of competitive effects is sufficient to substitute such step. She pointed out that there are a number of difficulties involved in a full fletched quantitative exercise of hypothetical monopolist tests. Market definition is still valid as conceptual as well as analytical framework for analysis.
Market definition in platform markets
Dr Christina Caffara and Oliver Latham from Charles River Associates tackled the third topic of the conference, “Market definition in platform markets”. In her opening remarks Dr Caffara agreed on the points made by Amelia Fletcher regarding emphasis on competitive constraints and went on to state that the binary nature of market definition is perceived by economists as somewhat problematic since markets are more complicated. There are, as Dr Caffara put it, “shades of grey and degrees of constraint” and so economists seek the kind of analysis that is able to capture that. Dr Caffara stated that there is no other market where this is more true than in digital markets which exhibit specific characteristics that inevitably complicate market definition. Dr Caffara observed that the inevitable default approach that has been opted is to fall back on functional definitions of markets by relying on technical characteristics. According to Dr Caffara, functional definitions may be simple to apply but are misleading and thus inappropriate for defining digital markets. Dr Caffara pointed out the importance of understanding the channels of competition and delved into some of the challenges that make it harder to think about substitution in digital market structures. Against this backdrop Dr Caffara presented other types of economic evidence that might be helpful in assessment of substitution in these markets. Oliver Latham presented the third challenge observed when dealing with digital markets, namely their two-sided nature and the issue of the broader competition between platforms for consumers “attention”. He added that functional differences are a source of differentiations but do not provide that there is a lack of substitutability. Dr Caffara took the case of Android as an example and concluded that in many cases involving digital markets it seems more productive to focus on whether there is a credible theory of harm that is plausible and that one ought to consider whether market definition is critical for proving an infringement in such cases. Digital markets pose a lot of difficulties and market definition is particularly fraught. Attempts to determine whether products are “in or out” will not work so caution must be taken going forward.
Kate Collyer from the UK Competition and Markets Authority (CMA), started off by concurring on the key challenges in defining digital markets identified by Caffara and found the proposed approach to focus on substitution patterns rather than functionality appealing. Collyer went on to raise the question of single-homing and market power in two-sided platform markets and how that is to be dealt with that in market definition: does that lead to a very narrow market definition? Collyer then presented the evidence the CMA relied on in the Just Eat/Hungry House merger case and presented other possible sources of evidence in order to further explore the topic raised by Caffara regarding other types of economic evidence that might be helpful in assessment of substitution in platform markets.
Market definitions in mergers
Dr Lucrezia Busa from the European Commission focused on the pros and cons of using the concept of market definition from the perspective of a competition enforcer. She elaborated on some of the challenges encountered when dealing with this concept in mergers and went on to present how the Commission dealt with such challenges. Busa highlighted some of the Commission’s recent practices involving sectors where market shares were deemed to be of limited informative value in terms of market definition. Busa put forth that market definition and the calculation of market shares have clear benefits and that they are an important screening tool for competition authorities. At the same time, Busa raised the point that it is important to keep in mind that the market definition has an intermediate role and that its interplay with the competitive assessment should be clear. Busa touched upon the role of supply-side substitutability in market definition and referenced the findings of the geographic market definition report conducted by Fletcher and Lyons. Busa then mentioned the Commission’s greater use of isochrones illustrating a more refined approach taken by the Commission in defining geographic markets. In her concluding remarks, Busa noted that the approach to market definition is in need of constant refinement since it is a case-specific and fact-based exercise. Busa then concluded that it is necessary to employ other, more sophisticated analytical tools or frameworks of analysis when dealing with more complex cases in some sectors in order to capture the real competitive dynamics and highlighted the importance of qualitative evidence to underpin findings while at the same time pointing out that quantitative tools and economic analysis are normally selected among those backed by economic literature.
The discussant Dr Sebastian Wismer of the Bundeskartellamt, brought additional perspectives to the points made by Busa from the view of a national competition authority. On the geographic market definition, Wismer elaborated on the approach adopted in the German hospital mergers and also mentioned that recourse to isochrones that account for actual behaviour has also been used method. At the same time Wismer stressed the importance of the market definition’s interplay with the competitive assessment in the area of merger control. Wismer reiterated the point made by Busa that market shares are helpful as a first step in a case in order to gain a holistic view but added that they can also be misleading when account is not taken of the market circumstances of the case at hand. As regards market definition in the presence of indirect network effects, Wismar chose not to focus on whether to define a market at all but rather on the question of how to define a market and in particular how many markets should be defined, one or separate markets. Wismer elarorated on the pros and cons to these two alternative approaches stating that a good rule of thumb would be to define separate markets for each side unless it is a transaction or matching platform. Wismer went on to stress that markets can exist if service is offered free of charge and informed of the latest amendments to the German merger control provisions regarding the merger thresholds. Wismer concluded by giving some thoughts on dealing with innovations by stressing that there are many competitive dimensions beyond price that are of high importance in many cases. Furthermore he concluded by raising the question of whether future markets or innovation spaces should be defined as well as how we ought to proceed if competitive concerns are identified but cannot be related to an existing or future product market.
Making market definition helpful
The last session of this year’s Pros and Cons seminar was by Joseph Farrell, professor of economics at University of California, Berkeley, who spoked through video conference from California. Professor Farrell started off the speech with a rather provocative comment that market definition has become a monopolist and accustomed to a “must-have” status in antitrust or merger control investigations without justified its superior efficiencies. He argued that requirement of defining a market can be harmful. So we (competition practitioners) need to introduce competitive discipline to market definition. This does not mean totally abandoning market definition. But how can we make it more useful?
Market definition is useful as an intellectual exercise to encourage clarity of thoughts. We now understand reasonably and clearly that for unilateral effects of mergers, one should ideally focus on diversion ratios, margins, and efficiencies. Thus if we can directly and perfectly observe diversion ratios and margins, market shares would not have any added value for the evaluation of unilateral effects.
We need to make market definition more supple, in the sense of more flexible and resilient, and more subtle, in the sense of thoughtful and not abandoning totally. The purpose of defining a market is more about identifying participants and less of one true path. We are now aware of a number of known glitches with the hypothetical monopolist test. One of those problems is the narrowest market principle. We need to rethink of it and move away to being vulnerable about it and sometime to downplay relevance of defining relevant market.
In short, we use market definition when it is useful and do not use market definition when it not useful. Sometimes it is sensible to assess competitive effects directly in such way we can improve overall performance of market definition. Professor Farrell wrapped up the speech referring a slogan, “Market definition is most useful when it is easy”. Market definition is not “always” and nor “never”. Let’s use market definition selectively.
Professor Lars Sørgard, Director General of Norwegian Competition Authority, concurred with Professor Farrell’s speech on the relevance of market definition. Professor Sørgard also commented that market definition is helpful for structuring the analysis. It is still useful analysing the Cournot style competition markets, bidding market, or electricity market.
Market definition is, however, an indirect test. As Amelia Fletcher argued, we can go directly to competition effects analysis. Professor Sorgard showed that market definition, if it is done properly, is actually closely related to the UPP (Upward Pricing Pressure) analysis, which is the modern method to directly assess unilateral effects.
A problem with market definition is that one should make up mind deterministically, either inside or outside of the market. He showed some examples that a deterministic delineation of a relevant market led to wrong conclusions. Professor Sørgard sees that we are going forward the right track, taking more consideration to direct assessment of competition effects such as UPP analysis.