Relevant Markets

Photograph courtesy of Alexander Antropov via Pixabay

e-Conomics harbours an expertise in methodologies for defining relevant markets. Most notably, founding member of the e-Conomics network, Nicolai van Gorp, is widely recognised as an expert in the assessment of relevant markets. Numerous parties have employed his expertise over the years.

Nicolai advised the European Commission on the revision of the 2014 Recommendation on Relevant Markets. This advice consisted of an extensive study analysing relevant markets in telecommunications at wholesale and retail level. Furthermore, the study concluded which of these markets (still) warranted ex ante regulation. The European Commission’s Recommendation of Relevant Markets 2014 adopted the conclusions of the study one-to-one.

Furthermore, e-Conomics has consulted on relevant market issues in various legal cases. The most recent cases related to the definition of relevant markets in the Belgian and Dutch telecom markets (see also the case on Joint dominance in Dutch telecoms).

Methodologies for defining relevant markets

Substitutability

The definition of relevant markets is based on assessing the degree for demand side substitutability and supply side substitutability. This determines whether products impose competitive constraints on each other.

Demand side substitutability is the degree to which consumers regard products a substitute for the focal product(s). For example, consumers might consider lemonade a substitute for cola. Hence, lemonade producers impose a competitive constraint on producers of cola.

Supply side substitutability is the degree to which producers of other products can redirect their production facilities to the focal product(s). For example, producers of sanitary towels may easily switch production towards diapers, and vice versa. In such cases, both products should be included in the relevant market.

SSNIP-test

The degree of substitutability is typically assessed using the Hypothetical Monopoly Test, or SSNIP-test.

This test starts with the assumption that the focal product is produced by a hypothetical monopolist. Next, one analyses whether the hypothetical monopolist can profitably implement a Small but Significant Non-transitory Increase in the Price (SSNIP).

In other words, the SSNIP-test examines whether customers would switch to alternative products. Alternatively, it examines whether the increase in price would persuade producers to redirect their production capacity. If either of these effects is large enough, the price increase would not be profitable. Consequently, alternative products impose a competitive constraint and should be included in the relevant market.

Qualitative evidence

The SSNIP-test is ideally based on quantitative evidence. However, courts have recognised that such evidence is not always available. Hence, they have accepted qualitative assessments of whether customers can and will switch in the event of a price increase.

The question of whether customers can switch is then based on a comparison of product features that are deemed essential for customers.

The question of whether customers will switch is often based on surveys asking customers what they would do in the event of a hypothetical SSNIP. Such surveys must be carefully designed to account for biases, most notably the hypothetical bias.

Cellophane fallacy

The SSNIP-test is susceptible to the so-called ‘cellophane fallacy’. When observed prices already are at the monopoly level, any change in the price would automatically result in a reduction of profits. Consequently, the SSNIP-test fails to identify substitutes.

Chain substitution

Many product markets comprise of niche products that differ slightly from each other, but not enough for them to be considered non-substitutes. That does not mean that all niche products are equal substitutes for each other.

For example, a Lada is not a substitute for a Mercedes. However, consumer may consider a Lada as a substitute for a Dacia, and a Dacia as a substitute for a Hyundai. In turn, consumers may regard Hyundai a substitute for a Volkswagen, et cetera. In such cases, there is a chain of substitute products that all belong to the same relevant market.

Indirect constraints

When defining relevant markets at wholesale level, the analysis may also need to account for indirect pricing constraints.

Indirect constraints exist when a SSNIP at wholesale level raises retail prices such that end users switch to alternative retail services using different wholesale products. This retail switch may cause demand for wholesale products to go down and thereby render the increase in wholesale prices unprofitable. In that case, wholesale products belong to the same relevant wholesale markets.

Indirect pricing constraints may not be strong enough for including different wholesale services in the same relevant market. However, that does not mean that these constraints are irrelevant for the remainder of the analysis. The subsequent analysis of dominance should still account for the competitive pressures imposed through these indirect pricing constraints.

Critical loss analysis

A critical loss analysis examines how many customers would have to switch for the price increase to be loss making. The analysis compares the critical loss of customers to the actual loss of customers. The latter is estimated on the basis of retail price elasticity. If the actual loss is greater than the critical loss, a price increase at wholesale level would not be profitable.

A critical loss analysis is typically applied in relation to indirect competitive pressures between different wholesale products.

Substitutes and complements

There may be discussion on complementary products being part of the same relevant market. In other words, is there a market for a bundle of products? For example, are there separate markets for left and right shoes? Or is there a single market for pairs of shoes?

The consensus is that there are separate markets when consumers can evade a price increase of the bundle by purchasing the individual products separately. This assumes of course that individual products are separately available (or that they become separately available following a price increase of the bundle).

So, the analysis examines whether (potential) stand-alone offerings exert competitive pressure on a hypothetical monopolist of bundles. As long as this is the case, bundles should be treated as a marketing strategy and not as a market.

Successive wholesale markets

The bundling discussion may apply to retail services (such as broadband and broadcasting services), but it may also apply to wholesale services. In particular, it may apply to successive wholesale services that complement each other to create a retail service.

Following the analogy with bundling, one may only define a single market for wholesale products when the bundle cannot be unbundled in response to a price increase. In other words, the wholesale client can neither produce its own wholesale services, nor can it purchase them from a third party. In that case, the markets for stand-alone wholesale services is non-existing.

The relevant geographical market

The relevant market also has a geographical dimension. The analysis of geographical markets is basically the same as the analysis of product markets. It focusses on demand and supply substitutability across regions. A more qualitative interpretation of the geographical market is an area across which the competitive conditions are the same.

Risks of false reasoning

Competition authorities and regulators have a tendency to work with starting hypotheses about the scope of the relevant market. From an analytical point of view, there is nothing wrong with that as long as the subsequent analysis aims at critically testing these starting hypotheses. In practice, however, it occurs that the analysis is not testing the hypothesis, but that it is reasoning towards confirming it. This happened, for example, in the most recent market analysis decisions of the Dutch and Belgian telecom regulators.

A single wholesale market for wholesale fixed access in the Netherlands

The joint dominance case in the Dutch broadband market (see here) contains a prime example of false reasoning in the definition of the relevant market.

In previous market decisions, the ACM concluded on separate successive wholesale markets. More specifically, the ACM defined a market for wholesale local access and several successive wholesale markets. These earlier market definitions were in line with the European Commission’s Recommendation on Relevant Markets.

In these previous decisions, ACM concluded that KPN held a dominant position in the market for local access. Essential for this conclusion was that cable networks cannot offer wholesale local access services. In addition, indirect pricing constraints in the local access market are too weak. Subsequently, ACM argued that an access obligation for KPN in this market would effectively restore competition in the above wholesale markets and the retail markets.

In its latest decision, however, the ACM aimed at proving joint dominance. Its intentions were to impose access obligations on both KPN and VodafoneZiggo. In order to argue joint dominance, however, ACM had to conclude on a single wholesale market for fixed wholesale access. In other words, it had to argue that there no longer was a difference between local access and other wholesale services. This conclusion ignores that access seekers can (and do) buy stand-alone local access service. They subsequently complement these with their own backhaul services to produce their own downstream wholesale services and retail services.

In the end, the court quashed the ACM’s decision because it could not substantiate joint dominance.

Separate wholesale markets for different networks in Belgian

In Belgium, the regulator aimed for a similar result as the Dutch regulator, but followed a different reasoning. However, the resulting analysis was just as much an example of false reasoning.

The Belgian regulator reasoned towards a market definition that avoided the joint dominance hypotheses, but still allowed it to impose access obligations on cable operators. It stated that cable and copper networks operate in different wholesale markets, each being a monopolist in their own market.

At the core, the regulator based this conclusion on the proposition that wholesale clients cannot easily switch between networks. This ignored the fact that, in the past, access seekers had indeed switched from one network to another.

In addition, the regulator claimed that indirect pricing constraints were not strong enough. The regulator based these conclusions on commonly used models. However, certain input variables about profit margins and price elasticities were poorly chosen and inconsistent with variables used in other Member States.

This Belgian market analysis deviated from the European Commission’s Recommendation on Relevant Markets. The Commission subsequently questioned the decision. However, the Belgian regulator defended itself by claiming that it made no difference how it defined the market. If it had defined a single wholesale market, the regulator argued, there would be joint dominance. This claim was not based on any robust analysis. Yet, the Commission went along. Subsequently, the Belgian court also accepted the regulator’s decision.

By the time the Dutch court had dismissed the joint dominance hypothesis in the Netherlands, mandated access to Belgian cable networks was already a fait accompli. Nevertheless, the Dutch verdict seriously questions the Commission’s acceptance of the Belgian notification and calls for a critical review of that decision.

Impact

The work of Nicolai has contributed to the development of European directives. In several cases, e-Conomics has identified and challenged false reasoning by authorities in their analysis of relevant markets.