Horizontal non-brand bidding agreements in Sweden: how may pro and anti-competitive effects influence an object assessment?

Introduction

On the 4^th of June, the Patent and Market Court (PMC) in case PMÄ 8348-25 ruled that a horizontal non-brand bidding agreement between online healthcare providers in Sweden did not constitute a restriction by object but instead had to be assessed on its effects (no link is provided since PMC judgments can only be requested from the PMC and are not published online). The Swedish competition authority (SCA) found that the agreement between Kry and Min Doktor to not bid on each other’s trademark keywords on Google search advertising was an object restriction. The PMC overruled this finding stating that the economic context of non-bidding agreements raised doubts as to whether they have the objective of restricting competition. The ruling may be appealed by the SCA.

This blog post will explain and analyse the main findings of the PMC. First, it will explain why the PMC relied on anticompetitive effects in the assessment of whether the agreement was an object restriction and why this is inconsistent with the case law. Second, it will explain how the PMC did not consider protection of trademark rights to be a valid justification for the agreement. Third, this post will show that the PMC relied on pro-competitive effects relating to advertising cost savings to categorise the agreement as an effects restriction of competition. Furthermore, it will also discuss to what extent these pro-competitive effects should have been permitted.

The PMC relied on the anticompetitive effects of the agreement to argue that it was not an object restriction

The PMC held that the agreement between Kry and Min Doktor clearly intended to prevent bidding on each other’s trademark keywords in online search advertising (p. 28 of the judgment). Nevertheless, considering case law from the Court of Justice such as Tondela and European Super League, it held that an assessment of the legal and economic context of the agreement was necessary to determine whether it had the objective to restrict competition (p. 30 of the judgment). According to the PMC, the fact that the conversion rate of consumer clicks on ads that were placed on rival keywords were very low, the agreement affected an insignificant parameter of competition and therefore it was not a restriction by object (p. 35 to 37 of the judgment). It relies on market data to reach this conclusion.

There are some issues that this finding raises. First, the PMC does not provide a clear legal basis for this claim that the conduct being an insignificant parameter of competition is relevant when determining whether the conduct is an object restriction. Nevertheless, it was required to address this argument which was raised by the parties. Second, and more crucially, it is unclear why this should be relevant when determining whether the agreement is an object restriction. The PMC is essentially stating that the effects of the agreement on competition are so minimal that the agreement should not be an object restriction. According to case law, an object restriction presumes anticompetitive effects arise, irrespective of whether they are minimal or not (See T-Mobile para 31). Furthermore, an examination of positive or negative effects is not necessary in an object assessment (European Super League para 166). This raises the question to if and to what extent should anticompetitive effects be relevant in an object assessment. Furthermore, the distinction between object and effects restrictions, which is already very blurry, becomes even less clear if this type of analysis is permitted.

The tension between trademark rights and competition law was not present in the case

The PMC held that using horizontal non-brand bidding agreements to protect trademarks could be a legitimate reason to enter into non-bidding agreements (p. 41 of the judgment). Nevertheless, it held that in the case, there was insufficient evidence to show that Kry and Min Doktor bidding on each other’s trademark keywords would raise issues relating to trademark rights (p. 41 of the judgment).

One of the fundamental tensions that may arise from non-brand bidding agreements is between trademark rights and competition law. By intervening under article 101(1) TFEU to prevent a dominant firm from protecting its trademark, this could have negative consequences for the incentives of firms to invest in their brands. It would allow other firms to free ride on the trademark holder’s investments. Therefore, had there been a legitimate concern related to trademark rights in the case, it would be hard to argue that the agreement had the objective to restrict competition. An effects assessment should be necessary.

The question of whether bidding on trademark keywords online constitutes an infringement of trademark rights was addressed in Google France. This case was concerned with the infringement of IP rights and not competition law. The Court of Justice held that it can only be an infringement where the average internet user cannot ascertain whether the goods or service originate from the trademark holder or from another undertaking (see para 99). Interestingly, the Commission in Guess addressed a vertical non-brand bidding agreement and used this judgment to support its finding that the agreement in question did not raise trademark concerns as there was no risk of confusion of the origin of the products (see para 117). In the present case, the average user is unlikely to have confused the origin of the services of Min Doktor when searching for ‘Kry’ and vice versa.

The PMC ultimately relied on pro-competitive effects but failed to properly assess whether they were significant and sufficiently certain

Due to the absence of any risks for infringing the respective trademarks in the specific case, the PMC inevitably had to rely on other reasons for why the agreement was not an object restriction. In its analysis of the economic context of the agreement, it referred to economic literature. It referred to studies by the Competition and Markets Authority (paper E), the Dutch Competition Authority, Motta and Penta and Simonov and Hill to demonstrate that it was not possible to conclude whether non-brand bidding agreements are motivated by pro-competitive reasons or not (p. 38 and 39 of the judgment). The economic literature cited referred to arguments relating to the saving of advertising costs and avoiding free riding on trademark investments. Hence, due to this uncertainty, it was not possible to place this agreement in the object category.

Although the PMC did not explicitly refer to the pro-competitive effects resulting from cost savings from advertising as being the main economic argument questioning the anticompetitive objective of the agreement, this is the only clear argument from the cited economic literature that would support this conclusion in the case. This is because as already explained, the conduct did not cause confusion for the average user and therefore the risk of rivals free riding on each other’s trademark is unconvincing.

The problem with the PMC’s use of economic research is that it did not carry out a specific assessment of the pro-competitive effects in the context of the case. Furthermore, following its finding that anticompetitive effects in the case were minimal and therefore that the agreement was not an object restriction, it should logically have also assessed whether the pro-competitive effects were sufficient to reach the same conclusion. In this regard, in Generics and Super Bock, the Court of Justice held that pro-competitive effects could be considered to determine whether the agreement is a restriction by object (see para 105 and para 36 respectively). In Generics, the Court also held that agreements that give rise to pro-competitive effects which are minimal and probably uncertain cannot call into question its objective to restrict competition (see para 108). In the case, whether the advertising costs were substantial enough and whether the savings were passed on to consumers was not clear. This is supported by the economic literature which the PMC cites in the case.

It is worth noting that there is a possible contradiction in the Court’s case law. In European Super League, which is a Grand Chamber judgment, the Court provided that when assessing an agreement’s objective to restrict competition, it is not necessary to examine or prove effects of the conduct, whether negative or positive (see para 166). This could be understood to contradict its findings in Generics which considered the positive effects of the agreement. As mentioned above, it also contradicts the findings of the PMC which looked at the negative effects.

Nevertheless, it might be argued that the Court of Justice in Generics and the PMC were not carrying out an effects assessment since they did not use detailed analysis to show that for example prices would increase or decrease. This is more convincing in the case of Generics as the Court provided a general assessment and did not refer to market data. Hence, the effects that the Court in European Super League is referring to might differ as they may require closer examination and proof. What this point reinforces is the problem that the line between object and effects assessments is excessively blurry. Furthermore, as raised by Giorgio Monti, the relevance of article 101(3) TFEU also becomes unclear if pro-competitive effects are considered for object assessments under 101(1).

Conclusion

The PMC’s judgment endorses an effects assessment when determining whether an agreement is a restriction by object since it argues that it affected an insignificant parameter of competition. This is not consistent with the Court of Justice’s case law and further blurs the lines between object and effects assessments. Furthermore, following its finding that minimal anticompetitive effects are relevant, it logically should have also considered whether the pro-competitive effects were sufficient to call into question the agreement’s objective to restrict competition. To what extent the advertisement cost savings were substantial and whether they were sufficiently certain is not clear. This is supported by the Court’s findings in Generics. However, assessing both the pro-competitive effects, but also anticompetitive effects, when determining whether an agreement is a restriction by object is also inconsistent with the Court’s findings in European Super League. Relying on pro-competitive effects also reinforces the problem that there is lack of clarity between an object and effects assessment under article 101(1) TFEU but also with the assessment under article 101(3).

PhD researcher at Uppsala University. Contact: [email protected]. The author has no connection to the case.