This paper – which can be found here – tries to identify when algorithmic price discrimination will be anticompetitive.
Price discrimination is not per se unlawful or anticompetitive; on the contrary, price discrimination may be efficient and lead to increased output. However, personalised pricing is commonly felt to be unfair – and it is undisputed (in Europe, at least) that some forms of price discrimination can be anticompetitive. This paper seeks to distinguish between those situations when algorithmic price discrimination is anticompetitive and those in which it is not.
The paper is structured as follows:
Section 2 looks at how price discrimination can harm competition. Two types of harm are identified: (i) primary line injury occurs where the supplier’s conduct discriminates against competitors in markets in which the supplier also operates; and (ii) secondary line injury takes place when a supplier discriminates between a number of its customers as against one another. While behaviour giving rise to primary line injury can harm the direct competitors of the supplier and have potentially exclusionary effects, practices amounting to secondary line injury instead lead to exploitative effects as the supplier gives preferential treatment to some customers and not to others when they are in competition with each other.
Section 3 describes personalised pricing. This section identifies the reasons why companies engage in price discrimination (i.e. ‘to extract as much as possible from a consumer’s maximum willingness to pay’); describes the various types of price discrimination identified in the literature (i.e. first-degree or perfect price discrimination, when the supplier charges each customer the maximum price that customer is willing to pay for the product; second-degree price discrimination, whereby products or services are offered in different packages from which consumers can self-select based on their preferences; and third-degree price discrimination, where the supplier can identify groups of customers by easily observable features, such as pensioners or students); and explains why technological developments facilitate ever-more sophisticated and perfect forms of price discrimination.
Section 4 looks at the welfare effects of price discrimination. From an economic perspective, price discrimination is mostly considered to be welfare-enhancing because it usually leads to increased output – which normally achieved by charging lower prices to customers who would otherwise have been priced out of the market. However, a supplier that can price-discriminate may also restrict output by raising prices to customers with a higher willingness to pay if the raise is such as to price these customers out of the market. As such, the overall effect of price discrimination on output and welfare is ambiguous. The challenge for competition law in this regard is to ensure that only discrimination that is harmful to consumer welfare is prohibited, while not precluding undertakings from engaging in discrimination that is welfare-enhancing.
With personalised pricing, the challenge is slightly different. Since the supplier is able to identify the willingness to pay of every customer, it has no incentive to limit supply to any customer as long as his or her willingness to pay covers marginal costs. In the event of a monopoly, the consequence of this is that higher prices will lead to overall welfare and output, but may not necessarily result into greater consumer welfare. This, in turn, will force competition authorities to choose between two legal standards that are usually presumed to be identical: overall welfareand consumer welfare. At the same time, it is important to note that personalised pricing may benefit consumers in oligopolies through increased competition. In order to poach customers from rivals, each firm will prefer to cut the price it sets to those consumers that it knows will otherwise not purchase from it. As a result, all consumers will face lower prices compared to the situation where no price discrimination is possible. In short, the overall effect of personalised pricing on consumer surplus is ambiguous, with the impact likely to vary from market to market.
Section 5 looks at the scope for competition intervention. This section begins with a review of EU law on price discrimination – and concludes that price discrimination can be unlawful in the context of various exclusionary abuses of a dominant position (e.g. margin squeeze, predatory pricing, tying, rebates and refusals to deal), but that it is unclear whether it can amount to a free-standing infringement on its own right.
At this point, the paper reviews the EU Treaty provisions and the case law on price discrimination under EU competition law, and argues that price discrimination towards end-consumers can amount to an infringement of EU competition law. After this, the author quickly looks at the normative desirability of competition law enforcement against price discrimination towards final consumers, and holds that; ‘The desirability of action against personalised pricing on the basis of competition law should (…) take into account whether there is an actual enforcement gap that is not already filled by other legal fields’.
Section 6 looks at these fields other than competition law that regulate personalised pricing. It identifies the following relevant legal regimes t:
- Data Protection – In order to engage in price discrimination, firms will typically collect personal data of individuals with a view to personalising prices, and thereby fall within the scope of EU data protection law. The author reviews the European data protection regime, and concludes that, in order to engage in personalised pricing, a company needs to have the explicit consent of the data subject (i.e. individual).
- Consumer Protection – Personalised pricing does not amount to an unfair trading condition under EU law. Nonetheless, the European Commission has clarified that a breach of the Unfair Commercial Practices Directive may occur in situations where personalised pricing is combined with certain misleading or aggressive commercial practices.
- Antidiscrimination Law – As with consumer law, personalised pricing may be prohibited in the context of certain discriminatory practices. However, it remains unclear to what extent a dominant undertaking can be held liable for differentiating or personalising prices among customers.
Section 7 contains an assessment of the suitability of competition remedies against personalised pricing. It is argued that the first line of defence against personalised pricing should be data protection and consumer protection law. As to competition law, the main challenge is to distinguish between harmful and welfare-enhancing forms of personalised pricing. As price discrimination is more likely to negatively impact consumers in monopolistic markets, it makes sense to limit enforcement actions to discriminatory behaviour of dominant companies as targeted by the abuse of dominance regime of Article 102 TFEU. The more difficult question, however, is how competition law should treat personalised pricing if its effects are not exclusionary. Following Chris’ paper of a couple of years back, the author suggests that a ‘relevant factor in examining the desirability of a competition intervention to address personalised pricing could be the extent to which the exploitation is sustainable or not’. Furthermore, competition interventions should ‘also be possible for exploitative behaviour that leads to ‘unfair’ outcomes directly harming consumers.’
This is a succinct, yet comprehensive, overview of the problems that personalised pricing are likely to pose to competition enforcers. I particularly enjoyed how the author assessed whether competition law is a suitable enforcement tool by reference to the wider regulatory framework – you may remember a number of previous posts where I’ve expressed the opinion that many of the issues raised by the digital economy are not, strictly speaking , competition issues. Instead, addressing them requires an overhaul of the wider regulatory framework.
Nonetheless, the issues that plague the development of such a framework also surface in this piece: it is unclear (as is the case with most competition pieces dealing with the digital economy) when competition law should intervene, and on what grounds. This can be seen in how the author falls back on the idea of ‘fairness’ as governing the enforcement of competition law. Even more explicitly, it is argued: ‘Consumers will generally not oppose to price discrimination that advances social goals (…) price discrimination that has profit maximisation as a sole goal and simply sorts consumers in order to be able to extract as much surplus from them as possible is unlikely to be found socially acceptable.‘ It goes without saying that profit maximisation is a legitimate – dare I say desirable – goal for companies under competition law, and that whether a social goal is laudable or not is a deeply political question. The same vague standard focused on achieving a ‘valuable social outcome’ seems to underpin the analysis of whether competition law should act when there are ‘regulatory gaps’ created by the other applicable regulatory regimes.
This is nothing new (see, to select just a few, my posts of 14 and 21 April, and 16 June 2017), but seeing this issue arise in paper after paper regarding the role of competition law and the new economy reinforces my sense that competition law is being used as a catch-all subsidiary regime for the failings of the wider regulatory framework. The real question which competition law is being asked to answer is what type of regulatory regime – and more widely, what type of society – do we want to see emerge from the digital revolution. Of course, the question of what role competition law should play in this world is an important one – but I have my doubts that we will be able to answer the most important questions by focusing on competition law alone.
P.s. of 18 August 2018: EU law on price discrimination has advanced since this paper was first published on SSRN, as the ECJ decided a case on secondary price discrimination in MEO. I highly recommend this decision, as well as AG Wahl’s Opinion.