Not surprisingly, AI-enabled price discrimination has been widely employed in both business-to-business and business-to-consumer relationships. Examples of the latter include marketplaces for airplane tickets, e-commerce, and travel services. In this context, questions arise on whether AI-enabled price discrimination is desirable in digital markets and how to tackle it if it is employed by dominant Big Techs as an abusive strategy. An economic analysis of the basic mechanism of AI-enabled price discrimination and its positive and negative effects may provide some answers. 

In economics, price discrimination occurs when identical products are sold at different prices under identical cost conditions or when non-identical but similar goods are sold at prices which are in different ratios to their marginal cost. Normally, it requires three conditions: some element of market power, barriers to take arbitrage from different groups of consumers, and a way to identify a consumer’s willingness to pay. In digital markets, amounts of consumer data and accurate algorithms as analytical tools allow undertakings to predict each consumer’s willingness to pay with increasing accuracy. In this case, AI-enabled price discrimination is close to the theoretical model of perfect competition.

From an economic perspective, price discrimination is not always undesirable. AI-enabled price discrimination can improve static efficiency, by increasing the number of units sold. This would happen if undertakings can tailor prices to consumers’ valuations and if arbitrage is not possible. Producer welfare would then increase, because it is always profitable for undertakings to serve consumers whole willingness to pay exceeds the marginal cost of production.  

The impact on consumer welfare will likely vary from market to market. In oligopolies, AI-enabled price discrimination may benefit consumers through intensified competition and thereby raise consumer surplus at the expense of industry profits. In monopolies, it could increase product affordability for consumers who have lower incomes or reservation prices. Nevertheless, price discrimination aims to capture as much consumer surplus as possible and consumer surplus might be entirely captured by a monopolist under some circumstances.

Similarly, its effect on dynamic efficiency tends to be ambiguous as well. On the positive side, the profits gained from AI-enabled price discrimination may increase incentives for undertakings to compete, by investing in innovation and reducing costs, which can benefit consumers directly and create positive externalities if other undertakings adopt and promote those innovations as well. On the negative side, undertakings may engage in rent seeking behaviour, by lobbying governments to introduce regulations that protect them from competition, rather than investing in innovation (see for an analysis of efficiency effects also the OECD background papers on price discrimination and personalised pricing in the digital era).

As such, the multi-dimensional economic effects of AI-enabled price discrimination require a more detailed assessment from a competition economics perspective. Competition economics typically distinguishes between exclusionary and exploitative effects in assessing the competitive effects of abuse, which respectively result in foreclosure of competitors and direct consumer harm. AI-enabled price discrimination draws more serious competition concerns if a dominant Big Tech abuses its market position through AI-enabled price discrimination.

On the one hand, AI-enabled price discrimination may impede effective competition once Big Techs abuse their dominance in digital markets. The stronger the market power, the stronger the possibilities to hinder market access by potential competitors which results in an even stronger position of dominance. On the other hand, when a Big Tech undertaking is very dominant in a market, it is much easier for such undertaking to directly exploit consumers as well as intermediate customers in upstream markets (and thereby create distortionary effects that harm consumers in downstream markets).

Since the protection of free competition and consumer welfare are objectives of competition law in the EU and China (which are the jurisdictions currently examined by the authors of this blog), it makes sense to consider competition law intervention. In the EU, price discrimination is specifically mentioned as an abuse of dominance under Article 102(c)TFEU where a seller applies “dissimilar conditions to equivalent transactions with other trading parties” and thereby places them “at a competitive disadvantage”. Laws outside of competition law such as the Platform to Business Regulation and the Digital Markets Act also impose obligations to undertakings to ensure contestable and fair markets in digital sectors. By contrast, in China, Article 17(6) AML challenges the discriminatory treatment of undertakings, while the Guidelines on the Platform Economy specify its application in digital markets. The recently enacted Recommendation Algorithm Regulations further prohibit undertakings engaging in discriminatory practices via recommendation algorithms. Nevertheless, also other cases where price discrimination amounts to an abuse of dominance are prohibited.

To conclude, AI-enabled price discrimination is not always undesirable in digital markets. In general, it is good for the economy, as it can increase static efficiency, and on some occasions, it can promote dynamic efficiency as well as boost consumer welfare. However, it may also lead to exclusionary and exploitative effects once Big Techs abuse their (very) dominant market positions. As such, competition law seems a proper instrument to step into digital markets to address these competition concerns. It is yet another question, however, to what extent the application of competition law to AI-enabled price discrimination in the EU and China is in line with economic theory. The answer to that question, and more particularly whether there is under-enforcement (false negatives) or over-enforcement (false positives), is beyond the scope of this blog.