Experts Discuss Algorithmic Pricing at CPI’s European Conference

At the inaugural European conference hosted by Competition Policy International (CPI) on March 22, 2024, titled “Dynamic Competition in Dynamic Markets: Charting the Path Forward?” held at SDA Bocconi School of Management, a panel of experts explored the dynamics of algorithmic pricing, innovation, competition, and their impact on market behavior.

The panel, moderated by Francesco Rosati, Partner at RBB Economics, featured Giuseppe Colangelo, Associate Professor of Law and Economics, University of Basilicata; Tobias Kretschmer, Professor of Management at LMU Munich; and Cristina Alaimo, Associate Professor of Digital Economy and Society at Luiss University.

Rosati initiated the discussion by probing the panelists to define algorithms and elucidate their role in competition and antitrust regulations, with a particular emphasis on algorithmic pricing and its potential ramifications.

Price discrimination within competitive markets has the potential to foster competition for the benefit of consumers. While monopolies may exploit this practice to extract consumer welfare entirely, in oligopolistic markets, it can promote efficiency and even bolster competition.”

Kretschmer drew an analogy, likening algorithms to power gamers in video games, driven by a single objective: winning. He emphasized that algorithms codify a firm’s objective functions, intensifying the pursuit of maximizing firm payoff without regard for conventional notions of fairness or moral obligations. Moreover, algorithms surpass human limitations in decision-making capacity and speed, resulting in more extreme outcomes.

It wouldn’t be surprising if algorithms eventually learned or devised methods of collusion. Humans engage in collusion not merely because of their humanity, but rather due to the incentive… of maximizing profits. If you instruct an algorithm to pursue the same objective, it is likely to deduce and execute similar strategies.”

Expanding on Kretschmer’s insights, Alaimo underscored the interplay between algorithms, data, and systems, highlighting the transformative impact of data availability and quality on algorithmic operations. She emphasized the symbiotic relationship between algorithms and data, where algorithms dictate user behavior, leading to the generation of data that, in turn, fuels algorithmic optimization. Alaimo stressed the complexity of algorithms, which now draw from diverse data domains, reshaping market dynamics and challenging traditional understandings of market competition.

The quality, complexity, and characteristics of data today significantly differ from what data firms previously handled.”

Giuseppe Colangelo highlighted the challenges posed by algorithms in antitrust enforcement, particularly in identifying collusion. He outlined three scenarios where algorithms might facilitate collusion, with the third scenario involving self-learning algorithms presenting a unique enforcement challenge due to its autonomous nature. Colangelo suggested potential regulatory solutions and discussed the effectiveness of algorithms in executing strategies like predatory pricing under Article 102. He concluded that while algorithms offered new challenges, existing antitrust rules could address these issues.

When addressing self-preferencing, the crux of the matter lies not in the algorithm itself but rather in the ethical consideration of whether self-preferencing constitutes a noble practice in the first place.”

The discussion moved further into the varied typologies of algorithms, ranging from recommendation algorithms in streaming services like Netflix to pricing algorithms in e-commerce platforms like Amazon. Kretschmer and Alaimo discussed how algorithms leverage data to tailor recommendations and pricing strategies, ultimately influencing consumer behavior and market outcomes.

In response to Rosati’s query on the impact of algorithms on market entry, panelists offered nuanced perspectives. While algorithms could potentially lower barriers to entry for startups with innovative ideas, they also necessitate access to vast datasets and specialized expertise, posing challenges for newcomers. Moreover, the proliferation of algorithmic tools raises questions about data access and technological capabilities, further complicating market entry dynamics.