Algorithmic Collusion

Abstract

The purpose of this thesis is to examine when firms participating in algorithmic collusion can be heldliable subject to TEUF art. 101(1), and how the use of pricing algorithms affect the pricing behaviourof colluding undertakings and the social welfare.The economic analysis finds that the use of pricing algorithms changes the structural marketconditions, hereunder the marked transparency and interaction frequency, and therefore algorithmiccollusion will be extremely stable and the participating undertakings will have no incentive to deviatefrom the collusion. The social net welfare decreases, as the decrease in consumer surplus is larger thanthe increase in producer surplus.The legal analysis finds that firms participating in an algorithmic Hub-and-Spoke conspiracy, can beheld liable subject to the prohibition in TEUF art. 101(1). Firms participating in algorithmic pricecoordination will likely only violate the prohibition in TEUF art. 101(1), if the algorithm is designed tosend specific signals of it’s intended future price, and one or more undertakings, by changing its price,accepts this signal, and the management of the undertakings was aware or should have been aware ofthe algorithm’s behaviour.Though algorithmic price coordination decreases the net welfare and leads to high, stable collusiveprices, these scenarios will in most cases not violate TEUF art. 101(1). The integrated analysis presentsthree proposals as to how such collusion may be regulated, 1) regulating the market transparencyand/or interaction frequency, 2) include algorithmic price coordination in the term concerted practicesin TEUF art. 101(1), 3) The Commission starts a scheme where the pricing algorithms are tested, andonly the algorithms which passes the test, may be used on the market.All the three proposals leads to lower prices and increased social welfare, but the second proposal hasthe least negative effects, and is therefore the preferred proposal.