An algorithm is a set of instructions that a computer programmer is fed to perform certain tasks. Pricing algorithms allow for constant adjustment and optimization of individual prices based on trial and error and through finding patterns from a great volume and variety of data, leading to optimal pricing. As companies collect more user data and algorithms have more opportunities to experiment on product differentiation and suggestion, pricing becomes more dynamic, differentiated and personalized.
Although the concept of algorithm pricing is not new, it has become common recently with the marketplace going digital. Companies, online retailers, transport, pharmaceutical industry etc are all using algorithms to set prices based on demand and supply. Amazon is a shopping giant whose algorithm can make or break other retailers. Recently, a Preliminary Report on the working of e-commerce by the European Commission found that nearly 500 third-party sellers on Amazon use algorithm pricing to charge customers based on various different criteria. Amazon gives itself an edge by not including the price of shipping on its own products. To get the same benefit, other sellers have to pay Amazon. From the point of view of competition, what matters is how these algorithms are actually used, and that although the use of these algorithms is not per se questionable but fair use of them should be ensured. The effects that algorithm pricing has on competition is two-fold. Not only does it infringe on an individual’s privacy by adjusting the prices based on the person’s data collected from various online sources, it also helps companies indulge in online collusion tacitly.
Section 3 of the Competition Act, 2002 prohibits anti-competitive agreements. However, this section fails to bring the collusion through algorithm pricing as it is never a direct collusion between companies but tacit collusion. Due to its online presence and nature, collusion through algorithm pricing does not need to be direct and can be brought about through the programme as it is designed in such a way. As pointed out in an OECD report, algorithms can be used to detect market threats, in the form of a new player or the increasing presence of an old player, very quickly which in turn help a company take a business decision preemptively. Most of the “tacit collusion” is not even visible to the naked eye. For example- suppose there are two websites which facilitate online booking of tickets. If the prices in one of them increase depending on more number of users on that particular website, the other one can immediately detect this change and increase/decrease its prices, depending on their market strategy. Again, if a particular individual has been availing a particular airline company’s services, the other airline companies might detect this through their algorithm and reduce their prices in order to attract that customer. These factors are commonly termed as market interaction and frequency of interaction.
The distinction between direct and tacit collusion shows how under a market having few sellers and transparent supra-competitive price strategies may be reasonable outcome of rational economic behaviour of each firm in the market. It is for this reason that tacit collusion or conscious parallelism falls outside the reach of competition laws on agreements between competitors. However, from a policy perspective, such a tacitly collusive outcome may not be desirable, as it confers firms the ability to significantly suppress output or raise prices to the detriment of consumers as much as an explicit agreement would.
This problem needs to addressed and it is become very relevant whether there the need to address algorithmic collusion should require a new definition of what is an agreement for antitrust purposes because identifying an “agreement” between competitors is a prerequisite to enforce the law against collusive outcomes.
Comments