Page 145 - วารสารกฎหมาย ศาลอุทธรณ์คดีชํานัญพิเศษ
P. 145

ฉบับพิเศษ ประจำ�ปี 2564



                            2. UBER-like Algorithms and Price Fixing





                    Algorithms and Price Fixing

                    Technological development makes it possible for a computer algorithm to operate
            price fixing schemes. Ariel Ezrachi and Maurice E. Stucke have developed four scenarios

            in which algorithms can fulfil this role: messenger, hub-and-spoke, predicable agent
            and digital eye.  These scenarios are based on the different roles an algorithm can play
                           8
            in the formation and implementation of collusion. The messenger scenario describes
            the algorithm taking the role of implementing, monitoring and policing a price fixing
            agreement that has been discussed and approved by humans.  In a hub-and-spoke
                                                                          9
            scenario, a single algorithm will determine the price of firms competing in the same
            market.  Predictable agent exemplifies the scenario in which different algorithms of
                    10
            independent firms predict a similar price outcome based on the market conditions it can

            observe.  In the digital eye, the algorithm’s role is to determine, completely on its own,
                    11
            what the best way is to reach a pre-determined goal, such as profit maximization. 12

                    Whereas the application of contemporary competition law to monitoring
            algorithms and the progress in machine learning does not yet allow for the implementation

            of predictable agent and digital eye, the hub-and-spoke scenario, conceivable in practice,
            is still has unsettled boundaries. Hub-and-spoke cartels “raise delicate questions
            pertaining to the line between perfectly legal information sharing between trading



                    8  Ezrachi and Stucke (2017), p. 1782. Digital eye has also been termed autonomous machine in an earlier
            version of the Ezrachi and Stucke’s work. Ezrachi and Stucke (2015), p. 7. The Secretariat of the Organization for
            Economic Co-operation and Development (OECD) merely acknowledges the different roles, but uses different
            names. The OECD distinguished between monitoring algorithms, parallel algorithms, signaling algorithms, and
            self-learning algorithms. OECD (2017), pp. 24-32. Niccolò Colombo has still used other terms for the four different
            roles: classical digital cartel, inadvertent hub-and-spoke, tacit algorithmic collusion and dystopian virtual reality.
            Colombo (2018), pp. 12-14.
                    9  Van Uytsel (2018), p. 157.
                    10  Ezrachi and Stucke (2017), p. 1782.
                    11  Van Uytsel (2018), p. 158.
                    12  OECD (2017), p. 30.



                                                                                             143
   140   141   142   143   144   145   146   147   148   149   150