Page 147 - วารสารกฎหมาย ศาลอุทธรณ์คดีชํานัญพิเศษ

ฉบับพิเศษ ประจำ�ปี 2564

above the competitive level.” Fourth, a third party develops a pricing algorithm that
19
is being sold to several competing firms, whether to the supplier or retail side of

the market. Fifth, similar to the previous version, but the suppliers or retailers are aware
that its price algorithm is being used by several competing firms. Sixth, a third party

develops and runs its own pricing mechanism to which other users subscribe to offer
their services for the price se byt the algorithm of the third party.

When Ezrachi and Stucke elaborated their taxonomy, they were less interested
in the first three versions of the hub-and-spoke. These versions are representations of

a classical hub-and-spoke. Hence, they should not be controversial in the enforcement
of competition law. The situation changes when the latter three versions are considered.
It is not clear what the position of competition law should be towards these versions.

Ezrachi and Stucke illustrate the fourth and fifth version with a reference to Boomerang
Commerce, a vendor of software helping online retailers to optimize their price setting.
It is conceivable that the same pricing algorithm of Boomerang is being used by several

competing retailers. Despite the decision to use pricing algorithms being made
independently, Ezrachi and Stucke are skeptical about the unawareness of the retailers
on what is happening in reality. If the retailers need to provide access to their data in
20
order to allow the algorithm to optimize the price setting, retailers must be aware that
the data of competing retailers will equally be used when such retailers decide on

the same third party vendor. The last version is exemplified with Uber. As “Uber’s algorithm
determine[s] for hundreds of competing drivers the base price for the trip, when to
implement a surge price, for which areas, for how long, and to what extent,” the question
21
is raised on whether this business model should be accepted from a competition law
perspective. Yet, an answer is not given. In response, Ezrachi and Stucke list elements

to take into consideration when looking for an answer. These elements relate to
alternatives and market power. In order to determine whether the legal use of an algorithm
could tip over in an anticompetitive use, it may be necessary to consider the ability of

19 OECD (2017), p. 27; see also Capobianco and Gonzaga (2017), p. 4.
20 Ezrachi and Stucke (2016), p. 49.
21 Ezrachi and Stucke (2016), p. 51.

145