Better things are in smaller packages! The Luxembourg Competition Authority clears an AI-based price-fixing arrangement on efficiency grounds
Introduction
Reliance on Artificial Intelligence (AI) to fix prices
may be regarded as a pro-competitive factor that comes into relevance to clear
a horizontal price-fixing arrangement if the use of AI leads to lower prices
for consumers. This is the take-away lesson from a decision recently handed
down by the Luxembourg Competition Authority (LCA) in the Webtaxi case[1].
The facts of the case
Webtaxi is a platform for the provision of booking taxis
service. Clients can book their taxis by phone, calling Webtaxi or any taxi
operators participating to the platform, and also through an app or the website
made available by Webtaxi. The booking requests are then matched by Webtaxi with
the available taxi drivers on the basis of geographical position and the other criteria
chosen by clients. The taxi fares are automatically calculated by the platform with
an algorithm that takes into account distance, time and traffic condition. Taxi
drivers and customers are not allowed to negotiate the fares determined by the
algorithm.
Upon the receipt of a complaint lodged by a
competitor, the LCA decided to open an antitrust investigation against Webtaxi
to determine whether the pricing policy applied by the platform has infringed Article
3 of Law of 23 October 2011 or Luxembourg Competition Law (LCL), corresponding
to Article 101 TFEU.
The decision of the LCA
Looking at the intermediation function between taxi
drivers and clients carried out by Webtaxi, the LCA qualified it as two-sided
market: it encompasses not only the relations between the platform and the
participating taxi operators but also the relations between taxi drivers and
clients. The taxi operators participating to the platform agreed, via Webtaxi, on
jointly determining the fares for the taxi rides allocated by Webtaxi. In the
LCA view, this was a horizontal price-setting agreement, amounting to competition
restraint by object in violation of Article 3 LCL. The next step of the LCA was
to establish whether that arrangement qualified for an individual exemption
under the four-limbs test set out in Article 4 LCL, corresponding to Article
101(3) TFEU. In that regard, the LCA addressed the following questions:
Does
the restrictive agreement increase economic efficiency?
The parties showed several efficiency gains coming
from the agreement, starting with the reduction in the number of empty taxi
rides. This more efficient allocation of resources would cut pollution caused
by circulation of taxis, thereby contributing to combat the greenhouse effect. The
platform would also enable taxi operators to offer a uniform, centralized and
wider offer of taxi services on a 24-hour 7-day basis.
Does
a fair share of the efficiency gains accrue to consumers?
The algorithm employed by Webtaxi might result in
lower prices for consumers. The fares determined by Webtaxi cannot be higher
than the fares showed on the taximeter of each taxi drivers. As the fares
determined by Webtaxi are immediately communicated to customers, they can
compare them with those indicated by taximeters. Moreover, the fares determined
by the algorithm are degressive as they decrease as distance increases. The Webtaxi
geolocalisation system might generate qualitative benefits for customers. Waiting
times may be lowered as the system matches clients with nearer taxi drivers. Moreover,
by regrouping several taxi operators the platform may increase the range of
taxi services available to customers. Lastly, the lower atmospheric pollution
resulting from the more efficient management of taxis would have positive
repercussion on the health of clients.
The major interest of customers in using the Webtaxi
platform is to have a fixed price in advance for a given ride. In absence of
uniform, pre-determined fares, the taxi driver offering the fare chosen by the
customer may not be the one closer to that client, which may put on risk the
business model followed by Webtaxi. Moreover, Webtaxi shields customers away from
the risks inherent in taxi rides that are entirely borne by the taxi operators
as in case of traffic jams that stretch travelling times without any fare increase
for customers. Should, instead, the fares be negotiated case-by-case such
efficiency gains would be unlikely to materialize. Hence, the LCA took the view
that the algorithm-based setting-price mechanism is necessary to ensure the
correct functioning of the Webtaxi platform.
Does
the agreement eliminate competition in the relevant market?
The health of market competition is unlikely to be
harmed by the arrangement in question, given that only 26% of the
Luxembourg-based taxi operators participate in the Webtaxi platform.
Having then established that the price-setting
arrangement met all the four cumulative statutory conditions, the LCA granted
the Webtaxi the individual exemption in Article 4 LCL closing the antitrust
proceedings.
Conclusion
Sometimes, better things are in smaller packages. Webtaxi that concerned a local service
supplier appears to be ones of those cases. The LCA took a correct
effects-based approach to assess the competition impact of a price-setting
mechanism using IA. Eventually, the LCA found that in this case the employed IA
was not a collusive but rather a pro-competitive factor. The LCA also took a broad
approach to identify which efficiency gains to consider when assessing whether
to grant an individual exemption. In that regard, the LCA took into
consideration economic and non-economic factors, such as the preservation of
the environment. Lastly, Webtaxi is one of the rare but not unique cases in
which the acting competition authority granted an individual exemption to a
serious competition breach as a price-fixing arrangement.
[1]
Conseil de la Concurrence, (Luxembourg Competition Authority), Decision of 7 June
2018 no. 2018-FO-01, Webtaxi Sarl, www.concurrence.lu. .
Comments