The advent of AI based technology has increased the use of pricing algorithms around the globe. However, along with several pro-competitive effects, algorithms are being used to facilitate collusion through many ways, one of which is the hub and spoke cartel. Hub and Spoke cartels are characterized as arrangements wherein entity at one level of economic chain, the “hub,” coordinates with competitors at different level, the “spokes.” Horizontal agreements between entities are penalized under section 3(3) of the Competition Act. However, there is no express sanction for an entity that indirectly facilitates the horizontal conspiracy. The Ministry of Corporate Affairs after recognizing the need to penalize entities for hub and spoke conspiracy has proposed to include the liability of a facilitator under section 3 of the Act.i
Recently, the National Company Law Appellate Tribunal (NCLAT) in Samir Agrawal v Competition Commission of Indiaii (2020) dismissed the allegations of existence of hub and spoke cartel between the cab aggregators and the drivers. The appeal in this case arose from the decision of Samir Agrawal v ANI Technologies Pvt Ltd (2018),iii wherein the Competition Commission of India (CCI) found no prima facie violation of Section 3(3) of the Competition Act on the part of Ola or Uber. This article seeks to analyse the decision of CCI and NCLAT with respect to existence of hub and spoke conspiracy vis-à-vis the anticompetitive effects of pricing algorithms.
During the course of proceedings before CCI, the informant alleged that Ola/Uber and drivers were indulging in hub and spoke conspiracy where the Ola/Uber was the “hub” and the competing drivers were acting as “spokes” in order to fix prices in the market. The CCI rejected the existence of hub and spoke cartel on the ground that the use of algorithmically determined prices by the platform (Ola/Uber) could not amount to collusion between the independent drivers as there exists no agreement between them to set prices or to delegate pricing power to the cab aggregators.
The informant appealed against CCI decision before NCLAT, which also dismissed the appeal. NCLAT held that the price determination by cab aggregators does not amount to price fixing as there is no collusion between the cab aggregators. NCLAT also refused to rely on the US Class Action Suit titled “Spenser Meyer v Travis Kalanick”(Kalanick) on the ground that business model of Ola and Uber in India doesn’t have an effect of restricting price competition among drivers observing that drivers have no connectivity or a platform or any other mode for sharing information.
Several studies in the literature have shown that pricing algorithms are being used as an intermediary by enterprises to achieve collusive outcomes.iv In the USA v David Topkins,v for the first time, the use of algorithm was recognized as a tool for the commission of price fixing offense. In hub and spoke collusion through algorithm, the competitors use the same algorithm or software to determine their pricing to optimize their pricing behavior in the market, which may lead to collusive outcomes.vi
In Kalanick, it was alleged that use of Uber’s pricing algorithm for charging the passengers restricted price competition amongst the drivers. The US District Court noted that where the parties to the vertical agreement knows that other participants are bound by identical agreements and their participation is contingent on this knowledge, then in such a case it may be considered that participants have entered into a horizontal agreement.vii Similarly, in Interstate Circuit Inc. v United States,viii the US Supreme Court held that there need not be an express agreement and an inference of parallel conduct can be drawn from the unanimity in conduct of the distributors.
In light of these cases, it can be said that the matters pertaining to the use of algorithms warrant detailed analysis, which begins with the acknowledgement of the possibility of a horizontal agreement. However, the NCLAT in this case completely disregarded the rationale of Kalanick despite factual similarity. NCLAT’s outright denial of the use of algorithm to effect collusion reflects its state of aversion to the complexity of technology.
In another landmark judgment, the Luxembourg Competition Authority (LCA) dismissed the anticompetitive allegations against Webtaxi citing pro-competitive use of pricing algorithms.xi The facts of that case are similar to that of the present, but the analysis is entirely different. In Webtaxi case, the LCA recognized the existence of price-fixing agreement; however, it went on to defend it on efficiency grounds by carefully examining the pricing algorithm factoring in several parameters. On the other hand, NCLAT denied the existence of horizontal agreement and opined that the traditional hub and spoke cartel concept does not fit into the present case without considering the various aspects of the cab aggregation model and price determination through algorithms.
The literature and the ruling by other jurisdictions indicate that such business models may be intrinsically collusive and easily facilitate price manipulation, and thus, should not be dismissed based on the absence of an express or formal agreement. Ezrachi and Stucke have suggested investigating into the design of the algorithm under more forgiving rule-of-reason standard.x The rule-of-reason approach requires careful examination of the working of the model to identify the possibility of any efficiency gains and benefits to the consumers. While looking into the efficiency gains, the authority may also consider the availability of alternative business models that provide similar benefits to the consumer but pose lesser risk of collusion. For instance, a competition enhancing alternative to pricing algorithm may be a bidding system, under which the nearest drivers can bid to ride above or below the suggested price and the customer can select a driver based on fare, rating and distance.
Digitalisation has led to several competition concerns. The proposal of Draft Competition Amendment Bill, 2020 to include hub and spoke cartels under section 3 of the Act holds a great significance in easing these new challenges. The approach of CCI might shift once the bill is brought into existence. Although there is still a long road ahead for digital market antitrust jurisprudence, it is surely a good first step towards a more vigilant regulation.
The author is an intern at CIRC. The views expressed are personal.
ii Samir Agrawal v. Competition Commission of India Competition Appeal (AT) No. 11 of 2019 (National Company Law Appellate Tribunal, 29 May 2020) <https://nclat.nic.in/Useradmin/upload/9123996565ed0ea3eec766.pdf> accessed 30 June 2020.
iii Samir Agrawal v. ANI Technologies Case No. 37 of 2018 (Competition Commission of India, 6 November 2018)
<https://www.cci.gov.in/sites/default/files/37of2018.pdf> accessed 30 June 2020.
iv Ariel Ezrachi& Maurice E Stucke, ‘Artificial Intelligence & Collusion: When Computers Inhibit Competition’ (Oxford, Working Paper CCPL (L) Paper no. 40, 2015) <http://ssrn.com/abstract=2591874> accessed 30 June 2020 (Oxford Working Paper).
v United States of America v David Topkins No CR 15-00201 WHO (United States District Court, 6 April 2015) <https://www.justice.gov/atr/case-document/file/513586/download> accessed 1 July 2020.
vi Organisation for Economic Co-operation and Development, ‘Roundtable on Hub-and-Spoke Arrangements- Background Note (DAF/COMP (2019) 14, 25 November 2019) <https://one.oecd.org/document/DAF/COMP(2019)14/en/pdf> accessed 30 June 2020.
vii Spenser Meyer v Travis Kalanick 2016 US. Dist. Lexis 43944, (29 July 2016) <https://casetext.com/case/meyer-v-kalanick-9> accessed 30 June 2020.
viii Interstate Circuit, Inc. v. United States 306 U.S. 208 (1939), <https://supreme.justia.com/cases/federal/us/306
/208/> accessed 30 June 2020.
ix Michele Giannino, ‘The Luxembourg Competition Authority exempts a price-fixing arrangement on efficiency grounds’ (Core Blog, 10 July, 2018) < https://coreblog.lexxion.eu/webtaxi-the-luxembourg-competition-authority-exempts-an-algorithmic-price-fixing-arrangement-on-efficiency-grounds/> accessed 12 July 2020
x Oxford Working Paper (n 4).