New Approaches to Cartel Enforcement and Spillover Effects in Brazil: Exchange of Information, Hub and Spoke Agreements, Algorithms, and Anti-Poaching Agreements

By Eduardo Frade & Vinicius Marques de Carvalho (VMCA)1

 

Once upon a time, cartels were made in smoke-filled rooms and cartelists were caught through good old wire taps. Smoking has long gone out of fashion, though. Then cartels started to be made in internet chat rooms and electronic communications, and cartelists were caught through email and text messages. These are also increasingly harder to find. Cartelists are much more careful, and collusion started to be exercised through much more subtle ways. Some are in fact so subtle that one needs to ask whether these are cartels at all. This article focuses on highlighting some of the new types of possible coordination schemes that have recently caught the attention of antitrust authorities worldwide.

Cartel enforcement has always been the core of competition policy in most jurisdictions. Conversely, countries have heavily invested in structuring leniency programs as an instrument to destabilize, uncover and prosecute cartel behavior. It was through leniency programs, mostly, that wire taps, emails and texts were first obtained in several of the most prominent cartel investigations around the globe.

But as antitrust enforcement is currently witnessing an apparent decrease in leniency applications worldwide, one may inquire if such programs have become less effective or too costly for applicants, if cartel behavior is actually diminishing or, maybe, if it is simply changing its contours.

As enforcement became stricter and fines became increasingly higher, it is safe to say that the typical hard-core price-fixing or the standard “hand-shaking” bid-rigging agreements have started to give way to more sophisticated arrangements. These are at once more complex, harder to identify, and harder to catch. As such, cartel enforcement has been undergoing a new wave of enforcement, as authorities widen the range of conduct falling within their radar screens. This movement can already be seen in jurisdictions with a more traditional and mature competition policy, such as the United States and Europe, but we can also see spillover effects from these trends in other countries, such as Brazil.

International experience has shown us that despite the still unclear features of this “new wave” in cartel enforcement, new approaches have been focusing on certain types of coordinated behavior. We identify here at least four such phenomena: the exchange of competitively sensitive information, hub and spoke cartels, pricing algorithms, and anti-poaching agreements. As overall attention shifts to these “new” forms of conduct, we anticipate that the focus of investigations, prosecution and even leniency applications will do so also.

Exchange of Commercially Sensitive Information

The exchange of commercially sensitive information has been on the radar of antitrust authorities for some while, especially due to its potential to promote and strengthen collusive behavior among competitors, to bring stability to cartel agreements and to supervise market behavior deviating from the content of the agreement.

Most recently, however, aside from being assessed as a means to promote hard-core cartel behavior, information exchange among competitors can be viewed as an independent form of conduct that, in many jurisdictions, is treated as a collusive practice in and of itself. As such, both the European Commission2 and the FTC3 have put the topic up for discussion, as has the OECD, which identified the possibility of an autonomous and independent form of anticompetitive conduct consisting of the mere exchange of competitively sensitive information.4 The 2019 ICN Cartel Workshop also had a panel entirely dedicated to the topic.5

In Brazil, the topic has not received much attention until very recently. CADE, the Brazilian competition authority, has recently initiated investigations into information exchange among competitors in both the market for after-market automotive parts (2016)6 and the market for brokerage of insurance and reinsurance in airline and airspace transportation (2019).7 The authority has also launched (and recently closed) investigations related to vertical exchanges of information in the payment methods market (2018).8

All signs point, therefore, to a more incisive approach to this type of conduct. We anticipate not only a sharp analysis of the boundaries of lawfulness of information exchanges, but also additional action by CADE and others, which have shown themselves to be keen to investigate such conduct.

Although the intent to investigate seems to be clear, certain key aspects are yet to be clarified. Among these are question marks regarding how such conduct should be treated –- either as per se violations (such as a traditional cartel) or as forms of behavior that need to be assessed through a rule of reason approach. Also, as in many of these new types of conduct, authorities need to clarify how any potential anticompetitive harm from such behavior will be viewed, considering that practices such as information exchange can sometimes also be neutral, bring efficiencies or even be pro-competitive.9

Hub and Spoke Infringements

Competition authorities have been pushing at the boundaries of the concepts of “traditional” cartels and are finding new types of anticompetitive behavior, such as so-called “hub and spoke” cartels. This increasingly common manner of cartelization involves the setting of commercial strategies and the exchange of competitive sensitive information between rival companies (the “spokes”) through a common supplier or a common client (the “hub”). Once again, unlike traditional cartels, such arrangements do not require any contacts or direct interactions between competitors themselves. The usual horizontal feature of collusion is tied to and executed through a vertically positioned agent. Hub and spoke infringements are also increasingly on the agenda of the current generation of enforcers.

Recently (in 2019), the OECD issued a background note, that highlights that practitioners seek clarification to identify explicitly the point at which a legitimate business activity turns into an unlawful horizontal agreement or coordinated conduct that can be subject to antitrust sanctions or even criminal prosecution.10

The increasing review of hub and spoke cartels comes, notably, although not exclusively, from technology such as pricing algorithms, price monitoring software or online platforms, that can be instrumental in promoting these structures.

Despite the rise of hub and spoke investigations in recent years, this type of conduct has been identified by authorities since at least 1939, in the well-known Interstate case, reviewed by the US Supreme Court.11 The antitrust authority of the United Kingdom (the “CMA”) has also been involved in this kind of investigation. Current enforcers all over the world are now ruling on cases involving this kind of infringement, such as those in the US and the national authorities of the European Union, including Belgium12 and Austria. Last year, the Indian authority also analyzed (and dismissed) claims against Uber and Ola regarding an alleged hub and spoke conspiracy.13

In Brazil, CADE’s General Superintendence has already started investigations that involve possible hub and spoke infringements,14 and has also recently highlighted the end of an investigation into alleged hub and spoke cartelization involving Uber.15 CADE’s Tribunal has not yet brought such cases to trial, but has, at least once, analyzed a related matter, in which one of the main discussions involved differentiating an alleged hub and spoke infringement from traditional infringing conduct such as resale price maintenance, or influencing uniform business practices – which goes to show one of the several difficulties involving the assessment and understanding of these new approaches.

Indeed, one of the many possible interesting discussions regarding hub and spoke is precisely pointing out the differences, if any, between such conduct and classical resale price maintenance in which the upstream player ends up provoking a downstream cartel between retailers or distributors. So far, the theoretical difference seems to be that in the hub and spoke case, the cartel is meant by the retailers, that use the upstream agent as a hub to execute the collusive scheme, whereas in the RPM case, the conduct is imposed by the upstream agent itself. If this is the case however –- which by itself is still somewhat unclear – are these conducts different merely depending on the subjective evaluation of who meant what? Is the harm not the same?

The challenges of reviewing this kind of hybrid vertical-horizontal conspiracy are not trivial. It is a hot topic, attracts attention from authorities, and awaits further clarification.16

Algorithmic Collusion

The discussion surrounding the impacts of new technologies and big data over antitrust practice and enforcement has led academics and practitioners to witness a growing concern regarding the potential of algorithms to promote and sustain coordinated behavior among competitors, either explicitly or through tacit collusion or parallel behavior.

The OECD,17 in a recent paper (from 2017), provides examples as to how to police this type of conduct, including assessing the role of algorithms in monitoring the performance of competitors; adjusting changes in demand and supply to facilitate parallelism (doing so even in the absence of explicit collusion) and providing signaling through price increases that may generate a similar reaction from competitors. Another potential risk is the use of machine learning to promote and sustain parallelism without express agreements. The U.S. FTC,18 the UK CMA,19 Portugal’s Autoridade da Concorrência,20 India’s Competition Commission and EU’s the DG Competition21 have also provided insights concerning possible developments in antitrust enforcement involving algorithms.

One interesting common feature of all such evaluations is that, in truth, no one is 100 percent certain of how to tackle antitrust issues insofar as algorithms are concerned. There is some consensus that such collusion may sometimes be very serious, but that algorithms may also be efficient. Consequently, there are still a lot of question-marks as to which types of cases should attract antitrust enforcement in this regard, and how the analysis should proceed. The efficiency side of such conduct seems to call for a rule of reason approach, also a feature that distinguishes it from traditional cartel assessments. 

So far, investigations have reached rather inconclusive decisions. For example, the Indian authority has closed a case against ride-hailing platforms involving claims that pricing algorithms promoted coordination between players without imposing fines, due to lack of evidence.22 Other cases that have raised concerns include allegations against Accenture’s auto parts pricing algorithm Partneo and its use by carmakers. As of yet this investigation has yet to reach any conclusion.23

In Brazil, cartels involving the use of collusion between “friendly” software are not new.24 New and controversial theories of harm involving parallel behavior between algorithms, however, have not yet been subject to any formal investigation. Nonetheless, considering that even consumers have increasingly started to notice such parallel behavior in certain markets, some say it is only a matter of time before the Brazilian competition authority, among as others, will decide to take a closer look at such alleged conduct.

Anti-poaching Agreements

Finally, another new and hot topic is the debate surrounding anti-poach and anti-solicitation agreements between competitors, with regard to employees. Ever since the emblematic decision in United States v. Adobe Sys., Inc., it has become increasingly clear for most commentators that conduct involving competition in labor markets (i.e. competition for workers) are on the radar screens of authorities. Anti-poaching agreements were also the subject of a panel in the recent 2019 ICN Cartel Workshop. 25

So-called naked no-poaching agreements may even be considered per se illegal, and therefore, in certain jurisdictions, subject to trial before criminal courts. These consist of agreements to restrict competition on the basis of salaries and hiring options devoid of economic rationale and specific efficiencies, that are not justifiable by the existence of a joint-venture, cooperative agreements, or other deals that require certain limitations on hiring power.

The sensitivity surrounding this topic has raised concerns not just in the US, where lawsuits (especially private litigation) have been growing,26 but also in Hong Kong,27 Japan,28 the Netherlands, the UK, France, Italy, and Spain.29 In Brazil, there are no ongoing public investigations or procedures by the competition authority assessing anti-poaching agreements under antitrust legislation, but if authorities follow the trends of their peers, as they usually do, such cases might be around the corner.

Conclusion

As we identify new manners of structuring coordinated or collusive behavior, and given the growing concern of antitrust authorities around the globe with new forms of cartels, we anticipate a shift of focus in cartel enforcement policy, and possibly in the type of conduct reported by leniency applicants. If that trend is real, Brazil and others will definitely follow on and we can expect other authorities to open investigations into such conduct soon, unless companies learn and adjust first. The tricky question is figuring out exactly what to do, since several of these forms of conduct first need be better understood and clarified by academics and enforcers.

Click here for a PDF version of the article


1 Eduardo Frade is Partner of VMCA (Sao Paulo), Guest Professor at the FGV Law Program and former General-Superintendent of CADE. Vinicius Marques de Carvalho is Partner of VMCA (São Paulo), Professor of Commercial Law at USP, Yale Greenberg World Fellow and former President of CADE.

2 See, for example, Case no. AT.40135 (the FX spot trading cartel), which comprised collusive behavior fueled by exchange of sensitive information between competing banks.

3 See, for example, “Information exchange: be reasonable” published by the FTC, available at https://www.ftc.gov/news-events/blogs/competition-matters/2014/12/information-exchange-be-reasonable.

4 Organisation for Economic Co-operation and Development (“OECD”), Information Exchanges Between Competitor Under Competition Law, 2010, p. 9.

5 See the breakout session 3 of ICN’s cartel workshop, available at https://icncartelworkshop2019.cade.gov.br/.

6 Administrative Proceeding no. 08700.006386/2016-53.

7 Administrative Proceeding no. 08700.000171/2019-71.

8 Administrative Inquiry no. 08700.005986/2018-66.

9 See, for example, the breakout session 3 of ICN’s cartel workshop. Available at: https://icncartelworkshop2019.cade.gov.br/. See also the decision of the Portuguese Autoridade de Concorrência, available at http://www.concorrencia.pt/vEN/News_Events/Comunicados/Pages/PressRelease_201917.aspx

10 See OECD’s background note in https://one.oecd.org/document/DAF/COMP(2019)14/en/pdf

11 Judgement of the US Supreme Court in Interstate Circuit, Inc. v. United States, 306 U.S. 208 (1939), paras 222, 226-227.

12 MATTIOLI, Evi. Hub and Spoke: Towards a Belgian Precedent? Journal of European Competition Law & Practice, v. 4, n. 4, abr. 2016, p. 261-266.

13 See Global Competition Review article at https://globalcompetitionreview.com/insight/gcr-q3-2019/1202826/when-robots-collude?fbclid=IwAR3QdUcAt9uSpAnzP-55-xyHMLzGQdOTsVS5fcjjtkXKorw_LCcayr1Qmns

14 See, for example, Administrative Proceeding No. 08700.008098/2014-71 (Defendants: Positivo S.A. and others).

15 Paper “Uber: collusion, or unilateral conduct?” published by Mlex and available at https://www.iiede.com.br/index.php/2018/12/30/alexandre-cordeiro-macedo-uber-collusion-or-unilateral-conduct/.

16 See, for example, OECD’s roundtable in December, 2019 to discuss hub and spoke arrangements in competition, available at http://www.oecd.org/daf/competition/hub-and-spoke-arrangements.htm See also Ascola Conference, available at https://univ-droit.fr/actualites-de-la-recherche/manifestations/32052-challenges-to-the-assumptions-at-the-basis-of-competition-law-workshop-on-competition-law-issues.

17 See OECD’s Background Paper on Algorithms and Collusion, available at http://www.oecd.org/daf/competition/Algorithms-and-colllusion-competition-policy-in-the-digital-age.pdf.

18 The FTC held hearings on “The Competition and Consumer Protection Issues of Algorithms, Artificial Intelligence, and Predictive Analytics” in 2018.

19 See CMA’s Pricing algorithms: economic working paper on the use of algorithms to facilitate collusion and personalised pricing, from October of 2018, available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/746353/Algorithms_econ_report.pdf.

20 See Autoridade da Concorrência’s Issues Paper Fair Play: com concorrência todos ganhamos, from July of 2019, available at http://www.concorrencia.pt/vPT/Noticias_Eventos/Comunicados/Documents/Issues%20Paper_%20Ecossistemas%20Digitais%20Big%20Data%20Algoritmos.pdf.

21 See the presentation Pricing algorithms and EU competition law provided by Szilvia Szekely, available at https://ec.europa.eu/competition/cartels/icn/szekely.pdf.

22 See Case No. 37 of 2018, Competition Commission of India

23 The case is cited on Ariel Ezrachi & Maurice E. Stucke. Sustainable and Unchallenged Algorithmic Tacit Collusion. Available at https://ssrn.com/abstract=3282235.

24 In 2005, CADE fined several companies involved in a cartel in the crushed stone market (“Cartel das Britas”). The investigation found that the companies had employed a sophisticated software to drive sales and enforce compliance with the agreements within the cartel. See Administrative Proceeding no. 08012.002127/2002-14. Another case involved the use, by airlines, of the software ATPCO (Airline Tariff Publishing Company), which published air ticket prices in advance, which led to collusive behaviors between the companies. See Administrative Proceeding no. 08012.002028.2002-24.

25 Available at https://icncartelworkshop2019.cade.gov.br/wp-content/uploads/2019/09/20190919_CWS-Draft-Agenda-v10-RevisedNMC.pdf. See also DOJ’s public workshop on competition in labor markets, available at https://www.justice.gov/atr/public-workshop-competition-labor-markets.

26 Most relevantly, see the DOJ/FTC joint document “Antitrust Guidance for Human Rights Professionals,” from October 2016. Recent cases include lawsuits brought against Jimmy John’s franchises, rail industry’s Wabtec, medical schools from Duke University and University of North Carolina, and Domino’s Pizza.

27 See Competition Commission Advisory Bulletin, April 9, 2018 “Competition concerns regarding certain practices in the employment marketplace in relation to hiring terms and conditions of employment.”

28 See, for example, the work of the Study Group on Human Resources and Competition Policy from the JFTC.

29 See, for example, the fines imposed in the Freight Forwarding cartel case.