Notes

Anticompetitive behaviors within the data economy (abuse and data, mergers in the data market, anticompetitive agreements and data). Now, anticompetitive behaviour within the data economy.

• The aim of the following slides is to show you the challenges brought about by the datafication to the traditional competition/antitrust analysis.
• As you know, to date, the debate as to whether the assessment of anticompetitive practices should change because of the datafication is quite lively.

The three pillars

1. Big Data and Abuse of Dominance

2. Big Data and Mergers

3. Big Data and Collusion (Cartels)

4. Big Data and Abuse of Dominance An abuse of dominant position takes place when:

• A company is in a dominant position
• Uses that dominant position to restrict competition in the market
• And competition/consumer's welfare/market efficiency is harmed
• To understand if an abuse of dominant position occurred we then have to:
  • Determine whether the company is in a dominant position
  • Determine whether its behaviour is anticompetitive, e.g. exclusionary (it excludes competitors from the market).
  • Determine that harm to competition/consumers' welfare/efficiency occurred.
• We know how to assess "dominance":
  • Step 1: identifying the Relevant Market (RM), where firms operate in terms of product/services and geographically.
  • Step 2: determining the Market Power (MP), a firm's share in the relevant market.

(In an exam you would have to firstly check whether a company is dominant, then verifying the conduct and also the effects on the market).

A case of abuse of a dominant position by a data company in a data market: the Cegedim case (2014) Cegedim was the main provider of medical information databases. Its clients used its databases through different software**.** Cegedim refused to license the data depending on the software a client used, mainly because behind this idea there was an agreement between Cegedim and the software producer. This was held to be an abuse of dominant position**.** Cegedim held 78% of the medical information database market, its refusal was exclusionary (i.e. they were excluding from the market those who were not accepting their conditions), i.e. it could restrict competition in the market of medical information DBs and market efficiency could be reduced. Therefore,

• Data companies that are in a dominant position in (their own data) relevant market
• And abuse of their dominant position
• Commit an "abuse of dominant position" or a "monopolization" according to Article 102 EU Treaty or Sherman Act Section 1.

2. Big Data and Mergers: Key Cases in Europe "The potential combination of DoubleClick's and Google data collections [...] would be unlikely to give the merged entity a competitive advantage that could not be matched by competitors, given that several competitors both run a search engine and offer ad serving ..." (2008) Reasoning behind

• Are the companies competitors? i.e.:
• Do the companies operate in the same relevant market?
  • Google's relevant market: search engine (makes money by selling ad)
  • Double click's relevant market: technology for online advertising (makes money selling their technology)
• No, they do not, therefore the merger should be authorized.

"There will continue to be a sufficient number of alternative providers to Facebook for the supply of targeted advertising [...] and a large amount of internet user data that are valuable for advertising purposes are not within Facebook's exclusive control".

"In this market [apps] any leading market position even if assisted by network effects is unlikely to be incontestable".

Data privacy concerns "do not fall within the scope of EU competition law" (2014). Reasoning behind

• Are the companies competitors? i.e.:
• Do the companies operate in the same relevant market?
  • FB messenger's relevant market: applications for smartphones
  • Whatsapp's relevant market: applications for smartphones
• Yes, they are, then: would the merge reduce competition in the market?
  • No, it would not, their combined market share is not wide enough.

"The combination of their respective databases does not appear to result in raising the barriers to entry/expansion for other players in this space, as there will continue to be a large amount of internet user data that are valuable for advertising purposes and that are not within Microsoft's exclusive control" - 2016.

Reasoning behind

• Are the companies competitors? i.e.:
• Do the companies operate in the same relevant market?
  • Microsoft's Relevant Market: primarily software solutions for CRM
  • Linkedin's Relevant Market: primarily professional social network services
• Yes, as there is a partial overlap in the secondary online ad-market.
• The overlap though is not of concern as the combined market share is not elevate.

While

• Acquiring a dominant position because of competition on quality/price (on merits) is good (procompetitive), what is bad is abusing it.
• Acquiring a dominant position through a merger or acquisition is bad (anticompetitive).

3. Big Data and Collusion (Cartels) Anticompetitive Agreements and Data: Examples

• Two companies agree on sharing the data that they use to feed the algorithms that they use to determine their products or services prices. (Illegal, this is a cartel, they cannot share the information regarding how the price is calculated.)
• A company develops and algorithm that reacts to its competitors prices on the market. (Legal, there is no agreement, it's just a reaction to others change in price).
• A company decides to make transparent the algorithm that it uses to determine its prices. (Legal, in order to be illegal, the authorities have to find an agreement)

In the end, it's only related to an agreement, if authorities can find two or more companies had a "meeting of minds" then they can be accused of Anticompetitive agreements.

An anticompetitive agreement takes place when:

• Two or more companies "agree" (explicitly or tacitly), i.e. "Meeting of Minds".
• The "agreement" has the object or the effect to restrict competition in the relevant market (agreement on price, quantity or quality).

Terms in place**, agreement** is a broad word: it goes from "expressed agreement" to "parallel practice".

An example: the Topkins case (2015), an anticompetitive agreement implemented through an algorithm:

• Topkins and his co-conspirators agreed to fix the prices of certain posters sold in the United States through Amazon Marketplace
• To implement their agreement, they adopted specific pricing algorithms for the sale of their posters with the goal of coordinating changes to their respective prices.
• In other words: they wrote computer code that instructed algorithm-based software to set prices in conformity with the agreement.
• They were found guilty of violating Sherman Act section 1.

There is more:

• "Classic" cartel, monitored digitally: like in the Topkins Case
• Digital Cartels
  • Companies code and train autonomously learning algorithms to adjust to competitor's prices and strategies.
  • Companies know that these algorithms are adopted.
• Meeting of minds is replaced by meeting of algorithms
• Tacit Algorithmic Collusion
  • A company designs and trains an autonomously learning algorithm to maximize profits.
  • In order to that, the algorithm learn to react to competitors' prices.
  • According to antitrust rules, this is not an agreement, therefore there is no violation.

"What business need to know is that when they decide to use an automated system, they will be held responsible for what it does. So, they had better know how that system works" - Margrethe Vestager.