With the developments in the field of technology, commercial life has ceased to be carried out in physical spaces and has taken its place in digital environments. Thus, it is not only the way of shopping that has changed, but also the competitive environment has begun to change. In fact, technology has created the opportunity for entrepreneurs to collect data about other actors in the market, process this data and make decisions that will provide maximum benefit in line with this processed data. Enterprises determine sales methods and sales prices by developing algorithms based on the data they obtain. In other words, with the digitalization of trade, it has become easier to analyze the current status of the market faster and more economically and to make new decisions according to changing conditions.

Algorithms consist of the same commands following each other and these commands perform the task of doing a certain job through a computer. On the other hand, companies use these algorithms as a profit maximization tool in order to determine the optimal price that will maximize their profitability in line with the supply and demand in the market by taking advantage of technological developments. Companies can also improve their products through algorithms. For example, e-commerce sites use consumers’ past shopping information and information about the pages they visit to make user-specific shopping recommendations to consumers, and thus users who buy a printer online will be relieved of the hassle of searching for a cartridge that is compatible with their printer.[1]

According to Article 4 of Law No. 4054 on the Prevention of Competition; agreements between undertakings, concerted practices, and decisions and actions of associations of undertakings that directly or indirectly aim to prevent, disrupt, or restrict competition in a specific goods or service market or that have or may have this effect are unlawful and prohibited. The purpose of the provision in question is to ensure that undertakings can act independently of each other.

In competition law, the agreement between undertakings depends solely on the existence of a consensus of will to act in a certain manner. There is no formal requirement for an agreement against competition. In cases where the existence of an agreement between undertakings cannot be proven, it is also necessary to look at whether there is a presumption of concerted practice.

In its decision dated 26.7.2007 and numbered 07-62/742-269 in the Hürriyet Newspaper, the Competition Board stated that in order for an act to be qualified as concerted practice, (i) there must be positive contacts between the parties, generally expressed verbally or in writing, including meetings, discussions, exchange of information or research, (ii) this contact must have the purpose of influencing market behavior and particularly eliminating in advance the uncertainty of an undertaking’s future competitive behavior, (iii) it must have the effect of creating or changing the commercial behaviors of the relevant undertakings in a way that will not be completely determined by competitive effects, and that the important issue is that undertakings learn the future behaviors of their competitors and that uncertainty in the market is eliminated.[2]

If the undertakings act in a coordinated manner to set a price above the market price, a collusive relationship is in question. The collusive relationship can be established based on an agreement or can be established as a secret collusive relationship without any agreement or communication. It should be noted that secret collusive relationship is not prohibited by Law No. 4054. However, this does not mean that secret collusive relationship has a legitimate basis.

It should be noted that algorithms can be used in a competitive or anti-competitive manner depending on the way they are used. Therefore, while it is possible to increase consumer welfare through algorithms, it is also possible to act against competition through a secret collusive relationship established by setting a price above the market price.

In a market where an anti-competitive agreement is implemented through algorithms, the actors who are parties to the agreement act against competition in accordance with Article 4 of Law No. 4054. Undertakings can reduce supply, set prices or share the market by illegally increasing the price through algorithms. The main objective in all these activities is to maximize profits by preventing competition. In addition, algorithms can easily detect whether the parties to the agreement are acting in violation of the agreement and also allow for rapid retaliation.

The Competition Board has not yet made a decision on anti-competitive agreements implemented through algorithms. However, there is a decision made on an enterprise managed by David Topkins in the United States. The subject of the case is an algorithm-based computer program used by the enterprise managed by David Topkins. The program in question collects competitor prices and determines the price of posters sold on Amazon within the framework of rules determined by the seller. David Topkins and his competitors agreed to increase the price of some posters sold on Amazon and used a specific pricing algorithm that can coordinate the change in the price of the posters in order to implement this anti-competitive agreement.83 The algorithm in question collects competitors’ price information and determines the price in order to ensure price coordination. In this way, the price balance in the market was determined above the competitive price. As a result of the case, David Topkins accepted the accusation and the file was closed.[3]

It should be emphasized here that the means by which the anti-competitive agreement was implemented does not matter in terms of violating Article 4 of Law No. 4054. As can be understood from the decision, not only human-made anti-competitive agreements but also anti-competitive agreements implemented through tools such as algorithms are in violation of Article 4. In other words, algorithms used as the implementers of an anti-competitive agreement operate here as a result of human will. Therefore, although the implementer of the anti-competitive agreement is an algorithm, the will regarding it belongs to humans. For this reason, the anti-competitive agreement will have its consequences.

Another point that needs to be emphasized is whether businesses will be held liable if algorithms that can make independent decisions act against competition, even though there is no human will for an anti-competitive agreement. Although a self-learning algorithm does not rely on a human will when it acts against competition, businesses should take all necessary measures to prevent an anti-competitive result by taking this feature of the algorithm into account. Otherwise, businesses may be liable for non-competition under competition law.

REFERENCE

Doğan, Algoritma ve Rekabet Hukuku: 4. Madde İhlallerinin Dijital Görünümleri, Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, 2017/2, p.389-433

Law on the Protection of Competition and Related Legislation


[1] Doğan, Algoritma ve Rekabet Hukuku: 4. Madde İhlallerinin Dijital Görünümleri, Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, 2017/2, p.389-433

[2] Doğan, Algoritma ve Rekabet Hukuku: 4. Madde İhlallerinin Dijital Görünümleri, Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, 2017/2, p.389-433

[3] Doğan, Algoritma ve Rekabet Hukuku: 4. Madde İhlallerinin Dijital Görünümleri, Galatasaray Üniversitesi Hukuk Fakültesi Dergisi, 2017/2, p.389-433