Site icon SCC Times

Changing Dynamics of Algorithmic Collusion: An Analytical Study

analytical study

Introduction

Algorithm collusion refers to the coordination between algorithms that are intended to work independently but end up working together to produce a result that is harmful or unfair. This can occur when algorithms are designed to make decisions that are influenced by each other’s behaviour, resulting in outcomes that are not in line with their original purpose. One example of algorithm collusion is when two algorithms that were designed to compete with each other end up working together to manipulate the market. This can lead to price fixing, reduced competition, and other harmful outcomes.1 Another example is when algorithms that are designed to provide fair and unbiased outcomes end up collaborating in ways that reinforce existing biases or discrimination.

This can occur when algorithms are trained on biased data, leading them to make decisions that perpetuate these biases. Algorithm collusion is a concern because it can undermine the transparency and accountability of algorithmic decision-making and have negative consequences for individuals and society. As a result, it is important to ensure that algorithms are designed and monitored to prevent collusion and to ensure that they are producing fair and ethical outcomes. This article analyses algorithmic collusion and examines India’s experience dealing with it. It also examines whether the Competition Act of 20022 (hereinafter “the Act”) is adequately equipped to examine algorithmic collusion and the positions of western countries in dealing with algorithm collusion.

How collusion works and occurs

Collusion refers to a situation in which two or more entities work together in a coordinated and secretive way to achieve a common goal. In the context of algorithms, collusion typically refers to the use of multiple algorithms in a coordinated way to manipulate the outcome of a decision-making process, often for some illicit or unethical purpose. Algorithmic collusion can occur in several ways, depending on the specific context in which the algorithms are being used. According to Ariel Ezrachi and Maurice E. Stucke, algorithms can be used for collusion in four ways:3

  1. Messenger scenario: In the messenger scenario, the market participants employ computers or a single algorithm as a means of collusion. Such an instance occurred in United States v. David Topkins4, when the conspirators opted to sell their posters using a single pricing algorithm to help make sure price parity. This agreement was found to be illegal. Similar findings were made in 1994 regarding the sharing of a computerised online booking system by six airlines, which allowed for collusive price setting and was thus declared anti-competitive.5 So, the messenger scenario is the situation in which the conspirators opt to utilise the algorithm to perform collusion.
  2. Hub-and-spoke conspiracy: This type of conspiracy refers to an agreement formed by vertical or horizontal players (spokes) via the use of a platform (hub), which is similar to an indirect agreement. A single price algorithm is utilised in this instance, as in the messenger scenario, but it is the algorithm developer who forces the players to collude. As a result, the agreement is focused on how the hub will be used.
  3. Predictable agent: In this type, there is no agreement between rivals. Each company unilaterally implements its own pricing algorithm, and as predictable agents, they keep an eye on and adjust to each other’s prices. Hence, even though competitors do not adopt the same algorithm, tacit collusion is still affected by programming algorithms to adjust to each other’s price.
  4. Autonomous machine: It involves self-learning learning algorithms that cooperate despite not being designed to do so in response to market data or price changes. According to the Organisation for Economic Cooperation and Development (OECD) Report, it is unclear how machine learning algorithms could arrive at a collusive result; but, once it has been established that market conditions are conducive to collusion, it is possible that algorithms learning more quickly than humans could achieve a cooperative equilibrium.

Algorithm collusion in India

Algorithmic collusion is a growing concern in India, particularly in the context of online marketplaces and online advertising. Algorithmic collusion can have a significant impact on competition and consumer welfare in India, as it can lead to reduced competition, increased prices, and reduced consumer choice. In India, the Competition Commission of India (CCI) is responsible for enforcing the Act, which aims to promote competition and protect consumers from anti-competitive practices. The CCI has the power to investigate and take action against companies that engage in anti-competitive behaviour, including algorithmic collusion. However, algorithmic collusion can be difficult to detect and prosecute in India, as it can occur in complex and interconnected systems and can be difficult to understand and prove.6 The rapid pace of technological change in digital markets also means that new forms of algorithmic collusion can emerge quickly, making it challenging for the CCI and competition authorities to keep up.

Despite these challenges, it is important for the CCI and other competition authorities in India to take steps to address algorithmic collusion and to promote fair and open competition in digital markets. This can include increasing their understanding of algorithms and their potential for harm, improving transparency in digital markets, and taking action against companies that engage in anti-competitive behaviour. By doing so, the CCI and competition authorities in India can help to ensure that the benefits of digital markets are shared fairly among all the participants and that the consumers are protected from harm.

Algorithm collusion can occur in various sectors, such as e-commerce, ride-hailing and social media. One example of algorithm collusion in India is the alleged collusion between ride-hailing platforms such as Ola and Uber to fix prices and control the market.7 In 2018, the CCI conducted an investigation into this matter against both the companies. However later, CCI rejected allegations of price fixing against both the companies and the same observations were upheld by National Company Law Appellate Tribunal (NCLAT)8. Similarly, in 2014, the CCI initiated an enquiry against price cartel in the Airline Industry.9 The Airlines have used third-party software to assist them decide, implement, and dynamically adjust the rates given to customers in real time, due to the remarkable growth of air travel and technology innovation over the years.10 Each of these software is built on a constantly evolving complex of algorithms that calculate the fares by taking into account variables including seasons, actual bookings, competitor pricing, etc., to determine the airfares.

Interestingly, CCI acknowledged the existence of software that was similar to or identical to that used by the airlines, but more importantly, it applied the very fundamental principle of the evidentiary standard needed to prove a cartel’s conduct using either direct or circumstantial evidence, the sector is competitive since many established firms are losing market share to a new or recent player over a period of 4 to 5 years due to considerable fluctuations in market shares of competing airlines.11 Furthermore, the CCI could not discover any concrete proof that other airlines had used the same software to set prices. In respect to the process of pricing determination, the CCI noted that the use of algorithms was only done to help with genuine price determination in an industry that requires dynamic pricing and was not done with the intention of implementing price cartel. The CCI noted that the involvement of a “human” element to decide the final prices indicated that the use of algorithms was only to facilitate genuine price determination in an industry that requires dynamic pricing and was not done with a view of implement price cartel.12

Regulations under Indian competition law

Section 3(3) of the Act13 prohibits collusion between enterprises. It states that any agreement between enterprises, decisions by associations of enterprises, or concerted practices by enterprises that have the effect of preventing, restricting, or distorting competition within India shall be considered anti-competitive. In addition to Section 3, the Act also prohibits abuse of dominant position under Section 414. A company that holds a dominant position in the market is prohibited from abusing that position by engaging in anti-competitive practices such as algorithmic collusion. CCI has taken a multi-pronged approach to curb algorithmic collusion and other forms of anti-competitive behaviour. The measures include market studies in various sectors, including e-commerce, to identify anti-competitive practices, including algorithmic collusion. These studies provide insights into the functioning of the market and help CCI to take appropriate actions to prevent and remedy anti-competitive conduct. CCI has also done collaboration with international agencies, including competition authorities of other countries, to share information and best practices for addressing anti-competitive behaviour.15

If the CCI finds that a company has engaged in algorithmic collusion, it can impose penalties and take other measures to restore competition in the affected market. The penalties can include a fine of up to 10% of the company’s average turnover for the preceding three financial years.16

Challenges ahead of CCI

Despite various efforts, the CCI will face several challenges in dealing with algorithmic collusion. Some of the key challenges are:

Difficulty in detection: Algorithmic collusion can occur in complex and interconnected systems, making it difficult to detect and understand the underlying causes. This can make it challenging for competition authorities to identify instances of algorithmic collusion and to take appropriate action.

Lack of transparency: Many algorithms used in digital markets operate behind the scenes, making it difficult to understand how they are making decisions and to identify instances of algorithmic collusion.17

Difficulty in proving anti-competitive intent: The Act requires evidence of anti-competitive intent, but algorithmic collusion can occur without any explicit intention to harm competition. This can make it difficult for competition authorities to bring cases against companies for algorithmic collusion.

Evolving technology: The rapid pace of technological change in digital markets means that new forms of algorithmic collusion can emerge quickly. This can make it challenging for the Act and its enforcement to keep up and to effectively address algorithmic collusion in a timely manner.

Despite these challenges, the Act can still play an important role in curbing algorithmic collusion in India. Competition authorities can take steps to improve their understanding of algorithms and their potential for harm, and to increase transparency in digital markets. Additionally, the Act can be used to address instances of algorithmic collusion that are identified, and to impose penalties and remedies on companies that engage in anti-competitive behaviour.

Position of foreign countries

Algorithmic collusion is a growing concern in the United States and Europe, as it can lead to anti-competitive outcomes in digital markets. In the US, competition authorities, such as the Federal Trade Commission (FTC) and the Department of Justice, have a role to play in addressing this issue. Similarly, in Europe, the European Commission (EC) is responsible for enforcing competition law and addressing anti-competitive behaviour. Both US and Europe have been taking a proactive approach to addressing algorithmic collusion in digital markets, including through the use of various tools and approaches, such as investigations, enforcement actions, and policy guidance. In recent years, there have been several high-profile cases in the US and Europe that have involved allegations of algorithmic collusion. For example, in 2016, the FTC settled with Google over allegations that the company had used its dominance in the online search market to harm competition.18 Similarly, in 2017, the EC fined Google €2.4 billion19 for using its dominance in the online search market to harm competition. To address algorithmic collusion, the competition authorities have used a variety of tools and approaches, including enforcement actions, investigations, and policy guidance. They have also been engaging with stakeholders, including industry, consumer groups, and academia, to gather information and insights on the issue and to promote fair competition in digital markets.

Conclusion

While the use of algorithms can bring significant benefits to businesses and consumers, it can also be used to facilitate anti-competitive behaviour, which can harm competition, consumers, and innovation. To address the challenges posed by algorithmic collusion, the regulators around the globe have taken several steps, including issuing guidelines, conducting market studies, providing leniency provisions, collaborating with international agencies, and conducting capacity-building programmes. To effectively address algorithmic collusion, the regulators will have to enhance their technical capabilities and develop specialised expertise to detect and analyse complex algorithms.

Additionally, greater collaboration with the private sector, academia, and civil society organisations can help them to stay ahead of the curve and effectively address the challenges posed by algorithmic collusion. Given the growing importance of algorithms and artificial intelligence in the digital economy, it is possible that we will see more cases of algorithmic collusion in India in the coming years. The CCI and other competition authorities will need to remain vigilant and take action when appropriate to ensure that digital markets remain competitive and that consumers are protected.


* IVth year student, BBA LLB, UWSL, Gandhinagar. Author can be reached at vinaysachdev8866@gmail.com.

1. Corporate Finance Institute, “Collusion: When Businesses or Individuals Influence Pricing in Particular Market Segments to their Advantage”, (9-1-2023) available at <https://corporatefinanceinstitute.com/resources/economics/collusion/> (last visited on 18-2-2023).

2. Competition Act, 2002.

3. Ariel Ezrachi and Maurice E. Stucke, “Algorithmic Collusion: Problems and Counter-Measures, Organisation for Economic Cooperation and Development”, (31-5-2017), available at <https://one.oecd.org/document/DAF/COMP/WD(2017)25/en/pdf> (last visited on 18-2-2023).

4. No. CR 15-00201 WHO (2015, US NDC).

5. Martin Tolchin, Six Airlines Settle Suit by Government on Fares, The New York Times, 18-3-1994, available at <https://www.nytimes.com/1994/03/18/business/six-airlines-settle-suit-by-government-on-fares.html> (last visited on 16-3-2023).

6. OECD (2017), Algorithms and Collusion: Competition Policy in the Digital Age, available at <www.oecd.org/competition/algorithms-collusion-competition-policy-in-the-digital-age.html> (last visited on 18-2-2023).

7. Samir Agrawal v. Competition Commission of India, 2020 SCC OnLine NCLAT 811.

8. Samir Agrawal v. Competition Commission of India, 2020 SCC OnLine NCLAT 811.

9. AZB & Partners, “Pricing Algorithms| CCI’s First Major Encounter with Assessing New-Age Collusions”, (15-3-2021), available at <https://www.lexology.com/library/detail.aspx?g=80f844ba-f9e1-4bbb-8268-0054179b7ef7> (last visited on 16-03-2023).

10. Alleged Cartelisation in the Airlines Industry, In re, 2021 SCC OnLine CCI 3.

11. Alleged Cartelisation in the Airlines Industry, In re, 2021 SCC OnLine CCI 3.

12. Alleged Cartelisation in the Airlines Industry, In re, 2021 SCC OnLine CCI 3.

13. Competition Act, 2002, S. 3(3).

14. Competition Act, 2002, S. 4.

15. Manas Kumar Chaudhuri, Armaan Gupta and Siddharth Bagul, “CCI Cnters Memorandum on Cooperation with Japanese Antitrust Regulator”, Khaitan & Co., (12-7-2021) available at <https://www.khaitanco.com/thought-leaderships/CCI-enters-memorandum-on-Co-Operation-with-Japanese-Antitrust-Regulator> (last visited on 19-2-2023).

16. Vivek Agarwal, “Competition Law Amendments: Of Penalties and Misses”, Livemint, (14-2-2023), available at <https://www.livemint.com/opinion/columns/competition-law-amendments-of-penalties-and-misses-11676313277888.html> (last visited on 20-2-2023).

17. Madhavi Singh, “Algorithmic Collusion in Flight Pricing in India”, (29-11-2018), available at <https://lawschoolpolicyreview.com/2018/11/29/algorithmic-collusion-in-flight-pricing-in-india/> (last visited on 22-2-2023).

18. OECD (2020), Abuse of Dominance in Digital Markets, <www.oecd.org/daf/competition/abuse-of-dominance-in-digital-markets-2020.pdf> (last visited on 22-2-2023).

19. OECD (2020), Abuse of Dominance in Digital Markets, www.oecd.org/daf/competition/abuse-of-dominance-in-digital-markets-2020.pdf (last visited on 22-2-2023).

Exit mobile version