Technology continues to have a major impact on the functioning of businesses in the ever-evolving commercial world. One of the increasingly popular recent developments is automated pricing driven by algorithms, machine learning and artificial intelligence. While this technological advancement promises efficiency and dynamic pricing strategies, it also raises significant competition law concerns.
Recently, there have been developments globally in this regard: The US competition authorities FTC and the Department of Justice's Antitrust Division (DoJ), the UK competition authority CMA and the European Commission issued a statement emphasising their commitment to protecting competition in the AI ecosystem.1 The aim of this statement is to ensure effective competition in the rapidly growing artificial intelligence ecosystem, guaranteeing fair and honest practices for both consumers and undertakings. The FTC and the DoJ issued a statement of interest in the case of hotels acting together through algorithms in room pricing.2 In addition, the DoJ is conducting an investigation into RealPage for engaging in exclusionary behaviour in the traditional multifamily rental market through the processing of sensitive information by artificial intelligence.3 The Turkish Competition Authority, has decided to open an investigation against Turkey's giant e-commerce platforms Hepsiburada, Trendyol and Amazon Turkey for implementing automatic pricing mechanisms.4 The Netherlands Authority for Consumers & Markets (ACM) conducted an sector inquiry on algorithmic trading in wholesale energy markets to deepen its understanding of this growing trading method.5 All these issues indicate that the impact of the use of algorithms or artificial intelligence in the markets on competitive dynamics will continue to remain on the agenda of the authorities. The purpose of this study is to evaluate the effects of the use of algorithms in energy markets from a competition law perspective within the framework of the sector inquiry report ("sector inquiry" or "the report") published by ACM.
Algorithms are essentially the design of the solution path by dividing it into steps to solve a specific problem or to achieve a goal. Currently, algorithms provide insight to users by processing large amounts of data quickly. Algorithmic trading in energy markets is crucial for trading companies and traders. Algorithmic pricing can become even more complex with the use of machine learning and artificial intelligence. Furthermore, algorithms lead to various efficiency gains or anti-competitive outcomes depending on the structural factors of the markets. Accordingly, they can maximise efficiency and profit margins and increase competitiveness by enabling undertakings to respond quickly to changes in market conditions, demand and competitor pricing. Also, information asymmetry may be reduced by algorithms that facilitate access to information for consumers. On the other hand, in the case of a horizontal agreement (cartel) or concerted action, the use of algorithms is likely to contribute to the maintenance of the cartel by facilitating the tracking/detection of deviations from the agreement or collusion.6 In this context, in the event that undertakings incorporate similar data in their analyses or conduct simulations on similar consumer populations while determining their strategies, it is probable that they will exhibit identical price movements, which can be assessed in terms of competition law.7
In the context of vertical relationships, the use of algorithms can also facilitate the determination of re-sellers' prices and the systematic monitoring and intervention of compliance with the most favoured customer clause.8 Furthermore, with the transition from an environment where store employees manually changes price tags to an environment where sophisticated computer algorithms swiftly calculated and changes product prices, the utilization of a singular algorithm by multiple competitors could potentially result in the emergence of a hub and spoke cartel.9
In July, the Netherlands Authority for Consumers and Markets ('ACM') conducted a sector inquiry in co-operation with the Netherlands Authority for Financial Markets ('AFM') on algorithmic trading in wholesale energy markets.10 The purpose of the sector inquiry published by ACM on the increasingly important algorithmic trading method is to investigate the general trends in the use of algorithms, the types of algorithms used, the motivations of market participants to participate in or avoid algorithmic trading, the possible impact on the wholesale energy market, and the procedures used by trading companies to ensure that algorithms work as intended. The increasing number of complaints received by the ACM regarding market abuses caused by algorithmic trading was also a factor in conducting this sector inquiry, in addition to comprehending the dynamics and evolving trends of the sector. Finally, it is intended to raise awareness of the new obligations for companies that conduct algorithmic trading under the EU Regulation on Wholesale Energy Market Integrity and Transparency ("REMIT"), which was updated in May 2024.11 ACM has an active obligation to monitor compliance with these rules.
As part of the sector inquiry, ACM conducted interviews with various market participants, algorithmic trading platforms, a technology provider specializing in surveillance services, and conducted a survey of a broader group of market participants. The sector inquiry is focusing on spot markets for electricity and gas trading. Spot markets include wholesale energy markets where energy is delivered on the same day (intraday market) or the next day (day-ahead market). Numerous participants in the spot markets have also engaged in trading in the derivatives markets and have utilized the insights gained from trading in these markets where appropriate. The rationale behind the emphasis on spot markets is the strong correlation with the physical delivery of traded commodities, which can result in distinct market dynamics and characteristics compared to derivatives trading.
The comprehension of algorithmic trading in the energy markets necessitates a profound comprehension of the fundamental trading mechanisms. In the energy supply chain, there are various links between production and consumption. The sector inquiry focuses on wholesale energy markets and analyses the activities of electricity and gas suppliers and producers. It is essential to maintain a continuous equilibrium between supply and demand, which is particularly noticeable in the case of electricity in energy markets. Wholesale energy trading usually takes place between exchange platforms, intermediaries or directly between the parties.
Algorithmic trading in energy markets is a process in which computer algorithms automatically determine trading parameters such as price and quantity, with limited human intervention. This trading method is integrated into broader trading strategies and will also cover manual trading. The revised REMIT defines algorithmic trading as the automatic selection of individual order parameters or order management.12
Three main types of algorithms are mentioned in the report:13
- Execution algorithms: Execution algorithms are used to execute a trading decision made outside the algorithm. Parameters of execution algorithms are set outside the algorithm, after which the algorithm places orders on the trading platform in an optimal way through a specified method, usually within certain price and volume limits.
- Signal generators: Signal generators are algorithms that – based on a set of inputs – signals trading opportunities or other supporting information for trading decisions to traders or execution algorithms.
- Trading algorithms: Trading algorithms differ from execution algorithms in one important respect: the algorithm additionally decides whether an order should be submitted on the trading platform or not. That is the case when signal generators are directly linked to execution algorithms.14
Different types of algorithms use various inputs. Execution algorithms operate using a set of parameters to execute orders in a specific way. These parameters may encompass the price and volume specifics of the order, along with various controls and limitations, such as maximum and minimum prices (price limits) and volume restrictions (volume limits)
Execution algorithms also use historical trading patterns. This can include technical analyses, such as the evaluation of the historical maximum volatility in a five-minute interval. Depending on their complexity, trading algorithms and signal generators may contain hundreds of parameters. These parameters can be derived from fundamental data such as weather information or data from Transmission System Operators (TSO) and market participants. Technical data includes historical and current prices and price fluctuations over various time periods.
The output of different algorithms will also vary. Both execution and trading algorithms generate orders that are sent to the trading platform. Signal generators provide forecasts and/or recommendations to manual traders or other algorithms. Forecasts may include expected prices, supply, demand, volatility or portfolio information.
The use of these algorithms varies across markets, with execution algorithms dominating the natural gas market, while all algorithm types are used equally in the electricity market. Algorithms can be applied to several trading strategies, such as minimizing price impact by spreading order volume over time, spread trading, which aims to profit from price differences, and market making, which aims to be present on both sides of the order book for most of the trading day.
According to the sector inquiry by ACM, the use of algorithms in energy markets is increasing and is expected to become even more widespread in the future. One of the key factors behind the growth of algorithmic trading is the energy transition. An energy transition is generally characterized as a major structural change in the supply and consumption of energy in an energy system. In this context, energy transition currently refers to the process of "transition to sustainable energy", or in other words "transition to renewable energy", to mitigate climate change. However, the increase in the transition to renewable energy causes an imbalance in the electricity supply. Consequently, renewable energy production becomes challenging to predict, and positions need to be balanced in a short period of time.
While the use of algorithms is widespread in the electricity spot market, it is comparatively less common in the gas spot market but is steadily increasing. The motivating reasons for algorithm use include efficiency, asset optimization, and risk mitigation, while the reasons for avoiding algorithm use include not having the necessary IT knowledge and competence as well as not seeing algorithm use as a necessity or not being interested in algorithm use in general. However, according to the survey results and interviews, automation and efficiency are the main reasons for using algorithms among market participants. In addition, algorithms provide additional benefits such as reducing the price impact on the market by splitting large volumes of orders into smaller pieces, data processing capacity and speed advantage. However, the cost of know-how, expertise, and IT investments make some market participants avoid using algorithms. In certain markets, it is becoming increasingly difficult to trade without algorithms due to some disadvantages associated with the lower speed of manual trading. The growing importance of the energy transition and renewable energy are key drivers supporting the use of algorithms and helping market participants manage their renewable energy portfolios.
According to the report, the impact of algorithms on the market varies during normal market conditions and periods of increased volatility. In normal conditions, algorithms improve market efficiency by increasing liquidity and facilitating price discovery. Thanks to their fast data processing capacity, they refine prices and narrow bid-ask spreads. However, during periods of high volatility, algorithms can pose serious risks and reduce liquidity. In sum, while the results of the sector inquiry reveal that algorithms increase liquidity and the competitiveness of market participants in energy markets, they also raise the risks of increased volatility and market manipulation. Moreover, the fact that algorithms are very similarly programmed can lead to significant price fluctuations in the market, and some respondents highlighted the risk of increased volatility. On the other hand, it is argued that if algorithms are well-programmed, it is possible that they include safeguards to prevent high volatility and can serve as a useful tool. Their sensitivity to manipulative information input and fast data processing capabilities raise the need for advanced analysis tools for market observers to detect such manipulations. While algorithms reduce market effects by breaking high volume orders into small pieces, they can make it difficult for manual traders to compete and negatively affect market transparency.
Although market participants in the sector inquiry indicated to ACM that they have compliance and risk measures in place for their algorithms, ACM did not assess the effectiveness of these procedures in the sector inquiry. According to market participants, the procedures for the development, testing and monitoring of algorithms are detailed and reviewed periodically. The algorithm development process begins with market and data analysis, followed by testing under different market conditions using realistic or historical data. Market participants keep documentation on the development and use of algorithms, such as the rationale and strategy, tests performed, programming code and log files. The majority of market participants utilize real-time monitoring tools to detect manipulative trading behaviors and unexpected movements of the algorithms. In addition, most algorithms have a "kill switch" that allows traders to stop all algorithmic trading activity when necessary. On the other hand, it is also stated that no matter how well-intentioned the compliance measures are, the risks of adverse behavior cannot be completely eliminated. As the complexity of algorithms increases, explainability and predictability become more difficult. Self-learning algorithms may make incorrect decisions in extreme market conditions. In this context, it is critical to monitor and control algorithms in order to ensure market order and security.
The revision of REMIT in 2024 imposed new responsibilities on market participants and national regulatory authorities engaged in algorithmic trading in EU wholesale energy markets. These changes aim to address the growing risks associated with algorithmic trading. Within this context, the prominent changes to REMIT can be listed as follows:
- Additional safeguards for fair competition and price formation in wholesale energy markets,
- An obligation for market participants from non-EU countries to appoint a representative in the EU Member State where they are registered to interact with energy regulators,
- Extending the scope of REMIT to apply to contracts and derivatives for the storage of electricity and natural gas (including LNG) in the EU, as well as energy products regulated by the Netherlands Authority for the Financial Markets (AFM).
These revisions require improvements to management, testing, and monitoring of trading systems, as well as the accurate documentation of information. The REMIT revision also clarifies the ACMs oversight powers and duties over algorithmic trading. In this context, ACM is authorized to request details and documents related to algorithms to verify compliance. Furthermore, market participants who are registered with the ACM are obligated to report their utilization of algorithms to the ACM and the EU Agency for Cooperation of Energy Regulators (ACER) The revision increases the regulatory powers regarding algorithmic trading in Dutch energy markets and aims to monitor market participants' compliance processes through a structured oversight process. Accordingly, the ACM plans to integrate information from market research on algorithmic trading into the oversight of the Dutch wholesale energy market and aims to provide comprehensive oversight in cooperation with partner regulators such as the AFM.
In general, the ACM will persist in regulating and monitoring the trading conduct of market participants, with a particular emphasis on adhering to algorithmic trading obligations. In this regard, the ACM will persist in collaborating closely with the AFM and other regulatory bodies.
Considering all of these factors, it can be inferred that the monitoring of algorithmic pricing in energy markets by various authorities indicates the seriousness of its consequences. Indeed, algorithmic pricing practices not only enhance market efficiency by facilitating price discovery, but also result in market transparency, which may lead to explicit or implicit agreements or coordinated behavior among competitors, potentially leading to anti-competitive effects.
CONCLUSION
ACM mentions the following key points in its sector inquiry report:
- Algorithms are getting used more and more by traders in energy markets. Many market players still use rule-based algorithms that set their own parameters, but more sophisticated self-learning algorithms are becoming increasingly common.
- Generation has become less predictable due to the energy transition towards renewable energy, and traders need to manage their positions at the last minute. In order to manage this process, the use of algorithms by many traders is increasing.
- The participants indicated that they possess diverse protocols for processing algorithms. These procedures encompass the development of algorithms, testing, data storage, establishing trading limits, implementing control procedures, and overseeing the use of algorithms. However, in the sector inquiry, the ACM did not examine whether these procedures are efficient and are actually applied.
- The ACM and AFM will continue to closely monitor whether transactions in the markets are occurring in accordance with the regulations. Specifically, the ACM emphasized that it will persist in monitoring and regulating algorithmic trading and market participants' compliance with obligations derived from the revised REMIT.
It is possible to say that electricity markets, which are not frequently on the agenda of competition authorities due to their generally regulated structure compared to many other markets, have recently started to become popular among the authorities. In terms of electricity markets, the increasing technology-oriented life in daily life continues to increase the demand for electricity. In accordance with this increase in electricity supply, renewable electricity generation facilities are being implemented in electricity generation due to the sustainability-oriented approach. However, this situation leads to an increase in the uncertainty regarding the supply of nature-dependent electricity. Today, the most popular topic where electrification has an impact is the electrification of automobiles. The widespread adoption of electric vehicles is anticipated to result in a substantial demand for electricity. Another issue in this regard is the infrastructure requirements that high demand will lead to. Electric cars are becoming more popular, which has led to new businesses like charging stations. To keep up with all the latest developments, algorithms are becoming more and more prevalent in businesses. This is because increasing renewable energy generation to ensure electricity supply from green sources increases uncertainty in electricity trading markets, and enterprises are trying to overcome these challenges by using algorithms. Furthermore, in order to cater to the projected surge in electricity demand resulting from electric vehicles, electricity distribution companies may use algorithms to optimize infrastructure investments. Similarly, electricity retail sales companies may utilize algorithms to maximize their effectiveness in the newly emerging market, determining matters such as investment and pricing for electric vehicle charging stations or emerging business areas.
The first thing to note is that it is undeniable that the use of algorithms in the electricity trading markets addressed by this sector inquiry yields efficiencies in terms of balancing processes in the electricity market. The sector inquiry report also considers the negative effects that may arise in the market due to the excessive use of algorithms by market actors in trading processes. As in other markets, the algorithm usage may lead to horizontal collaborations or coordinated actions in electricity trading markets, a potential risk that should not be overlooked. Furthermore, it may also be expected that the processing of data within the dominant incumbent electricity supply companies by means of algorithms to be used in newly emerging service areas (charging stations market, etc.) may also be expected to be discussed within the framework of abuse of dominant position. As a result, it would be advantageous to examine the purposes and consequences of the application of algorithms correctly and take steps to mitigate any possible violations of competition law, without waiting for the competition authorities to take action.
Footnotes
2. FTC and DOJ File Statement of Interest in Hotel Room Algorithmic Price-Fixing Case – 28.03.2024
4. Turkish Competition Authority Investigation Numbered 23-49/940-M – 15.11.2023
5. ACM: Use of Algorithmic Trading on the Energy Market Has Increased Strongly – 17.07.2024
6. Pelin Teber Karabudak, "Algoritmik Stratejiler Yoluyla Rekabete Aykırı Anlaşmalar", p.23
7. Ulrich Schwalbe, "Algorithms, Machine Learning, and Collusion" Journal of Competition Law and Economics 4, vol. 14 (2018): 568- 569.
8. Pelin Teber Karabudak, "Algoritmik Stratejiler Yoluyla Rekabete Aykırı Anlaşmalar", p.26
9. EZRACHI, A. and M.E. STUCKE (2016), "Virtual Competition" – The Promise and Perils of The Algorithm-Driven Economy, Harvard University Press, Cambridge, Massachusetts. p. 585
10. ACM Sector Inquiry Report: Algorithmic Trading in Wholesale Energy Markets (acm.nl)
11. EU Regulation on Wholesale Energy Market Integrity and Transparency ("REMIT") – 07.05.2024
12. According to REMIT, algorithmic trading refers to trading in wholesale energy products where a computer algorithm automatically determines the individual parameters of trading orders, such as whether to initiate an order, the timing, price, or quantity of the order, or how the order is managed after it is placed, with limited or no human intervention. This includes the use of any system for directing orders to one or more organized markets, or for processing orders or confirming orders, or for post-trade processing of transactions that does not involve determining any trading parameters.
13. This distinction is made by the ACM solely for the purpose of the sector review and does not necessarily align with the scope of the definition of algorithmic trading under the revised REMIT. It stated that any additional guidance on the scope of algorithmic trading would be coordinated by ACER.
14. ACM Sector Inquiry Report, p.16
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
