It is now clear that online markets’ promises have only been partially realised. A few retailers control the online market. Online prices are sometimes slightly lower than in-store prices, but this is not always the case. Furthermore, the price of a product can vary greatly between online retailers. Prices on websites like Amazon can change throughout the day and are sometimes significantly higher than the suggested retail price. Online markets are clearly not as competitive as some initially believed.
Pricing algorithms are increasingly being used by online retailers. Rather than a human setting prices, a computer programme can rapidly monitor market conditions, including the behaviour of competing retailers, and adjust prices autonomously in near real-time. The fact that these developments have occurred concurrently—more responsive pricing behaviour, as well as large price differences that consumers may pay for identical products—contradicts initial expectations about online market competition.
In our daily lives, we all use pricing algorithms, whether we realise it or not. Whether it is to order a taxi, shop in an online marketplace, book a hotel for our next vacation, or fill up the car with gas. Pricing strategies are implemented using powerful algorithms hidden behind well-known platforms such as Amazon, Uber, and Google. These new tools are an essential part of the globalisation process, as they are inextricably linked to market digitalisation, which “triggers a domino effect that promotes wider use of algorithms in an industry”.1
What is a pricing algorithm?
A pricing algorithm is essentially a computer software that automatically changes prices depending on recent and historical data about demand, costs, or competitors’ prices. The price of airline tickets was one of the first businesses to adopt algorithms. The number of markets affected by the adoption of algorithms has changed drastically with the growth of online markets, and significant funds have been invested to enhance price algorithms in a number of different ways.
Numerous theories exist on the effects of pricing algorithms. Most economists concur that algorithms with price adjustments based on costs and/or demand situations have the potential to boost efficiency. However, there is rising worry that additional elements of pricing algorithms may result in a reduction in competition and increase prices.
Pricing algorithms generally fall into two categories:
(i) Algorithms which are developed by businesses to set the prices for products which they produce and sell to consumers. Generally, they are produced by larger companies with the resources and expertise to develop them.
(ii) Algorithms which are developed by specialist algorithm development firms. They do not specifically tailor their algorithm to one product or market, and instead licence their algorithms for other companies to use. These are sometimes bundled with a broader suite of “business intelligence” services.
Algorithmic pricing differs from conventional approaches in that prices are set manually by a pricing manager or analyst. This is because pricing algorithms operate without human interference.
First, algorithms can let merchants establish pricing in accordance with computer-encoded rules. The rules may be predetermined, or they may change over time in response to factors like machine learning techniques or human input. As we will describe below, the ability of enterprises to adhere to these regulations in the short term can have significant effects on both prices and competition.
Second, algorithms can absorb more data and make adjustments more quickly than conventional pricing techniques because computers are faster at doing complex calculations.
In order to change pricing depending on information that comes in at a high frequency, retailers frequently configure their algorithms to run at regular intervals—once per day or once per hour. Recent sales, inventories, or outside elements like weather predictions are some pieces of information that could affect prices. Algorithms have the potential to deliver goods more effectively than traditional pricing because of their capacity to react to rapidly changing conditions.
The phenomenon’s recent emergence and particular relevance in the economy today made it a topic of particular interest to write about, the multidimensional nature of pricing algorithms is even more fascinating. Pricing algorithms are evolving at the intersection of data, market power, and competition law, as well as multiple other areas of law such as intellectual property, privacy, data protection, and consumer protection. As a result, pricing algorithms are particularly difficult to envision, frequently necessitating not only legal but also economic and technical approaches.
The potential inputs into a pricing algorithm could be any piece of information that would be relevant to price formation, for example:
competing firm’s prices;
firms’ past pricing/profit/revenue data;
individual customer information, including their purchase or browsing history or other indicators;
market information such as competitors’ stock levels (e.g. whether it is in-stock or not, or more detailed information if this has been made publicly available by competitors);
external information such as weather patterns; and
firm costs, such as production, storage, and fulfilment.
Algorithms can process this information using a set of simple rules, such as price matching the competitor with the lowest price. In this case, the algorithm does not benefit from having past data to draw from. This is because the algorithm does not “learn” from past experiences, but simply chooses prices based on pre-set rules. In spite of the benefits of algorithms outlined above, there is a growing opinion in competition policy literature which raises concerns about the potential of algorithms to cause consumer harm. One of the main theories of harm relates to the possibility that pricing algorithms might lead to collusive outcomes, with consumers paying higher prices than in a competitive market.
Competition law is a body of legislation designed to prevent market distortion caused by anti-competitive business practices. Competition law is also known as anti-trust law in the United States, Canada, and the European Union.
The goal of competition law is to ensure a fair marketplace for consumers and producers by prohibiting unethical practices intended to gain a larger market share than would be possible through honest competition. Anti-competitive practices not only make it difficult for smaller companies to enter or succeed in a market, but they also result in higher consumer prices, poorer service, and less innovation.
Anti-competitive behaviour includes, among many other examples, predatory pricing, which occurs when a monopoly or oligopoly charges an exorbitant price for something that consumers have little choice but to buy; price fixing, which involves collusion between would be competitors to set similar prices for products; bid rigging, which involves collusion to determine the winner of a contract in advance; and dumping, which involves selling a product at such a low price that consumers have no choice but to buy. Despite the fact that each country has its own unique laws, competition law typically forbids these behaviours.
Effects of pricing algorithm on competitive interplay
As previously stated, the implications of pricing algorithms have significantly challenged traditional market dynamics, introducing a very powerful technological power based on artificial intelligence. Pricing algorithms can benefit the entire competitive process by intensifying it and achieving its initial procompetitive promise by answering customers’ expectations in an ever more advanced and precise manner. These advantages, however, must be balanced against the potential anti-competitive effects of pricing algorithms (collusion, monopoly, price discrimination, etc.), the so-called “perils” of the algorithm-driven economy.
Both the significant advantages of pricing strategies for customers as well as the procompetitive implications of pricing algorithms on the economic process must be emphasised.
Pricing algorithms clearly help businesses, but they also help consumers by giving them access to fresh, improved, and more specialised goods and services. This is mostly because some pricing algorithms, such those used in price comparison websites, offer improved transparency. By decreasing the information asymmetries and transaction costs that previously existed between sellers and customers at the expense of the latter, such greater transparency will assist consumers in making more informed decisions.
Transparency generally results in the efficient use of resources. Customers may now examine a wider range of options, and if the quality-price ratio of one product is unsatisfactory, they can choose to move to a different product from a rival company. Customers gain from this technique since it helps to reduce their search expenses. Instead of having to travel from one location to another to compare pricing for a certain item, they can now get all the necessary information on online platforms. In this way, knowledgeable customers are less likely to be vulnerable to monopolistic pricing or higher prices. The user-defined criteria that allow customers to select the highest price they are willing to pay for an item or to view the overall user rating help to some extent.
Digital comparison websites that use pricing algorithms are then qualified to play the part of “digital butlers” or “digital half,” guiding the clients in making judgments about what to buy. For instance, personalisation algorithms go a step further to align the recommendations provided by online marketplaces and to tailor the outcomes to the needs and interests of the user.
These concurrent changes in businesses or platforms put more pressure on suppliers to innovate and compete to maintain their market dominance, which in turn directly encourages them to do so. By placing businesses under ongoing pressure, this beneficial mechanism creates dynamic efficiencies, enhancing market efficiency. In fact, algorithms are supporting improved supply and demand adequacy by increasing market transparency since they can stop unmet demand, and supply excesses and keep the market in a state of continual balance. As a result of the customer’s ability to quickly choose the provider or product that best meets their demands thanks to these internet tools, businesses and suppliers may now adjust in order to manage and optimise their inventories. It enables effective resource allocation.
As more firms and industries adopt pricing algorithms, there is growing concern that, rather than collusion dying, new forms of collusion will emerge. Pricing algorithms may act beyond the reach of the law by using subtler means, favouring tacit collusion, or amounting to price discrimination, as identified by the competition literature. Despite the numerous potential pro-competitive justifications for the use of pricing algorithms, there is concern that they may, inadvertently or otherwise, lead to anti-competitive market outcomes. When pricing algorithms rely on proprietary data, for example, they may result in unwanted forms of price discrimination or increased barriers to entry. Furthermore, as we will see below, the use of pricing algorithms may result in another major concern: algorithmic collusion.
At a high level, there are at least four different ways in which pricing algorithms can lead to collusion, each with varying degrees of practicality:
Explicit algorithmic collusion
According to a 2017 EU e-commerce sector inquiry, the majority of online retailers use algorithms to monitor competitor prices, with roughly two-thirds using algorithms to automatically adjust prices in response.
However, the increasing prevalence of automated pricing may make it easier for competing managers with malicious intent to implement a price agreement. Instead of constantly discussing and calibrating joint pricing behaviour, they can now use simple algorithms.
Algorithmic hub-and-spoke collusion
A “hub-and-spoke” structure is another way pricing algorithms can undermine competition. In this case, a common supplier (the “hub”) coordinates the prices of downstream competitors (the “spokes”) without requiring these downstream competitors to form a horizontal agreement among themselves.
While illegal, building a solid case based on allegations of hub-and-spoke collusion is generally more difficult than explicit horizontal collusion because it requires proof that the downstream “spokes” competing with each other are aware of the likely collusive consequences of giving up their pricing autonomy.
Tacit algorithmic collusion
Collusion is not always obvious. Pricing algorithms may also enable firms to unilaterally implement strategies that prevent aggressive pricing in the market, resulting in a tacit collusive outcome that is nearly impossible to prosecute.
However, reaching a stable but silent agreement on high prices is difficult. Firms’ cost structures and inventories differ, and new firms may enter the market and demand may fluctuate—factors that destabilise a tacit agreement to keep prices high.
Simultaneously, the practical feasibility of tacit human collusion due to algorithms should not be overlooked.
Autonomous algorithmic collusion
The greatest concern, however, may arise when algorithms can learn to optimally form cartels on their own—not through instructions from their human masters (or some irrational behaviour), but through optimal autonomous learning (i.e. “self-learning” algorithms). If such an outcome occurs, it may be difficult to prosecute because businesses using such algorithms may be unaware of what strategy the algorithm has learned.
Various approaches taken by competition authorities in Europe and the United States to address the use of pricing algorithms
European Commission approach
The European Commission (EC) has taken a proactive approach to addressing the use of pricing algorithms by firms. The EC has identified the potential anti-competitive benefits that can be gained from using pricing algorithms to manipulate markets and has sought to address this through a number of different measures.
The first measure taken by the EC was the adoption of the Guidelines on Vertical Restraints in 2013. These guidelines provided guidance to firms on how to comply with EU competition law when using pricing algorithms to set prices. The Guidelines state that firms should not use pricing algorithms to engage in anti-competitive practices such as price fixing or resale price maintenance, and that firms should ensure that their pricing algorithms do not lead to any form of market foreclosure.
In addition to the Guidelines on Vertical Restraints, the EC has also adopted a number of specific measures to address the misuse of pricing algorithms. In 2017, the EC launched an investigation into the use of pricing algorithms by e-commerce platforms to determine if they were engaging in anti-competitive practices. The EC found that some of the platforms were using pricing algorithms to limit the ability of their competitors to compete on price and imposed a fine on the platforms in order to address this behaviour.
The EC has also recently adopted a new regulation, known as the Digital Markets Act, which seeks to address the potential anti-competitive effects of the use of pricing algorithms. The Digital Markets Act will require firms to provide more transparency on their pricing algorithms, and to ensure that their algorithms do not lead to any form of market foreclosure.
The EC has also recently proposed a new competition tool, known as the Digital Markets Act ex ante regulation, which will allow the EC to impose binding commitments on firms that are found to be engaging in anti-competitive practices through the use of pricing algorithms. This will enable the EC to take more direct action against firms that are found to be using pricing algorithms to manipulate markets and reduce competition.
United States Department of Justice approach
The United States Department of Justice (DOJ) has also taken a proactive approach to addressing the use of pricing algorithms by firms. The DOJ has identified the potential anti-competitive benefits that can be gained from using pricing algorithms to manipulate markets and has sought to address this through a number of different measures.
The first measure taken by the DOJ was the adoption of the Antitrust Guidelines for International Enforcement and Cooperation in 2018. These guidelines provided guidance to firms on how to comply with US competition law when using pricing algorithms to set prices. The guidelines state that firms should not use pricing algorithms to engage in anti-competitive practices such as price fixing or resale price maintenance, and that firms should ensure that their pricing algorithms do not lead to any form of market foreclosure.
In addition to the Antitrust Guidelines, the DOJ has also adopted a number of specific measures to address the misuse of pricing algorithms. In 2019, the DOJ launched an investigation into the use of pricing algorithms by online travel agencies to determine if they were engaging in anti-competitive practices. The DOJ found that some of the agencies were using pricing algorithms to limit the ability of their competitors to compete on price and imposed a fine on the agencies in order to address this behaviour.2
The DOJ has also recently adopted a new rule, known as the final rule on price transparency, which seeks to address the potential anti-competitive effects of the use of pricing algorithms. The final rule will require firms to provide more transparency on their pricing algorithms, and to ensure that their algorithms do not lead to any form of market foreclosure.
In short, competition authorities in Europe and the United States have taken a proactive approach to addressing the use of pricing algorithms by firms. The EC and the DOJ have identified the potential anti-competitive benefits that can be gained from using pricing algorithms to manipulate markets and have sought to address this through a number of different measures. These measures include the adoption of guidelines, the launching of investigations into firms, and the introduction of rules and regulations that seek to provide more transparency on pricing algorithms and ensure that they do not lead to any form of market foreclosure. By taking these measures, competition authorities in Europe and the United States are seeking to ensure that firms do not use pricing algorithms to engage in anti-competitive practices, and that consumers benefit from increased competition and reduced prices.3
Countervailing competition law issues
Tacit coordination is a market outcome that is anti-competitive without the need for explicit communication between competitors. The following sections discuss why algorithmic pricing may increase the likelihood of tacit coordination.
Concerns about increasing data availability and the use of pricing algorithms extend beyond their potential to exacerbate collusion. A second set of concerns is that, in conjunction with the expansion of “Big Data,” they may result in personalised pricing. Personalised pricing is defined as pricing in which a business uses data/information about people’s desires or characteristics to set different prices to different buyers based on what the business believes the buyers are willing to pay.
Personalisation of pricing can often be advantageous; for instance, the capacity to give targeted discounts may make it easier for new entrants to compete, especially in markets with high switching costs, and may even increase output. Personalised pricing, on the other hand, may occasionally cause harm to consumers.4
Ezrachi and Stucke describe three major ways that algorithms can result in the formation of a tacit coordinated pricing outcome: hub-and-spoke, predictable agent, and autonomous machine. The first way that algorithms can lead to tacit collusion is when sellers use the same algorithm or data pool to determine price. If multiple competitors use the same pricing algorithm, they may react similarly to external events such as changes in input costs or demand.
Furthermore, if competitors are aware or can infer that they are using the same or similar pricing algorithms, firms will be able to better predict their competitors’ reactions to price changes, which may help firms better interpret the logic or intention behind competitors’ price setting behaviour. Widespread knowledge and use of common pricing algorithms may thus have a similar effect to information exchange in reducing strategic uncertainty, which may help sustain (but does not always result in) a tacitly coordinated outcome.5
Anti-competitive agreements are those that have the intention of, or have a real impact on, preventing, limiting, or mutilating competition in any market in India. Such arrangements cover not only agreements, but also decisions made by associations of persons/entities, as well as the direct of gatherings acting in agreement.
Algorithms in computers have transformed the way we trade and will continue to do so at an increasing rate. Although the development of fast-moving, digitalised markets has many advantages, algorithms also change the dynamics of competition and may limit it. Our discussion focused on four types of algorithmic-supported collusion. We identify instances where algorithms facilitate conscious parallelism and are unlikely to be challenged under current laws as the most difficult from both a legal and enforcement standpoint.6
The potential separation between the algorithm’s actions and those of its human designers and operators raises concerns about the ability to assign liability to the algorithm’s operators, who may escape scrutiny due to the unpredictable nature of self-learning. Concerns about the rule of law include how to distinguish between express agreement and accommodating behaviour, as well as increased subjectivity over whether and when computers “agreed”.
To what extent should humans be held accountable for low probability or difficult to predict events? There is a greater risk of such conduct being ruled not anti-competitive if there is no human intent and it is immoral.
In a digitalised universe where the ethical texture of the law is irrelevant, game speculations will be demonstrated continuously until an objective, predictable result is recognised. Given the simple concept of these markets, algorithms may change market dynamics and encourage tacit collusion, higher costs, and a more prominent wealth imbalance.
Firms may have a strong incentive to shift pricing decisions away from humans and towards algorithms in such a scenario. Humans will most likely wash away any ethical concerns by rejecting any relationship and obligations between them and the computer.
† BA LLB (Hons.) student at Maharashtra National Law University, Mumbai. Author can be reached at email@example.com.
1. Anne-Sophie Thoby, “Pricing Algorithms & Competition Law: How to think optimally the European competition law framework for pricing algorithms?”, Competition Forum, 2020, art. n°0009, <https://competition-forum.com/wp-content/uploads/2020/12/art.-n%C2%B00009.pdf>.
2. Hariharan Venkateshwaran, “Anticompetitive Nature of Pricing Algorithms”, International Journal of Legal Developments and Allied Issues, Vol. 7 Issue 1 (01/2021) <https://thelawbrigade.com/wp-content/uploads/2021/01/Hariharan-Venkateshwaran-IJLDAI.pdf>.
3. “Algorithmic Collusion: Can the Competition Act Protect Against Self-learning Algorithms?”, IndiaCorpLaw <https://indiacorplaw.in/2022/01/algorithmic-collusion-can-the-competition-act-protect-against-self-learning-algorithms.html>.
4. Zach Brown and Alexander Mackay, “Are Online Prices Higher Because of Pricing Algorithms?”, Brookings Edu <https://www.brookings.edu/research/are-online-prices-higher-because-of-pricing-algorithms/>.
5. Ivy Wigmore, “Competition Law”, TechTarget <https://www.techtarget.com/whatis/definition/competition-law>.
6. Algorithms and Collusion: Competition Policy in the Digital Age, OECD (2017), <https://www.oecd.org/daf/competition/Algorithms-and-colllusion-competition-policy-in-the-digital-age.pdf>.