CHAMWE KAIRA
Pricing algorithms are increasingly creating risks for market competition by enabling companies to coordinate prices and charge consumers different rates based on their willingness to pay, according to research presented at the 2026 Advancing Competition in Africa Biennial Conference held in Windhoek.
University of Hong Kong Professor Thomas K. Cheng examined how automated pricing systems are challenging traditional competition laws, particularly where companies use algorithms to maintain higher prices without direct agreements between competitors.
The research found that while pricing algorithms can reduce transaction costs and allow businesses to respond quickly to changes in demand, their adoption has also been associated with higher prices and reduced consumer benefits in several industries.
A 2024 study by Assad and other researchers found that the adoption of pricing algorithms in Germany’s petrol retail market resulted in prices that were 9% higher in non-monopoly markets and 28% higher in duopoly markets.
In the healthcare sector, a 2023 study by Brown and MacKay found that the adoption of algorithms by five major over-the-counter medicine retailers increased corporate profits by 9.6%, while reducing consumer surplus by 4.1%.
Research by Calder-Wang and Kim in 2024 also found that increased use of pricing software in multi-family housing was associated with higher rents and lower occupancy rates, suggesting imperfect coordination among landlords.
Cheng outlined three categories of algorithmic collusion, highlighting the challenges competition authorities face in identifying and addressing anti-competitive conduct involving automated systems.
The first category, facilitative collusion, involves businesses deliberately using software to implement and monitor an existing price-fixing arrangement.
An example is the US case involving poster sellers on Amazon Marketplace, known as United States v. Topkins.
The second category, coordinative collusion, occurs when competing businesses use a shared third-party algorithm that collects confidential pricing information and centralises price-setting decisions.
This issue is being contested in US rental market litigation.
The third category, autonomous collusion, involves self-learning algorithms independently identifying that maintaining higher prices through tacit coordination is more profitable than competing.
This can happen without direct communication between businesses or explicit instructions from their management.
The research also examined personalised pricing, where companies use consumer data to estimate how much individual customers are willing to pay and adjust prices accordingly.
Cheng argued that dominant companies could use this capability to engage in targeted predatory pricing or impose product ties selectively on consumers who are less sensitive to price changes.
Such practices could make it easier for companies to exploit market power, potentially lowering the traditional thresholds used to identify anti-competitive abuse.
The presentation further considered whether consumers could use their own algorithms to negotiate better prices or organise group purchases.
However, the findings indicated that buyer-side negotiation tools have limited ability to conceal consumers’ willingness to pay because sellers can analyse counteroffers, purchasing patterns, timing and acceptance decisions.
Group-purchasing algorithms also offer limited protection when they bring together consumers with similar valuations, as the combined group can still reveal its overall price sensitivity.
The research suggested that consumers could conceal their purchasing preferences more effectively by deliberately changing their online shopping behaviour, including abandoning shopping carts or refusing offers they would otherwise accept.
These actions could make consumers appear more price-sensitive and make it harder for seller algorithms to accurately estimate their willingness to pay.
