Anti-competitive practices online

Experience of China

In December 2024, China's State Administration of Market Regulation (SAMR) opened an investigation against U.S. GPU maker Nvidia for violating the acquisition terms¹ of Mellanox Technologies for more than $6.9² bn, which SAMR authorized in 2020. Information on the exact reasons for the charge has not been released, but back when this deal was approved in 2020. SAMR identified that the merger could impact the global and Chinese market for GPUs, private network interconnect equipment and high-speed Ethernet adapters. SAMR has therefore approved the transaction subject to the following conditions:

1) Unreasonable trade restrictions should not be imposed on the sale of GPUs and highspeed networking equipment on the Chinese market, and the market should be supplied on the basis of fairness and non-discrimination.

2) Compatibility of NVIDIA GPUs, as well as Milos high-speed networking devices must be ensured with products from competing manufacturers, including Chinese manufacturers.

3) The open originating software code for peer-to-peer and aggregated communications for the Milos High Speed Network Interconnect Appliance must be maintained so that other vendors can support interoperability of the appliances.

Experience of South Korea

South Korea's Fair Trade Commission³ (KFTC) has fined Kakao Mobility and its subsidiary the taxi-hailing services KakaoT for $10.2 mn.

On the cab market, Kakao offers general taxi-hailing services (as an aggregator, similar to Yandes.Taxi) and specialized taxi-hailing services under the franchise of its subsidiary Kakao T Blue (for example, Uber and TADA operated under the franchise). It was revealed that from 2019, the company, in order to exclude competing taxi operators working under the Kakao franchise, began blocking taxi bookings for drivers who work for rival companies. The company reasoned that taxi calls were being duplicated on multiple platforms, on Kakao and those of competing operators, creating confusion for passengers. Kakao therefore required the 4 competing franchised taxi operators to pay their drivers a commission for using Kakao, or the company needs to sign a partnership agreement with Kakao. This allowed Kakao Mobility to collect information including trade secrets of its competing franchisees, data on users, cab drivers, and trips.

The KFTC considered that on the one hand, Kakao Mobility entered into such contracts and collected confidential information for its own business strategy, for example, Kakao could manipulate the platform's algorithms to prioritize franchised taxis over non-franchised taxis. This allowed Kakao Mobility to hold a dominant position.

On the other hand, a competing franchised taxi operator is forced to enter into a partnership agreement or else it will not be able to receive taxi call information from Kakao Mobility, which has a market share of over 50%. KFTC estimates that Kakao Mobility's practices have led to an increase in its market share from 51% in 2020 to 79% in 2022.

Experience of Italy

In December 2024, the Italian National⁴ Authority for Market and Competition published the results of an investigation into the use of dynamic pricing algorithms in passenger air transport on domestic routes to and from Sicily and Sardinia.

Airlines use dynamic pricing algorithms when prices for the same service (or product) change based on demand and other information. The frequency of price adjustments can vary for each individual transaction, including depending on the data available to the airline. Dynamic⁵ pricing can lead to price discrimination, where a monopolistic company offers different prices for the same goods or services to different consumers or groups of consumers in order to make more profit. The companies processed information on the number of bookings, load factor, flight occupancy rate, previous booking trends on the same flight, route, departure dates, number of days before departure, etc.

Information on market structure and trends was also analyzed, such as the market shares held by competitors and their prices. To obtain such information, all companies used the same Infare platform that releases publicly available information on competitors' flight prices. However, such data was rarely used to determine prices for each flight but was necessary for the companies' analysts for comparative analysis and assessment of their own position on the market.

The Competition and Market Authority concluded that the information collected, including from competitors, is not used for price discrimination. Changes in airfares are mainly influenced by trends in actual and expected demand estimated based on historical booking data for the same flight or using forecasting models, for example, if actual demand is higher than expected demand or the occupancy rate of a flight is also higher than the expected rate, the system determines a shift to higher fare classes.

The Authority for Competition and Market concluded that the collected information, including from competitors, is not used for price discrimination. Changes in airfares are mainly influenced by trends in actual and expected demand estimated on the basis of recent booking data for the same flight or using forecasting models, e.g., if actual demand is higher than expected demand, or the occupancy rate of a flight is also higher than the expected rate, the system determines a shift to higher fare classes.

It is also revealed that airlines use in the mechanism of price dynamization a classical system of fare formation due to a predetermined fare grid, within which the price of each flight varies over time, opening and closing orders for each predetermined price class randomly, although based on the data collected by the companies, the distribution of the number of seats in each fare class can occur.

Nevertheless, no evidence of discrimination was found regardless of the type of device used to purchase tickets or the operating system used to search for tickets. Although one company was found to change its pricing policy when the same user logs in repeatedly, or when logging in to an airline's website via a flight aggregator (like Google Flights). As browsing history or information about other user activities is accumulated, the user may be offered a personalized price (rather than a randomly generated price within the fare grid), but more often than not it was a price “frozen” for the user, which does not change up or down, for example, each time the user browses the prices again, which also does not cause price discrimination.

Russia’s experience

In November 2024, FAS and the Ministry⁶ of Transport released a joint statement attempting to regulate airfares. Airlines were recommended to describe approaches to the pricing of passenger transportation in economy class, including the methodology for determining minimum, maximum and intermediate fares with justification of the method of their pricing. It is also proposed to increase the range of fare groups with a detailed description of each of them, to work out approaches to the implementation of sales and price reduction measures. Such measures should increase the transparency of passenger air transportation pricing.

From the monitoring issue No. 12, December 2024. Download the full issue (PDF) · issue page at the Gaidar Institute