What was novel about the case though, was that the conspirators also agreed to adopt specific pricing algorithms for their products with the goal of coordinating changes to the online prices. Topkins himself programmed his company’s algorithm-based software to set its prices. The software operated by collecting competitor pricing information for a specific product sold on Amazon Marketplace and applying certain pricing rules set by the seller. Topkins was charged with and pleaded guilty to violating Section 1 of the Sherman Act as his acts amounted to an unreasonable restraint of interstate commerce, and paid a US$20,000 fine. While the use of such software in e-commerce was rolled into the mix of conspiratorial conduct here, the case was ultimately decided on clear anti-competitive violations since there was clear evidence of the conspiracy and multiple price discussions between competitors.

Fast forward to 2017, the European Commission (EC) has in February launched an investigation into the e-commerce online retail practices of consumer electronics manufacturers Asus, Denon & Marantz, Philips and Pioneer. The investigation centres around manufacturers’ restrictions on online retailers’ ability to set their own prices and will likely also consider the widespread practice of both manufacturers and retailers using algorithm software which monitors the price of retailers across the internet and automatically adapts to leading competitors.

Other automated practices facilitated by technology like geo-blocking based on the consumer’s location or country of residence, and price discrimination on the basis of location have also merited recent EC investigation.

Algorithms & Automation

The cases above bolster what we already know – that whether we like it or not, we are living in a dystopian age of science fiction where machine-learning algorithms and automated decision-making rule our lives.

Without needing human involvement in cartel conduct, machines can now be programmed to automatically monitor competitors’ behaviour online, trawl for price discrepancies and automatically adjust and reset prices in real time, the net effect of which is coordinated outcomes and prices amongst competitors.

Interesting competition issues arise – like the definition of collusion (which traditionally would require ‘concurrence of wills’) between machines, or questions of robotic agency, and whether anyone should be punished for such conduct. If automated agents collecting data cannot “meet” or necessarily “exchange information” in a way human cartels could, can they be guilty of anti-competitive conduct? On the other hand, would the gathering of market intelligence in the way only these machines can ultimately reduce cost across market segments by improving efficiencies and promoting transparency?

Perspectives from the Regulators

Antitrust regulators are already tackling these issues. The Chairman of the Competition Commission of Singapore (CCS) Mr Aubeck Kam suggested in his keynote speech at the Antitrust in Asia Conference last year that while the Asian market was very diverse with different stages of development, regional competition regulators should develop a common approach on the growing use of machine algorithms in e-marketplaces. The CCS also initiated the ASEAN Competition Policy and Law Programme last year to bolster knowledge and expertise in this area, and has since embarked on the development of a regional handbook on e-commerce to guide ASEAN Member States on identifying and addressing potential challenges on competition law enforcement relating to e-commerce. The handbook is slated to launch in August this year, along with a forum for senior competition officials from ASEAN Member States to discuss competition-related issues on e-commerce.

While the efforts to develop a harmonised approach across jurisdictions are certainly important and necessary, the Chair of the Competition Markets Authority in the United Kingdom has recently also cautioned that antitrust enforcers may not currently have adequate tools to investigate whether and how the growing number of online algorithms affect competition.

It is not hard to see how this could be the case. Machine algorithms, in particular those based on deep learning techniques, process data in a manner so opaque that it might be practically impossible to explain how they reach decisions. Indeed there have been recent calls for tech conglomerates to be more transparent on their use of algorithms in collecting, selecting and evaluating data, although this information is jealously guarded by the titans of tech. Will the day soon come when tech corporations will be ordered to disclose information on how their algorithms operate, in order to prove that they were not designed to allow or facilitate collusion and anti-competitive behaviour?

A Brave New World

In one generation, we have moved from RoboCop at the movies, to robo-drivers, robo-advisers, robo-chefs and robo-baristas in real life. Competition and regulatory agencies across the world have the difficult job of seeking ways to ensure a competitive and fair framework with the robots in charge.

As Singapore grows as an e-commerce hub and with Amazon and Alibaba’s imminent battle for a piece of the growing pie here, regulators and participants in our online marketplaces alike will continue to have to keep ahead and grapple with evolving digital practices which may restrict or distort fair competition with the use of technology in our rapidly changing online platforms.