Pricing Algorithms and Its Implications under the Competition Act 2010
28 June 2018
Studies have shown that there is a current evolutional shift in the global business landscape where emerging and thriving businesses are combining the availability of big data with advanced technological tools, such as artificial intelligence, algorithms and machine learning to improve their users’ experience, automate processes, predict market trends, and analyse the behavior and preference of their targeted audience for better marketing strategy.
This phenomenon is blooming across the world, especially the Southeast Asian countries, and Malaysia is poised as one of the leading nations to catch the trend.
This article aims to explore the untapped legal implication of pricing algorithms under the Competition Act 2010, in particular, its anti-competitive risks such as price fixing and collusion.
An algorithm, technically, is a sequence of rules, commands, or orders when one applies in precise that achieves the exact result one desires. They are the hidden pervasive mechanisms that have automated most complex and yet repetitive and time-consuming task in our daily life.
With the advancement of computer science, attempts and efforts are made to bring algorithms to a whole new level, from replicating human minds, to making instantaneous independent decision with speed and precision that a human cannot match, and more importantly, predicting the likelihood of future outcomes, such as stock prices movement, financial market, and future pricing for e-commerce businesses.
In the e-commerce market, pricing algorithms are algorithms sophistically designed to forecast price changes, based on the analysis of relevant price products data, in enabling online businesses to react spontaneously in matching the real-time or future price movement imposed by competitors, such as ticket fares and hotel rooms.
The perceived benefits of pricing algorithms are to increase price transparency, allow constant price adjustment, reduce time spent in searching for the best online deal, and to ultimately increase overall customers’ experience and decision-making.
However, there are two sides of a coin. The increasing demand for pricing algorithms in the e-commerce market will also create an overly transparent environment that creates anti-competitive effects. To name a few, pricing algorithms will reduce the incentive for competitors to reduce their prices, as pricing algorithms can instantaneously react to real-time price changes. For the similar reason, it removes the temptation for the members of cartel to deviate from the agreed price, as any deviation will be detected and retaliated immediately, and as a result, pricing algorithms will also become a powerful tool for cartels to administer collusion. Also, another major competition concern is when ‘companies start sharing the same dynamic pricing algorithm, which may be programmed not to compete against other firms, but to set anti-competitive prices. Such algorithms would allow companies not only to collude, but also to have their prices automatically reacting to market changes without the need to engage in further communications’, stated by the OECD.
Interestingly, in 2015, the United States Department of Justice has prosecuted David Topkins, with a fine of $20,000 being imposed for using the pricing algorithms to engage in price-fixing activities. ‘TOPKINS and his co-conspirators agreed to fix, increase, maintain, and stabilize prices of the agreed-upon posters. In order to implement this agreement, TOPKINS and his co-conspirators agreed to adopt specific pricing algorithms for the agreed-upon posters with the goal of coordinating changes to their respective prices. In furtherance of the conspiracy, TOPKINS wrote computer code that instructed Company A’s algorithm-based software to set prices of the agreed-upon posters in conformity with this agreement. For the purpose of reaching agreements on prices, enforcing adherence to the agreements reached, and monitoring the effectiveness of the pricing algorithms, TOPKINS and his co-conspirators collected, exchanged, monitored, and discussed information on the prices and sales of the agreed-upon posters.’
Topkins is a noteworthy case where the pricing algorithms were specifically programmed for price-fixing activities. This article, however, anticipates that the day will come where machine-learning algorithms will nonetheless be self-directed to engage in collusive activities without the need of human’s instruction or interference.
Even though it is still too premature to positively conclude the anti-competitive effects of pricing algorithms in the Malaysian e-commerce market, however, the legal implication of pricing algorithms has already raised the alarm in other jurisdictions, and the Malaysia Competition Commission must be prepared to face the uphill task of deciding whether it is a form of infringement under the Competition Act 2010 where the algorithms self-instructed to engage in algorithmic collusive activities in the absence of an actual anti-competitive agreement between enterprises. In such case, the definition and notion of an agreement ‘any form of contract, arrangement or understanding, whether or not legally enforceable, between enterprises, and includes a decision by an association and concerted practices’ may need to be revised in addressing the concerns of algorithmic collusion in this modern era of digital age.
1. The MIT Encyclopaedia of the Cognitive Sciences | 2. The United States Department of Justice | 3.Algorithms and Collusion: Competition Policy In The Digital Age