Algorithms and Big Data: An antitrust perspective

By Karan Chandhiok and Lagna Panda, Chandhiok & Mahajan

There has been a growing interest among antitrust authorities across the globe on how companies with market power collect and process data. Use of algorithms, machine learning and otherwise, and Big Data are being studied keenly by antitrust authorities, particularly on the issues of likelihood of collusion and “personalized pricing”.

Karan ChandhiokPartnerChandhiok & Mahajan
Karan Chandhiok
Chandhiok & Mahajan

The UK government recently initiated an open consultation under the Digital Competition Expert Panel to gather evidence on “the state of competition in the digital economy”. The French and German antitrust authorities have also launched a joint project to examine the impact of use of algorithms on competition.

Pricing effect: Much of the interest of competition law authorities in Big Data and algorithms stems from the fear and anticipation that big companies have the capability to employ sophisticated algorithms to process data and bring about personalized pricing. That’s not to say that companies do not already engage in differential pricing. Third-degree price discrimination, i.e. charging different prices to different consumer groups based on factors like demography, is a standard business practice in various industries in India — airlines offer discounts to senior citizens and laptop manufacturers offer discounts to students. The question, therefore, is whether differential pricing, in particular, personalized pricing, can and should be a competition law concern. In India, this question has particularly come up in the context of e-commerce marketplaces.

Personalised pricing in its truest form would be nothing but perfect price discrimination where companies are able to charge the maximum price based on each consumer’s willingness to pay (WTP), thereby extracting all available consumer surplus.

Lagna PandaSenior associateChandhiok & Mahajan
Lagna Panda
Senior associate
Chandhiok & Mahajan

To engage in perfect price discrimination, businesses need to be able to precisely estimate and predict the WTP of each consumer for a specific (brand of) product. However, even with extensive data and computing power, the likelihood of companies being able to engage in perfect price discrimination is very low.

A monopoly may at best be able to engage in an imperfect price discrimination since the nature and quantum of information required to implement perfect price discrimination is not easy to get. However, in case of an oligopolistic market, or a perfectly competitive market, competition is likely to keep the profit maximisation agenda of companies in check.

This is particularly relevant for markets like India where offline traders still cater to a large number of consumers. Against this backdrop, price prediction and estimation become difficult. Also, business decisions cannot ignore the possibilities of consumers getting alienated or sellers getting priced-out.

Non-price benefits: Even if one were to assume that companies could implement perfect price discrimination, access to information for consumers has become fairly easy because of technological progress. In a market with transparent pricing, a consumer is armed with necessary information and can take a reasoned decision on whether to purchase a product at a personalized price. In addition, online platforms have increased consumer choice, with a greater variety of products available on a single online marketplace when compared with an offline store.

Although competition authorities are wary about the use of Big Data and algorithms to “individualize” pricing, achieving it remains difficult, as the process of predicting each individual’s maximum paying capacity is complex even for a company with sufficient data and computing power.

Therefore, not every form of use of Big Data and algorithms by companies is likely to have competition concerns. On the contrary, some uses may result in price and non-price benefits. In fact, some companies have begun using Big Data and algorithms to take strategic business decisions. For instance, Netflix uses algorithms to personalise recommendations for its viewers, and also curates its shows based on consumer preference data. Such use of data promotes customer choice, thereby increasing consumer welfare.

The new-age business models spurred by Big Data and algorithms are still nascent and competition in these markets is beginning to play out. Any interference in such nascent markets could thwart innovation and technological progress. Even the Competition Commission of India, in a recent order, recognized the possible market efficiencies and consumer benefits that these new technologies and markets can provide, and noted that any intervention in such nascent markets must be carefully crafted.

Karan Chandhiok is a partner and Lagna Panda is a senior associate at Chandhiok & Mahajan.

C-524, Defence Colony
New Delhi – 110 024

Mumbai | Bengaluru

Contact details
Tel: +91 11 4163 0033
Fax: +91 11 2433 9075