Updated: Aug 1
Coca Cola’s utility increases on a hot day compared to other days which was noticed as a profit maximization opportunity and in 1999, the company considered the implementation of temperature-sensitive vending machines that would alter the price of the product as per the intensity of heat in its surroundings. This sums up the idea of dynamic pricing, which various corporations have been employing while retaining the demand-supply equilibrium in the nucleus. Algorithms, on the other hand, are just one of the extremely effective means to harness the concept of dynamic pricing which translates to Algorithmic Pricing. It is applied for constantly altering the offer price depending upon the consumers’ personalized data sets, derived mostly by past activities and recent trends. In the most generalist sense, Algorithms can be traced back to the time before the invention of computers. Simply put, it denotes a set of rules that ought to be repeated, in a specified sequence to accomplish the given task. With the overall technological revolution, Algorithms have become more and more sophisticated due to the application and evolution of Big Data, Machine Learning and Artificial Intelligence (AI), which have made it particularly efficient in creating useful data repositories, making predictions, decision making and achieving goals within time. Especially e-commerce websites like Amazon and Flipkart have been able to achieve exponential growth due to this inexpensive piece of technological innovation.
The present-day Algorithms are capable of performing complex tasks like predictive analysis which helps a corporation to appraise consumer behaviour, forecast risks, highlight the new competition and estimate demand, all at once and synchronized to capitalize on profits. Governments have also been using it to gauge criminal behavioural patterns, intensities and possible occurrences of the crime, by applying it to different variables like period and location. Due to the expansive nature of Algorithm Pricing, it has been subject to scrutiny in recent years. Domestic regulators, as well as International Organizations, have been conducting research into possibilities of intrusions into privacy, business ethics and competition regulations caused by such Algorithms.
This analysis will critically examine the regulatory concerns associated with Algorithm Pricing in isolation while acknowledging that growth of the e-commerce industry on whole has indeed led to innovation, connectivity and overall progress.
It is a policy issue: International Trade Law
The body of International Trade Law consists of treatise, principles and customs which regulate the transactions of two or more private sector parties belonging to different states. The World Trade Organization (WTO) and the Organization of Economic Cooperation and Development (OECD) have been expressing concerns of violation of privacy rights and competition policies. How the WTO and OECD give teeth to these concerns is yet to be seen since at International level and in private contracts, these organizations can only emphasize States to take action depending on their accessions. For instance, the prices of essential goods are regulated as per the General Agreement on Tariffs and Trade (GATT) which is consensual. Although the silver lining is that these organizations play a vital role in establishing standards for the conduct of business which will be dealt with in detail in further paragraphs. International organizations and technological innovations have always gone hand in hand, as for instance, the United Nations Commission on International Trade Law (UNCITRAL) played an important role during the internet regulation phase wherein its efforts helped recognize electronic contracts and records, in policies of individual states. In 1998, WTO along with the support of other States, granted a moratorium on duties for products of electronic transmissions, thereby facilitating easy trade in digital products.
The Algorithm Pricing mechanisms adopted by e-commerce companies gather huge amounts of consumer and competitors’ data which is then sorted out to understand seasonal preferences and then flash appropriate offer prices. The international community is specifically concerned since a major chunk of the population is unaware that such processes are being employed to gather their data and use it contrary to their interests. The European Union (EU) Courts’ technological approach asserts the highest importance to interpret provisions in a way, so as to further the goal of consumer welfare. If such processes cannot be accommodated into a specialised law, then it can be accommodated into general welfare provisions through the channel of interpretation. The functioning of this Algorithm Pricing (AP) goes against consumer welfare as the consumer is subjected to imperfect knowledge about the prices being charged to him. A strong welfare setting would require that a consumer knows, not only what he is being charged but also what other consumers, with identical/similar status, are being charged. One would argue that similar friction prevails in an offline setting as well, but the underlying point here is that consumers lack preliminary knowledge of dynamic pricing on e-commerce platforms. Consumers must be expressly made aware of the variables which determine a personalised price setting for different individuals or just the fact that dynamic pricing exists. It is difficult to prove the illegality of the use of the algorithms itself or the manner in which they are being processed, but it is necessary to investigate the nature of actions like targeted advertisements, collection of consumer data and preferences, disclosures made, responsiveness to price offers and past purchases trajectories. Ultimately, it is for the respective States to decide, as a matter of policy, the extent to which these intrusions should be allowed. It becomes an increasingly important issue since foreign companies’ activities related to data mining in the domestic sphere can lead to remittance of vital information out of the country which, in exceptional circumstances, may lead to strategic concerns for a state.
Simply by selling the same products at different prices to different people is not illegal. Issues arise only when two or more market players expressly agree to raise, reduce or stabilize their prices, in order to defeat natural competition. One would say that in the absence of an express agreement, Algorithmic Pricing could at the most, lead to tacit collusion with the competitor to alter the price, which is outside the purview of Competition law. Regulators here have raised concerns that the sheer vastness of capabilities of Algorithmic Pricing can lead to an understanding, whereby it can teach itself to manipulate prices anti-competitively. Hence, even without an agreement or human intervention, Big Data analytics might supplement reactive pricing in a way which will eventually teach itself to arrive at a cooperative equilibrium with other competitors' algorithms. In effect, this cooperative behaviour leads to conscious and express coordination which at all times, prevents unsatisfied demand and excess supply. Such processes then deprive the consumer of the lower price which it could have gained from the result of the competition despite the e-commerce platform being able to afford a reduction in price below the threshold. The OECD has recognized these activities as potentially illegal and asked to keep a check on developments in AI to determine if it can take business decisions which will finally bring Algorithmic pricing under regulatory scrutiny.
Another concern is that due to these Algorithms processing data every second, they act as a watchdog to quickly identify a new entrant or competition in the market. Such identification is then used to offer competitive prices which the entrant can never afford to match. There is no denying the fact that consumers derive huge benefits from competitive pricing but the big picture shows that those benefits are not derived equally by all consumers, which goes to the root of collective justice.
Establishment of standards
International Organizations are in a better position to establish standards to deal with exploitative aspects of Algorithm Pricing. The following reasons suggest why such Algorithms are unfair which may not necessarily be illegal but surely essential for setting standards. The rationale for establishing standards lies in the fact that Algorithms being used for pricing are neither transparent nor ubiquitous which makes it unfair and highly violative of the well-established social conventions one would follow in an offline retail setting. It is a matter of moral legitimacy that a transaction must not only involve consensus but also informed consent, which is absent in an online transaction where a consumer would not have indulged with the seller if it was aware that the e-commerce platform waived off its willingness to lower the prices by colluding.
Data-driven Algorithms assess consumer information at a granular level which in contrast is not available to the consumer leading to plausible exploitation. When one consumer is offered a price for a product, it should act as a threshold for other consumers to determine the value of the product. Consumers could have come to another valuation (lower) for the same product which was denied to them without knowledge of the same. Due to the opaqueness of the Algorithm Pricing strategies, behavioural economics suggests that consumers are being treated unfairly. Privacy concerns related to these activities are huge, as everything from past behavioural activities to location tracking to capacity and willingness to pay is tapped, stored and utilized at the expense of the consumer itself making Algorithm Pricing a fit case for regulation by well-defined standards.
Transparency serves as a separate calling for regulation of Algorithm Pricing as it is this aspect from which all other issues relating to Algorithmic Pricing arise. Generally, transparency for such a technological process can be achieved with an adequate evaluation of the software system, monitoring its activity and subjecting it to self-regulation as a reporting entity. Although the ground reality is not so simple, considering that transparency was to be imposed then even submission of complex and long algorithmic program codes would be extremely challenging to decipher in order to unwrap some real substance of unfairness. Proving illegality will altogether be more difficult as, in a sandbox environment, achieving the result of the desired price to prove guilt will depend on all requisite variables falling perfectly in the right place and time in a sequential and synchronized manner. There is little realization of the fact that providing source codes will not lead to transparency unless the rationale behind particular codes are made known to the authorities which is an impossible task as companies have a stronger right to secure their trade secrets. Regulators also need to be mindful about the fact that any policy brought does not become too conducive for the existence of the industry itself.
Some of the possible regulations which can be adopted are restricting volatility of price fluctuation by introducing cap limits, tapping into competitor’s data could be declared to be illegitimate and regulation of the algorithm designs from the inception. Although these restrictive policies can be adopted, the regulators, domestically and internationally, will have to take a cautious approach to avoid excessive regulations and undertaking supervisory burden for an industry which has generated great wealth for the society. An appropriate regulatory structure should involve definite laws which could be information technology law, competition law or intellectual property rights. If several agencies need to be involved then a structure of coordination to avoid overlapping needs to be in place. International Trade Law should fortify its substantive laws to cover situations dealing with foreign companies in other jurisdictions. Ultimately, the regulators do not have concrete proofs of privacy or competition law breaches but have serious concerns about the possibility of illegalities being committed by Algorithmic Pricing. Regulators are ready with cavalry to engage in meaningful regulation of Algorithmic Pricing but which innovation will be the last straw to break the camel’s back is to be seen.
 Seele P, Dierksmeier, Mapping the Ethicality of Algorithmic Pricing: A Review of Dynamic and Personalized Pricing, J Bus Ethics (2019), https://doi.org/10.1007/s10551-019-04371-w  Thomas Gehrig, Oz Shy and Rune Stenbacka, 'A Welfare Evaluation of History-Based Price Discrimination' (2012) 12 Journal of Industry, Competition & Trade  CPI Talks, Interview with Antonio Gomes of the OECD, May 2017, https://www.competitionpolicyinternational.com/cpi-talks-interview-with-antonio-gomes-of-the-oecd/  WTO, Substance of Accession Negotiations, Handbook on Accession to WTO, https://www.wto.org/english/thewto_e/acc_e/cbt_course_e/c5s2p3_e.htm  World Trade Organization, How Do We Prepare for Technology Induced Reshaping of Trade, World Trade Report 2018, https://www.wto.org/english/res_e/publications_e/wtr18_4_e.pdf  Christopher Townley, Eric Morrison & Karen Yeung, Big Data and Personalised Price Discrimination in EU Competition Law, King’s College London, Legal Studies Research Paper Series, Paper No, 2017-38, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3048688  Algorithm Pricing and its Effect on Competition Law, Society of International Trade and Competition Law, https://nujssitc.wordpress.com/2018/02/23/algorithm-pricing-and-its-effect-on-competition-law/  Refer Supra Note 5  OECD, Algorithms and Collusion – Background Note by the Secretariat, Directorate for Financial Enterprise Affairs, DAF/COMP(2017), https://one.oecd.org/document/DAF/COMP(2017)4/en/pdf  Refer Supra Note 6  Refer Supra Note 5  Deven R Desai and Joshua Kroll, Trust but Verify: A Guide to Algorithms and the Law, Harvard Journal of Law and Technology Vol 31, https://jolt.law.harvard.edu/assets/articlePDFs/v31/31HarvJLTech1.pdf  Refer Supra Note 9