Why does my cab cost more than yours?

The widespread effects of technology-based personalised pricing.

When I was forced to spend my days commuting in the Mumbai traffic in cabs , I became intrigued by the way cabs were priced. One day's cab would cost me ₹300 while for a friend it would cost  ₹550. The price which should be standard across people was behaving erratically but quickly, I started noticing these erratic pricing trends everywhere from airlines , to discount coupons which made me think about pricing differently. 

Price discrimination refers to charging different prices of a product or service for different customers, different , different quantities and different time.. The major objective for this practice is to take the opportunities to increase revenue and ultimately maximize profits. 

Arthur Pigou’s 1920 theorisation of the three degrees of price discrimination

While second- and third-degree price discrimination is widespread in today’s world, for most of history, first-degree price discrimination felt like an extremely distant hypothetical because it was practically impossible for companies to observe and accurately estimate every single consumer’s willingness to pay for their products. But with AI and data’s  grand entry into the commercial market, this Black Mirror-like phenomenon is very much coming true. 

AI needs a lot of data to operate effectively. The shift from cash to electronic payments has enabled businesses to collect what's been called a "gold mine" of information. For example, Mastercard says its data lets companies "determine optimal pricing strategies." So much information is collected when you pay electronically that in 2024 the Federal Trade Commission issued civil subpoenas to Mastercard, JPMorgan Chase and other financial companies demanding to know "how artificial intelligence and other technological tools may allow companies to vary prices using data they collect about individual consumers' finances and shopping habits." Experiments at the FTC show that AI programs can even collude among themselves to raise prices without human intervention. 

Degree	What the firm does	What it needs to know
First-degree (perfect) price discrimination	Charges each individual the maximum they are willing to pay	Each consumer’s individual willingness to pay
Second-degree price discrimination	Price varies according to quantity, version or package chosen rather than the identity of the buyer	How consumers sort themselves into different options
Third-degree price discrimination	Consumers are divided into observable groups, with different prices for each group	Which group a consumer belongs to and how price-sensitive that group tends to be

Personalized pricing could increase efficiency by reducing deadweight loss that arises due to the constrains on some mutually beneficial trades under uniform pricing.

Firstly, when firms can only charge a single price to all consumers, some buyers whose willingness to pay is higher than the marginal cost but lower than posted price are excluded from the market, these unrealized transactions trigger deadweight loss because they could have improve social welfare. Suppose a service costs ₹100 to provide, but the firm charges everyone ₹300. A customer willing to pay ₹200 will not buy it, even though their willingness to pay exceeds the cost of providing it. The missed transaction creates ₹100 of deadweight loss: ₹200 − ₹100.

If the firm offered that customer a personalized price of ₹150, the customer would gain ₹50 in consumer surplus and the firm would earn ₹50 above its marginal cost. The transaction would therefore create ₹100 of total surplus that uniform pricing left unrealised.

 In this case, personalized pricing could be more socially desirable because it allocates goods more concisely and allows a larger set of efficient transactions to take place . However, once personalized pricing is widely used, consumers would become market participants instead of passive recipient to market signals, they may intensify search efforts, delay purchases, switch to other sellers, or make purchases through intermediaries to avoid higher prices . Then, the conditions that originally made personalized pricing efficient may no longer hold. 

Despite the efficiency argument, personalised pricing has costs that are easy to overlook. The first question that arises is who exactly benefits from personalised pricing and it’s effects on firm behaviour. Even when more trades take place, the extra value does not need to be shared fairly. If a firm can estimate the most each customer is willing to pay, it can charge close to that amount, and buyers lose the savings they would have enjoyed under a single price. Perfect price discrimination can eliminate deadweight loss by allowing every mutually beneficial transaction to take place, but the firm captures the consumer surplus that would otherwise have gone to buyers. The market may become more efficient without consumers necessarily becoming better off.

A second problem is who ends up paying the most. The customers most likely to be charged high prices are often those with the least time or information: people in a hurry, people who do not compare prices, and loyal customers who never switch. Shoppers who are careful and well informed can protect themselves, while those who are not may pay more for the same product. 

      Third, personalized pricing depends on collecting large amounts of personal data. To set a price for one person, a firm must study their location, browsing habits, purchase history and behaviour. Consumers may not know what is being collected or how it affects what they pay, and because the price is private, they cannot easily compare it with anyone else's. 

The practice can also weaken competition. Firms with the most customer data can price most accurately, which helps them win more customers and gather still more data. The more data . Smaller rivals, lacking that information, cannot target individual customers in the same way, so the competition that would normally push prices down may never develop.

Finally, as noted earlier, consumers do not stay passive. Buyers may try to get around their assigned prices by using other accounts, devices or intermediaries, which wastes time and effort and can distort demand. If they come to believe the system is working against them, they may lose trust in sellers and buy less, which shrinks the market that personalized pricing was meant to serve. 

What should India’s response be ? 

While I am an optimist for how technology can revolutionize the way firms price their products , I do vehemently believe in effective guardrails to define the limits of algorithmic pricing. 

The first step , in my opinion, should be greater price transparency. Platforms should disclose the main factors influencing a fare and make it clear when personal data, rather than just demand and availability, affects the price. Consumers do not need access to an algorithm's source code, but they should be able to understand why they are being charged differently.

Secondly , India should place reasonable limits on how personal data is used for pricing. Consumers should know what information is collected and how it influences the prices they see. Particularly sensitive information, or data that reveals financial vulnerability, should not become a tool for extracting higher payments.  I do also believe that India's consumer protection framework should address demonstrably deceptive or discriminatory pricing practices.

All in all , I think technology based personalized pricing can become a reality , after all we all like it when products are priced exactly according to our wishes . However , this may also turn into a nightmarish reality if our data truly is exploited  to its core , necessitating a strong regulatory framework which is enforced strictly.  

References

• Bergemann, Dirk, Benjamin Brooks, and Stephen Morris. “The Limits of Price Discrimination.” American Economic Review 105, no. 3 (2015): 921–57.

• Calvano, Emilio, Giacomo Calzolari, Vincenzo Denicolò, and Sergio Pastorello. “Artificial Intelligence, Algorithmic Pricing, and Collusion.” American Economic Review 110, no. 10 (2020): 3267–97. https://doi.org/10.1257/aer.20190623.

• Central Consumer Protection Authority. Guidelines for Prevention and Regulation of Dark Patterns, 2023. Notified November 30, 2023. Summary at https://www.jsalaw.com/newsletters-and-updates/ccpa-issues-guidelines-for-prevention-and-regulation-of-dark-patterns-2023/.

• Digital Personal Data Protection Act, 2023 (Act 22 of 2023), India.

• Federal Trade Commission. “FTC Issues Orders to Eight Companies Seeking Information on Surveillance Pricing.” Press release, July 23, 2024. https://www.ftc.gov/news-events/news/press-releases/2024/07/ftc-issues-orders-eight-companies-seeking-information-surveillance-pricing.

• MIT OpenCourseWare. “Price Discrimination.” Lecture slides, 14.271 Industrial Organization I, Fall 2022. https://live.ocw.mit.edu/courses/14-271-industrial-organization-i-fall-2022/mit14_271_f22_lec3slides.pdf.

• Pigou, A. C. The Economics of Welfare. London: Macmillan, 1920.

• Mankiw, N. Gregory. Principles of Economics. Chapter 15, “Monopoly.” PowerPoint slides by Ron Cronovich. South-Western Cengage Learning, 2009. https://slidetodoc.com/chapter-15-monopoly-economics-n-gregory-mankiw-principles/.