Will big data lead to price discrimination?

Many reasonable concerns about the rise of Big Data—the collection of very large amounts of data generated by digital activity—seem to be coming to a head due to worries about privacy as more and more information is collected on individuals. One particular concern is that businesses may use collected data to charge different prices to different individuals depending on the information businesses have access to. But could potential price discrimination fueled by big data be more benign? A recent analysis by the President’s Council of Economic Advisers looks into that very question.

Despite the connotation of the word discrimination, price discrimination is not necessarily a bad thing. Consider movie tickets. Movie theatres often charge differential prices to different groups for showings. Customers over 65 get senior discounts, for example, because theatres are using age as a proxy for sensitivity to prices as retirees often have a tighter budget than working people. By giving the elderly a discount, the theatres get more business and increase their profits by getting customers you might not have attended at the regular price. And those customers gain because they go to a movie that they wouldn’t be able to attend at the higher price.

This type of price discrimination is what economists call third-degree price discrimination, the kind where prices are based on demographics groups. Other examples include when e-commerce companies or travel reservation websites adjust their prices according to users’ computers or phones. Second-degree discrimination is when the price of a product goes down the more of it a consumer purchases, which happens all the time in a market economy.

But big data may give rise to the kind of price discrimination known as first-degree discrimination. In these cases, businesses can tailor prices to ensure each customer pays the highest price she is willing to pay for the product. The Council of Economic Advisers report uses the example of cars as a consumer good where prices are individualized. Anyone who’s bought a car at a new or used car dealership knows the sticker price is just a starting point and one-on-one haggling is expected. The use of this kind of discrimination when businesses have access to data on customers, such as their individual earnings profile and credit history, is fairly clear.

But here’s perhaps more telling example: If book sellers know that a potential customer has been sharing articles on a certain topic on social media they might be able to charge a higher price for a related book. The book seller gets a higher profit off the sale, but the customer might be left a lower consumer surplus as they were charged a higher price than a market targeted at the whole population might bare.

After talking to researchers and others using big data The Council of Economic Advisers report says the use of first-degree price discrimination doesn’t seem to have caught on. Internet companies appear to have held off on personalizing prices for now. But the report does note that companies do seem to be using tests on websites to figure out the demand curves for different goods. While not exactly price discrimination, this process does provide companies with information on what prices the market can bare.

So big-data-fueled price discrimination of the first degree doesn’t look like a reality in the present. But we are just at the beginning of the use of these kinds of large data sets for business use. In the meantime, as the authors suggest, policy may be better focused on enforcing existing regulations concerning privacy, consumer protection, and discrimination. Given the speed of technology, policy makers should keep an eye on this area.

This article is published in collaboration with The Washington Centre for Equitable Growth. Publication does not imply endorsement of views by the World Economic Forum.

To keep up with Agenda subscribe to our weekly newsletter.

Author: Nick Bunker is a Policy Analyst with the Washington Center for Equitable Growth.