FTC to Examine if Companies Raise Prices Using Consumer Surveillance

Agency is seeking information from eight companies about the use of AI and other tools to customize prices based on shoppers’ individual traits

By Dave Michaels & Inti Pacheco of The WSJ. 

One issue in this article is price discrimination. After some excerpts, I will discuss what price discrimination is, how it can increase profit and why it is not necessarily a bad thing.

"The Federal Trade Commission is seeking information about how artificial intelligence and other technological tools may allow companies to vary prices using data they collect about individual consumers’ finances and shopping habits.

The FTC said its study aims to reveal the inner workings of personalized pricing, a way of varying prices down to the individual level that has long been the holy grail of marketing. The commission says the algorithms and models that drive pricing strategies are opaque and may rely on surveilling consumers’ online footprints."

"Revionics said that its software recommends optimal pricing to retailers but doesn’t use any individual consumer data. Its software can recommend price decreases as well, the company said. A Mastercard spokesman said the company was reviewing the FTC’s request and would cooperate with the inquiry. Other companies that received the FTC’s request didn’t respond to messages seeking comment."

"Price discrimination among different consumer groups is a common strategy in marketing. Companies try to maximize profits by offering prices that are acceptable to different groups of consumers, defined by attributes such as age, socio-economic status and their affinity for certain goods and services.

But online advertising and data analytics have made it easier to target smaller groups of consumers. The FTC’s reference to “surveillance pricing” suggests it believes that consumers don’t understand how their personal information may be used to tailor prices to them. 

Some of the FTC’s concern stems from the growth of a similar concept, dynamic pricing, which allows companies such as grocery stores and fast-food restaurants to quickly vary prices during busy periods. Consumers have rebelled against dynamic-pricing in some contexts, such as when fast-food operator Wendy’s considered it, but it has long been used by airlines, hotels and ride-hailing services."

"Executives at retailers such as Lowe’s, Kroger and GoodRx have talked about using consumer data to deliver personalized pricing to shoppers. The companies said they mostly do so by offering discounts through loyalty programs. 

Lowe’s finance chief Brandon Sink said last month during a conference that the goal of the company’s DIY loyalty program is to increase repeat visits to its stores and to be able to market products on a more personalized basis." 

Businesses can make more money if they charge different prices to different buyers instead of just one price to everyone.

Charging different prices to different groups of customers based on their ability and willingness to pay (a discount) is price discrimination. Buyers with a lower price elasticity of demand will be charged a higher price.

If the firm were to charge the same price to each group, they would actually make less profit since they would end up violating the rule which says "choose Q so that marginal revenue (MR) = marginal cost (MC)."

Suppose a firm has two groups of customers, A and B, shown below. Group A's demand is generally less elastic. The blue line is demand. Green is MR. The flat line is both MC and ATC (average total cost). Having ATC = MC is not realistic but it simplifies the explanation.

The darker red lines just show us how to find P and Q for each group. Profit is Q*(P - ATC).




Now group B


Group A profit) 8*(24 - 8) =128

Group B profit) 12*(20 - 8) = 144

Total profit = 128 + 144 = 272

What if they charge both groups 22?

Group A profit) 9*(22 - 8) =126  (at Q = 9, MR does not equal MC)

Group B profit) 10*(22 - 8) = 140 (at Q = 10, MR does not equal MC)

Total profit = 126 + 140 = 266

So there is less profit (266) if they charge the same price to each group than if the price discriminate (272).

Economist Robert P. Murphy gives an example of how there is nothing wrong with price discrimination:

"the granting of special pricing for certain groups need not harm the groups paying full freight.

For example, if a movie theater in a small town were barred from giving child, student, and senior discounts — and instead had to charge one ticket price for all customers — it might not be able to stay in business. It would hardly help the middle-aged adults to have a "fair" pricing policy with no theater in town. This example shows the pitfalls in thinking about "the cost" of providing a seat in a movie theater and deriving the "fair" price that a theater ought to charge all customers."

See The Economics of Coupons and Other Price Cuts

Also see Price Discrimination by Tejvan Pettinger. A key passage is:

"Price discrimination will enable some firms to stay in business who otherwise would have made a loss. For example price discrimination is important for train companies who offer different prices for peak and off-peak. Without price discrimination, they may go out of business or be unable to provide off-peak services."

Tyler Cowen said

"The point of price discrimination is to sell more goods and services while taking in more profit; the low demanders can pay a lower price, yet the company still sells for a higher price to the high demanders.  If output goes up, social welfare usually does too."

See Progressive taxation as price discimination. 

Also

"The web has also, notes Cowen, facilitated price discrimination, which typically gives low prices to people with low time values. Since time values and income are highly positively correlated, price discrimination “is usually an egalitarian development.”"

See A Love Letter to Tyler Cowen by David R. Henderson.