Sunday 7 February 2021

What If Price Robots Could Read Our Mind?

 


Let me start by planting a scenario in your head. Jack and Jill go to the supermarket an hour apart, and they see some pineapples for sale for £1 each. Jack really loves pineapples and would willingly pay £2 for each one. Jill likes pineapples but would only pay £1.30 for each one. Both Jack and Jill each buy three pineapples, and the supermarket takes in £6 of sales. If there was a magic price-setting robot that could adjust the price of pineapples to exactly the most that each customer would willingly pay, then it would charge Jack £6 for three pineapples, and Jill £3.90, meaning supermarket sales of £9.90 rather than £6. In the real world, the supermarket doesn't know in advance what Jack and Jill will each pay, so they both get charged £1 per pineapple, even though they'd both willingly pay more.

Now picture a new scene: this time it’s the year 2045, and the price system is structured rather differently. Whereas once upon a time we would all expect to pay the same price for a pineapple on a supermarket shelf, or a music album on Amazon’s website, these days, in the year 2045, the basis for what individuals are charged for all goods and services is based on their own past consumer habits. In other words, just as the products Amazon shows us in the suggestion bar are based on past consumption, prices too may be dynamically adjusted based on how much we are thought to value something.

The current price system is built on an imperfect approximation of a weighted average of revealed preferences in society. But it does mean that we all get different value for different things. This is revealed in the prices we are willing to pay, and the levels of consumer surpluses enjoyed. That’s why a healthy economy caters for Tom’s love of Star Wars, Dick’s love of model railways and Harry’s love of snooker. Each may hate the other person’s favourite thing, but prices account for all the information signals, as individuals look to maximise their own utility, and are a best approximation of how much to charge for something.

The downside of the price system for suppliers is that there is an awful lot of consumer surplus out there (the difference between the price you’d pay and the actual price), which means prices do not accurately reflect individual demand curves. This may possibly change in the age of the Internet and its concomitant ‘big data’ programs, algorithmically designed to track our every move, purchase, taste and interest. An algorithm that can have a much better idea about our demand curves can tailor bespoke prices for individuals. This may mean that a Star Wars retailer would offer their merchandise to sci-fi geek Andy for a higher price than it would Harry the Star Wars-hating snooker fan. In fact, this is happening already is some areas of retail. There are data mining devices that can evaluate your desire for a holiday (if you’ve visited the page multiple times in a day for example) and adjust its prices accordingly.

While this will probably never be an exact science, as individual utility gathering will be difficult to determine with precision, I can well imagine a time in the future when the majority of prices are much more closely aligned to what people are willing to pay. And while this will benefit suppliers, it will no doubt eat into our consumer surpluses, and therefore reduce the amount of value created in society - that is, unless a very adaptive human psychology can keep us one step ahead of the game.

/>