Don’t judge a book by its cover price

A book tagged at $23m is a long way from the dream that the internet would usher in an era of price transparency

Last spring, a young biologist tried to buy a copy of a common but out-of-print reference work, Peter Lawrence’s The Making of a Fly. Amazon offered 15 used copies at reasonable prices – and two new copies, the cheapest of which was $1,730,045.91, plus $3.99 shipping.

Michael Eisen, an evolutionary biologist at UC Berkeley, heard the tale and tried to figure out what was going on. It can’t have been a prank – there were two sellers involved, both with thousands of satisfied customers. Eisen heard that many prices on Amazon were set by computers, and suspected this might be the reason.

The next day, both prices had risen to around $2.8m. By the end of the day, the higher-priced copy was on offer for $3,536,675.57. Eisen began to figure out what was happening. One seller, profnath, would set its price to fractionally undercut the best price available, once per day. The other seller, bordeebook, would discover this after a few hours and reset its price to be 1.270589 times higher than profnath’s price. Over time bordeebook would keep raising its price and profnath would shadow the price rises, always undercutting them. Eventually, the spiral stopped – presumably a human intervened – but not before bordeebook was offering The Making of a Fly for a mere $23,698,655.93, plus $3.99 shipping.

Profnath presumably had a copy of the book and was trying to offer the highest price it could, consistent with undercutting everyone else. Bordeebook is more puzzling, but Eisen has an ingenious theory. He thinks bordeebook didn’t have the book at all. Instead, it was trying to generate interest from buyers who didn’t mind paying a bit more for dealing with a reputable seller (bordeebook had more than 100,000 satisfied customers on record).

The only trouble was, if somebody ordered the book, bordeebook had to get hold of it – hence the reason for charging the prevailing market price plus mark-up.

Of course $23,698,655.93, plus shipping, is a lot to pay for a book – a long way from the dream that the internet would usher in an era of perfect price transparency, in which consumers would discover the cheapest products and prices would inevitably fall to the cost of production. (Not everyone making this case in the dotcom-bubble era realised that it was inconsistent with the other touchstone principle of the time, that internet companies would be insanely profitable.)

In fact, it’s far from obvious why price transparency would lead to lower prices. The problem was studied by the 19th-century French mathematician J.L.F. Bertrand. Bertrand realised that with two competitors offering identical goods with transparent prices, all customers would flock to the cheaper offering. Each company could capture the entire market by undercutting by a penny, and the process of undercutting would only stop when prices fell to the cost of production.

Like any economic model, Bertrand’s is an oversimplification. And a small tweak turns his prediction on its head: simply repeat the process indefinitely. The competitors will benefit hugely if they can find a way to raise prices; not to $23m, perhaps – but to the level a monopolist would charge. And price transparency, which seemed to be the customer’s best friend, becomes her worst enemy.

Here’s why: with two competitors charging monopoly prices as part of an understanding, there is no incentive for either to cut prices. None. The price cut will be instantly observed and matched by the competitor. Far from inexorably lowering prices, transparency can ensure they stay high.

The only thing that will bring prices down is a new entrant. And in many industries breaking into an established market is not easy. Transparency can be excellent for lowering prices – but it is not wise to take that for granted.

Also published at ft.com.