The 2020 Nobel prize in economic sciences rewards work on an ancient form of transaction that has acquired new complexity and urgency in the modern age: the auction.
Insights in auction theory made by Paul Milgrom and Robert Wilson, both of Stanford University in California, have found applications ranging from the pricing of government bonds to the licensing of radio-spectrum bands in telecommunications.
Diane Coyle of the University of Cambridge, UK, says that the Nobel, announced on 12 October, will be widely welcomed. “These two not only did foundational work themselves”, she says, “but also inspired cohorts of younger researchers.”
Economist Preston McAfee of Google agrees. “I, and thousands like me, use the fruits of their work on a daily basis to make markets work better — to improve pricing, to manage incentives, to facilitate decision-making, to increase efficiency.”
Their research has intersected with computer science and communications engineering to lay the foundations for many online platforms, Coyle adds.
Economist John Kagel of Ohio State University in Columbus, USA, called it “an outstanding selection”.
Online platforms such as eBay have raised public awareness of some of the complexities of auctions. There are many ways to stage them: for example, in a so-called “English auction” the item on offer simply goes to the highest bidder; whereas in a “Dutch auction” the selling starts from a high price, and bidders submit the price they are willing to pay.
But bidding is affected by many more factors that might reduce the seller’s final profit, cause losses for the winning bidder, create inefficiencies of allocation, or harm the public good. The work of the two laureates has helped to reduce these problems and to suggest new, more efficient ways for auctions to be conducted.
One problem is that different bidders can have different degrees of knowledge about an item for sale. For example, in a property auction, all bidders for a property will have access to some public information such as its resale value. But other kinds of information — such as hidden structural damage — will be private and not known to everyone.
A bidder who does not have such information might end up overpaying if they want to buy the property. They might be able to infer what others know about the value if bids are public – and people start to drop out – but not if bids are private.
In the late 1960s and 1970s, Wilson showed what happens to prices and profits in auctions when bidders have different degrees of private information.
Furthermore, if information about a property is highly uncertain — if the nature of the neighbourhood is rapidly changing, say — that could make buyers cautious and reduce the seller’s profit. In the 1980s, Milgrom — a former doctoral student of Wilson’s — developed models (partly in conjunction with Robert Weber of Northwestern University) that showed there is then an incentive for sellers to gather and share expert information with bidders, within different auction formats. The predictions of how such public information helps prevent losses to sellers and increases their revenue have been born out by experiments, says Kagel.
A spectrum of options
Auctions can be more complex when the goods for sale are divisible into parts or batches — for example, when governments sell licenses to companies bidding to operate in energy, telecommunications or transportation markets. One issue for such auctions is that sellers are vulnerable to collusion between buyers to keep the buying price down. Wilson’s work in the 1970s helped to identify these problems and to design new auctions to avoid them, for example in markets for electricity provision.
The sales of items might also be interdependent. A classic example in the 1990s was the sale of radio-frequency bands to telecom companies for mobile-phone networks — which many countries decided was best done through auctions.
If rights to frequency bands were simply auctioned region by region, a national telecoms company couldn’t be sure of acquiring the same frequency everywhere. And the value to them for one region would depend on whether they could buy the same frequency band elsewhere. The resulting patchwork of coverage would be inconvenient for users too.
To tackle such problems, Milgrom and Wilson (and independently, McAfee) devised the simultaneous multiple-round auction (SMRA). Here, bidders can place bids over several rounds of bidding. This gives them a chance to glean something about others’ private information while bidding, creating fairer and more efficient outcomes.
This approach was used in 1994 for auctioning telecom licences in the United States, and has been adopted in Canada, India, and several European and Scandinavian countries. Milgrom has also devised other formats that ease some of the shortcomings of the SMRA.
“Unlike many theoreticians, Wilson and Milgrom brought their work to the real world, and transformed government policies toward auctions around the world,” says McAfee.
“There was no question that these two would win the Nobel prize at some point,” says economist Paul Klemperer of the University of Oxford. “It could have happened at any time in the past 20 years.”
“One could even imagine Paul Milgrom having a second Nobel prize,” he adds, for his work in information economics and industrial organization. Milgrom has given a Nobel acceptance speech before: in 1996, as a stand-in for William Vickery, who died three days after the announcement of his prize for laying the foundations of auction theory in the 1960s.