Developed by William Vickrey (Counterspeculation, Auctions, and Competitive Sealed Tenders, 1961), Roger Myerson (Optimal Auction Design, 1981), Paul Milgrom and Robert Wilson (the FCC spectrum auctions, 1994), and Paul Klemperer (Auctions: Theory and Practice, 2004).
Sets the rules for how bidders compete (open or sealed, ascending or descending, first price or second), so bids reveal true value and resist collusion.
One license, four bidders, and what three auction formats make the winner pay
A software maker sells a five-year regional license. Set what each of four distributors privately thinks it is worth, and how far below value bidders shade a sealed bid when the winner pays its own bid. The same bidder wins every format; the price does not. In the open format and the second-price sealed bid the winner pays about the runner-up's value, so honest values set the price; in the first-price sealed bid, the price rests on how much bidders shade. Illustrative.
An example
A software maker wanted one exclusive distributor per region, on five-year licenses. In the first region four distributors were interested, and the head of channel sales planned to ask each for its best offer and negotiate with the top two. A board member pointed out that each distributor would shade its offer and expect a second round, so the process would reward whoever bluffed best.
The team wrote down the goal first: the license should go to the distributor that valued it most, at a price close to that value, without the four comparing notes, since the same firms would bid for six more regions. They ruled out an open ascending auction, where repeat bidders could signal across regions. They chose a sealed second-price format, with a reserve price set from the old reseller's margins and a published rule that the winner pays the second-highest bid. Bids came in at values close to each firm's own projections. The winner paid $7.5 million, more than the best negotiated offer the team had expected.
Why it works
A sale's rules decide what bidders say. Under a rule where the winner pays its own bid, every bidder has a reason to understate its value and to guess what the others will do, so the outcome rests partly on guessing skill. Under a rule where the winner pays the next-highest bid, understating can only lose the item and never lowers the price, so the best bid is the true value.
Designing the auction means choosing among these effects on purpose. Open formats reveal information as they run, which builds trust and helps bidders who are unsure of the value, and also lets repeat bidders signal to each other. Sealed formats hide bids and resist collusion. A reserve price protects the seller when competition is thin.
The thinking behind it
William Vickrey William Vickrey, an economist, was the first to analyze auctions with game theory, in Counterspeculation, Auctions, and Competitive Sealed Tenders (1961), in the Journal of Finance. The paper founded auction theory. The sealed-bid second-price auction, in which bidding one's true value is the best strategy, carries his name. He shared the 1996 Nobel Memorial Prize in Economic Sciences with James Mirrlees.
Roger B. Myerson Roger Myerson, a game theorist at Northwestern and later the University of Chicago, derived in Optimal Auction Design (1981) the auction that maximizes the seller's expected revenue, a founding result of auction theory. He shared the 2007 Nobel Prize in economics with Leonid Hurwicz and Eric Maskin for laying the foundations of mechanism design, the study of rules built to produce a chosen outcome.
Paul Milgrom Paul Milgrom, an auction theorist, developed A Theory of Auctions and Competitive Bidding (1982) with Robert Weber and later wrote Putting Auction Theory to Work (2004). With Robert Wilson, Preston McAfee, and John McMillan he worked on the simultaneous multiple round auction that the US Federal Communications Commission adopted in 1994 to sell spectrum licenses. He and Wilson shared the 2020 Nobel Memorial Prize in Economic Sciences for improvements to auction theory and inventions of new auction formats.
How to run it
- Decide the goal. Choose among revenue, putting the item with whoever values it most, speed, and resistance to collusion, and rank them.
- Size up the bidders. Note how many there are, how differently they value the item, whether they meet again in other sales, and how sure each is of the value.
- Choose the format. Open formats help uncertain bidders and build trust; sealed formats resist signaling; second-price rules invite true bids.
- Set the reserve and the rules. Fix a floor below which the item does not sell, and decide what bidders see and when.
- Publish the rules before bidding. State the format, the payment rule, and the tie-breaks in writing so bidders can plan honestly.
- Simulate when stakes are high. Run the format against plausible bidder values to see who wins and at what price.
- Watch repeated rounds. Look for patterns such as bids that end in the same digits or regions that never overlap, and change the format if they appear.
Where it goes wrong
The first failure is a vague process presented as negotiation. Bidders who cannot tell how the winner will be chosen bid defensively and save their real numbers for later rounds that may never come. Pick a format and say what it is.
The second is ignoring repeat play. Bidders who meet across several sales can share the spoils, each staying out of the others' regions, and an open format makes the signals easy. If prices across lots look oddly orderly, switch to sealed bids or combine lots.
The third is forgetting the reserve. With only two serious bidders, a second-price auction can sell a valuable item at a low price. Set a floor from what the item is worth to the seller.
When to use something else
- Mechanism Design. When the rules being designed are broader than a sale, such as a budget process or an internal market, and must make honest reports the best strategy.
- BATNA and ZOPA. When there is one buyer and one seller and the deal will be negotiated between them.
- Game Theory. When the question is how a few parties will respond to each other's moves, before any sale format is chosen.
Where it sits in SOLVE eX
Go to the source
- Counterspeculation, Auctions, and Competitive Sealed Tenders, Journal of Finance, 1961
- Optimal Auction Design, Mathematics of Operations Research, 1981
- Simultaneous multiple-round auction design for the FCC spectrum sales, 1994
- Auctions: Theory and Practice, 2004