Vickrey auction

A Vickrey auction is a type of sealed-bid auction, where bidders submit written bids without knowing the bid of the other people in the auction, and in which the highest bidder wins, but the price paid is the second-highest bid. The auction was created by William Vickrey. This type of auction is strategically similar to an English auction, and gives bidders an incentive to bid their true value.

Vickrey's original paper considered only auctions where a single, indivisible good is being sold. In this case, the terms Vickrey auction and second-price sealed-bid auction are equivalent, and are used interchangeably. However, when either a divisible good or multiple identical goods are sold in a single auction, these terms are used differently.

The most obvious generalization to multiple or divisible goods is to have all winning bidders pay the amount of the highest non-winning bid. This is known as a uniform price auction. The uniform-price auction does not, however, result in bidders bidding their true valuations as they do in a second-price auction unless each bidder has demand for only a single unit. A generalization of the Vickrey auction that maintains the incentive to bid truthfully is known as the Vickrey–Clarke–Groves (VCG) mechanism. The idea in VCG is that items are assigned to maximize the sum of utilities; then each bidder pays the "opportunity cost" that their presence introduces to all the other players. This opportunity cost for a bidder is defined as the total bids of all the other bidders that would have won if the first bidder didn't bid, minus the total bids of all the other actual winning bidders.

For example, suppose two apples are being auctioned among three bidders.

• Bidder A wants one apple and bids $5 for that apple.
• Bidder B wants one apple and is willing to pay $2 for it.
• Bidder C wants two apples and is willing to pay $6 to have both of them but is uninterested in buying only one without the other.

First, the outcome of the auction is determined by maximizing bids: the apples go to bidder A and bidder B. Next, the formula for deciding payments gives:

• A: B and C have total utility $2 (the amount they pay together: $2 + $0) - if A were removed, the optimal allocation would give B and C total utility $6 ($0 + $6). So A pays $4 ($6 − $2).
• B: A and C have total utility $5 ($5 + $0) - if B were removed, the optimal allocation would give A and C total utility $(0 + 6). So B pays $1 ($6 − $5).
• similarly, C pays $0 (5 + 2) − (5 + 2) = $0.

Vickrey auctions are much studied in economic literature, but are not particularly common in practice. One market in which they have been used is stamp collecting. eBay's system of proxy bidding is similar, but not identical, to a Vickrey auction. A slight generalized variant of a Vickrey auction, named generalized second-price auction, which is different from the VCG mechanism, is known to be used in Google's and Yahoo!'s online advertisement programmes. NYU Law School uses an iterated version of the Vickrey auction model for its course registration lottery.

Properties

Self-revelation/incentive compatibility

In a Vickrey auction with independent private values (IPV) each bidder maximizes his or her expected utility by bidding (revealing) his or her true valuation.

Ex-post efficiency

A Vickrey auction is decision efficient (the winner is the bidder with the highest valuation) under the most general circumstances; it thus provides a baseline model against which the efficiency properties of other types of auctions can be posited. It is only ex-post efficient (sum of transfers equal to zero) if the seller is included as "player zero," whose transfer equals the negative of the sum of the other players' transfers (i.e. the bids).

Weaknesses

Despite the Vickrey auction's strengths, it has shortcomings:

• It does not allow for price discovery, that is, discovery of the market price if the buyers are unsure of their own valuations, without sequential auctions.
• Sellers may use shill bids to increase profit.

The Vickrey–Clarke–Groves (VCG) mechanism has the additional shortcomings:

• It is vulnerable to bidder collusion. If all bidders in Vickrey auction reveal their valuations to each other, they can lower some or all of their valuations, while preserving who wins the auction. [1]
• It is vulnerable to shill bidding with respect to the buyers.
• It does not necessarily maximize seller revenues; seller revenues may even be zero in VCG auctions. If the purpose of holding the auction is to maximize profit for the seller rather than just allocate resources among buyers, then VCG may be a poor choice.
• The seller's revenues are non-monotonic with regard to the sets of bidders and offers.

The non-monotonicity of seller's revenues with respect to bids can be shown by the following example. Consider 3 bidders A, B, and C, and two homogeneous items bid upon, Y and Z.

• A wants both items and bids $2 for the package of Y and Z.
• B and C both bid $2 each for a single item (bid $2 for Y or Z), as they really want one item but don't care if they have the second.

Now, Y and Z are allocated to B and C, but the price is $0, as can be found by removing either B or C respectively. If C bid $0 instead of $2, then the seller would make $2 instead of $0. Because the seller's revenue can also go up when bids are increased, the seller's revenues are non-monotonic with respect to bids.

Proof of dominance of truthful bidding

The dominant strategy in a Vickrey auction with a single, indivisible item is for each bidder to bid their true value of the item.

Let v_i be bidder i's value for the item. Let b_i be bidder i's bid for the item.

The payoff for bidder i is $\begin{cases} v_i-\max_{j\neq i} b_j & \text{if } b_i > \max_{j\neq i} b_j \\ 0 & \text{otherwise} \end{cases}$

The strategy of overbidding is dominated by bidding truthfully. Assume that bidder i bids b_i > v_i.

If $\max_{j\neq i} b_j < v_i$ then the bidder would win the item with a truthful bid as well as an overbid. The bid's amount does not change the payoff so the two strategies have equal payoffs in this case.

If $\max_{j\neq i} b_j > b_i$ then the bidder would lose the item either way so the strategies have equal payoffs in this case.

If $v_i < \max_{j\neq i} b_j < b_i$ then only the strategy of overbidding would win the auction. The payoff would be negative for the strategy of overbidding because they paid more than their value of the item, while the payoff for a truthful bid would be zero. Thus the strategy of bidding higher than one's true valuation is dominated by the strategy of truthfully bidding.

The strategy of underbidding is dominated by bidding truthfully. Assume that bidder i bids b_i < v_i.

If $\max_{j\neq i} b_j > v_i$ then the bidder would lose the item with a truthful bid as well as an underbid, so the strategies have equal payoffs for this case.

If $\max_{j\neq i} b_j < b_i$ then the bidder would win the item either way so the strategies have equal payoffs in this case.

If $b_i < \max_{j\neq i} b_j < v_i$ then only the strategy of truthfully bidding would win the auction. The payoff for the truthful strategy would be positive as they paid less than their value of the item, while the payoff for an underbid bid would be zero. Thus the strategy of underbidding is dominated by the strategy of truthfully bidding.

Truthful bidding dominates the other possible strategies (underbidding and overbidding) so it is an optimal strategy.

Use in network routing

In network routing, VCG mechanisms are a family of payment schemes based on the added value concept. The basic idea of a VCG mechanism in network routing is to pay the owner of each link or node (depending on the network model) that is part of the solution, its declared cost plus its added value. In many routing problems, this mechanism is not only strategyproof, but also the minimum among all strategyproof mechanisms.

In the case of network flows, Unicast or Multicast, a minimum cost flow (MCF) in graph G is calculated based on the declared costs d_k of each of the links and payment is calculated as follows:

Each link (or node) $\scriptstyle e_k$ in the MCF is paid

p_k = d_k + MCF(G − e_k) − MCF(G),

where MCF(G) indicates the cost of the minimum cost flow in graph G and G − e_k indicates graph G without the link e_k. Links not in the MCF are paid nothing. This routing problem is one of the cases for which VCG is strategyproof and minimum.

In 2004, it was shown that the expected VCG overpayment of an Erdős–Renyi random graph with n nodes and edge probability p, $\scriptstyle G \in G(n, p)$ approaches

$\frac{p}{2-p}$

as n, approaches $\scriptstyle \infty$, for $n p = \omega(\sqrt{n \log n})$. Prior to this result, it was known that VCG overpayment in G(n, p) is

$\Omega\left(\frac{1}{np}\right)$

and

$O(1)\,$

with high probability given

$np=\omega(\log n).\,$

References

• Vijay Krishna, Auction Theory, Academic Press, 2002.
• Peter Cramton, Yoav Shoham, Richard Steinberg (Eds), Combinatorial Auctions, MIT Press, 2006, Chapter 1. ISBN 0-262-03342-9.
• Paul Milgrom, Putting Auction Theory to Work, Cambridge University Press, 2004.
• Tech Ho, "Consumption and Production" UC Berkeley, Haas Class of 2010.

External links

• Paper on the history of Vickrey auctions in stamp collecting