advanced 12 min read
Social sciences · Topic
Auctions
probability theory · optimization · convex optimization · information theory
Auction theory analyses how different bidding rules affect revenue, efficiency, and bidder behaviour. The revenue equivalence theorem unifies first-price, second-price, English, and Dutch auctions under symmetric independent private values, while Myerson's optimal mechanism design characterises the revenue-maximising auction.

Auction Formats

Format Winner pays Dominant strategy
Second-price (Vickrey) Second-highest bid Bid true value $v_i$
First-price Own bid Shade bid below $v_i$
English (ascending) Dropout price Stay until price $= v_i$
Dutch (descending) Stopping price Stop at optimal shade

In the Vickrey auction, truthful bidding is a weakly dominant strategy: winning at the second price gives $v_i - b_{(2)} \geq 0$ regardless of others’ bids.

Revenue Equivalence Theorem

Under symmetric independent private values drawn from a common distribution $F$ on $[0,\bar{v}]$, any auction that (i) allocates to the highest type and (ii) gives a type-0 bidder zero expected surplus yields the same expected revenue to the seller. For $n$ bidders with values i.i.d. $F$, the equilibrium expected revenue equals:

\[\mathbb{E}[R] = n \int_0^{\bar{v}} v \, f(v) F(v)^{n-1} dv - \int_0^{\bar{v}} F(v)^{n-1} dv \cdot \text{(boundary term)}\]

The first-price equilibrium bid for uniform $F$ on $[0,1]$ is $b^*(v) = \frac{n-1}{n} v$.

Myerson’s Optimal Auction

Myerson (1981) showed the revenue-maximising auction uses virtual valuations $\psi_i(v_i) = v_i - \frac{1 - F_i(v_i)}{f_i(v_i)}$. The optimal rule allocates to the bidder with the highest non-negative virtual valuation and sets a reserve price $r^$ solving $\psi(r^) = 0$. For uniform $F$ on $[0,1]$, the optimal reserve is $r^* = 1/2$ regardless of the number of bidders.