State the revenue equivalence theorem assumptions. Prove first-price sealed-bid yields…
- User
State the revenue equivalence theorem assumptions. Prove first-price sealed-bid yields the same expected revenue as second-price Vickrey because bidders shade bids to match expected second highest.
Assistant
This is a question about auction theory, specifically the Revenue Equivalence Theorem (RET). Let me break this down carefully.
Revenue Equivalence Theorem Assumptions
The standard assumptions for the RET are:
- Symmetry: All bidders are ex-ante identical (have symmetric value distributions)
- Risk Neutrality: Bidders are risk-neutral
- Independent Private Values: Each bidder's valuation is independently drawn from the same distribution and is private information (known only to that bidder)
- No budget constraints: Bidders can bid up to their true valuation
- Monotone bidding strategies: Equilibrium bidding functions are monotone increasing
- Participation is exogenous: All bidders must participate (though this can be relaxed)
Proof that First-Price Sealed-Bid Yields Same Expected Revenue as Second-Price Vickrey
Let me work through this carefully.
Setup:
- n bidders
- Values drawn independently from distribution F with support [0, v̄]
- Bidder i has value vi
Second-Price (Vickrey) Auction:
In the Vickrey auction, the equilibrium bidding strategy is:
- Bid truthfully: bi(vi) = vi
This is a dominant strategy equilibrium. When you bid your true value:
- You win if vi > all other players' values
- You pay max of other players' bids = second highest value
Expected payment when winning:
- E[Payment | Win with value v] = E[v(2) | v(1) = v] where v(2) is second highest value
First-Price Sealed-Bid Auction:
In first-price, if you win, you pay your own bid. So you must bid strictly less than your value or you get zero surplus.
Let β(v) be the equilibrium bid function for value v in first-price auction.
In equilibrium (symmetric), a bidder with value v will:
- Win if their bid exceeds all competitors' bids
- Pay their own bid β(v)
For a bidder with value v to be willing to use strategy β(v), they must be indifferent at the margin. By monotonicity of β, winning occurs precisely when v exceeds all other values, which happens with probability F(v)^(n-1). The indifference condition requires that the expected surplus from bidding β(v) equals zero at the boundary, which pins down the bid function through the differential equation that relates β(v) to the distribution F.
The equilibrium bid function must satisfy a key property: the expected payment by the winner should match the expected value of the second-highest bid, just as in the Vickrey auction. This equivalence between first-price and second-price auctions reveals that despite their different payment rules, they generate the same expected revenue and allocate the object to the highest-value bidder.
For a bidder with value v, the expected payoff from bidding β(v) is the surplus times the winning probability: U(v) = (v - β(v)) · F(v)^(n-1)
The optimal bid function emerges from the condition that the payoff's derivative with respect to valuation equals the winning probability. Integrating this relationship from 0 to v determines the full structure of β.
To find expected revenue, I need to consider what the winner actually pays. In a second-price auction, the winner pays the second-highest valuation rather than their own bid, which creates a different revenue structure than the first-price case.
The equilibrium bid in first-price auctions reflects this: a bidder with value v shades their bid down to match the expected second-highest value conditional on being the highest bidder. This bid shading—β(v) = E[v(2) | v(1) = v]—ensures that both auction formats generate equivalent expected revenues, a result known as revenue equivalence. the bid-shading intuition. In a second-price auction, bidders bid truthfully and pay what the runner-up offers. In first-price, I shade my bid down to match what I expect the second-highest bidder would bid in that same auction. Since the runner-up bids their true value in second-price, my shading in first-price offsets this by the same expected margin, yielding identical expected revenue across both formats.
Revenue Equivalence Theorem: Assumptions and Proof
Assumptions
The Revenue Equivalence Theorem holds under:
- Symmetric bidders: All bidders have valuations i.i.d. from the same distribution F
- Risk neutrality: Bidders maximize expected payoff linearly
- Independent private values (IPV): Each bidder's valuation is privately known and independent of others
- Monotone bidding: Equilibrium bidding functions are increasing in valuation
- Efficient allocation: The bidder with the highest value wins
- Regular distribution: CDF F satisfies standard regularity conditions
Proof: First-Price Equals Second-Price Expected Revenue
Second-Price (Vickrey) Auction Equilibrium
In Vickrey auctions, bid truthfully is a dominant strategy:
- Winner pays: the second-highest value
- Expected revenue:
First-Price Sealed-Bid Equilibrium
With symmetric bidders, let denote the equilibrium bid function.
Bid-Shading Formula: A bidder with valuation shades their bid to equal the expected second-highest value conditional on having the highest value:
Derivation of bid shading:
The expected surplus from bidding as if value is is:
where is the probability of winning (your value exceeds others).
By envelope theorem:
Integrating from 0 to with :
The first-order condition for optimal bidding yields the bid function satisfying:
Revenue Equivalence
Expected revenue in first-price:
Substituting the bid-shading formula and using the law of iterated expectations:
Therefore:
Intuitive Explanation
Auction Bid Payment Expected Value Vickrey Bid truth () Pay 2nd place truthful bid () First-Price Shade to expect 2nd place () Pay own shade bid The key insight: Bidders shade their first-price bids down to what they expect the runner-up will value. Since the runner-up bids truthfully in Vickrey, this equalizes expected revenue.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.