The Revenue Equivalence Theorem (RET) answers a surprisingly clean question: does it matter which auction format a seller picks? Under the right conditions, the answer is no. All standard auction formats produce the same expected revenue. This result is foundational because it shifts attention away from format choice and toward the factors that actually move revenue, like reserve prices, bidder participation, and information structure.

Assumptions and Implications

The theorem holds under four key conditions:

Risk-neutral bidders who maximize expected payoff
Independent private values (IPV): each bidder knows their own valuation, drawn independently from the same distribution, and doesn't care about others' valuations
Symmetric bidders: all bidders draw values from the same distribution $F(v)$
Standard allocation: the item goes to the bidder with the highest value, and a bidder with the lowest possible value expects zero surplus

When these hold, any auction mechanism satisfying the allocation and zero-surplus conditions yields identical expected seller revenue. This covers first-price sealed-bid, second-price sealed-bid (Vickrey), English (ascending), Dutch (descending), and even all-pay auctions.

The practical takeaway: if the assumptions hold, the seller's format choice is irrelevant to expected revenue. That means other design levers, such as setting a reserve price or attracting more bidders, matter far more than whether you run an English or sealed-bid auction.

Mathematical Formulation

Let $n$ bidders each draw a private value $v_i$ independently from a common CDF $F(v)$ on $[0, \bar{v}]$ . Define $P(v_i)$ as the probability that bidder $i$ wins when their value is $v_i$ .

The expected payment of a bidder with value $v_i$ is pinned down by the allocation rule alone:

$E[p_i(v_i)] = v_i \cdot P(v_i) - \int_{0}^{v_i} P(x) \, dx$

This formula comes from integrating the incentive compatibility (truth-telling) constraint. The first term is the expected value of winning; the integral subtracts the "information rent" the bidder earns from having a value above the minimum.

The seller's expected revenue sums over all bidders:

$E[R] = \sum_{i=1}^{n} E[p_i(v_i)]$

The core logic: since the payment formula depends only on the allocation rule $P(\cdot)$ and the boundary condition (lowest type pays zero in expectation), any two mechanisms with the same allocation and the same boundary condition must produce the same expected revenue. That's the entire theorem.

Applying Revenue Equivalence

Assumptions and Implications, Auctioning risk: the all-pay auction under mean-variance preferences | Economic Theory

Comparing Auction Formats

Revenue equivalence links the four standard formats in pairs:

First-price sealed-bid ↔ Dutch auction. In both, the winner effectively pays their own strategic bid. In a Dutch auction the price drops until someone claims the item, which is strategically identical to choosing a sealed bid. Both yield the same expected revenue.
Second-price sealed-bid (Vickrey) ↔ English auction. In both, the winner pays something tied to the second-highest valuation. In a Vickrey auction, you pay the second-highest bid directly. In an English auction with private values, bidders drop out at their true values, so the winner pays just above the second-highest value. Again, same expected revenue.
Across pairs. The RET says first-price and second-price formats also yield the same expected revenue, even though the realized payments differ. In first-price auctions, the winner pays less than their value but shades their bid; in second-price auctions, the winner bids truthfully but pays the second-highest bid. These effects exactly offset in expectation.

The theorem extends to all-pay auctions too. Every bidder pays their bid regardless of winning, but the expected revenue to the seller still matches, because bidders adjust their strategies accordingly.

Optimal Auction Design

The RET provides the baseline from which optimal auction design begins. If all standard auctions yield the same revenue, a seller who wants to do better must break one of the theorem's conditions.

Myerson's optimal auction theory builds directly on this insight. It characterizes the revenue-maximizing mechanism subject to:

Incentive compatibility: bidders find it optimal to report their true values
Individual rationality: bidders participate voluntarily (non-negative expected surplus)

Key results from Myerson's framework:

Setting an optimal reserve price above the seller's own valuation excludes low-value bidders and extracts more surplus from high-value ones. The optimal reserve is the same across formats satisfying the RET conditions.
The seller may benefit from discriminating among bidders if their value distributions differ (asymmetric bidders), for instance by giving weaker bidders a bidding advantage.
More complex settings may call for bundling items or using non-standard pricing rules to boost revenue beyond what any standard single-item auction achieves.

Limitations of Revenue Equivalence

Assumptions and Implications, 13. Procurement Management – Project Management – 2nd Edition

Relaxing Assumptions

Each assumption, when violated, breaks equivalence in a predictable direction:

Risk aversion. Risk-averse bidders bid more aggressively in first-price auctions to reduce the risk of losing. They don't change behavior in second-price auctions (where bidding truthfully is dominant regardless). Result: first-price auctions generate higher expected revenue than second-price auctions when bidders are risk-averse.

Affiliated (correlated) values. When bidders' valuations are positively correlated, open formats like the English auction reveal information during bidding. Bidders update their beliefs upward as they observe others staying in, which supports more aggressive bidding. The linkage principle (Milgrom and Weber, 1982) formalizes this: auction formats that reveal more information tend to generate higher revenue. The ranking becomes English > second-price sealed-bid > first-price sealed-bid.

Asymmetric bidders. If bidders draw values from different distributions, the allocation may no longer go to the highest-value bidder in all formats, and the equivalence breaks down.

Budget constraints, participation costs, and collusion all introduce further deviations. For example, participation costs reduce the number of active bidders, and the magnitude of this effect can differ across formats.

Extensions and Generalizations

Reserve prices: The RET extends to auctions with reserve prices. The optimal reserve is the same across formats satisfying the theorem's conditions, and it equals the value at which the seller's marginal revenue from an additional bidder type equals zero.
Multi-unit auctions: Revenue equivalence generalizes to settings where multiple identical items are sold, under appropriate conditions. The Vickrey-Clarke-Groves (VCG) mechanism is the multi-unit analogue of the second-price auction and serves as a benchmark here.
Interdependent values: Researchers have extended the theorem to cases where a bidder's value depends partly on others' private signals, though the conditions become more restrictive.

Relevance of Revenue Equivalence

Practical Applications

The RET is best understood as a benchmark rather than a literal prediction. Real auctions almost always violate at least one assumption, so the theorem tells you where to look for revenue differences across formats rather than claiming those differences don't exist.

In practice, the theorem has informed auction design in high-stakes settings:

Spectrum license auctions (FCC): designers used RET-based reasoning to focus on participation incentives and information revelation rather than obsessing over format alone
Timber auctions (U.S. Forest Service): empirical studies have found approximate revenue equivalence between sealed-bid and oral formats in some timber sales, consistent with the IPV model
Government procurement: the theorem guides decisions about sealed vs. open bidding by highlighting which assumption violations are most relevant

Empirical Evidence and Future Research

Empirical results are mixed, which is exactly what the theory predicts once you account for assumption violations:

Studies of U.S. Forest Service timber auctions found rough equivalence in some settings, supporting the IPV model for those markets.
Online advertising auctions and other settings with asymmetric bidders or correlated values show systematic deviations from equivalence.

The gap between theory and data points toward productive research directions: developing auction models that incorporate risk aversion, affiliation, budget constraints, and participation costs simultaneously, and testing them against the increasingly rich datasets available from online platforms and government auctions.

2,589 studying →