Strategic Economics

Mechanism Design and the Revelation Principle

Stop guessing how people will game your rule and start designing the rule so that the behaviour you want is the behaviour that pays. Mechanism design is the engineering discipline behind auctions, take rates, quota plans, and marketplace policy.

  • Expert
  • 17 min total
  • 15 chapters

What decision this helps you make: Whether to change a rule you control — a bid format, a commission plan, a supplier selection process, a marketplace policy — and how to check, before you ship it, that honest participation is the profitable move rather than a favour you are asking people to do you.

What this topic is

Mechanism design is game theory run backwards. Ordinary game theory takes the rules as given and asks what people will do. Mechanism design takes the outcome you want — the efficient allocation, the honest forecast, the revenue — and asks what rules produce it once everyone plays selfishly. The revelation principle is the result that makes the search tractable: whatever any clever set of rules achieves in equilibrium, some simple rule achieves the same thing by asking every participant to state their private information truthfully.

Why it matters

Every rule you write inside a business is a mechanism, whether you designed it as one or not. A commission plan is a mechanism. A supplier selection process is a mechanism. A marketplace ranking algorithm is a mechanism. In each case, the people subject to the rule know something you do not — their real costs, their real pipeline, their real quality — and they will report it in whatever way the rule rewards. You do not get the behaviour you intended; you get the equilibrium of the rule you actually shipped.

Who should learn it

Anyone who sets rules other people optimise against: founders designing marketplace policy, sales leaders writing quota plans, procurement teams choosing between a tender and a negotiation, product managers building ranking and pricing systems, and operators who have watched a well-meaning incentive produce exactly the wrong behaviour.

What you will understand

  • What a mechanism is formally — an allocation rule plus a payment rule over reported types — and why writing yours down that way exposes the leak
  • The revelation principle, what it lets you skip, and the four assumptions that quietly carry it
  • How to test a proposed rule for incentive compatibility with arithmetic rather than hope
  • The impossibility results that bound what any rule can achieve, so you stop searching for a design that cannot exist

Prerequisites

Common misconception

"If I can just explain the rule well enough, people will use it as intended." The rule's intent is invisible to the equilibrium. Participants respond to the payoff surface, not the memo, and they find the profitable deviation faster than you can find it in review — often without any conscious intent to game anything, simply by each doing the locally sensible thing. The second, subtler misconception is that the revelation principle means real mechanisms should ask people to reveal everything. It means no such thing. It is an analytical device that lets a designer search over simple mechanisms; the one you actually ship can and usually should be indirect, because asking a bidder to hand over their entire valuation requires a level of trust in you that they will not extend.