Skip to content

Implementation Theory

Why It Matters

Designers often specify what outcomes they want without checking if any mechanism can actually achieve them. Implementation theory separates wishful thinking from structural possibility. If your desired outcome isn't implementable, no amount of clever mechanism design will achieve it. You must either change your goals or accept imperfect implementation.

The Key Move

Start with your desired outcome. Ask: "What information would a mechanism need to achieve this? Can participants be incentivized to reveal that information truthfully? Does the outcome require agents to take actions against their self-interest?" If the answer is yes to the last question, the outcome isn't implementable without changing incentives or enforcing compliance. Look for impossibility results that cover your setting - they tell you what you can't achieve.

Classic Application

Efficient trade with private valuations (Myerson-Satterthwaite theorem). Buyer values item at V_b (private), seller values it at V_s (private). Social efficiency requires trade if and only if V_b > V_s. But no mechanism can guarantee efficient trade while maintaining individual rationality (voluntary participation) and budget balance (no external subsidies) and incentive compatibility (truthful revelation). It's structurally impossible. You must sacrifice one goal.

Surprising Application

Feature prioritization in product development. Product wants to build features users value most (efficiency). But users overstate feature value if they think it affects prioritization (no incentive compatibility). And you can't force participation in feedback (need individual rationality). And you can't pay users for accurate estimates (budget balance). The goals are incompatible. Real prioritization methods sacrifice something - typically incentive compatibility (accept biased feedback) or efficiency (use proxy metrics like usage, not stated value).

Failure Modes

Ignoring impossibility results (trying to achieve goals that are provably impossible). Assuming all reasonable-sounding goals are jointly achievable (many aren't). Over-applying theory to settings where assumptions don't hold (implementation theory assumes sophisticated strategic reasoning that real agents may lack). Treating impossibility results as reasons to give up rather than constraints to work within (knowing what's impossible clarifies which tradeoffs to make).

Go Deeper

Myerson and Satterthwaite (1983), "Efficient Mechanisms for Bilateral Trading." Maskin (1999), "Nash Equilibrium and Welfare Optimality." Jackson (2001), "A Crash Course in Implementation Theory" - excellent survey.