Mechanism Design: How to Engineer Rules Under Which Self-Interest Leads to the Common Good

If the principal-agent problem explains why direct guardianship is impossible, then mechanism design shows a way out: to design not a command but rules.

Intuition: the reverse engineering of institutions

Ordinary game theory asks: "what outcome will this game produce?" Mechanism design inverts the question: "what rules should we set so that rational egoists, in equilibrium, arrive at the socially desired outcome on their own?" (NobelPrize.org).

Formalism and the revelation principle

Leonid Hurwicz introduced the notion of incentive compatibility and showed the limits of what is achievable when information is dispersed among participants. Roger Myerson's revelation principle sharply simplifies the problem: without loss of generality, one can restrict attention to mechanisms in which participants find it advantageous to report their private information truthfully (Optimal Auction Design).

Implementability à la Maskin

Eric Maskin formalised implementation theory: under what conditions (Maskin monotonicity) a mechanism implements the desired social function in all equilibria, not just in one convenient one. This is a safeguard against a "good" rule generating "bad" equilibria as well.

Applications

An exemplary application is optimal auction design: how a seller maximises revenue without knowing buyers' valuations. The same logic underlies the allocation of benefits and subsidies, tenders, platform algorithms, and school placement (VoxEU/CEPR).

Conclusions: rules instead of commands

For the theme of paternalism this is a key turn: instead of commanding a person to behave "correctly," the institution designs rules in which self-interested choice itself leads to the common goal. This is "soft" paternalism of the highest order — power designs not the outcome but the structure of the game.