Principle of Revelation and Impossibility Theory

Boundaries of the Possible: What Cannot Be Achieved by Rules

Mechanism design faces fundamental barriers. Even with the most clever rule design, there are things that cannot be implemented with strategically thinking agents. Impossibility theorems delineate what is fundamentally unattainable.

This is not just abstract mathematics: the Arrow and Gibbard–Satterthwaite theorems explain why an “ideal” voting system does not exist, and allow for conscious choice of which imperfections are permissible.

Arrow’s Impossibility Theorem (1951, Nobel 1972)

Context: n ≥ 2 voters, m ≥ 3 alternatives. A “rational” aggregation of individual preferences into a social choice is needed.

Four axioms: 1. Completeness and transitivity of the result. 2. Pareto Principle: if everyone prefers x over y → the social ordering prefers x over y. 3. IIA (Independence of Irrelevant Alternatives): social choice between x and y depends only on individual preferences x vs y, not on the positions of other alternatives. 4. Non-imposition: there is no predetermined winner.

Arrow’s Theorem: The only rule that satisfies all four axioms is dictatorship: there exists one agent d whose preferences always coincide with the social choice.

Meaning: It is impossible to simultaneously have a “reasonable” aggregation of preferences with three or more alternatives. Every real voting system violates at least one axiom.

Condorcet’s Paradox: With three voters with cyclical preferences (A>B>C, B>C>A, C>A>B): majority voting yields A>B (2:1), B>C (2:1), C>A (2:1) — cycle! There is no “Condorcet winner,” transitivity is broken. This illustrates violation of transitivity under majority rule.

IIA Violation: The Borda method (sum of ranks) — intuitively attractive — violates IIA: adding a third alternative changes the relative ranking of the first two.

Gibbard–Satterthwaite Theorem

Context: n ≥ 2 agents with strict preferences, m ≥ 3 outcomes. Social choice function f: Lⁿ → X (L — set of strict orderings).

Theorem (Gibbard 1973, Satterthwaite 1975): Any SCF whose range contains at least three outcomes and which is not a dictatorship, is manipulable: there exists an agent i and preferences P such that i can improve their outcome by submitting false preferences.

Corollary: In any non-dictatorial voting system with three or more candidates, there exist incentives for strategic voting. “Strategic” choices are not a pathology, but a mathematical inevitability.

Numerical Example (Strategic Borda Voting): Three voters, three candidates. With honest preferences, candidate A wins (Borda sum of ranks). But if voter 3 changes the stated order, candidate C wins, whom 3 prefers over A. The strategy “yields a result.”

Mechanism Design with Limited Rationality

Classical mechanism design presupposes perfectly rational agents. The new direction — behavioral mechanism design — takes into account real cognitive limitations.

Examples: Nudge — gentle prompting: changing the default option alters behavior without changing incentives. Pension accumulations: opt-out (default participation) vs opt-in (must actively join) → drastically different participation levels. Simplifying choice: with too many options, agents make random or default choice (paradox of choice). The optimal mechanism may limit the number of options.

Optimal Regulation of Monopoly

Problem: The regulator knows: monopoly is efficient (θ=θ_H) or inefficient (θ=θ_L). The regulator does not know the type. Wants to maximize social welfare subject to a budget constraint.

Optimal mechanism: Contract menu (q, s): the efficient monopoly chooses (q*, s*) with q* = socially optimal output; the inefficient chooses (q**, s**) with q** < q* (distortion downward — deliberate reduction in output to “screen out” the inefficient type). The efficient monopoly receives an informational rent (paid for the “type”). The inefficient produces less than the optimum but receives no rent.

Borda Alternatives: If the IIA axiom is dropped, rating systems can be constructed. Alternative voting (ranked-choice, IRV) is used in Ireland and Australia — a compromise between manipulability and practicality. Approval voting is more resistant to strategic behavior and is used by several scientific organizations. Quadratic voting (Glen Weyl) introduces “buying” votes at a quadratic cost and allows for the consideration of preference intensity.

Voting Theory in Electoral Law and Corporate Governance

Voting theory is used in designing electoral systems and corporate governance. Arrow’s impossibility theorem (1951) proves that no preference aggregation system can simultaneously satisfy the reasonable axioms (Pareto, IIA, no dictator), which explains inevitable compromises in choosing an electoral system. The majoritarian system (winner-takes-all) in the USA and UK creates incentives for a two-party system (Duverger’s law) and is insensitive to minorities. Proportional system (Israel, Netherlands) provides representation but promotes coalition instability. Borda count and alternative voting are used in professional organizations (academic rankings, nominations) and in several countries (Australia, Ireland). In corporate governance, the voting structure of the board of directors and shareholders is critically important: cumulative voting protects minority shareholders by allowing concentration of votes on one candidate. The weighted voting system (dual-class shares) gives company founders (Google, Facebook) control with minority economic rights. Strategic manipulation of voting (Gibbard–Satterthwaite theorem) proves that any non-dictatorial voting system, when there are more than two choices, is subject to strategic behavior.

Quadratic voting (E. Posner, Weyl) is an alternative to classical mechanisms: each voter receives a “vote” budget and pays a quadratic price for each additional vote. This allows for accounting of preference intensity, not just order, and is resistant to strategic behavior in a certain sense. This mechanism is being tested in decentralized autonomous organizations (DAO) on the blockchain for collective decisions on resource allocation.

Assignment: (a) Borda voting: 3 voters with preferences A>B>C, B>C>A, C>A>B. Find the winner. Is there a “Condorcet winner”? (b) Prove that with majority voting (pairwise comparisons), Condorcet’s cycle arises. (c) Which voting rule would you choose for elections among 4 candidates? Explain the trade-offs (manipulability vs IIA vs Pareto).