Single-crossing property

Single-crossing property
Name	Single-crossing property
Field	Economics; Political science; Decision theory
Introduced	1970s
Notable for	Comparative statics; median voter theorems; mechanism design

Single-crossing property The single-crossing property is a condition used in Economics, Political science, Choice theory, Mechanism design, and Voting theory that restricts how preferences or payoff differences vary across types, locations, or parameters. It underpins comparative-static results in models associated with scholars and institutions such as Anthony Downs, Kenneth Arrow, James Buchanan, Avinash Dixit, and Jean Tirole, and finds applications in contexts studied at London School of Economics, Princeton University, Harvard University, Stanford University, and Massachusetts Institute of Technology.

Definition

The single-crossing property requires that for two alternatives A and B and a one-dimensional index t (type, income, ideology), the sign of the difference in utility or payoff between A and B changes at most once as t varies. Seminal uses appear in work by William R. Clark, Dennis Mueller, Gary Becker, Amartya Sen, and Kenneth J. Arrow in comparative studies associated with Cowles Foundation and Centre for Economic Policy Research. The condition is central to results related to the Median voter theorem, Hotelling's law, Downsian spatial models, and comparative-static proofs presented in textbooks by Hal Varian, Paul Samuelson, and Roger Myerson.

Mathematical formulation

Formally, let u_A(t) and u_B(t) be payoff or utility functions indexed by t ∈ T, where T may be a subset of the real line as in analyses at University of Chicago or Yale University. The single-crossing property holds if the function Δ(t) = u_A(t) − u_B(t) is monotone in t or has at most one sign change. Mathematical treatments are found in work by Milton Friedman, Leonid Hurwicz, Hervé Moulin, Eric Maskin, and John Harsanyi. Equivalent formulations use increasing differences as in Topkis's theorem and the theory developed at Northwestern University and Columbia University. In mechanism-design contexts studied by Roger Myerson and Eric Maskin, the property is often expressed via the monotone likelihood ratio property (MLRP) linking distributions over types as in research at Princeton University and University of Pennsylvania.

Examples and applications

Examples include voters’ preferences in spatial models pioneered by Anthony Downs and William Riker, firms’ entry decisions studied in Joe S. Bain-style industrial organization, and labor-supply choices in analyses by Gary Becker and Jacob Mincer. Applications extend to auction design by Paul Klemperer and Martin Shubik, contract theory by Oliver Hart and Bengt Holmström, and regulatory models at Federal Trade Commission and Federal Communications Commission. Empirical contexts include policy positioning in elections like the United States presidential election and candidate placement analyses by Svennson-style researchers, geographic models related to Hotelling's law and Sara LaLonde-style urban studies, and insurance selection problems addressed in studies affiliated with National Bureau of Economic Research and World Bank projects.

Properties and implications

When the single-crossing property holds, comparative static results often yield monotone strategies in games and monotone mechanisms in principal–agent problems studied by Roger Myerson and Eric Maskin. It underlies uniqueness results in equilibrium analyses by John Nash and comparative predictions in models advanced at London School of Economics and Cowles Foundation. In voting, single-crossing preferences combined with unidimensional ideological spectra produce Condorcet winners analogous to conclusions in the Median voter theorem and stability properties linked to work by Kenneth Arrow and Amartya Sen. In auctions and bargaining, the property facilitates revenue comparisons and incentive-compatibility proofs used by Paul Milgrom and Robert Wilson.

Relation to other preference restrictions

Single-crossing relates to increasing differences and the monotone likelihood ratio property (MLRP) often contrasted with quasi-concavity assumptions in work by Frank Ramsey and John von Neumann. It differs from single-peakedness studied by Anthony Downs and strategy-proofness explored by Kenneth Arrow and Amartya Sen; while single-peakedness restricts ranking over a policy line, single-crossing restricts how differences in utility move across types as in analyses by Geoffrey Brennan and James Buchanan. Connections to supermodularity and complementarity appear in treatments by Paul Milgrom and John Roberts.

Identification and estimation in empirical work

Empirical tests and estimation strategies for single-crossing rely on observed choices across types, often using panel or cross-sectional data as in studies by Angus Deaton, Esther Duflo, Joshua Angrist, and Guido Imbens. Researchers at National Bureau of Economic Research, Institute for Fiscal Studies, and Brookings Institution implement nonparametric tests, structural estimation, and instrumental-variables approaches to detect monotone variation in Δ(t) across observed covariates. Applications include estimating preference heterogeneity in voting using datasets from the American National Election Studies, labor supply estimation using Current Population Survey data, and insurance demand estimation in studies by AER-publishing scholars.

Extensions and generalizations

Generalizations include multi-dimensional single-crossing conditions, weak single-crossing, and probabilistic variants used in robust mechanism design investigated at MIT and Princeton University. Multidimensional extensions link to work on partial orders by Hervé Moulin and lattice-theoretic approaches by Donald Topkis; stochastic and Bayesian adaptations relate to research by Eric Maskin, Roger Myerson, and Kenneth Arrow. Recent advances connect the property to algorithmic mechanism design in collaborations between researchers at Stanford University and Carnegie Mellon University.

Category:Economics