Voting Experiments: Measuring Vulnerability of Voting Procedures to Manipulation
|Title:||Voting Experiments: Measuring Vulnerability of Voting Procedures to Manipulation|
|Journal or Publication Title:|
AUCO Czech Economic Review, 5, 3, pp. 324-346
|Uncontrolled Keywords:||Voting, manipulation, information, computation-based simulations|
A minimal reduction in strategic voter’s knowledge about other voters’ voting patterns severely limits her ability to strategically manipulate the voting outcome. In this paper I relax the implicit assumption made in the Gibbard-Satterthwaite’s impossibility theorem about strategic voter‘s complete information about all other voters’ preference profiles. Via a series of computation-based simulations I find that vulnerability to strategic voting is decreasing in the number of voters and increasing in the number of alternatives. Least vulnerable voting procedures are Condorcet-consistent procedures, followed by elimination procedures, while most prone to manipulation are the simplest rules. Strategic voting is vulnerable both to an absolute and relative reduction in amount of information.