Working Paper: NBER ID: w21943
Authors: Laurent Bouton; Aniol Llorente-Saguer; Frédéric Malherbe
Abstract: We study the information aggregation properties of unanimous voting rules in the laboratory. In line with theoretical predictions, we find that majority rule with veto power dominates unanimity rule. We also find that the strategic voting model is a fairly good predictor of observed subject behavior. There are, however, cases where organizing the data seems to require a mix of strategic and sincere voting. This pattern of behavior would imply that the way majority rule with veto power is framed may significantly affect the outcome of the vote. Our data strongly supports such an hypothesis.
Keywords: Voting rules; Information aggregation; Unanimity; Veto power; Experimental economics
JEL Codes: C92; D70
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| majority rule with veto power (D72) | information aggregation (D83) |
| veto (D72) | fewer mistakes (C52) |
| unanimity (D70) | more mistakes (C52) |
| veto (D72) | significant reduction in type II errors (C52) |
| framing of majority rule with veto power (D72) | voting outcomes (D72) |
| framing of voting actions (K16) | behavior (C92) |
| veto (D72) | better information aggregation than unanimity (D80) |
| deviations from model predictions (C52) | impact outcomes under unanimous rules (D79) |