Manipulability of majority relation-based collective decision rules

Abstract

In the problem of aggregation of rankings or preferences of several agents, there is a well-known result that reasonable social ranking is not strategy-proof. In other words, there are some situations when at least one agent can submit insincere ranking and change the final result in a way beneficial to him. We call this situation manipulable and using computer modelling we study 10 majority relation-based collective decision rules and compare them by their degree of manipulability, i.e. by the share of the situation in which manipulation is possible. We found that there is no rule that is best for all possible cases but some rules like Fishburn rule, Minimal undominated set and Uncovered set II are among the least manipulable ones.