Hugh Warren is a member of the ERS Council.
It is one of the precepts of preferential voting systems that a later preference should neither help nor harm an earlier preference. The purpose of this paper is to show that the Condorcet system of preferential voting infringes this precept.
Consider an election for one seat in which there are 3 candidates:
AB 16 AC 18 BA 16 BC 17 CA 15 CB 17The Condorcet method for the election yields the following results:
(Incidentally, the Single Transferable Vote, which amounts to the commonly called Alternative Vote in this case, would exclude the lowest, C, and hence would elect B.)
If the paradox is resolved by electing A, then, if instead of voting AC Voter X had voted AB, Candidate B would have beaten Candidate C, and accordingly by the Condorcet method Candidate B would have been elected. Therefore changing the second preference of Voter X from C to B works to the detriment of his first preference A.
If the paradox is resolved by electing B, then, if instead of voting BA Voter Y had voted BC, Candidate C would have beaten Candidate A, and accordingly by the Condorcet method Candidate C would have been elected. Therefore changing the second preference of Voter Y from A to C works to the detriment of his first preference B.
If the paradox is resolved by electing C, then, if instead of voting CB Voter Z had voted CA, Candidate A would have beaten Candidate B, and accordingly by the Condorcet method Candidate A would have been elected. Therefore changing the second preference of Voter Z from B to A works to the detriment of his first preference C.
Therefore, no matter how the paradox is resolved, the precept that a later preference should not harm an earlier preference is infringed.