Condorcet voting systems (Executive summary)

  1. Condorcet voting systems are more complicated than range voting (and the best Condorcet voting systems are much more complicated). This probably already is sufficient to make them politically unadoptable.
  2. Range voting is more expressive than Condorcet voting.
  3. Condorcet voting systems are not usable on many of today's voting machines. But range voting is usable on every voting machine in the USA, including noncomputerized ones (and usually easily), right now, without modification and without reprogramming. That makes range more adoptable with less pain and expense.
  4. The fundamental idea (by Condorcet) which lies behind these systems, also is satisfied by range voting, i.e. there is a (nontraditional but arguably better) way to define "Condorcet system" that (a) is compatible with all previous applications of the term to voting systems found in all previous political science books, and (b) under which range voting is also a "Condorcet system."
  5. The traditional Condorcet property sounds good at first but has numerous disadvantages such as the free the slaves vote, the "DH3 pathology," and Fishburn's counterexample to Condorcet being a "good" thing and Condorcet self-contradiction example.
  6. All Condorcet systems exhibit "favorite betrayal." In a 3-way election like Bush v Gore v Nader 2000, voters are tempted to exaggerate their good and bad opinions of Bush and Gore by artificially ranking them first and last, even if they truly feel Nader is best or worst. If that happens, then Nader cannot win. The result of that, over time, would be self-reinforcing 2-party domination, causing Condorcet not to be much if any improvement over plain plurality voting.

    There is good reason to believe (or at least suspect) that this pathology will happen with every Condorcet system, but that it will not happen with range voting. For example about 90% of Australians vote in this exaggerated manner on their rank-order ballots.

    Meanwhile, with range voting, even if every voter exaggerates and ranks Gore=99, Bush=0 (or the reverse) then Nader still has a very realistic possibility of winning (without any ties required) so there is no obvious reason we will get 2-party domination.
  7. Counterintuitively, we can prove that (under reasonable assumptions about strategic voter behavior) Approval and traditional-Condorcet voting actually are not in conflict (no-conflict theorem) and it is plausible that range and approval voting both will actually be more likely in practice to elect honest-voter Condorcet winners, than "official" Condorcet methods! [Summary of the model.]

    And because strategic range voters generally vote approval-style, the same would be true of range voting elections with strategic voters. In other words:

  8. Even if you don't quite totally buy all that, we think the logical force of this still should be enough to convince you that, in practice, one cannot expect any great advantage for Condorcet methods over the much simpler range voting system.

More details

Reply to Maskin & Sen 2016

Return to main page