[EM] Election-Methods Digest, Vol 53, Issue 53

[Greg]

> If I remember correctly, Abd wrote that, in every
> IRV election for public office ever held in the
> USA, the IRV winner was identical to the plurality
> winner. Doesn't that mean that -- when we apply
> your logic -- plurality voting always elects the
> right winner?
>
> Markus Schulze

Markus, I don't recall whether Abd wrote that, but the statement is
incorrect. Two counterexamples:

1) The first ever IRV election for public office in the United States
(to my knowledge) for the Mayor of Ann Arbor, where the Republican
received 49% of the first choices to the Democrats 40%, with the
remaining 11% for the Human Rights Party candidate. The Human Rights
Party candidate was eliminated, and the Democrat emerged the winner
with barely over 50%. I think it's safe to say the Democrat would have
won under Condorcet as well.

2) The most recent IRV election in the United States in Pierce County,
Washington. Sean Bunney received 35% of the first choices for County
Executive with Pat McCarty in second with 26%, but McCarthy won in the
final round. An analysis of the ballots shows McCarthy was the
Condorcet winner.

Still, I actually *do* think the plurality will select the right
winner most of the time. But it has perhaps a more important problem:
the relatively small chance of a "spoiled" election creates a
tremendous nomination incentive that discourages third parties from
running, having their issues heard, and competing is a genuine way.
IRV allows third parties to participate and paves the way for
proportional representation.

Greg