[EM] IRV in San Francisco

[Justin Sampson]

On Fri, 12 Nov 2004, Eric Gorr wrote:

> At 10:11 AM -0500 11/12/04, Warren Schudy wrote:
>
> > 1) Did the ballot only allow each voter to give the top three choices?  
> > I suspect that restriction would significantly decrease the
> > effectiveness of IRV.

Yes, three choices. The City Charter says that allowing full ranking is
preferred, but that three choices is the minimum if the Department of
Elections can't figure out how to support full ranking.

> Well, it will cause IRV to fail the Independence of Clones Criterion and
> thereby be subject to a spoiler effect again.

Doesn't IRV suffer from spoiler effects anyway?

> What I found interesting is that a claim made to get the voters to
> approve IRV was that it would eliminate the spoiler problem.

I believe the main arguments in San Francisco were (1) avoiding the extra
cost of December run-off elections and (2) avoiding the low turnout of
December run-off elections. There was also some partisan analysis that the
turnout in December was more "conservative" than the turnout in November,
so that "progressive" voters would be stronger with IRV.

Justin