Subject: Re: [EL] Republicans got sigs for Nader in 2004 only in Michigan
From: Rob Richie
Date: 10/25/2010, 10:14 AM
To: "Scarberry, Mark" <Mark.Scarberry@pepperdine.edu>
CC: Election Law <election-law@mailman.lls.edu>

Mark is right that Larry's concern is possible with Condorcet voting ( but not with instant runoff voting unless the Nader-type candidate has more core support than a major party, forcing the major party's elimination before the final runoff).

For instance, you could have these ranking groups

49% - Bush-Nader-Gore
48% - Gore-Nader-Bush
3% - Nader (with 2% Nader Gore-Bush and 1% Nader-Bush-Gore)

In this situation, 51% of voters support Nader over Bush and 51% of voters support Nader over Gore. Under Condorcet rules, Nader wins as the Condorcet candidate.

Note that these results aren't so far-fetched. While Charlie Crist is doing quite well in the polls in Florida, he would be nearly unbeatable under a Condorcet system no matter who well he was doing in first choice support -- because he almost certainly would be the middle choice of nearly every Republican and every Democrat.

I think this analysis points to some ways that Condorcet voting falls short of IRV, at least from an IRV advocate's perspective.

1) Condorcet would more often result in the election of darkhorses that were totally off the radar, but for that reason were put in as a lot people's middle choice without being hardly anyone's first choice. "IRV elects leaders" is one slogan some IRV advocates use, with an understanding that to win an IRV election, you do need to earn a following of people who really want you there as a first choice, even while needing to provide that you aren't so polarizing that a majority would prefer your top opponent.

2) With Condorcet there could be a kind of "tyranny of the centrist" -- something some people indeed want, but would drain legislatures of more polarizing, but important legislators of the left and right. IRV would elect centrists when they can get in the top two against the left and right,requiring demonstration of a healthy degree of core strength and that some people really want them there. For more on this view, see Alec Slatky's blog on our site from this summer:
http://www.fairvote.org/why-the-condorcet-criterion-is-less-important-than-it-seems

3) As Mark points out, the possibility of having to resolve a Condorcet cycle (A beats B, B beats C, C beats A) mean that it's inevitable that indicating a second choice can result in the defeat of your first choice (a problem with the Borda count, for example). I think it's problematic any time you create direct incentives for people to truncate their ballot.

Note that IRV is used for countless NGO/faculty/student elections and has withstood the test of experience. For at list of NGOI's in the US using it, see:
http://www.fairvote.org/organizations-and-corporations-using-instant-runoff-voting

- Rob Richie, FairVote


On Sun, Oct 24, 2010 at 4:15 PM, Scarberry, Mark <Mark.Scarberry@pepperdine.edu> wrote:
There are alternatives other than instant runoff voting to our usual first past the post system (as I'm sure most of the list members know). Larry's concern could be real under one of those other approaches.
 
Consider the Condorcet approach. I'm sure a lot of list members know a lot more about it than I do but here's my understanding: Under the Condorcet approach a candidate who would defeat each of the other candidates in head to head competition is the winner. That seems a sensible result.
 
If a Condorcet voting approach were used (and if we limit consideration to three candidates), then there would have been, in effect, three contests in each state: Bush v. Democratic nominee, Bush v. Nader, Nader v. Democratic nominee. Voters would have been asked to vote for their preferred candidate in each of those three head to head matchups. If, in head to head voting, Nader beat Bush (because Democratic voters flocked to him as their second choice and because some Republicans preferred him to Bush) and Nader beat the Democratic nominee (because Bush voters flocked to him as their second choice and because some Democratic voters preferred him to the Democratic candidate), then Nader would be the Condorcet winner. It seems highly unlikely that Nader would have won either of those head to head contests, but I suppose it is possible he might have won them both.
 
Does anyone support such a Condorcet approach? Is it used anywhere in political contests?
 
One negative is that it can be cyclic, thus producing no winner: Democratic candidate defeats Bush, Bush defeats Nader, Nader defeats Democratic candidate. Thus there would be no Condorcet winner. But a cyclic result may be unlikely, and if it occurs there could be a tie-breaker like IRV or first past the post.
 
Also, the Condorcet approach can get complex if there are a lot of candidates, with the number of head to head matchups becoming large.
 
I've suggested using a Condorcet approach for voting by faculty on faculty candidates, with a Borda system backup (in which votes are weighted--1 for first choice, 2 for second, etc., each voter must vote on each candidate, low total score wins) to deal with cyclic results. Is that a bad idea as list members see it?
 
Mark Scarberry
Pepperdine

From: election-law-bounces@mailman.lls.edu on behalf of Rob Richie
Sent: Sun 10/24/2010 12:03 PM

To: Larry Levine; Election Law
Subject: Re: [EL] Republicans got sigs for Nader in 2004 only in Michigan

Larry, I'm afraid you don't understand instant runoff voting.

Nader only could have won in 2000 if he finished ahead of one of the major party candidates in first choice rankings -- just like in a regular runoff election. Nothing dark or sinister about it!

Rob

On Sun, Oct 24, 2010 at 2:38 PM, Larry Levine <LarryLevine@earthlink.net> wrote:
Fool around with this one for a minute. Could instant run off voting have made Nader President in either 2000 or 2004? He would have been the second choice of voters in both major parties, neither of which would have wanted to vote for the other major party. This could be looked at from the stand point of state-by-state results or a national popular voter.
Watch out for the "reformers".
Larry
[snip]

_______________________________________________
election-law mailing list
election-law@mailman.lls.edu
http://mailman.lls.edu/mailman/listinfo/election-law




--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"Respect for Every Vote and Every Voice"

Rob Richie
Executive Director

FairVote  
6930 Carroll Avenue, Suite 610
Takoma Park, MD 20912
www.fairvote.org rr@fairvote.org
(301) 270-4616

Please support FairVote through action and donations -- see http://fairvote.org/donate. For federal employees, please consider  a gift to us through the Combined Federal Campaign (FairVote's  CFC number is 10132.) Thank you!