[EM] Fighting the good fight against IRV

[Michael Welford]

I saw yet another recommendation for IRV. It appeared in the Hightower
Report. Jim Hightower is a left-wing populist from Texas. ( I know that
description sounds contradictory. ) In response to the recent
presidential election debacle, he recommends a number of electoral
reforms. Most were good, including abolishing the electoral college and
using proportional representation.

Unfortunately the first recommendation is IRV. In support of IRV he
gives a hypothetical example wth Bush, Nader and Gore running, in which
Ralph Nader is eliminated first after a respectable showing and Gore
wins when votes are transfered.  In response, I sent him the following
e-mail.

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------

This is in response to your Hightower report article on electoral
reforms. I agree on most most of them, but I have some reservations
about Instant Runoff.

The problem ironically enough is that IRV can be a little too friendly
to third parties. The example you give for a strong Nader showing is a
reasonable example of what could happen in 2004 under IRV.

Now consider what might happen 4 years later. We imagine that Gore has
drifted to the right to pull in conservative voters and succeeded only
alienating the left wing. Polls show Bush likely to take 40% of 1st
place votes in a (re)rematch with Nader and Gore getting about 30% each.
It's likely that Gore would beat either of the other candidates in a
head-to-head match. Nader supporters vote for their favorite confidant
that the spoiler problem has been solved. The result is that Gore gets
the lowest number of first place votes and is eliminated. There are two
possible outcomes at this stage, both of them bad. No matter whether
which candidate wins the runoff their victory is tainted. And the vote
lists exist that prove that Gore would have beaten any other candidate.

But that's not the worst of it. If Nader is just a little behind Gore in
1st place preferences, a small number Bush supporters might dishonestly
put Nader at the top of their lists just to eliminate Gore. ( I  suppose
most Bush supporters would be too honest to do this. That's what the
manipulaters will count on. )

It gets worse if there is also a party of the extreme right in the race.
This can give rise to both ends against the middle strategy.

The problem is that IRV can discriminate against centrist candidates in
a basically unfair way. And I think centrists will be smart enough to
figure that out.

Many of us favor a method that will always pick a candidate that can
beat any other candidate, if such a strong candidates exists. And the
voting lists needed for IRV give the information to identify them. Our
problem is that we can't agree on what to do when there is no such
candidate. ( I personally would revert to plurality in these rare
cases.)

If you want to know more you do a web search on voting and Condorcet. Or
reply to me and I'll share some of my bookmarked URLs.

I like the other electoral reforms just fine.

Keep up the good work.

Mike Welford