[EM] Sunday reply to Bill Lewis Clark

[Richard Moore]

--- In election-methods-list at yahoogroups.com, "Bill Lewis Clark" 
<wclark at x> wrote:
 > I'll grant you that IRV would present some nasty problems down the 
road --
 > some of them arguably more troublesome than those Plurality causes 
-- but
 > there's no reason IRV couldn't be used as a stepping stone to later
 > changes to an even better system.

I disagree -- IRV would make a poor stepping stone.

IRV has only two positives over lone-mark plurality: It is cloneproof 
and it never chooses the Condorcet loser. Now the fact that it never 
chooses the Condorcet loser meant that in a 3-way race, if two 
candidates are preferred by a majority of voters to the remaining 
candidate, then one of those candidates will win, so neither of the 
first two candidates can be considered a spoiler. The third candidate 
could be a spoiler in some cases, if he/she has enough first place 
support to cause the Condorcet winner to be eliminated. But even 
neglecting that possibility (which after all could be indicative of a 
weakness in the CW's support), there is no guarantee that adding a 
fourth candidate would not spoil the race between the top two. In fact 
any claim that IRV removes the spoiler problem is worthless when more 
than 3 candidates are in the race. Even in 3-candidate races the 
problem is only made more subtle than it is in lone-mark plurality voting.

Passing IRV only seems likely if it is sold on the false advertising 
claim that it fixes the spoiler problem.

So 5 years after IRV is passed, if a better solution is proposed, the 
public will very likely be divided into two major camps. The first 
camp will say, "IRV was supposed to fix the spoiler problem, but look 
how it failed. How can we trust this new-fangled system to be any 
better?" The second camp will say, "Why do we need a new system? We 
fixed the spoiler problem 5 years ago!"

So the future of election method reform after IRV would not be very 
promising.

  -- Richard