Alternative Vote – fairer winners are only sometimes different winners

Posted on 21 April 2011 by Paul Gregory

(Time to dust off this under-used blog, because some things need more than 140 characters.)

I’m been getting increasingly angry at the crap spouted by the No camp ahead of May’s referendum on improving the voting system.

Their latest argument could be summed up as “It’s not fair if the person who wins the first round doesn’t win the match.”

Of course, in First Past The Post (FPTP), there is only one round. The candidate with the most votes wins.

But the two points of AV are
a) voters can rank candidates in order of preference
b) the winner needs 50% of votes

Any system with a guaranteed first-round winner would effectively be the same as FPTP:
a) voters can vote for one candidate only
b) the winner needs the highest number of votes

So, BECAUSE the first-round winner isn’t necessarily the winner, AV has
a) greater opportunity for voters to vote for their real first preference
b) a better match win condition

However, this does NOT mean that the first-round winner never wins the match.

If a candidate gets 50% of first preferences in the first round, there is only one round and the first-round winner wins.
A natural consequence of this is that if there are only two candidates, there is only one round and the first-round winner wins.

If the win condition hasn’t been met, there is a further round, in which the last-placed candidate is eliminated and any second preferences are added to the other candidates.

It is by this process that someone other than the person who won the first round might win – it won’t always, because these preferences could just as easily go to the first round winner – but sometimes it will.

If AV didn’t *sometimes* give a different result, there would be no point changing to it, would there?

AV will only ever give a different result in situations where the first-round winner was not a fair winner. Without a true 50%+ majority, a first-round winner has not got enough proven support. The second preference counting is not about justifying the first-round winner’s legitimacy, it’s about finding the legitimate winner. Which might be the first-round winner.

That’s all just the vote *counting* side of things. Because voting preferences are being counted, voters no longer need to discount their true first preference if they fear it would be a wasted vote. Instead of trying to guess how everyone else will vote, they can trust the system to sort that out. No more wasted votes, no more split votes.

It may take a while for everyone to grasp this, but the fact that the first-round winner might not win the match means that people can vote differently in the first-round than they do today. That might even mean that candidates from parties that don’t feature well in national polls or in historic data will get enough first preference votes to win in the first round. So everyone’s happy.

Different winners? Sometimes. Fairer winners? Always.

Should the person who wins the first round win a match, or the person in the lead in the last round after eliminations?

End note: The "No" camp won. I'm still bitter.