If the first past the post (FPTP) voting system did not exist, we would not dream of inventing it. Apart from the simple case when there are only two candidates, FPTP has the fundamental flaw that it requires me, the voter, to second guess how other people are going to vote. Conversely, if I vote without thinking what other voters are likely to do, I am simply being irrational.
The recent Oldham East and Saddleworth by-election provides a straightforward example. For simplicity, ignore all candidates except those from the three main parties. As a card-carrying Conservative, I obviously wanted the Conservative candidate to win. However, the general election result and polling data indicated that there were not enough Conservative voters to get our man elected. For me to vote Conservative in such circumstances is rational only if I absolutely do not care whether Labour or the Liberal Democrats take the seat if the Conservatives cannot have it.
If I am not indifferent to choosing between Labour and the Lib Dems, then it is silly for me to vote Conservative when there is no reason to think the Conservative candidate can win. Instead, the logical decision is to vote for either Labour or the Lib Dems, depending on which of them I prefer. In Oldham East and Saddleworth I would have voted Lib Dem.
In every FPTP election held in Britain where more than two candidates stand, many voters have to make exactly this type of unpalatable choice.
The sensible approach is what our own Conservative Party does when choosing a leader. We have a ballot (in our case of the MPs) and if someone has the support of more than 50% of the electorate, we declare him elected. If no candidate has majority support, we eliminate the least popular candidate and vote again, until we have a winner. Of course in the case of the Party leader, there is the wrinkle that the last two go forward to face a different electorate, but that does not change the principle. The Conservative Party for its own leadership election has designed a system where voters do not need to attempt to second guess what other voters are going to do.
Holding repeated ballots is expensive, and I cannot imagine anyone wanting to do it when electing the House of Commons. However, the mathematical substitute is the AV system. I have learned recently that the Americans call it the “Instant Run-off System” which is actually a much better name, since it explains how the system works, whereas AV is a complete misnomer. However, in the UK we are stuck with the name. If nobody’s candidate preferences change from ballot to ballot as candidates are eliminated, AV gives the same result as holding repeated ballots, eliminating the least popular candidate each time, but AV achieves that without the expense of repeated ballots.
Many organisations use AV for internal elections, including my former firm of PricewaterhouseCoopers LLP when electing its senior partner.
Apart from being more rational, AV also has the advantage of giving more information about true levels of party support. If in Oldham East and Saddleworth I tactically vote Lib Dem, nobody will ever know that there was a voter who really wanted the Conservative candidate to win. AV allows us to see the true level of support for each party, as well as getting to the result that best fits the electorate’s preferences.
There are of course some arguments against AV, but none that I find convincing:
Some voters will fail to understand it, or are incapable of numbering, say, eight candidates without losing count. That is true, but we cannot keep FPTP just because some voters are incapable of coping with anything more complicated than putting one cross on a ballot paper.
A person’s fourth preference (say) is less important than his first preference. This argument denigrates the voter; if I cannot have my preferred Conservative candidate elected, why should I be ignored in the choice between the Labour and the Lib Dem candidates. It is far better to have me rank the candidates by preference than to force me to pretend to be a Lib Dem supporter by FPTP tactical voting.
AV will lead to more coalitions. I don’t see why it should and nor does the recent paper Worst of Both Worlds: Why First Past the Post no longer works from the IPPR. Even if it did, I am not convinced that having coalitions is bad; for every country with coalitions that is badly governed one can think of another that is well governed, and coalitions prevent lurches from one extreme to the other. More immediately, AV does not automatically help minor parties. For example if a constituency is divided Con/Lab/Lib Dem 40/40/20 with every Labour and every Conservative voter giving the Lib Dem candidate their second preference, that will not help the Lib Dem at all since he will be eliminated on the count of first preferences.
To summarise, I think the strongest argument for AV is that it is the system organisations choose for electing a leader where the electors themselves have decided on the system, as is the case with both MPs in the Conservative Party and partners in PricewaterhouseCoopers LLP.