As we near the potential date of a referendum on AV, the rhetoric is becoming more strident. The smoke of war obscures the battlefield on which the contest is being fought, with consequences for the clarity of the argument. My fear is that we object to AV on the wrong grounds, and in so doing, make it sound more appealing than it actually is!
The decision to refer to the proposed system of voting as the Alternative Vote has caused great confusionin the first place, not least amongst policy makers, campaigners and august politicians of all hues. It should have been made clear from the beginning that what was on offer was Optional Preferential Voting as utilised in state elections in Queensland and New South Wales. This is fundamentally different to the Alternative Vote system used in the Australian Federal Election since voters under OPV have the option , but not the obligation, to express their preferences for each candidate in rank order. Under Australian-style AV, each candidate on the ballot paper must be ranked by preference. This would be unacceptable in this country, since many would baulk at having to express a preference of any sort for the BNP or Socialist Workers. To avoid that, and for no doubt other reasons, it is OPV that is being offered to the British people – and the fact that it is optional, not compulsory, has profound consequences in my view for its impact.
Does it change the outcome?
The most recent state election in Queensland (2009) saw only 3 seats of the 89 contested won by a candidate (after preferences had been taken into account) who had not led after the first round of voting. Similarly, in the New South Wales election of 2007 saw only 2 of the 93 seats ‘switch’ after preferences were taken into account.
It is worth examining the circumstances of the five seats:
- Barron River (Qld) – the second preferences of the Greens broke disproportionately to Labor, allowing the Labor candidate to overtake the Liberal National (LNP) candidate who had ‘won’ on the ‘first round’.
- Chatsworth (Qld) – Green second preferences once again allowed a Labor candidate to narrowly defeat the LNP candidate who had ‘won’ on the ‘first preferences’.
- Nanango (Qld) – quite different in that an Independent candidate had to rely on the second preferences of the ALP candidate to defeat the LNP candidate (from the Bjelke-Petersen dynasty, legendary in Qld politics) who had ‘won on first preferences’.
- Dubbo (NSW) – similarly to Nanango, Labor second preferences enable an Independent to overhaul the National Party candidate who had received the most first preferences.
- Lake Macquarie (NSW) – Liberal voters helped deprived Labor of a seat when their second preferences split 90-10% in favour of an Independent (and local mayor) who had been substantially behind on the first preference count.
In summary, there is no distinct pattern emerging of either one side or another benefitting. Each division where the eventual winner did not take the most first preferences has its own story of ‘why’ the result happened to happen. It cannot be said that OPV favoured one side or the other in a system very like ours with a strong centre-left and centre-right party.
It is perhaps also worth comparing this with the 11 out of 150 electorates in the Australian Federal Election where the eventual winner did not win the most ‘first preferences’. Had the ‘first preferences’ operated as under first past the post, this would have seen a net gain of 8 for the Liberal/National Coalition, and a very different outcome. So it is right to argue that AV can affect a tight election’s outcome, but less certain that OPV would.
Given there is a New South Wales state election imminent, it will be fascinating to see if this argument still holds true after they have voted!
Does it guarantee that the winning MP will have the support of half his electorate?
It is a frequent statement by supporters of AV that it will ensure the winning MP has the support of half of their voters. Given we do not have compulsory voting in the UK, this cannot be achieved in any case. Australia does have compulsory voting, so is easily achievable. In any case, it is highly unlikely that all MPs would have the support of 50% even of those who did go and vote in our non-compulsory system.
In Queensland in 2009, 17 of 89 seats saw the winner elected with the support (once preferences had been taken into account) of fewer than 50% of those who had expressed first preferences (‘formal votes’). Similarly, in New South Wales, 22 of the 93 seats were won with the support of fewer than 50% of those who had cast ‘formal votes’.
Why this occurs is due to ‘plumping’. In the state electorate of Sydney in New South Wales, 76% of Liberal voters had their votes ‘exhausted’ by the time their candidate had to drop out in the third round of preference assignation. That was 6,821 out of 8,235 Liberal voters who failed to express even a second preference, let alone a 3rd, 4th, 5th or 6th .
Even more interesting is the state electorate of Indooroopilly in the suburbs of Brisbane in Queensland. The result after first preferences was:
- Liberal National 11,570 44.45%
- Labour 6,907 26.54%
- Green 6,749 25.93%
- Other 802 3.08%
This, surely, would be a perfect example for supporters and detractors of AV alike of where Green voters’ second preferences would switch en masse to the Labor candidate, and see them to victory. The exclusion of the 4th placed candidate saw the Green come to within just 42 votes of the Labor candidate – not enough to overhaul therefore, so the Green candidate dropped out and their preferences reassigned. The 6,996 votes were then distributed as follows:
- Re-distributed to Labor 3,146
- Re-distributed to LNP 1,090
- ‘Exhausted’ 2,760
Thus some 40% of Green voters failed to express a second preference, ‘plumping’ only for their own candidate. The result for Indooroopilly saw the LNP candidate defeat the Labor candidate by 2,707 votes – only slightly fewer than the total of ‘exhausted’ Green votes. The victorious LNP candidate had the support of ‘only’ 49.52% of those casting a ‘formal vote’.
This is crucial example because of the weight attached to Robert McIlveen’s well-argued pamphlet for Policy Exchange against AV. In that report, he cites the example of how the LibDems would have won Reading East on an AV system. It is a good, well-worked example of how the system could unfold if we used the Australian federal system of AV. However, because it is based on a British Election Study of asking voters to fill out an AV ballot paper where they are told to record all preferences, it does not replicate the system we would actually have in this country, and it does not take account of plumping. For example, I would be surprised if every single Labour voter in Reading East would utilise their second preference in the first place.
I would argue that, at the moment, we already have an informal implementation of Optional Preferential Voting, which we refer to as ‘tactical voting’. Voters filter in their mind, before they go to vote, who is most likely to achieve the political outcome they want. Voters who wanted to keep the Conservatives out, voted for the candidate most likely to achieve that. In safe LibDem seats, there will always be a rump Labour vote, just as there will always be a rump Conservative vote in any LibDem/Lab marginal. The key point is that the winnowing process has already occurred before any ballot is cast. OPV merely allows the voter the opportunity to record those preferences on the ballot paper rather than in a discussion round a breakfast table with the family, for example.
Does this mean I am in favour of AV? Not at all. I believe the referendum to be a necessary part of the coalition agreement, though not something I would have sought. I do not believe changing our electoral system in this way will do anything to improve the quality of democracy. It will not produce ‘better’ results. It may produce a handful of ‘different’ results than if FPTP applied, but as we can see from Queensland and New South Wales, those ‘different’ results will be few and far between and are unlikely to affect the overall outcome, nor guarantee that the magic ‘50% support’ threshold is reached.
There are far bigger constitutional arguments to be had than this sideshow. Whichever system triumphs, the fundamentals remain the same for those contesting elections.