Andrew Rawnsley in The Observer today has written a fairly ridiculous article in which he complains that FPTP supporters spend too much of their time being negative about AV, and uses the second half of the article to be negative about supporters of FPTP.
But let’s pass over this lack of self-awareness and give him a bit of credit; he does lay out succinctly what are supposed to be the advantages of AV:
I do think it would be a fairer and more appropriate electoral system for contemporary Britain. It will be a worthwhile improvement if MPs have to gather some form of support from at least half of the voters. The parties will be impelled to engage with more parts of the country than just a minority of marginals and it will pay MPs to connect with more parts of their constituencies.
I’d like to address these points a bit more seriously than I did yesterday.
First: the claim that AV is more appropriate for contemporary Britain. By Rawnsley’s lights this may well be the case, but what is so different about contemporary Britain? If AV is appropriate now for the reasons he gives, it has always been appropriate, and FPTP is a bastard system that has always been unfair and unrepresentative. And if this is true of FPTP in Britain, it is true of FPTP everywhere and at all times. Rawnsley does not address why, then, most democratic countries use FPTP.
Second: that MPs will have to gather support from at least half the voters. This is probably right, for certain values of ‘support’ and ‘half.’ For one thing, nothing in AV necessitates that an MP will have the support of half the constituency electorate; only that s/he will have the support of half of those who actually voted. If we examine, for instance, the Oldham and Saddleworth by-election in January, we see that less than half of the electorate turned out. Already we are not going to have the majority of voters represented.
Furthermore, if we redistribute the votes from all but the top four parties (Lab, Lib, Con, UKIP) to Labour, for the sake of simplicity, their share of the votes cast goes from 42.1% to 49.3%—not enough to win under AV.
Now we have to play a little game. Which party would UKIP voters place as their second preference? Probably Conservative. If we distribute the UKIP votes to the Conservatives, our top three totals are as follows:
Lib Dem: 31.9%
Still no majority. So we have to redistribute the Conservative votes to their second preferences, and the UKIP votes to their third preferences. I have a difficult time believing that either of these groups of voters would choose Labour as anything but their last wish, although some might have chosen parties already eliminated in earlier rounds. However, let us say that most of these votes would go to the Lib Dems next, so we’ll add another 15% to the Lib Dem share, giving us:
Lib Dem: 46.9%
Um, crap. We’re on our last two candidates. We’re certainly not going to start fiddling around with the second preferences of the Lib Dem voters, as that would be utterly absurd. But if we don’t, then we don’t have a candidate with at least 50.1% of the votes cast. So AV does not necessarily deliver MPs support from at least half the voters who voted.
Now obviously this is a rough and dirty calculation, that doesn’t take into account tactical voting. There may have been Conservatives who voted Lib Dem in the hope of keeping Labour out, as the Lib Dem candidate very nearly won in the general election. But that still doesn’t solve the problem that the Labour candidate, even if you very generously allow all of the smaller-party votes to them, would most likely not have achieved a 50.1% majority under AV. What happens in that case? A run-off, which would be less fair even than FPTP, since it would deny a good 50% of the voters to select a candidate they truly agreed with and thereby restrict voter choice even further? Who knows.
Third: the claim that AV would impel parties to engage with more of the country than just marginal constituencies. This, I fear, is a silly belief. Seats are ‘safe’ because a majority of the voters in those constituencies firmly and regularly support one party. Let us examine what is generally considered to be the safest seat in Britain: Bootle in Merseyside, which has been held by the Labour party since 1945. In that time, the Labour candidate has never won with less than 50% of the votes. The closest was in 1955, when the Labour candidate won with 52% of the vote. Since then, the Labour share has been well over 60%, once even as high as 83%. Other parties in Bootle simply do not have a chance, nor would they even under AV. This seat is not ‘safe’ because FPTP delivers a skewed result in which some people’s votes don’t count. This seat is ‘safe’ because the vast majority of voters there like Labour. And the fact that it’s ‘safe’ doesn’t appear to affect voter turnout; turnout in Bootle is no worse than anywhere else on average.
I’m not going to spend my whole afternoon trying to discover whether these facts are similar for all ‘safe’ seats, but I imagine they probably are. In which case, even under AV, I have a hard time believing Labour would bother directing its campaign effort at Bootle. Labour is in no danger whatsoever of being ousted from Bootle.
Parties will always campaign the hardest where they have to work the hardest to get elected, and I see no reason to think this would not also be the case under AV. It seems that the real complaint here is that parties don’t engage on a broad national level during election time, which is a pretty bizarre complaint if you ask me. A particular candidate need only bother engaging with the voters in his own constituency anyway; they’re the only ones who can vote for him. If his party as a whole focuses its campaign efforts elsewhere, that’s between him and his party, not between him and his voters. AV won’t solve this.
Fourth: AV will pay MPs to connect with more parts of their constituencies. If this is true, then MPs are damned stupid and there is simply no saving them. What candidate does not try to get the largest share of the vote possible? What candidate does not, already, try to increase his existing majority? In short, what candidate is not trying his damnedest to win? Show me this person. No, really: show me.
AV, therefore, is not a solution to minority majorities, safe seats, or MPs who don’t give a damn about the voters. It does not even make every vote ‘count,’ as there is still only one winner. At best, it is an opportunity for those who feel unrepresented by any major party to give their support to the candidate they dislike the least, and a chance to indicate that a vote for a party doesn’t necessarily equate with wholehearted support for that party.
But FPTP already permits people to send these signals. And send them the voters do. What, then, is the advantage of AV? That we would have a better way of quantifying these signals?
I mean, imagine how this would pan out. My MP is Chuka Umunna. Here is how he won:
Lib Dem: 35.8%
All others: 3%
Let’s assume the Conservatives all give Lib Dem as their second preference (I have a hard time believing any real Conservative voter would choose Labour anything but last.) Under AV, it’s therefore possible that Chuka Umunna would have lost. But could anyone then claim that Chris Nicholson was supported by more than half the voters? No. A true statement would be that Umunna was supported by less than half the voters. Since the FPTP vote already delivers this message, there is no need for AV to send the signal. Another true statement would be that more than half the voters preferred someone other than Umunna. Again, we can see this from the FPTP vote. But why should this mean Nicholson deserves to win, when clearly more voters actively want Umunna than passively reject him? Those 42.8% of people who really want Labour are effectively disenfranchised by the fact that the Conservatives get to vote twice. How is this ‘fair’?
Or let’s examine it another way. Lots of people vote Lib Dem for tactical reasons—in the case of Streatham, probably as a vote against Labour. Under AV, the theory goes, people wouldn’t need to do this. Suppose a third of those Lib Dem votes were actually Conservatives in sheep’s clothing. Under AV, our first-preference result might look like this:
Lib Dem: 24%
All others: 3%
Let’s be generous and say that of those 24% of voters whose first preference is the Lib Dems, a third choose Labour second and two thirds choose the Conservatives. In which case Chuka Umunna wins, and AV delivers a result absolutely no different from FPTP: Chuka Umunna becomes MP for Streatham. Everything else an AV vote might tell us is academic.
If I studied the results from every constituency in the 2010 general election, I could probably show this again and again: AV would deliver an obviously unfair result, or one exactly the same as FPTP, or one where nobody manages to secure more than half of the votes. The more marginal the seat, the more likely an unfair or inconclusive result; the safer the seat, the less difference AV would make.
FPTP means that a candidate can win with less than half of the votes. Admittedly this is not great. But at least it means he was wanted by more voters than any other candidate was. At least it means every voter had exactly one vote of equal weight. Why should we reject this system for one in which some people’s half-hearted second preferences are held equal to others’ whole-hearted first, and may not even then deliver a conclusive result? That is not ‘fair’. That is the bastard system.