AV could bring about a Tory landslide, increase the North-South divide and bring back two-party politics

Very little joined-up thinking appears to have been done about what the real-world political impact of AV would be if the British people collectively decide to introduce it at next May’s referendum. Here’s how I think it could pan out:

Scenario No. 1: By spring 2015, the coalition is still in place and is popular, having balanced Britain’s books and being seen as having set in motion a ‘revolution’ (one of Cameron’s pet terms) in the way English public services are delivered, involving greater community participation and decision making. The beneficiaries of this in electoral terms will be the Conservatives, and AV will help them to a massive landslide, which is a well-known defect of that system – although I note that the Electoral Reform Society has now dropped the list of negative features of AV, including its potential to exaggerate landslides, from its summary description of AV.

How would AV contribute to a Tory landslide? AV would do nothing to diminish the Conservatives’ dominance of the South of England (excluding London): they poll between 40% and 55% in most constituencies in this part of England; and they would easily amass enough second preferences where they needed them, including from Lib Dem first-preference voters, to hold these seats. In addition, AV could allow the Conservatives to virtually wipe out Labour in the South, as Labour pluralities could be overturned by the effect of Tory and Lib Dem first-preference voters awarding their second preferences to each others’ parties.

In the Midlands and the North, Labour pluralities could also be overturned by the same effect, the main beneficiaries being the Tories, as they would mostly finish in second place on first preferences, and so Lib Dem second preferences would mostly transfer over to the Conservative candidates (that is, assuming that this pattern of second-preference voting takes place, which depends on how popular the coalition is, and whether there is any formal or informal pact between the coalition parties to encourage this sort of voting).

Net result: massive Tory majority.

Scenario No. 2: In this scenario, the coalition’s popularity has dipped and Labour’s has risen. This would result in a stark North-South divide, with Labour making virtually a clean sweep in the Midlands and North of England (to say nothing of Scotland and Wales), while AV again did nothing to eat into the Tories’ dominance of the South.

In this instance, in the North and Midlands, AV would transform Labour first-preference pluralities into majorities, as disgruntled Lib Dem supporters of third-placed candidates would award their second or third preferences to Labour ahead of the Conservatives. In the South, any Labour resurgence would simply split the opposition, while the Conservatives would still consistently poll upwards of 40% on first preferences and so would in most cases easily garner enough second preferences to win. So you’d end up with an even more starkly polarised England, and indeed Britain, than under FPTP, with the Tories virtually wiped out from the Midlands upwards and westwards, and Labour still making no in-roads in the South.

Two-party politics: What both of these scenarios have in common is a return to two-party politics. I can’t see any realistic scenario for how AV could improve the Lib Dems’ electoral performance, despite this being a much-touted advantage of the system for that party. If anything, AV would exercise even more of a two-party squeeze against the Lib Dems. In the South of England, under both my scenarios, even if they finished second on first preferences, they would not command enough Labour second-preference support to be elected. In the Midlands and the North, they would finish mostly third – so their votes would be transferred to Labour or the Tories.

Existing Lib Dem seats would be under threat from any swing to either of the other parties – although AV does offer some protection to the Lib Dems here in that, so long as their candidates finish at least in second place on first preferences, AV gives them a chance of winning in the final run-off against either the Labour or Tory candidate.

Conclusion: Even if you don’t accept either of my specific scenarios, or that voting patterns will pan out the way I suggest in the presence of those scenarios, AV will do nothing to remedy some of the most damaging political consequences of FPTP: disproportionate Conservative dominance of the South of England; polarisation between the South and the rest of Britain; electoral elimination of third and smaller parties. If anything, AV will make these things worse.

And in terms of the strategy for moving forward from AV to PR – such as that strategy is, because it seems built more on hope than expectation – how on earth do supporters of PR see that as being advanced by a system capable of generating massive landslides for the Tories or Labour; one which consolidates those parties’ dominance of particular regions; and one which destroys any hope of plural politics? It seems to me that those parties will simply say, ‘Thank you very much and goodbye’.

Advertisements

11 Responses

  1. Interesting points David, but I think that you are missing the point of our electoral system. We don’t have PR, and nor is it on offer. Therefore to compare your imagined results of AV against a PR system (i.e. national support levels) isn’t really valid. FPTP doesn’t give proportional results, nor does AV. All this referendum is about is the best constituency electoral system.

    AV, like FPTP, is designed solely to give you the most popular local candidate. If that means in the South that most people in a constituency want a Tory or Lib Dem, then that is what they will get under AV, unlike under FPTP. Suggesting that we shouldn’t switch to AV because some constituencies with a minority Labour vote (less than 50%) would lose their MP is both silly and anti-democratic. The majority should have the choice of their MP, not a minority.

    There are dozens of other scenarios that could happen, from a Lib Dem resurgence to a split of the Tory party over the EU.

    Or, more likely, things will pretty much stay the same and the sky won’t fall in…. and we will have a much better electoral system.

    In the unlikely event of one of your scenarios coming true, then the point you are missing is that the result reflects the majority voting patterns. That’s what people actually VOTED for. It may not be proportional, but it does mean that locally every constituency MP is supported by at least 50% of the voters. And that’s what democracy is all about.

    In contrast, the 1997 Labour party landslide, under FPTP, was often achieved with minority support in constituencies.

    • Thanks for your comment, Anthony. I’ll number my responses for convenience:

      1) I agree that to judge AV by the standards of a PR system would be misleading, but I still think it’s a valid criticism of a voting system if it does deliver substantially disproportional results. Clearly, the intention isn’t to introduce PR in the UK at this stage. What I’m saying is that replacing a disproportional system with another disproportional system that may even exaggerate the disproportionality of FPTP results will not help to produce more representative and plural politics, and a fairer geographical distribution of that representation, which many of the supporters of AV hope it will.

      2) You say that AV will always hand the victory to the most popular candidate, who obtains 50%+ of the vote. This is not true. AV treats all the preferences of all the voters as carrying equal weight. For instance, the fourth preference of one voter that is transferred to the second-placed candidate is regarded as one vote for that candidate, equal to all the other first, second, third, fourth and subsequent preferences that are for that candidate. However, this inconsistently disregards all the second and subsequent preferences of those whose first preference was for one of the top-two candidates (assuming that, in most seats, it will come down to a race between two candidates). It is quite mathematically possible, and in UK political terms quite likely, that a losing candidate could amass more preferences of any flavour (first, second, third, etc.) than the ‘winning’ candidate under AV. But if all preferences are equal, why are some counted and some not?

      I know that the reason they’re not counted is because of the transfer process used by AV. But it does mean that the non-first preferences of more voters are ignored than are used. The fairer way to allocate non-first preferences would be, in the event of no majority of first preferences, to then add every candidate’s second preferences to see if they have amassed a majority of first + second preferences, and so on until someone does have a majority or, in the event of more than one candidate obtaining a majority, awarding the victory to the candidate with the largest majority at that point in the process. This method is called Bucklin voting (there’s an article on it in Wikipedia).

      This method also eliminates the flaw in AV whereby it’s also quite possible for candidates to win on less than 50% of the vote if enough voters do not indicate any preference for the two candidates remaining in the race. As I argue elsewhere, the recent election of Ed Miliband to the Labour leadership was in fact an example of that, which was glossed over by ignoring exhausted ballots (i.e. those not expressing a preference for any of the candidates still in contention) from the percentages at each stage.

      3) So your argument that the aggregate result of an AV election would be a reflection of what the majority in each constituency actually want does not hold up, both because, in some seats, there will be losing candidates who will have garnered more preferences (including first + second preferences in some instances) and hence a larger majority than the supposed winner; and because in other seats no candidate will win a majority. You say that this is better than FPTP. I would concede that, in a majority of instances, the result would probably be fairer; but to say it would always be fairer relies on AV’s assumption that all votes are equal (i.e. some voters fourth preferences for a candidate carry the same weight as others’ first preferences), such that an AV ‘majority’ is an unqualified majority of people who’ve voted ‘for’ that candidate to an equal degree. In any case, consider the following: parties A, B and C poll 40%, 26% and 24% of first preferences, but party C wins because enough voters for parties finishing below C put C as their second preference to ensure that C goes into the run-off against A, and first-preference B voters transfer their votes to C, allowing C to win. This sort of scenario has happened in real-world AV elections and could easily promote Labour in the South, and Lib Dems or the Tories in the North, in contrast to my scenario. That might even out the North-South divide, but it doesn’t really seem fair, does it?

  2. 1) I agree that it is fair to criticise both AV and FPTP for being disproportionate. Studies of AV have shown that in some cases such as 1997, the result might have been slightly less proportional (by about 30 seats) under AV, but that often it will be more proportional (like in 2010). Overall, it is pretty much the same as FPTP.

    Therefore, in a comparison of the two, proportionality really isn’t a factor.

    2) The problem is that you are thinking in terms of ‘preferences’ rather than voting rounds. AV is an instant run-off system. Those ‘preferences’ are in reality just a quick way of saying which party you would like to vote for *if* they have a candidate.

    If your ballot moves on to your second choice, it is as if your first choice candidate never stood in the election.

    Under FPTP, if there are ten candidates in my area, I might like to vote for UKIP, the EDP, the Greens and then the Lib Dems in that order (purely as an example!). If there are just three candidates, my vote might go to the Lib Dems. Does that mean that the Lib Dems were my first choice? Of course not. Not only that, I might vote Tory to stop Labour getting in. This is the fallacy of FPTP supposedly representing the first choice of voters.

    Electoral choice under a constituency based system is an illusion. FPTP is designed for a two party system – in nearly every constituency there are really only two candidates who could ever win. Occasionally there are three way marginals, but these are rare.

    All the other candidates standing in those constituencies are distractions. Anyone voting for them is wasting their vote under FPTP because they have no chance of being elected. What this means is that the election is decided solely by those voting for the two main local parties, and those choosing anyone else are effectively disenfranchised.

    If we could remove those time-wasting candidates and just allow people to choose from the two (three) who could possibly win, then we would have a real picture of the electoral landscape, and every vote would count.

    And that’s (sort of) what AV facilitates. While still allowing you to vote for your favoured candidate, it also means that you don’t necessarily waste your vote. Your vote will move on to one of the viable winners. AV allows *everyone* to take part in the decision about who is elected, not just the supporters of the big parties.

    And let’s not forget that AV is just an improved version of FPTP. It only kicks in if no one secures 50% of the vote immediately. This has to be better than FPTP which allows people to be elected with just 33% of the vote or so.

    WRT Bucklin voting, that’s not on offer, so it’s not even worth you thinking about it. It is AV or nothing 🙂

    As for ‘exhausted’ ballots – if people choose not to fill in all of the boxes, they are simply saying that they do not wish to vote for any of the candidates on offer. It is no different to people not turning up on polling day because they don’t support any of the three or four candidates on offer under FPTP. It is their choice to make.

    3) The example you have given is extremely rare, but again, it is no different to a FPTP election in which only candidates A and C stood – C would be elected because they are more popular.

    And one can also get ridiculous theoretical results under FPTP too – if there are ten candidates, it is possible for someone to be elected with 11% of the vote for example. There will always be some mathematical extremes in any electoral system.

    Britain is moving more and more towards a multi-party system. The EU elections showed that 42% of people voted for a party other than the big three. FPTP only works properly in a two party state, so it is entirely unsuitable for modern Britain. AV, on the other hand, allows people to express their support for their favoured party without wasting it.

    There are a whole range of other advantages of AV over FPTP that you haven’t mentioned in your article:
    http://isupportav.co.uk/av-is-better/

    • Anthony, thanks for your second response. Here’s mine.

      You say: “Those ‘preferences’ are in reality just a quick way of saying which party you would like to vote for *if* they have a candidate. If your ballot moves on to your second choice, it is as if your first choice candidate never stood in the election.” Too bad if you actually do prefer that candidate, then. It’s as good as saying that vote doesn’t count, as is implied by your other statements:

      – “If we could remove those time-wasting candidates and just allow people to choose from the two (three) who could possibly win, then we would have a real picture of the electoral landscape, and every vote would count. And that’s (sort of) what AV facilitates.” So anyone other than the big three (or four in Scotland and Wales) might as well not bother standing then.

      – “As for ‘exhausted’ ballots – if people choose not to fill in all of the boxes, they are simply saying that they do not wish to vote for any of the candidates on offer. It is no different to people not turning up on polling day because they don’t support any of the three or four candidates on offer under FPTP. It is their choice to make.” Yes, but then if you make out that they haven’t in fact voted at all (which is what your analogy suggests), then the 50%+ threshold reached by the winning candidate will be mere puff in many cases: based on ignoring the fact that some voters have not voted for either of the final two but have in fact turned out and voted.

      “The example you have given is extremely rare, but again, it is no different to a FPTP election in which only candidates A and C stood – C would be elected because they are more popular.” Yes, but who would win if only A and B stood? Possibly B. And who would win if only B and C stood – possibly B. B would in this instance be the Condorcet winner: the candidate, as you imply, who would win in a straight run-off against any other. It’s a well-known defect of AV that it often fails to find the Condorcet winner. There are many single-winner systems that do pick the Condorcet winner (see here). Plus there are other single-winner methods that resolve the question of who should win in the absence of a majority of first preferences for any candidate more fairly than AV, including range voting methods such as my own ARV system.

      “AV, on the other hand, allows people to express their support for their favoured party without wasting it.” Yes, but you’ve just said that voters might as well not bother voting for them if they’re not one of the top-three parties. So it’s just giving people more of an emotional outlet and allowing them to send a bit of a message to politicians about what they ‘really’ think; but ultimately, the net effect is the same: they end up having to vote for the same party that they would have had to vote for tactically under FPTP.

      So AV really is a case of ‘plus ça change’ . . ..

      • “Too bad if you actually do prefer that candidate, then. It’s as good as saying that vote doesn’t count, as is implied by your other statements:”
        Yes, except that the first preference vote will be recorded for your favoured party, allowing them to see their real level of support.

        “So anyone other than the big three (or four in Scotland and Wales) might as well not bother standing then.”
        Yes, that’s exactly what I am saying FPTP makes happen. AV is better because it not only allows everyone to decide the winner (not just the core supporters), it also allows smaller parties to see (and thus build up) their real support.

        WRT to your ARV system, it is all irrelevant. None of that is on offer. It doesn’t matter if there are alternatives that may or may not be better. All that matters is that AV is significantly better than FPTP. Could it be further improved? Of course. It is screaming out for a bolt on top-up system to make it PR.

        “So it’s just giving people more of an emotional outlet and allowing them to send a bit of a message to politicians about what they ‘really’ think; but ultimately, the net effect is the same: they end up having to vote for the same party that they would have had to vote for tactically under FPTP.”
        Yes, although the net effect is improved by ensuring 50% support. Ultimately, everyone elected is still going to be from one of the local major parties until we change to PR.

        This is the reality of what is on offer – it is a significant upgrade to the existing FPTP system, but isn’t going to change much in terms of the political landscape in Westminster… at least in the short term.

      • “it is a significant upgrade to the existing FPTP system, but isn’t going to change much in terms of the political landscape in Westminster… at least in the short term.”

        Well, I’ll agree with you on the latter half of that sentence, at least! I accept that a discussion of ‘alternative alternative voting systems’, as I put it, is academic given that only AV is on offer. That doesn’t mean one should stop looking at better alternatives, to help people make up their minds whether they want to settle for something as woefully inadequate as AV, and to inform the discussion in the event of an inconclusive referendum or a vote in favour of AV, followed by the inevitable (?) disappointment at the outcome of the first AV election!

  3. Hi Anthony – you state ” AV is better because it not only allows everyone to decide the winner (not just the core supporters), it also allows smaller parties to see (and thus build up) their real support.”

    AV does NOT allow EVERYONE to decide the winner – up to 49% of voters may still be ignored. It allows MORE voters to decide who shall NOT get in. Also, we do not need to change the voting system to see how much support smaller parties get – opinion polls can do this.

    It is perfectly legitimate to state that AV is no more proportional than FPTP and, when a landslide occurs, it will make things more disproportional. This alone is a rational reason for democrats to vote against AV.

    David is essentially correct.

    • AC2011 I think that you are misunderstanding the point about everyone getting to take part. Under FPTP, anyone who votes for a candidate other than one of the top two (or occasionally three) is essentially not taking part in the election – their vote won’t influence who is actually elected. The result would be no different, or even potentially different, had they not voted.

      Under AV, those same people will (assuming no candidate gets 50% immediately) be able to transfer their vote to one of the proper contenders, and have a say who is actually elected to represent them. In this case, their vote CAN influence the result.

      Opinion polls don’t do this at all. Opinion polls tell us about voting intentions… which is simply a reinforcement of the result we actually get. It is very rare to see an opinion poll about who people would actually like to vote for given a free choice.

      Don’t forget that something like 20% of votes under AV are tactical – that’s 20% of people spending their vote on a party that isn’t their first preference. Hence we have no real idea of what the true landscape of political opinion looks like after an election under FPTP.

      I know that we have spoken before about this issue of ‘landslides’. It is 50/50. Sometimes AV will be more proportional than FPTP, sometimes less, but we are only ever talking about a redistribution of 30 seats or so. It is not a major consideration, and one of the main arguments used by the NO campaign is that AV is generally more proportional then FPTP and thus is more likely to lead to a coalition Government in the future.

  4. […] could bring about a Tory landslide, increase the North-South divide and bring back AV could bring about a Tory landslide, increase the North-South divide and bring back two-party poli… Very little joined-up thinking appears to have been done about what the real-world political […]

  5. As with a lot of things, Britain is a hybrid of European and American Systems. We have dual metric and imperial usage. First past the post like America, but people more informed and willing to vote for alternate parties outside the two main parties. This isn’t surprising given that we exported our system to the North American colonies. I’m not arguing that our system is superior, just in this historical context we haven’t had many coalition governments as is generally the rule across Continental Europe.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: