Meh to AV

There is a referendum on voting systems in the UK next month. Personally, I do not think that AV is all that much better than FPTP; both systems fail to be proportional. On the other hand, the only truly proportional system would be to have every person voting on every issue, something that would be ridiculously complicated. So, I am placing myself fairly in the “Meh to AV” campaign.

What I do care about, however, is politicians (and others) misleading or lying to the public for their own ends. The AV debate has included an impressive quantity of “fear, uncertainty and doubt” (‘FUD’) being spread around by both sides – something that is not particularly surprising that the arguments for both systems are particularly weak. Rather than focussing on making a case for either side, in this post I am going to investigate the main claims being made by both sides.

For an analysis of the voting systems themselves, I recommend the excellent post by Tim Gowers, Professor of Mathematics at Cambridge’s DPMMS, Fields Medallist and prize-winning writer/editor. Some of the arguments presented there are included below.

What is “First Past the Post” (FPtP)?

The voting system we currently use in the UK is called “First Past the Post”. Under this system, members of an electorate vote for one candidate and the candidate with the most votes wins. The name is supposedly an analogy to horse-racing (in which the first horse to pass the post wins) but, rather confusingly, in FPtP there is no actual post to reach. Instead, it is more like a long-jump competition, where the candidate who gets furthest wins.

What is the “Alternative Vote” (AV)?

Under the proposed voting system, AV, the electorate list as many of the candidates they wish in order of preference. There is no requirement to list more than one candidate, but if multiple candidates are listed, there must be a preference given (i.e. 1st, 2nd, 3rd…). All the ballot papers are examined and the highest preference vote is counted for each one (in the first round, either a 1st preference, or only preference). If one candidate has more than half the vote, that candidate wins. If no candidate reaches this number, the candidate with the lowest number of votes (A) is eliminated and the votes are recounted.

The second round is counted in exactly the same way; for each ballot paper, the vote goes to the highest preference candidate; the difference being that on some ballot papers (obviously not that many), the highest preference candidate will be the 2nd choice (the first no longer being a candidate). Any ballot papers where there is a highest preference for the eliminated candidate but no lower preference are removed and not counted in this round. Once the votes are all counted, again if a candidate has more than half the votes, they win. If not, the lowest candidate is eliminated and the votes are counted a third time. This repeats until one candidate passes the 50% mark (ironically, in AV there is a post, and the first candidate to pass it wins).

Eventually one candidate must get more than 50% of the vote; in the worst case, when only two candidates are left. If there are only two candidates, the one with the most votes will have more than half. This shows that AV and FPtP will return identical results when there are only two candidates, and so the supposedly ironic point often made that the AV referendum is being conducted under a FPtP system is somewhat meaningless – in either case the result would be the same.

The Arguments For AV

There is some sort of list of arguments for AV provided by the Yes to Fairer Votes campaign, on their What’s Wrong with FPTP page.

  • First Past the Post fails the basic test of fairness, by letting many MPs speak for the majority with support from the few.

As with many of these arguments, vague words such as “fairness” are used without being defined. While it is true that under FPtP a candidate can win with a rather small percentage of the vote (less than 32% in the Oldham East election in 2010) one candidate must win and it may be that a significant proportion of the other voters could have listed that candidate as their second preference. Without knowing the lower preferences of the voters, there is no way to know how the election would have turned out under AV.

  • The most popular candidate wins [under AV]?

Again, without defining popular, it is hard to work out whether or not this is true. Under FPtP, there is nothing to prevent the candidate the most people support winning (for example, my current MP was elected with nearly 60% of the vote). It could be argued that while FPtP may not elect the “most popular” candidate, AV may simply elect the “least unpopular”. In any constituency election there must be one winner, and unless one candidate has complete support, some people will be unhappy with the outcome.

  • A strong constituency link?

This argument suggests that “any claims to a constituency link under First Past the Post look laughable.” Again, this rather depends on how “constituency link” is defined. Under either system, there will be those who are unhappy with their elected MP. However, which ever voting system is used, the preferences of each constituent will be the same, so the “link” they feel with their MP will be unaffected by changing the system. The only difference would be if a “less unpopular” candidate was elected, some of the constituents might feel slightly more connected with their MP. Quantifying this difference would require an examination of how elections would have turned out under the different systems; if AV is implemented, this will hopefully be investigated at the next election.

  • Delivering strong government?

This is one of the most puzzling arguments put forward by the “Yes” campaign; that because FPtP returned a coalition government at the last election, it is bad. Obviously, AV can also return coalition governments, and FPtP can return strong, single governments (as it has done for most of the last 100 years here). In addition, there is an implicit assumption that “strong governments” are a good thing, which is debatable. Personally, I would suggest that the increased likelihood of “weaker” coalition governments is a good point of AV.

  • Kicking out unpopular governments?

This argument suggests that FPtP makes it hard for unpopular governments to be removed from power. However, no reasoning is given to suggest that AV makes this any easier. Weaknesses of FPtP are not enough alone to suggest moving to AV; it should be shown that AV is better. It is possible that the ease of removing an unpopular government will depend on the popularity of the government, not the voting system available.

  • A guardian against extremism?

Under this point, the “Yes” campaign suggests that AV will be less likely to result in extremist parties being elected – in particular, they mention the BNP. Interestingly, the “No” campaign argues that FPtP will be worse for the BNP as well. Logically at least one side must be wrong here. Personally, I am inclined to suggest that AV will make it harder for extremist parties, wherever they lie. Under FPtP, a candidate can be elected with the full support of 35% of the population, but with the other 65% being firmly opposed to them, as that majority vote was split. Under AV, there is a greater chance that, due to second and third preferences counting, the majority opinion will win over, but this is not guaranteed.

  • One person, one vote? With First Past the Post all votes are not created equal.

Again, this is a slightly misleading point, mainly due to a lack of definition for “equal”. Under FPtP, each voter gets one vote, and all votes are weighted equally. In theory, every vote is worth the same as every other. In theory. In practice, we end up with safe seats, where an impossibly large swing is needed to a change in party (such as the 35% swing needed in my constituency). This does not mean that my vote is counted fewer times than anyone else’s, merely that the chances of my vote making a difference are much smaller. Obviously, safe seats are a consequence of voter intentions, not any particular voting system (although under a truly proportional system, each vote would be reflected in the overall result), however, it is argued that switching to AV would significantly reduce the number of seats.

For the record, under AV, people may have a different number of votes, in some sense. Each voter will get up to one vote in each “round” of voting, provided at least one of their preferred candidates was still in that round. This means that the number of votes (or rather, the number of times each voter’s preference is counted) will vary between constituencies, and there is a chance that, if a voter has indicated preferences for few enough candidates, their ballot paper will not make it to the final result. Of course, under FPtP, the only votes that make a difference are the ones for the winning candidate (or even the number of extra votes the winning candidate received than the runner-up). Whether this system is fairer depends on the definition of fair and the evidence used.

In summary, while there are some arguments presented for AV, they do seem to be rather vague and poorly defined, and not particularly useful for basing a decision on.

The Arguments Against AV

The “No” campaign seems to be slightly better organised, and provides very clear arguments against AV on their Why Vote No page.

  • AV is costly – The change to AV will cost up to an additional £250 million.

This is, by far, the easiest and clearest point to pick up. This statement, while true, is impressively misleading, so much so that there have been legal threats over it. The actual cost of switching to AV is estimated by the No campaign themselves to be around £26 million in voter education, (a rather questionable figure, as discussed by Professor Gowers in the article linked above). The other costs they are including in that £250 million are £82 million on the referendum (which is, of course, a “cost” of the No campaign as much as the Yes campaign), and £130 million in electronic vote counting machines (which the government has no plans to acquire).

Switching to AV will require a level of education, but possibly not that much; the instructions are rather simple (“put a 1 in the box for your top candidate, a 2 for the second and so on until you do not care any more”). No new ballot papers will be required, nor will any electronic counting machines be needed. The cost is likely to be between £5 million and £30 million – which is considerably less than the estimated £80million cost of running an election – this is not a significant cost (particularly when one considers that in the opening days of the recent conflict in Libya, the US and UK fired over 100 missiles, each costing around £1 million).

  • AV is complex and unfair

Oh look, the word “fair” is creeping in again without a definition. The argument here seems to be that “under AV the candidate who comes second or third can actually be elected.” This is probably the most stupid argument either side has presented. The candidate who wins under AV is the candidate that comes first, by definition. The essential argument is that the candidate who wins under AV might have come second or third under FPtP; i.e. AV produces different results to FPtP. Well, yes. That is why we are thinking about switching to it. The argument for AV is that this different result better reflects the opinions of the voters. Whether or not it is is impossible to determine from the evidence presented.

This point also argues that “just three countries in the world” use AV. This is slightly misleading as there are other systems that use this vote, just few at a national level (for example, the Conservative and Labour parties both use something similar to AV for electing their leaders). This does not mean that every other country uses FPtP, though (although many do for some of their voting); there is a useful table of which country uses which voting system on Wikipedia.

  • AV is a politician’s fix – AV leads to more hung parliaments, backroom deals and broken promises

It is quite likely that AV will lead to more hung parliaments and coalition governments. Of course, this does not lessen the number of backroom deals, we get plenty of those under FPtP, it may just mean that the backroom deals will have to involve more or different parties. Another way of putting this, though, would be that AV leads to more compromises and less adversarial politics. Whether or not this is a good thing is left as an exercise for the reader. While the current coalition may have made some unpopular choices regarding spending (arguably necessary, given the previous few governments success at bankrupting the country) they have managed to avoid passing laws as terrible (in my view) as the previous governments, with each member of the coalition tempering the other.

The No campaign also has a list of advantages of FPtP on their Why our current system is better than AV page;

  • It creates strong governments

As our current, weak (but often effective) government shows, this is not the case. Australia (which has used AV for a very long time) has had plenty of majority governments and Canada (which uses FPtP) often has minority governments. The strength of a government is dependent on the opinions of the voters, not on the voting system. The No campaign also suggests that this leads to “accountable governments”, forgetting that a government, once formed, has near absolute power (until the next election) and, if backed by a strong majority in the House of Commons, is completely unaccountable. One merely has to look at the success of “votes of no confidence” to see how accountable governments are; with only one successful vote since 1925.

  • It’s fair

Hmm… fairness again.

Under this heading, the No campaign suggests that “supporters of fringe parties can end up having their vote counted several times” under AV. As discussed above, while this may be the case, every time one person in a constituency has their vote counted, everyone else will as well (provided they have listed enough preferences). If anything, fringe party voters will have their votes counted fewer times as it is more likely all their preferred candidates will be eliminated before the final round. It is not the case that different people get a different number of votes, more than between rounds some voters are forced to change their vote due to their preferred candidate being eliminated.

  • It’s simple to understand and easy to implement

This is true. However, it is also true for AV as well. As discussed above, both the voting and counting mechanisms for AV are pretty straightforward; there’s no need for special counting machines, or huge education campaigns. If the No campaign actually believes that the average UK citizen cannot cope with numbering candidates by preference, that is a sad indication of their faith in the people and our education system.

  • It excludes extremist parties

And the BNP get mentioned. With one data point for how AV is a bad thing. This was discussed at length earlier – it probably is not the case. One hopes that the reason the BNP and other extremist parties do not get elected is because they do not have sufficient public support, not due to the voting system. On the other hand, arguably the voting system should not be the only reason they do not get elected – if a significant percentage of the population supports such extremists, surely they should be represented?

  • It’s the most widely used system in the world

There was a wonderful retort to this given by Ian Hislop on the BBC show Have I Got News for You, where he pointed out that the majority of the world are starving, and thus it would be a good idea for us not to eat either. There are many other counters to this sort of argument (“if everyone else jumped off a cliff would you follow…”), but they are left as exercises for the reader.

So, that concludes the No arguments. Again, some are fundamentally flawed, others are badly defined or unclear. There is nothing particularly convincing.


As can be seen, there are significant flaws in the arguments put forward by both sides in the debate. One only needs to look at each side’s “myth-busting” pages to see that most of their arguments are easily countered. In short, if you want to see logical and rational arguments for or against AV, neither main campaign is worth looking at. It seems that even when there are easy arguments for and against something, politicians will still turn to what they know; hyperbole, dramatic language and FUD.

For completeness, there are a couple of other “No” sites that might be of some interest; the No to AV, Yes to PR campaign and the AV Referendum Site (both of which seem to present arguments AV contradicting some of the other arguments presented above – see comments below on the latter group), and there are some arguments on the Electoral Reform Society’s AV page. I would suggest that the latter is the only one of the three worth reading, but all should be taken with a heavy degree of scepticism.

Despite my “meh to AV” stance, I plan to vote for AV in the referendum. I have my reasons, but I was not swayed by any of the arguments given to me by the above sites.


11 Responses to “Meh to AV”

  1. AV2011 Says:

    Duke – thanks for the mention. We advance 5 arguments for a NO vote (from a perspective of PR) – none of which are contradictory. Did you have something else in mind? The five reasons we cite are:

    #1. AV is not proportional and can exaggerate landslide elections. In 1997, for example, it is probable that Blair would have had a majority of over 200 seats with AV. Landslides are the worst breaches of electoral justice, giving enormous political power to a party without the equivalent mandate from the electorate. Worse, if AV supporters are to be believed, such disproportional results would gain greater legitimacy under AV because most MPs would have been elected with a majority of voter support.

    #2. AV exaggerates the tendency of our current system to direct voters into a two-sided competition. Smaller parties, such as the Greens or UKIP, are no more likely to be elected than today. In fact, because the threshold is raised to 50%, it may actually be more difficult for these parties to be represented. For example, the Greens “illegitimately” won Brighton Pavilion with 31% of the vote in 2010; had the election be conducted under AV it is entirely possible the Greens would not have have reached 50%.

    #3. AV is likely to derail reform. A brutal dilemma appears unavoidable. If AV proves durable, another majoritarian and disproportional system will be entrenched for decades, as it has been in Australia. If it does not, then the next step is far more likely to be back to the familiarity of first-past-the-post (FPP). Where there’s a cultural tradition of FPP the political reflex is to gravitate back to it. Two examples: the western provinces of Canada re-implemented FPP after using AV for 30 years; a majority of Australian voters appear to prefer a return to FPP, according to some opinion polls.

    #4. AV gives minorities greater electoral leverage without democratic accountability. A minority party can barter with larger parties, urging its supporters to give the larger party their second preferences via its how-to-vote card in return for policy concessions. If the minority votes are necessary in electing the MP then the party can continue to lobby the MP under threat of withdrawal of support at the next election. Smaller parties should be heard but transparently and after receiving a mandate, which would happen under PR or even FPP.

    #5. AV fails a basic test of electoral consistency: increased support for a candidate should always increase the chances of that candidate winning. With AV, increased support can sometimes lead to a candidate losing! This happens when supporters switch their first preferences and change the order of elimination, which can be crucial in determining the winner under AV. We believe this consistency is important for a single-member electoral system. FPP, for all its many flaws, has this consistency; AV does not.

  2. Duke Says:

    My first issue with AV2011 is that it seems to be a “No” site, but doesn’t make it obvious anywhere – I don’t have a problem with groups presenting only one side of the argument, but it should be clear that that is the case. Anyways, on to the ‘five’ reasons – I will deal with them separately, in different comments:

    #1 – AV is not proportional. True. However, neither is it a fruit. If the vote was between AV or a banana, it not being a fruit might be relevant, but choosing between AV and FPtP, it is completely irrelevant as FPtP isn’t a fruit either. In much the same way, proportionality is irrelevant to the debate.

    With regard to landslides, obviously one must be careful when guessing how elections would have turned out under AV as we do not have the information. The data on the 1997 (and 1992) elections are examined by Professor Gowers, and it is clear that the results under FPtP are pretty insane (Lib Dems gained 1% of the vote, and doubled their seats, etc.).

    Of course, the question is not whether AV would have resulted in a landslide or not, but whether or not this would have reflected the will of the public. That is what voting is all about, after all. From what I recall of the public mood in the mid 90s, the Conservatives were really quite unpopular, and a landslide may well have reflected that.

    In any case, we should not take elections in isolation – it may be that under AV the ’92 election would have had a substantially different outcome, leading to the same ’97 result being less of a swing. While the ’97 might have been made “less representative” under AV, it may be that that election is the exception; particularly looking at the figures for ’92 and ’05. One of the advantages of switching to AV now would be that we could actually get some real data on how FPtP and AV compare.

    • AV2011 Says:

      Hi Duke – Any criterion of “will of the public” necessarily involves proportionality so how can the most disproportionate result of all (the AV landslide) reflect the “will of the public”? It is one thing to say in 1997 the country wanted a Labour government and quite another to say the country wanted a super-landslide majority (esp given it would have got around 43% of 1st prefs).

      The AV landslide, as I call it, is likely to happen when an unpopular government faces the nation under AV. And AV is likely to be worse than FPTP in this respect. Most academics agree on this point.

      If you are trying to rationally assess both systems you must be prepared first to accept in principle that AV may be worse than FPTP in some respects. This is one of them. While I can see that you are clearly rationalistic in your approach, this answer smacks of pure pro-majoritarian apologetic!

  3. Duke Says:

    #2 – On driving things to a two-party system; FPtP is designed for a 2-party system, it works quite well for that. Speaking as an adviser and governor of a smaller party, I am aware that the chances of smaller parties getting seats may well be lower under AV than FPtP; that is a consequence of having single-MP constituencies as much as anything else, though – and is why the smaller parties tend to do significantly better in the various ‘proportional’ elections (most notably, the European Parliamentary elections).

    As for the Greens winning a seat with 31% of the vote, the question here, again, is whether or not that was “fair” or “right” – I imagine that had that been the BNP winning a seat with 31%, this would have been given as a strong argument *for* AV, not against. Do we really want a voting system where a party can win with a mere 30% of the vote, even if they are a “nice” (or perhaps “harmless”) party like the Greens?

    Extremist parties may well be worse off (in terms of number of seats) under AV, but from an extremist party point of view, the advantage of AV is not winning seats, but getting more votes on paper; it may encourage individuals to select an “extreme” party as their first preference, safe in the knowledge that their vote won’t necessarily be entirely wasted (as it would be under FPtP). These extra votes give us a much-needed indicator of where our support is, and where we can focus our efforts for local, EP or even general elections.

    • AV2011 Says:

      You can do the latter with an opinion poll, Duke. As Dr Lundberg says, on our site, small parties winning seats is important for any the PR transition. Yes, the Greens deserve its seat in Brighton – they got 1 seat on 3% of the national vote. UKIP got no seats on 5%. FPTP is awful, AV could make it worse.

      Here is the CLASSIC example of what is at stake. Duke wants the electoral system to better reflect the views of voters in, say, Brighton than the nation as a whole. Surely, it is better from the national perspective that Greens take Brighton on 31% rather than no representation at all?

  4. Duke Says:

    #3 – Derailing reform. Now this is a tricky one and very hard to judge on. On the one hand, if the population votes “No” this will be used to deflect any future attempts at electoral reform. It has taken over 100 years to go from a Committee recommending switching to AV for elections with more than two candidates to having a referendum on it – if the referendum fails, it could be another century before we get another chance to change. On the other hand, if the vote succeeds, we have all the issues raised by AV2011; that people will not want further reform any time soon.

    Personally, I’m not convinced by this argument. Keeping FPtP is a good way of maintaining the status quo. Past election results suggest that, on the whole, FPtP is what has kept the two main parties in power – parties elected under (and perhaps, because of) FPtP are unwilling to move away from it (hence the Conservatives campaigning so hard against it). The only reason we have even had the option of debating voting reform is the unusual result of the last election, and the Liberal Democrats being able to force a vote. It may be a very long time before such an opportunity arises again – we should make the most of the one we have now. At least a “Yes” vote would indicate that the general public are happy and support change.

    • AV2011 Says:

      I disagree, Duke – but as you suggest this is more of a value judgement.

      IF AV fails, I expect to see movement for PR in the House of Lords or locals. Cameron wants the boundaries changed – he will stay in power until this is done. The LibDems can bring the Government to an end so it is possible that PR for HOL or Locals will result in the aftermath of a NO vote. And PR will REALLY show the level of support for minor parties and then voters will compare PR (HOL, locals or both) with general elections and will – I suspect – demand PR there too, after time.

      A YES vote for AV would kill this, in my view. Any change to the electoral system – whether local, HOL or House of Commons – must be made on the principle that it increases proportionality. That is our position.

  5. Duke Says:

    #4 – Undemocratic leverage. Again, it may be the case that AV will give minor parties some small amount of power over the larger ones, but I do not see this as a bad thing. One only has to look at the work of the coalition to see that a large party can be tempered by a smaller party with reasonable success. Is it so wrong that a minor party’s views can shape policy, or do we want to stick to a 2-party system? Having seen the difference between the very adversarial politics we have here, and the more “compromise”-based politics of other groups (such as the EP), I’m not convinced that our system is the better one.

    At the moment, we already have a system where the larger parties will expand their policies in an attempt to woo supporters of minor parties, and the minor parties will recommend which larger one to vote for (particularly in constituencies they are not contesting; here I can speak from experience). It isn’t particularly undemocratic – if anything, the opposite, as it means that the views of more people will be reflected in the policies of government (as the large parties will be trying to choose their policies to maximise support; gaining one or two extra votes at one extreme may well cost a dozen or more at the other).

    As for transparency; in any system you will have groups (parties, trade unions, religions) telling people which way to vote, and bartering with candidates, all without being open – it is unfortunate, but a consequence of democracy. I don’t see how AV makes it significantly worse.

    • AV2011 Says:

      IF AV decreases transparency, this cannot be good for democracy, in my view.

      I agree that there is lack of transparency now – but I do think AV could make it worse. AV could be called the “coalition vote” because it brings electoral pacts to every constituency prior to a vote being cast. The extent of the deal and the pressure exerted post-election we may not know.

  6. Duke Says:

    #5 – Increased support => increased chance of winning. I think this argument falls flat on its face when one looks at FPtP and the overall votes. Between ’92 and ’97, the Liberal Democrat vote fell by 1%, but their seats went from 20 to 46. Again, between ’05 and ’10, Lib Dem vote percentage went up 1%, but their seats went down from 62 to 57. Under FPtP, all sorts of odd things can happen with vote percentages and number of seats – whether AV is better at this remains to be seen.

    Obviously, the argument presented concerns individual constituencies, rather than the national situation, so this is slightly different. I would quite like to see a possible simulation of this (i.e. theoretical numbers to show how it works) but for now, I think it is worth noting that this sort of “tactical” voting may or may not be worse than the tactical voting we get under FPtP. All that matters, and that should matter, is the overall result and how that reflects the wishes of the electorate. Whether or not AV does this better than FPtP is a very complicated question – why I have tried to show on this page is that the arguments given by both sides do not really go any way to answering it.

    • AV2011 Says:

      Duke – the argument, as you well know, concerns the constituency level – after all this is what AV is all about: the constituency result.

      You are quite wrong to suggest this is a form a tactical voting. People honestly change their votes to a candidate and the order of elimination is changed and that candidate (who received more preferences) is defeated when without the additional support would have one. We present an example on our website.

      So, I agree that the official campaigns provide no information for someone to make an informed decision. And I also acknowledge that there IS a case for AV – it is just that – all things considered – it is not quite as strong as the one against it, in my view. This is why AV2011 was set up (it was originally intended as an impartial site but we could not sustain the position).

      However, as a rationalist, I am keen to understand where we contradict ourselves as, of course, we would rectify this immediately.

      “All that matters, and that should matter, is the overall result and how that reflects the wishes of the electorate.”

      I agree totdally. This is why I believe in proportional representation. All the evidence strongly suggests that AV is no more proportional and in one circumstance, the landslide, it even more flagrantly distorts the “will of the people”. This alone should make anyone suspect that AV is the right step. I’m surprised you do, Duke, but you may have other reasons.

      Let me finish by praising your blog, though. I may disagree with its conclusions, but I can at least say you decision to vote YES is based on an informed decision. You appear to realize that AV is not going to end safe seats, not going to make MPs work harder, and is not going to be any fairer (at the national level or even in many constituencies).

Leave a Reply

Please log in using one of these methods to post your comment: Logo

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

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s