My Definitive View of the AV Debate

I’ve written two previous pieces on particular aspects of the Alternative Vote (AV). Now that the referendum and date of 5th May are confirmed, I’ve put together a more comprehensive view of the debate and its current arguments. If anything I haven’t considered here comes up, I will try to address it. If you want to skip to the core argument of why AV is a fundamentally more democratic system, regardless of any other considerations, go here.

Why think about it at all?
On the face of it it may seem that there is little at stake in the coming referendum on the voting system for the UK House of Commons. I have to say that this was my first thought on the matter when the plan for this referendum was announced as part of the Conservative/Lib Dem coalition agreement. In my case this is because I believe in a consensus democracy in which national policy decisions flow as much as possible from the equal input of all citizens in the decision-making process. I believe in such a system for two reasons – firstly that in the long-run it is likely to end up making the best decisions for everyone, and secondly because it is the system most likely to avoid sections of the populace, who feel unable to get a fair hearing, attempting to get attention in other more unpleasant ways. I would contrast this sort of democracy to ‘elected dictatorships’ where a government is chosen at intervals with the power to more or less ignore the wishes of the electorate (or at any rate those that are not its confirmed supporters) for its period of office.

In a system dominated by political parties, consensus democracy requires that the representation in Parliament matches as well as is practicable the proportion of votes those parties have received. In other words the electoral system should be a proportional one. The Alternative Vote as it is envisaged here is unlikely to significantly affect the proportionality of the electoral system. Under some scenarios, it may be more so, under others less. So for me it is in a sense ‘a miserable little compromise’ as Nick Clegg has notoriously described it. For others the issue may be that they cannot really see what relevance a technical change in the electoral system can make to their lives. It’s not something they’ve thought about, or really think worth devoting much mental energy to even now, when they have the chance to vote on it.

A small change maybe – but a significant one

In the light of the reality of this referendum, facing voters with simple Yes/No choice for a change from the current First Past the Post (FPTP) system for elections to the UK Parliament to an Alternative Vote (AV) system, I’ve had a rethink. This change in the electoral system is quite a small change – that is true – but should not be dismissed for that reason. Small changes may be better than big changes. They are more likely to be accepted and to stick. It is a matter of logic that all continuous change begins with a small change! So the issue is not whether it is a small change, but whether it is a significant one in relation to its likely cost.

The benefits of AV depend on what exactly democracy means to us. We don’t have a system, for various reasons, in which we all have a say in the actual decisions taken by our Parliament and the government formed from its members. We have a representative system, in which individuals are selected to represent us in a forum where they put their electorate’s point of view and listen to the points of view of other representatives before coming to collective decisions. These decisions are hopefully thought by the majority of the forum’s members to represent the best interests of the nation’s citizens. If we believe in democracy, as opposed to some form of paternalist meritocracy where we select MPs to make decisions not according to our views of what we think best for ourselves, but according to their views of what is best for us, then how the voting system converts our individual choices into the choice of candidate is very important. It is in fact particularly important when we have a non-proportional system, since the likelihood is low that the party make-up of Parliament is going to be an exact match with the overall make-up of voters’ party preferences.

In a non-proportional election, each constituency is voting for an individual voice in Parliament rather than an element in such a party-voter alignment. We should not therefore see a Westminster general election primarily as a vote for an administration, but as 600+ local elections for voices to speak for us in the formation of an administration and in the policies that administration adopts. Our representative should be judged on both of these counts; an MP who is seen to play a part in obstructing the formation of an effective government should deserve approbation as much as one that advocates policies against his or her statements to their electorate. If we and they take this point seriously the problem of ‘coalition haggling’ need not greatly concern us.

Why is AV more democratic than FPTP?
Consider an election where the primary issue is a proposal to build two supermarkets on two local parks – North Park and South Park. Forty per cent of the voters have gardens and don’t use either of the parks – they want to see both supermarkets built. Thirty-one per cent of the voters have no gardens and use North Park – they would ideally like to see a supermarket built on South Park, but would prefer one supermarket built on North Park rather than building on both parks. The remaining twenty-nine per cent of voters also have no gardens and use South Park, but would prefer one supermarket built on South Park rather than building on both parks.

There are three candidates in the election. Candidate A supports the building of both supermarkets. Candidate B supports the building of a supermarket on South Park. Candidate C supports the building of a supermarket on North Park. In the election those without gardens all vote for candidate A; the North Park users vote for candidate B and the South Park users vote for candidate C.

Under FPTP, candidate A is elected and both supermarkets are built against the wishes of 60% of the electorate who would have preferred that one supermarket was built on either of the two parks, leaving them with some open space to use. Under an Alternative Vote election, North Park users would place a second preference for candidate C and South Park users a second preference for candidate B. Candidate A would win 40% of first preference votes, not enough to immediately be declared the winner. Candidate B would win 31% of first preferences, and candidate C 29% of first preferences. Candidate C would be eliminated from the count and his or her second preference votes would be re-distributed. Since these would all go to candidate B, candidate B now wins the election with 60% of the votes. One supermarket is built on South Park. Everyone has access to a supermarket and to open space. Crucially from the point of view of democracy, there is no single alternative outcome to this election that a majority would have preferred.

Now of course, most elections of MPs are not decided by single issues, and we might surmise that given information about the preferences of the voters and a willingness to negotiate beforehand, the two groups of park users and their candidates might have come to an agreement with one of the pro-park candidates withdrawing or an agreement on tactical voting taking place. But the example illustrates the possibility for FPTP to elect one candidate when the majority of voters would have preferred to be represented by one of the other candidates. Whatever else might be said about such an outcome it seems to conflict with some basic idea of democracy. So we should be aware that when, for example, David Cameron objects to AV on the grounds that ‘it could mean that those who are courageous and brave and may not believe in or say things that everyone agrees with are pushed out of politics and those who are boring and the least controversial limping to victory’ he has an argument, but if not actually an argument against democracy as a method of choosing political representatives, it’s pretty close to one.

This recognition that the election of a candidate who is not the preference of the majority is pretty unacceptable explains why important elections such as that for the Conservative Party leadership are frequently organised as run-off elections in which there are repeated rounds of voting with elimination of the least popular candidate in each round until one candidate commands a majority of the votes. The Alternative Vote is simply a streamlined version of this sort of election in which it is possible to specify on one ballot how we would vote in each possible round. If we have expressed a preference for all of the candidates, then whoever are the top two candidates in the final round (in this case of counting rather than voting) our preference between those two candidates will be counted.

The trade-off between a run-off election and an AV election is that the former is more time-consuming (at least for the voters) and the latter does not allow the changing of preferences between rounds. In the absence of new information the latter doesn’t seem much of a loss, and you can’t do this under FPTP either.

Some claims about AV

1. It is less proportional than FPTP.
Given that neither FPTP or AV is devised to be proportional, arguing this either way seems a bit pointless. Under some scenarios one turns out to be more proportional, under others the other one does. In practice, it doesn’t seem to make a lot of difference. In any case, if we have a specifically non-proportional system it becomes even more important that individual MPs are as representative as possible of their constituents.

2. It leads to more hung parliaments than FPTP.
Again, since AV is not a proportional system it depends on which scenarios we look at. Comparing the UK with Australia, there isn’t much evidence for this in practice. Why in any case are hung parliaments necessarily a worse phenomenon than large Parliamentary majorities based on a minority of the vote? After 5 days of negotiations in May last year we managed to avoid a government few Labour or Liberal Democrat voters would have preferred (a pure Conservative administration) and one no Conservative voter would have preferred (a Labour/Liberal/Others coalition). What is important is the attitude of MPs to working together in the interests of the majority of their constituents. Ensuring that they were the true preference of a majority surely enhances this.

3. It gives ‘more votes’ to supporters of small parties such as the British National Party (BNP), Greens or Monster Raving Loony Party (MRLP).
Surprisingly, this is a common argument with those who really ought to know better – including two beneficiaries of the Conservative Party’s run-off leadership elections, William Hague and David Cameron. If AV gives more votes to some voters than others then the Conservative leadership elections gave more votes to those MPs whose favourites were eliminated early in the ballot rounds than to those who supported Hague and Cameron from the beginning. In fact this is nonsense in both cases. Under a run-off election everyone has a vote in each round. If a candidate is eliminated in an early round those that voted for that candidate can vote for a different candidate in the next round. In effect this is no different from the rounds of counting in AV in which the votes for an eliminated candidate are transferred to another candidate. If your first preference candidate is not eliminated your votes are still counted in every round, but instead of being transferred remain with that candidate.

A related argument is the one that supporters of small parties can ‘determine the outcome’ of an AV election. In the sense that their transferred votes may decide between the top two candidates this is true. But unless a candidate has garnered a fair number of first preferences to start with, they are most unlikely to win. In any case this is another anti-democratic argument. Many may abhor the views of the BNP, or think the Greens or MRLP daft, but democracy insists their supporters have as much right to a telling vote as anyone else.

4. It’s ‘not fair’ that the candidate getting the most first preference votes is not the winner.
This is the incumbent politician’s special pleading to retain the system that allowed him/her to get elected. ‘Fairness’ to the candidates is irrelevant. Fairness to the voters is what matters. Is it fair that a candidate can be selected when a majority of voters would prefer one of the other candidates? AV is fairer because it uses more information to determine the candidate most representative of the voters’ choices.

5. A change to AV would make an early change to a full proportional system less likely.
Maybe. But it’s difficult to see why. Neither a future Labour nor Conservative government winning alone are likely to introduce PR. The Liberal Democrats are unlikely to form a government on their own in the foreseeable future, so the most likely scenario for reform is going to be piecemeal under successive coalitions involving the Liberal Democrats. Under those circumstances a referendum on progression to PR following a successful AV referendum, perhaps in two parliaments’ time, seems more likely than an attempt at a one-step reform following a failed AV referendum.

6. It’s very expensive.
The ‘No’ campaign have come up with a figure of £250 million as the cost of introducing AV, which seems to include the £80 million cost of the referendum itself (not much to be done about that now!) as well as £130 million for electronic counting machines that no-one has confirmed will be required (Australia has managed AV elections for 90 years without them). It’s also worth noting that were electronic counting machines to be introduced their cost would replace the cost of manual counting, not be in addition to it. Excluding the latter leaves a cost of adopting AV of around £40 million for voter education. That’s a one-off payment of less than £1 per UK citizen. If the new system genuinely improves political representation that doesn’t seem a huge price to pay.

7. It’s complicated.
AV is a bit more complicated than FPTP, both for the voter and in working out the results. This is inevitable because AV uses more information about voter preferences to choose an MP. But for the voter who is only interested in continuing to express a single vote for one candidate all they have to know is that they must put a 1 for that candidate instead of an X. If their candidate is one of the top two then their vote counts, if not then it is discarded. This is in effect no different to what would happen to that vote under FPTP. To vote more effectively, candidates just need to be numbered in order of preference. Is this too complicated for the vast majority of voters? Given the frequency with which numbered preferences are used in voting in other democratic organisations, market research and so on, it shouldn’t be. If Australians can manage it, shouldn’t we be able to? The procedure for counting the votes after the election is actually quite straightforward (pdf), if a bit more time consuming. How to explain how the result is arrived at? There are successive rounds of counting preferences in which the bottom candidate is knocked out each time, leaving two left at the end. The one of those that wins has more voters preferring that candidate to the other than vice-versa. Sounds fair, doesn’t it?

2 thoughts on “My Definitive View of the AV Debate”

  1. Diarmid, I fear your example vote is a biased one, and therefore doesn’t make the definitive rational case you’re aiming for.

    This is because you have constructed things so that the outcome the system delivers is a good-sounding one.

    It is hard to sympathise with the option A voters, because they have gardens and what they want – 2 supermarkets – deprives others of open space.

    What if you ran the same maths and voting intentions with a different option A? Option A voters have no gardens either. Unlike B and C, they also have no cars and cannot drive to a supermarket across town. They need a local supermarket. Also they are poor, so they need the competition two supermarkets would provide to drive down prices. B and C voters are better off and can afford local specialist shops and can drive when they need a supermarket. They differ, as you have it, in that they want the supermarket in different parks.

    In this scenario, 40% of the voters need local supermarkets. The better off 60% would prefer one supermarket in their respectively preferred (differing) locations.

    With these options, it seems fairer/more democratic if the “simple majority” wins (ie Candidate A). At least, it seems unfair that relatively comfortable B and C voters can “club together” to prevent less-comfortable A voters getting something they need.

    I’m deliberately using emotive language here – because I think first preferences can be VERY strongly-held – and AV has no way of taking this into account.

    Am I being unfair?

  2. Hi Dominic,

    Sorry I didn’t spot this comment before – it seems to have got lost in the spam. I guess it’s a bit academic now, but none the less I think it’s an important point.

    Of course you are right that AV may produce outcomes that don’t seem fair once all criteria are taken into account, but this is true of any voting system. It is why voting is not all there is to democracy. Consider the possibility that 99 people would be made ‘better off’ by the death of 1 (innocent) person. If it was just a matter of numbers then we wouldn’t think the death of that person was wrong.

    In comparing voting systems purely as voting systems we have to assume that all outcomes are fair on other grounds. This is what economists refer to as the ‘ceteris paribus’ condition. I tried to apply this condition to my example. What you describe sounds unfair, but it would be even more unfair if the A voters were actually the better off ones and they won against a split B and C vote.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.