Praesidium

Saturday, April 30, 2011

FPTP Double Counts Some Votes

In my last piece, I argued that what matters when we’re thinking about a fair and democratic electoral system is not how many votes each person casts, but how many votes they have counted. I argued that, on this understanding, AV respects the maxim ‘one person, one vote.’ I didn’t criticize FPTP as such, allowing that it also satisfied this requirement. In this post, however, I want to show that FPTP actually double counts some votes and is therefore unfair.

For simplicity, I’m going to focus on an example where we have only three candidates, who I will call A, B, and C, and twelve voters. Let’s assume that (first preference) votes are allocated as follows:

A: 5 votes
B: 4 votes
C: 3 votes

Under FPTP, A is declared the winner, because A has more votes than anyone else. But what does this mean? Does it mean that A has a majority? The answer, evidently, is not in the strict sense, since a majority means ‘more than half’ and A only has five of the twelve votes, which is less than half.

It is sometimes said that A has a ‘relative majority.’ That is, A has a majority over B (five to four) and also a majority over C (five to three). Note, however, that when we compare A only with B, we ignore those who voted for C. In saying A has a majority of five to four (over B), we are only counting nine votes. Those three votes for C are excluded here (though, of course, they come into play when A’s votes are compared to C’s – there it is B’s votes that are ignored).

Saying that A has a majority – in this sense – over each of B and C taken separately does not show that A has a majority over the two of them together. It might be that all those who voted B would also prefer C to A (i.e. their full preferences were B > C > A) and similarly those who voted C prefer B to A (i.e. C > B > A). In this case, a majority (seven people) prefer B to A and a majority (seven people) prefer C to A.

How then does FPTP declare A the winner? The problem, it seems, is that FPTP either ignores some people’s votes – as votes for C are ignored when simply comparing A and B – or it double counts some. Think again of the comparison between A and B. We might imagine the votes ‘cancelling out’ until, after four votes on each side have been cancelled, A wins because there is still a remainder (one) in favour of A. To think that A also defeats C, however, we have to allow that each vote for A can cancel not only one of B’s votes but also one of C’s.

Someone who voted B doesn’t, I think, have good reason to accept defeat when only first preferences are known. They might acknowledge that B trails A by five to four at this stage, but they can rightly reply that they are not yet shown to be in a minority when we do not know how the other three (C voters) would rank A and B.

Now, it might be that C voters would prefer A to B. If this is the case, then AV would declare A to be the winner. This is a case, however, where AV and FPTP would agree. Most cases would probably be like this, so there’s little need to argue between them. If A is picked by both FPTP and AV then, uncontroversially, A should be the winner. (Remember, what we’re concerned with is who should win.)

To bring out the difference between FPTP and AV, we need a case where they come apart. Therefore let us assume that all C voters prefer B to A. In this case, B can reasonably complain about any electoral rule (such as FPTP) that awards the election to A. Here B is preferred to A by a majority of the electorate (seven of twelve), so surely the idea of equal votes and majority rule tells us that B should win.

The only way we can say that A ought to win is if we illegitimately infer that because A has a majority over each of B and C taken separately (five to four and five to three, respectively) then A also has a majority over the two of them together – but this is not so, since in this example we have assumed that a majority would actually prefer either B or C to A in a two-horse race (seven to five in either case). A only wins if votes for A are counted twice, first as defeating votes for B and then again as defeating votes for C.

As I argued last time, there’s nothing unfair about counting second preferences. Under AV, each person has only one vote counted (the three C supporters have their second preference for B counted instead of a vote for C, only once C has been eliminated). This is in stark contrast to FPTP, where as I’ve just argued A only has a majority if either some votes (those for B or C) are ignored or if those for A are counted twice.

Democracy is about responding to the people’s preferences, so surely it’s more democratic to have full information about people’s preferences. Imagine that A and B had tied in the first round (say, four against four, with three for C and one abstention). How could this be resolved?

One possibility would be some random device – such as the drawing of straws or flip of a coin. That procedure is actually deployed in the case of at least some tied elections. An alternative, however, is to break the tie by appealing to voters’ preferences – we already know those of the eight people who voted for either A or B, but we could ask either the one abstainer and/or the three who had voted for C which, out of A and B, they would have voted for had they had to. This is what AV does and surely, since it responds to people’s preferences, that is a more democratic way to break the tie between A and B.

But, if the second preferences of C voters (between A and B) are the most democratic way to break a tie between A and B, why shouldn’t they also come into a close contest? Once again, supporters of B have no reason to accept five to four as a defeat, if three people’s preferences between A and B have not yet been considered.

If C voters prefer A then, fair enough, B is in the minority (eight to four). But that is never shown under FPTP. AV will establish, once and for all, whether it is A or B that has a majority. If A, then it agrees with FPTP. But if B is preferred to A by seven of the twelve voters, then surely it’s more in keeping with democracy, majority rule, and equal votes to declare B the winner.

Labels: , , ,

Thursday, April 28, 2011

AV and 'Multiple Votes'

Continuing my series (1, 2) in the run-up to the electoral reform referendum on May 5th, this time I want to look at another claim on the No2AV leaflet I have in front of me: “One person should have one vote. That’s fair.” (See also here.) Again, I find this flawed on several counts.

Firstly, one natural reading of this is simply that everyone should have at least one vote, and there is obviously no conflict between that and AV. Presumably, however, their intended meaning is that everyone should have exactly (or perhaps at most) one vote. They claim that “The AV system will mean the end to equal votes,” presumably meaning that expressing a second or third preference amounts to having two or three votes.

‘One person, one vote’ is certainly a rhetorically effective slogan, but again it needs to be analysed more carefully if we are to understand its meaning and implications. The appeal of the slogan is that it expresses political equality, but there’s no particular reason why it should be ‘one person, one vote’ rather than, say, ‘one person, five votes.’

Indeed, in the Scottish elections (to be held on the same day as the referendum) each person will cast TWO votes. I don’t mean a vote for their MSP and a vote in the referendum. Rather, the Scottish parliament consists of one lot of MSPs elected on a constituency basis with a second lot elected by PR, so each voter will get two ballot papers and cast one vote on each. I don’t hear anyone protesting that this is undemocratic.

Of course, the No2AV campaigner could say that there’s a difference between having two votes on two different ballot papers and having two votes on a single ballot paper. This merely highlights another ambiguity in the slogan though. We don’t think that each person should only have one vote in their lifetime. We wouldn’t disenfranchise someone now because they already voted five years ago. Hence we don’t literally enforce only one vote, but ‘one vote per ____’ where that blank needs to be filled in to specify how often each person should get their one vote.

The No2AV campaigners would have us believe that this blank should be filled in with per election, but I haven’t heard any good argument as to why we should favour that over per *round* of vote counting, as happens under AV.

As I hope I made clear in my original explanation of AV, no one exerts more influence than anyone else. The point of people expressing second and third preferences is that these ‘alternative votes’ are counted instead of their original (first preference) vote, if their first preference is eliminated. The principle is essentially the same as in a Single Transferable Vote (STV) system which, as the name implies, gives each person only one vote – the difference being that they are allowed to transfer it. (AV is in fact STV when electing only a single person per constituency.)

We could, in fact, replicate these results by having multiple elections. That is, if no one of five candidates standing wins an overall majority, we might eliminate the fifth placed candidate and hold another election with only four candidates standing, and so on until one wins a majority. Then it would be quite clear, I take it, that each person only has one vote per election. The point of AV though is that it saves the need for multiple rounds of elections, by having voters specify all of their preferences to begin with, so we can calculate what would have happened had there been a four-way election instead of a five-way one. (This also of course means that turnout is held constant and prevents tactical vote-switching between rounds.)

‘One person, one vote’ doesn’t tell us whether people should have one vote per election or one vote per round of counting. That’s the problem with slogans – since they’re not arguments, they can’t guide us in cases of ambiguity.

Here’s another example of ambiguity where simply asserting ‘one person, one vote’ doesn’t help. It might be objected that our current system fails to respect ‘one person, one vote’ because children are persons but aren’t given the vote. We might propose to remedy this by giving parents votes to cast on behalf of their children. Is this in keeping with the requirements of ‘one person, one vote’ because it gives children votes or contrary to it because it gives parents more than one vote? Either interpretation is possible, so the slogan alone is no help here. We need to get beyond the simple slogan and explore the reasons behind it, which might help to adjudicate between these two possibilities.

The same is true when it comes to AV or FPTP. Each system reflects a different understanding of ‘one person, one vote.’ FPTP focuses on how many votes one casts, whereas AV focuses on how many votes one has counted. The essence of AV is that everyone’s vote is counted once in each round of voting. Expressing a second preference does not give one more influence overall, it merely means that one can still have influence once one’s first choice has been eliminated. That second preference is counted (if at all) instead of one’s first preference, not as well.

Now when I’ve pointed this out to No campaigners, some have acknowledged that each person only has one vote counted, but still held that it’s objectionable for people to cast more than one vote, even if only one is counted. But this doesn’t seem plausible to me. What’s important is surely how many votes you have counted, not how many you cast.

Suppose we allow everyone to cast one vote, but then a significant section of those votes (either selected at random or perhaps those of some particular group) are simply discarded without being counted. Surely that wouldn’t be democratic. The ability to write marks on bits of paper isn’t what’s at issue here; it’s influencing the political process that matters. Thus, it matters that your vote is counted, not merely that it is cast.

Someone might say that’s unfair of me. We needn’t assume that it’s either voting once or counting once. It might be that both of these matter. Thus they could say that everyone has to count once (and only once) but also that people should only cast one vote. AV respects the first requirement, but not the second.

But I don’t see why we should think that each person casting one vote ought to matter, independently of each person counting once. Suppose we had a reform that allowed each person to cast as many votes as they liked, though only their first would be counted. (Leave aside questions about how this might be enforced.) I wouldn’t see anything wrong with that. You could go down to the polling station and mark as many ballots as you like, but it wouldn’t give you any more influence than me. In other words, it doesn’t seem to matter how many votes you cast, provided only one is counted.

For the sake of completeness, we can also imagine another example, in which everyone casts one vote but some people’s votes are counted twice. That, I take it, would be objectionable and undemocratic. So everyone casting one (and only one) vote doesn’t ensure equality; we have to make sure that each person is only counted once.

Let’s recap. Each person casting one vote isn’t sufficient for equality, if some of those votes are counted twice or not counted at all. Nor is it necessary, since we could allow people to vote more often, but only count their first vote. Thus, it seems that what matters is that each person’s vote is counted once and not more. It does not matter whether people cast more than one vote, provided each person only has one vote counted.

So, even if we interpret AV as people casting more than one vote, albeit only conditionally, it doesn’t violate the supposedly democratic requirement of ‘one person, one vote.’ One person should have one vote *counted*. That’s perfectly consistent with AV.

Note I’m not saying that ‘one person, one vote’ is inconsistent with FPTP. We might have an argument as to whether ‘wasted votes’ for minority candidates are really counted. But my aim here is merely to counter an argument against AV, not to offer positive argument for it. My claim is merely that ‘one person, one vote’ – when properly interpreted – gives us no reason to favour either system over the other.

Moreover, as I pointed out at the outset, this slogan isn’t actually particularly compelling to begin with; the Scottish elections illustrating that one person might have more than one vote in both senses.

The aim of AV is to ensure that the candidate elected is preferred to his or her rivals by a majority of voters. What’s undemocratic about that?

Labels: , , ,

Tuesday, April 26, 2011

Under FPTP The Loser Wins!

I’ve already written one piece in which I set out my position on the coming AV referendum, but I’m so annoyed with some of the bad arguments that I hear coming from the No2AV side that I felt the need to tackle them individually. I’ll get to ‘one person, one vote’ later but, for now, I’ll focus on the claim that ‘under AV the loser can win.’ This is often presented through the medium of sports analogy, for instance the leaflet I have in front of me shows four runners crossing the finishing line of a race – one clearly in front, but with another marked as ‘the winner under AV.’ This isn’t really an argument as such, but rhetorically I think it’s very effective, so it’s worth showing what is wrong with this.

Firstly, AV will never elect the one who comes last on first preferences, because they will be first eliminated. In a two horse race, FPTP and AV are equivalent, so campaign posters that focus on such cases, such as a boxing match, are misleading. Nor can the one who comes ‘last’ in a race win.

Secondly, the sporting analogy is somewhat flawed. In many sports, the winner is determined not in a single contest, but over a series of contests. The Olympic gold medallist, for instance, is the person fastest in the final, with the qualifying rounds deemed irrelevant. Like in AV, it is not the initial results that matter, but the final standing.

As I pointed out in my earlier post, the name First Past The Post is actually importantly misleading, because (unlike in a race) there is no fixed winning post. Therefore showing runners crossing a fixed finishing line is disingenuous. AV does have a clearly defined winning line, namely getting a majority of votes remaining at any round of counting. Consequently, from the perspective of AV the winner is indeed the first to cross that line and the problem with FPTP is that it takes a snapshot of the race at some arbitrary point, such as after the first lap, and concludes that whoever is leading then is the winner, though they have not yet reached any previously specified finishing line.

The more general problem is that the slogan ‘the loser wins,’ common in the No2AV campaign, presupposes that who is the loser and who is the winner has already been determined. The supporters of FPTP are implicitly appealing to FPTP criteria (who gets the most first preference votes) to determine a winner and loser and then asserting that AV picks the wrong winner because AV doesn’t agree with FPTP. But this is circular reasoning that assumes precisely what is at issue.

FPTP and AV sometimes disagree, that’s the point of electoral change. Supporters of AV could similarly argue that under FPTP, the loser, according to AV standards, sometimes wins. That’s what shows they are different systems. The disagreement is not about who wins, but who should win. Proponents of AV think that, where the two systems disagree, the candidate picked out by AV is the one with the better claim to win.

Again, a sporting analogy is useful, provided that it is properly handled. There isn’t usually much disagreement about who wins a football match – the team that scores more goals than the other. There is, however, course some debate about how best to settle ties – e.g. whether to play extra time, replays, penalties, etc. This is more analogous to the voting referendum. It may be that the team that wins after extra time would not have won had the match gone straight to penalties. We cannot, however, reject either of these methods of determining a winner simply on grounds that they disagree – or we would, on grounds of consistency, be forced to reject both.

One problem of course is that we do not know who would have won penalties if they did not in fact take place. (Similarly, we do not know who would have won under AV if voters were never asked for their second preferences.) Nonetheless, football league tables can also serve as a useful illustration. Some leagues use two points for a win and others three points for a win, though these can produce different results.

Let’s look at the current English Premier League. At the moment (26/04/11), Manchester United would lead Chelsea, whether two or three points for a win. But that need not be the case. Imagine that, at the end of the season, we see the following results:

1sth: Manchester United (W 24; D 6; L 8) – 80 points
2nd: Chelsea (W 22; D 13; L 3) – 79 points

Interestingly, United win the league in this (hypothetical) example, despite losing more games than Chelsea. Why? Because they’ve also won more, having fewer draws. With three points for a win, it’s better to have a win and a loss than two draws. What if we used two points for a win? Then Chelsea would come out in front:

1st: Chelsea (22 x 2 + 13) = 57 points
2nd: Manchester United (24 x 2 + 6) = 54 points

(For a real world example, see here.)

If we calculate the table one way, then one team wins, but calculate it the other way and the results are different. That’s because we’re using a different means to determine the results. It’s no good simply asserting that ‘the loser (according to one method) wins (when a different method is followed)’ – again, that’s the whole point of a different method.

The debate we need to have is over who should win. In the football case, we need to decide whether a win and a defeat should count for more, less, or the same as two draws. In the political case, we need to consider whether it’s better to be first choice of a large minority of the voters (though perhaps widely detested beyond those) or to have widespread support of second or third votes, even if perhaps fewer people’s first preference.

We shouldn’t take the sports analogy too far. I’m not saying that if you think teams should get three points for a win that you ought to support FPTP (or vice versa). My point is that the debate is over who should win, so it’s no good for either side to pretend that that is already settled.

Don’t forget, under FPTP the loser (according to AV) can win...

Labels: , , ,

Thursday, April 21, 2011

Leaning Towards AV (Though it Wouldn't be my First Preference)

In about two weeks’ time, voters in the UK will be given an historic opportunity to change the electoral system. The referendum motion asks whether the existing First Past The Post (FPTP) system should be replaced with the alternative known, conveniently enough, as the Alternative Vote (AV) system.

Unfortunately, while this is a potentially monumental issue, much of the media coverage that I’ve seen has been frankly very poor, with a lot of bad arguments and negative campaigning on offer from both sides. I thought I’d put my PhD in politics to use, by trying to offer something more sensible – though I hope that the following review won’t be too long or technical.

** Preamble **

Cards on the table time: I’m inclined to vote for the switch to AV. I say that at the start, to make clear that this won’t be an entirely neutral and dispassionate piece. Nonetheless, I’m not strongly in favour of AV, so I’ll try to resist advocacy. I think there are good arguments on each side, so if someone finds the arguments for FPTP more convincing that’s fine. My concern is that people vote based on good, informed argument, rather than on the basis of some of the bad arguments that I’ve heard thrown around.

Why the hesitancy? Well, I think it’s worth noting that NO voting system is perfect. Thankfully, our choice is this referendum is made simpler by the fact that we need not consider all of the many voting systems that have been devised. Our choice is simply between two: FPTP and AV. Hence, there’s no need to consider, for example, Proportional Representation (PR). I’ll define FPTP and AV in a moment, but first one further point is in order.

The referendum question is actually about which system we should use. In deciding how to vote in the referendum, I know some people who are swayed by other considerations, such as giving Nick Clegg a bloody nose or which option they think will favour future reform (some think AV might lead towards PR in future, while others think that this change will stifle further reform). I’m also going to set these considerations aside and focus, so far as I can, simply on the intrinsic merits of FPTP and AV.

** Defining the Alternatives: FPTP **

First Past The Post (FPTP) is probably the simplest option to understand, perhaps in part because it’s the system currently used in UK elections so already familiar to most of us. It works like this: each voter places a vote next to one candidate of their choice. After all votes are cast, they are counted. The candidate who gets the most votes is declared the winner.

It should be noted that the name – First Past The Post – is slightly misleading. There is no fixed winning post that candidates must cross. It would be possible in theory for someone to win with a very small share of the vote – just 10% for instance. Suppose we had eleven candidates competing in the election. If they are very close in support, it might be that one wins 10% of the vote and the other ten each win 9% of the vote. In this case, the former is declared the winner. This is true even though only one tenth of the electorate voted for them.

There are two points that we should notice, in particular. Firstly, the winner need not have (and in fact in UK elections rarely does have) the support of a majority (i.e. over half) of the voters. Secondly, the person with the most votes wins, even if deeply unpopular with everyone else. I’ll return to that point in a moment, but first it’s time to present the alternative.

** Defining the Alternatives: AV **

The alternative vote system allows voters to rank candidates in order of preference. Voters may, if they wish, rank only one. If everyone did that, then the system would in effect work just the same as FPTP. It also allows voters to indicate a second choice, third choice, and so on. These are their ‘alternative votes.’ In effect, it allows someone to say ‘ideally, I want X, but if not X, then I want Y’ and so on – until, if they wish, they have ranked all candidates.

Once voters have expressed their preferences, the first round of counting begins by looking only at people’s first preferences. If any candidate has a majority of votes (over half), then that candidate is duly elected. If no one has a majority, then counting proceeds to a second round. In this case, the candidate with the smallest share of first preference votes is eliminated from the running. Votes for that candidate are then redistributed – any that did not express a second preference are eliminated, while those that did are now transferred to that second preference candidate. These votes are counted equally along with first preference. In effect, the voters concerned, having had their first choice candidate eliminated, are asked for their alternative choice out of those remaining.

(In French presidential elections, there are actually two separate votes. First, everyone gets to vote for whoever they like. Then a second vote is held between the two candidates who come first and second, to see which of these has majority support. This involves only two stages, but requires a second election. The aim of AV is to allow voters to express their alternatives on a single ballot paper, avoiding the need for multiple rounds of voting.)

This process is continued, with the least popular candidate eliminated in each round, until one candidate has over half of the votes remaining active in that round. Since this may ultimately result in only two candidates left in the running, it is almost certain that one will have a majority. (The only exception is if the final two end up exactly tied, 50% each – but ties are also possible under FPTP.)

** An Illustration **

These definitions may be clearer if accompanied by an illustration of the two systems in practice. I’ll use a simplified, non-political example. Suppose we have a club with 100 members trying to decide where to go on a day out. Let’s assume that there are four options: beach (B), football (F), museum (M) and shopping (S).

1) Ten people rank the options as follows:
B > F > M > S
That is, they prefer the Beach to anything else. They like going to the Football less than the Beach, but more than either a Museum or Shopping. And, finally, they prefer the Museum to Shopping.

2) Nine more rank:
B > M
These people also prefer Beach to anything else. They express no preference, however, between the Football and Shopping (though it is assumed, since their second preference is for the Museum, that they prefer Museum to either of these; they are simply indifferent between the other two).

3) Twenty-one people rank:
F > B > M > S

4) Twenty-five people rank:
M > F > B > S

5) Thirty-five people rank:
S > M > F > B

Under FPTP, these people can only vote for their first or most preferred option. That is, the people will vote for B, B, F, M, and S respectively. These are then totalled up, to show how many people support each option. The results are as follows:

Beach: 19
Football: 21
Museum: 25
Shopping: 35

Hence it is decided that the group will go shopping, because this option has more votes than any of the other options.

Note, however, that while Shopping was the most preferred option of 35 people (just over a third of the group), it was the *least* preferred option (solely or jointly) for all of the others. These people, presumably, would be very dissatisfied with this outcome.

Thankfully, there is a remedy at hand. Suppose the twenty-one people in the third group (those who rank F > B > M > S) realise what is likely to happen. They know that the Football is not that popular, so votes for Football are likely to be ‘wasted.’ (This is analogous to those who prefer a third or fourth party, but know that either Labour or Conservatives are likely to win.) Since these people still prefer Museum to Shopping, they may decide to vote tactically. That is, they could vote for Museum, rather than the Beach, even though this is not their real first preference. Then Museum would have 46 votes (to 35 for Shopping) and so win.

AV is, in effect, a way of avoiding the need for such tactical voting. It allows people to still express their real first preference – for Football – yet also to have a say when the choice comes down to Museum or Shopping. Because of this, however, the counting of votes will be slightly more complicated. In the first round, counters will look only at first preferences, so the first round will look like this:

Beach: 19
Football: 21
Museum: 25
Shopping: 35

The same as before. No option has a majority (over half) of the votes, so counting will now go to a second round. The Beach – as the least popular option – will be eliminated. Rather than effectively disenfranchising those who had voted for the Beach, though, the AV system says that we should ask how they would have voted from the remaining alternatives. (This is done without the need for a second, three-way vote, because we asked voters for their second preferences in the initial round of voting.)

Ten of those who voted for Beach would have voted for Football (their second preference) had the Beach not been an option, while the other nine would have voted for the Museum. Thus these voters are reallocated, according to these second preferences. This gives us:

Football: 31
Museum: 34
Shopping: 35

As we see, things are now very close. Had the Beach not been an available option, Shopping would still have won under a FPTP, but only by a single vote. Still, however, no option has majority support. It looks to be between Shopping and Museum, which are one and two in the votes, so Football is eliminated as now being least popular. But again to declare Shopping the winner would ignore the fact that it is the least popular option of almost everyone else. Almost two-thirds of people, in this example, would prefer either Football or the Museum to Shopping. This is confirmed when voting goes to a third stage.

Here, Football is eliminated and those whose votes are currently counted as for Football are transferred to their next preference, to see how they would vote between Museum and Shopping. As it happens, these are voters from the first and third groups above, all of whom prefer Museum to Shopping. (For some of these, Museum is their second preference and for some their third.) Thus, when these votes are reallocated, we get the following result:

Museum: 65
Shopping: 35

So Museum wins!

Sorry for the slightly lengthy explanation, but I think it goes a long way not only to explaining the difference between FPTP and AV but also the relative merits of the two. FPTP only looks at people’s first vote. Consequently, those in a minority are either effectively ignored or forced to misrepresent their true preferences by voting strategically. AV, on the other hand, looks at all of a person’s preferences. This means that an option, such as going to the Museum, can win in virtue of being a widely popular second choice, though it was not the most popular first choice. Political parties would, therefore, have incentives to appeal widely to as many voters as possible, even if they were unlikely to be first choice amongst those voters. Picking up second or third votes might still be enough to allow them to win.

Now let’s look at some of the arguments in more detail…

** Arguments Against AV, Considered and Rebutted **

One argument offered for FPTP is that it’s simpler. Voters only need to vote for one party, rather than ranking their whole preference ordering. I don’t think this shows much faith in voters. The added complexity of AV lies largely in vote counting, not in the act of voting. All voters need to be able to do is order their preferences. If you can count to three, then the chances are that you can rank three alternatives, for instance Conservative > Lib Dem > Labour. It should be added that voters are not *required* to rank all of the alternatives on offer. It’s fine to express preferences between your top three candidates, but then no further, in effect abstaining if the choices were to come down, say, to UKIP or BNP. If this is too difficult for most voters, then we probably ought to reconsider whether we want political decisions made by such idiots!

A second argument I’ve heard is that AV means giving some people more votes than others. This isn’t really true. Each person only has one vote counted at any given stage. The point is that people can also express a second preference, which might be counted *instead of their first* should their first be eliminated. These people aren’t given any more influence though, because those whose first preference is still in the running are still having their first preference counted. It seems ludicrous to suggest that someone who gets their second preference has more power or influence than someone who gets their first preference!

In fact, I think it could be argued that AV better ensures equality between voters, since it allows all voters (if they wish) to have their preference counted between whatever candidates are in the running. FPTP, as already explained, provides incentives for strategic voting. That is, savvy voters may realise that they can better serve their preferences by voting for a candidate who is not their genuine first preference. This behaviour is sometimes criticized as dishonest, though I am not sure I would go that far – why shouldn’t voters be able to use their vote as they wish? The problem, however, is that it means those who know how to ‘play’ the system can get more out of it than those who do not. The naïve bumpkin who simply votes for her genuine first preference may effectively waste her vote, when she could have been better served by voting for her second preference, to ensure that the candidate she detested did not win. This seems more inegalitarian to me. AV removes the advantage is strategic voting, so all people can express their genuine preferences and have them counted equally.

A third argument worries not simply about the fact that some people get their second preference counted, but about who these people are – namely, fringe minorities. Suppose, for instance, we had the following scenario:
49% vote Labour
48% vote Conservative
3% vote BNP, with second preference for the Conservatives
In this case, Labour would win under the FPTP rule, but under AV the 3% BNP voters would be reallocated to the Conservatives and they would win. It is worried that this makes fringe minorities potentially pivotal.

I think there is some merit to this concern, but it seems to be overstated. The electoral result doesn’t depend simply on the 3% BNP ‘fringe.’ The Conservatives only win, in this example, with those votes, but they also have 48% support in their own right. Moreover, it should be remembered that FPTP simply encourages tactical voting. The BNP supporters might simply have voted Conservative to begin with and we’d never have known. At least AV allows voters to express their true preferences and to have influence on the final choice.

Also it should be noted that a few fringe voters are unlikely in practice to make a pivotal difference. Where we have three (or four) major parties, it is likely to take the elimination of one of these before a winner is decided. Consider:
35% vote Labour
35% vote Conservative
30% vote Lib Dem
5% vote BNP, with second preference for the Conservatives
In this case, I have assumed more BNP support (5% rather than 3%). Even so, the elimination of the BNP leaves us with 40% Conservative, 35% Labour, and 30% Lib Dem, so still no overall winner. Who wins will depend on the second preference of (in this case) Lib Dem voters, not (only) the BNP voters.

These are some of the arguments that I’ve heard most commonly presented in the media. As should be obvious, I think that they’re all bad. AV isn’t too complicated (in fact it’s often used in many elections, including in student societies), doesn’t mean that some voters count for more, and doesn’t mean that electoral results will be determined by fringe minorities. There are, however, some better arguments against it. For reasons of balance, let me consider one of the more serious.

AV asks voters to rank their preferences in order, but an ordering gives us no information about how strong someone’s preference is. One voter might be almost indifferent between the candidates she has numbered 1 and 2, while another may have a very strong preference for his number 1 and consider his number 2 merely the best of a bad bunch. Consider this case:
Forty people rank A > B > C and much prefer A to either B or C, both of which they detest.
Thirty-one people rank B > A > C, but are almost indifferent between B and A.
Twenty-nine people rank C > B > A, but are almost indifferent between B and A.

Under AV, C would be eliminated and then B would win, 60 votes to 40. This is so, even though no one has a strong preference for B over A, but 40 people have a strong preference for A over B. We should be wary, this reasoning suggests, of reading too much into second preferences. This is true, but it should be noted that it is hardly a glowing endorsement of FPTP either. People may not be equally satisfied, under an FPTP system, with the alternatives on offer and, as we have seen, many may in fact vote strategically in any case. As I said at the outset, AV isn’t perfect – no system is. While this seems like a potential problem with AV, it doesn’t show that FPTP is any better. Where our choice is between the two, I’m still inclined towards AV.

** Conclusion **

As I said at the outset, I’m not totally confident in my preference for AV, because there are problems with it, but I don’t think these should blind us to the (possibly greater) problems with FPTP just because we’re more accustomed to them.

I’m uncertain about the referendum because, given the many factors that we might consider, I’m open to saying that one’s vote in the referendum *shouldn’t* be determined simply by which, of FPTP and AV, is the better system. (For instance, perhaps one should think about which will better serve the cause of future reform.) Nonetheless, if the question were simply which is better out of the two, then I’m much more confident that AV is preferable to FPTP and that’s the way I’m inclined to vote in the referendum.

My aim isn’t to convince others, but to encourage more constructive debate. Vote how you like, provided that you’ve thought seriously about the issue and have genuine reasons, rather than bad arguments, for your choice. I’d welcome constructive disagreement in comments!