Thursday 6 January 2011

On electoral systems, and as dull as it seems

So I return to trying to blog after several months and my first post is half an essay about electoral reform. I’m such a huge geek. Sorry.

On a side point, it's great to be in the Labour Party at the moment if you're interested in electoral reform -- as the only major political party without a set position there's actual debate, and this is a Good Thing if you're a politics debate. I'm not talking just about debate among those who actually have much power either, but rather at all levels. It's interesting how on this issue there's genuine and open disagreement and debate that I don't think would exist if the party had a set line. This isn't necessarily because people feel they are not allowed to disagree openly if the party had an official line: rather because it's all too easy for people with a political affiliation, no matter how lowly or low-profile their position in the party, to fall back on the party line as a default. They don't mean to, but to do so is just natural psychology.

I'm not going to try to set out any kind of exhaustive argument as to why I think AV is preferable to FPTP. This is because as a certified amateur philosopher, it's not my job to actually come to any useful conclusions on anything. My skills lie mainly in making fine and dull but possibly useful distinctions, so this I will do.

There are four broad kinds of reasons why people can like electoral systems: they can like that they produce certain kinds of results -- solid majorities, or coalitions, for instance; there can be benefits based on way it encourages those elected to behave –- that they fight for a particular locality, or that they try appeal to a broad range of people, for instance; they can like that the results are proportional; and they can like that a system is clearly understood by the public.

Now the first kind of reason is a complicated category, and perhaps something I’ll revisit in a later post. I think to some extent they’re very different kinds of arguments to the others in that they’re dependent upon assumptions of voting patterns as well as the structure of the electoral system. For example, somebody who is for a system of PR may present an argument about the damage that very large majorities can cause. Presumably, then, they would change their mind in the situation where voting patterns were such that PR actually granted larger majorities than other systems (ie where one party is overwhelmingly popular, but has a dreadful distribution of votes). This is a very different kind of argument to the one that presents proportionality as good in itself. Anyway, I will possibly write about these more at some other time because they’re very interesting.

Nor am I going to talk about the second kind of reason. Frankly they’re just too complicated. How politicians act is pretty inscrutable at the best of times, let alone when considering what would happen in hypothetical electoral systems.

The third and fourth kind of reasons are stalwarts in electoral reform debates, and two varieties of them are being used to criticise AV at the moment.

Some point to the larger majorities that Labour would likely have got under AV in the 1997 and 2001 elections (somewhere between 213 and 245 compared to the actual result of 196 in 1997, for example). I understand the reluctance that those who favour a more or less proportional syste, have for voting for a system that can produce results appearing even more disproportionate. But it’s important to note an assumption made in this particular argument that isn't made explicit.

The key to this is to note that the major losers from a switch to AV in 1997 would have been the Conservatives. The reason for this is that while a significant portion of the population still voted Conservative in 1997, on the whole those who didn’t really really didn't want them in government. They wouldn’t have picked up many second preferences at all, and in Wales and Scotland they would have been a lot of people who put them fourth or lower.

When we say that a system is proportional we mean that there is a relatively straightforward relationship between what the voters voted for and what they got. But “what the voters voted for” is, of course, dependant on the ways in which they are allowed to vote. If you compare a result only to what first preferences were you’ll get a different level of “proportionality” than if you compare it to first and second preferences, and different again for first, second and third, and so on.

To judge proportionality by looking only at people’s first preferences were (ie who they voted for in 1997) is to make an assumption that first preferences are what matters when proportionality is at stake. It is, then, to make an implicit judgement in favour of systems like FPTP over multiple preference systems like AV.

When people say that AV would produce less proportional results in 1997, what they mean is that it is less in proportion to what people’s first preferences were. But is this really a surprise? Isn’t the whole point of AV that it’s not just first preferences that count?

The take home message is that proportionality isn’t as simple as it looks. Don’t let those who argue that AV produces more disproportionate results than FPTP get away with it – by assuming the primacy of first preferences they’re begging the question!

The other argument in favour of FPTP as opposed to AV is clearly of the fourth kind I described above, in that it’s about public understanding of the electoral system. They argue that with FPTP people readily understand how it was that the candidate who won did so: they got more votes than anyone else. This argument ignores that there is more than one facet to public understanding. It’s undoubtedly important that the public understand how we get from a pile of ballot papers to a candidate and eventually to a Government. But it’s more important that the public readily understand how to get what they want out of an election, how to make the most of their vote. If the public don’t understand how they need to vote to get what they want there really is no point in democracy.

The primary problem with FPTP in this regard is that there is barely a single constituency where supporters of one major party or other shouldn’t consider voting tactically since their own first preference has no realistic chance of winning. The level of analysis required by voters in order to make the best use of their vote in this case is absurd. You need knowledge of previous election results, how this could have changed in the past few years, and how likely other voters in your constituency are to vote tactically themselves.

The ability of AV to eliminate the need to vote tactically is such a huge step forward in clarity that, for me, it more than compensates for the added complexity in calculating the result. Oh, and those who argue that less intellectually able or less politically interested voters won’t be able to rank candidates are seriously underestimating the public. In any case, the worst case scenario is that candidates at the top of ballots receive a slight benefit from the tendency to rank from top to bottom among candidates that aren’t differentiated in the voter’s mind.

So the second take-home message is that the clarity of voting systems in the public understanding is more complicated that it first seems: it’s not just important that the public understand how the system works, but (primarily) important that they understand how it is best used.