In the aftermath of David Cameron’s achievement of both a clear victory and overall majority, the polls again appear to have been wrong. But I wonder if this is right. I do wonder whether there is an argument – hard to prove mind you – that it was the polls in the run up to the election that led to the result: at least in part. That is, the polls were right at the time about people’s voting intentions (voters were telling the truth) but as people noted what the polls were telling them about other people’s voting intentions and digested the implications, it caused them to vote differently to the way that they had intended.
This might represent a political version of reflexivity; individuals’ decisions are, in politics like finance, affected by the decisions of people around them and, indeed, by their expectations of what others are going to decide. A French friend cites the case of the Front National in France. The party, as well as its core following, has a floating constituency of people who will vote for them as a protest vote but who will not vote for them if they fear they are likely to win elections. That is, these floaters watch what other voters are likely to do and accordingly adjust their voting.
As we know, there are feedback loops in all sorts of areas and in human activity it leads to the game theory dimension. What others do affects you and what you do affects others. Therefore what you think others will do will affect what you will do and so on. Feedback loops occur in dynamic systems and there were a couple of ingredients in the UK election that created stronger feedback loops and, if you believe this line of argument, produced an outcome that was so different to what was expected.
The first, of course, is that the UK election was, by the UK’s modest standards, a complicated multi-party affair. There is, after all, less game theory in US elections since you largely have only three choices: Democrat, Republican or abstention. In the UK, we have several kingdom-wide parties – the traditional three plus UKIP – a bunch of regional parties, as well as a kingdom-wide single issue party. This was true of the previous UK election but there was so much less coverage of the smaller parties.
The second ingredient, I think, was that voter preferences were more complex reflecting the weightiness of the debate. Even in a multi-party system, if voters have only strong positive single preferences, there is not a lot of gaming. You want one particular party and that is that. But I suspect that many voters had more complex preferences. They may have positively wanted A but also negatively not wanted B. Moreover, what if the preferences were multiple and negative preferences were stronger than the positive. They may have wanted a lot of A, or a bit of both C and D, but definitely not, under any circumstances, B.
Which points up that an election is a distinct dynamic system. It is dynamic between elections but then stops being dynamic on election day. So, it was reasonable for voters to express their positive single preference to pollsters because that is what they were asked to do. Then, however, as election day neared, voters saw what options were actually available, given others’ preferences, and adjusted. I may have wanted A and definitely not B. But that was not available, given others’ preferences. So, in order definitely not to get B, I had to forego A and get a lot of C.
This flurry of last minute activity resembles, to an extent, an Ebay auction. The item goes on auction to close at some point in the future. Buyers might at the outset make some modest bids but most of the action takes place in the run up to the auction closing – the moment the system stops being dynamic. I think what I am saying is that it should be no surprise that the polls were consistent, and apparently wrong, in their predictions until the last moment.
So what? A narrative has emerged in which the Conservatives were always going to win really and that Ed Milliband had always been a disaster for Labour, and voters had simply fibbed to the pollsters. That is, a Conservative victory was knowable in advance. May be. May be in part. But I think there was a lot of real surprise, it was hard to know in advance, it was the outcome of a complex game amongst millions of voters and so the subsequent lionizing of David Cameron and demonizing of Ed Milliband is probably overdone.
The Oblique View 