The Blog

Swing or Roundabout?

In many respects, the way in which we talk (and perhaps the way in which we think) about general elections is decades out of date. We bend our understanding of modern events to fit the language that was coined to describe the events of the past and sometimes, even if we are aware of this, we are in danger of being led astray.

In many respects, the way in which we talk (and perhaps the way in which we think) about general elections is decades out of date. We bend our understanding of modern events to fit the language that was coined to describe the events of the past and sometimes, even if we are aware of this, we are in danger of being led astray.

A classic example of this is the way we talk about the shift of votes between parties. We look at the 2010 general election, when the Conservatives gained more votes than the Labour Party, we calculate the ground that Labour has made up since then as a net percentage of the vote so that we can extrapolate to infer the position in individual constituencies, and we still call this change a "swing". This makes it easy to forget that most of the change that has taken place is, these days, unlikely to involve many people who voted Conservative five years ago now having decided to vote Labour instead. Today, most of the movement is between the two big parties and the smaller parties, and of people who voted in 2010 not being sure that they will turn out this time, or vice versa. To say nothing of the people moving between the smaller parties. Rather than thinking of a swing, it is probably more useful to picture a roundabout.

Every month our Political Monitor polls ask our sample of potential voters how they voted in 2010 as well as how they would vote tomorrow, and the comparison between the two paints a fascinating and far more complex picture of political change than was the case (or was thought to be the case) in the 1950s, when "swing" was perhaps a less misleading name for what was taking place. Aggregating our polls over the first quarter of this year (to give ourselves a big enough sample size to make our figures reasonably precise), we find that our "headline" voting intentions put Labour 2 percentage points ahead of the Conservatives, 35% to 33%. Since the Conservatives won the last election by 37% to 30%, this constitutes a swing of 4.5%, and it would be easy to assume that what has happened is simply that 4% of voters have switched from the Conservatives to Labour (with Labour picking up that stray 1% from somebody else). Simple, but wrong.

In fact, we find that only about 1% of the public have switched from the Conservatives to Labour and, moreover, that an almost identical proportion have done exactly the opposite and switched from Labour to the Conservatives. That is not where Labour's advantage comes from at all. Nor would it be accurate to suppose that, since the Conservatives have suffered a net loss in support while Labour has gained, Conservatives have proved less loyal to the party they supported in 2010 than Labour supporters. Of those who tell us they voted Labour in 2010, 57% still support Labour and say they are certain to vote; for the Conservatives, the figure is the same, 57%. And of those who are still with the same party but not sure they will vote this time? Still no real advantage for Labour, 19% of their 2010 vote compared to 17% of the Tory vote, which is a difference so small that it may be no more than sampling error in the survey. " (In our infographic below, these voters who say they are not certain to vote are counted as "might not vote" rather than including them in any of the parties' totals.)

In fact, those figures might be taken as an extraordinary achievement by the Conservatives. Think about the state of the Labour Party in 2010 under Gordon Brown and who voted for it. At its lowest ebb in years, it must be a pretty fair assumption that most of those who stuck with Labour in 2010 were pretty committed to the party, hard-core Labour voters who would not easily be persuaded to go anywhere else. And yet the Tories have proved as good at hanging on to the loyalties of those who voted for them last time as Labour has.

What about the rest of their votes? About 12% of the Conservative voters from 2010, one in eight, are now intending to vote for UKIP instead; 4% have switched to Labour, 2% to the Lib Dems, and 2% to the Greens, while 6% don't know or won't say what they are going to do. Labour has lost less to UKIP and more to the Greens - in fact about 5% of its support has gone to each of those parties, and a further 5% to the Tories; while 6% of former Labour voters, too, are don't knows. (And 2% of Labour's support has switched to the SNP, which does not make a big splash in the national figures but amounts to a positive tidal wave in Scotland.)

But the much more dramatic, and potentially decisive, movement has of course been of the Liberal Democrats. Less than a third of their 2010 supporters, 29%, say they will vote Liberal Democrat again (and a quarter of those are not sure they will vote at all): 25% are now planning to vote Labour, 14% Green, 11% Conservative and 5% UKIP. But there are also a substantial number of don't knows, 14%. This is where Labour has gained most of its advantage over the Conservatives in the last five years (although they also benefited from higher shares of the much smaller numbers abandoning the nationalist parties and the Greens than the Tories did).

And what about those who admit they didn't vote five years ago, are were too young to have the chance? The biggest number, of course, are probably going to fail to vote again. But around 8% of the public say they didn't vote last time but are now going to vote for one of the big two parties - again, Labour has the advantage here, 5% to 3%. Whether they do, in fact, get to the polls this time remains to be seen, of course.

Does all this matter in the scheme of things, except to understand the minutiae of the swingometer? After all, it's the numbers that matter in the end, isn't it, not where they came from? Actually it matters a lot if you want to understand what politicians are trying to achieve (or should be trying to achieve) in their campaigning. George Osborne's budget, for example, is not really about trying to attract Labour voters to the Tories or stop wavering Tories from switching to Labour, for all that the political debate over it takes the form mainly of a confrontation between government and opposition. It is the leakage of Tory votes to UKIP, the potential for weakening the determination of Labour supporters to get to the polls, the possible gains from the Liberal Democrats, that an "election budget" really needs to address.

Swing is a very useful concept, by far the most useful single figure encapsulating the impact of all of this on the political balance of the country and enabling us to understand national trends in terms of their relevance for individual constituencies in the election. But never forget that it is now a misleadingly simple name for a very complex business.



NB: Table includes only those who are currently "absolutely certain to vote" and are therefore included in Ipsos MORI's "headline" voting intention figures