How to lose vote share without losing a single voter.

Last week there was an excellent post by David Herdson over at PoliticalBetting considering the argument that 35% represents Labour’s new ‘core vote’.

It’s a well argued post, and while I disagree with the conclusion, his points are valid. I just think that the positives that David lays out are balanced by other risks.

My view is based on the data I’ve discussed before which suggests around a third of current Labour supporters aren’t convinced by the party on some key measures and that in all recent elections opposition vote share has declined in the final year before polling (yes, even Michael Howard and William Hague).

That means I don’t feel one can confidently speak of a firm floor to Labour support, even if you agree that 2010 LibDems provide a strong floorboard. The damp may set in elsewhere.

The relative importance of these difference elements are a matter of judgement, as is are the unknown factors – how the economy will perform, any unexpected events and so on. This is why I think there’s a good chance the election will turn on small differences. That meant one point in David’s post made me really think.

As he says

“Can Labour actually fall any further?  Bar a point or two at most, the only way the figures could decline further is if other parties start eating into those who voted Labour in 2010, or into the Yellow-to-Reds – or if people from either of those groups sit it out altogether.”

I want to focus on David’s ‘point or two‘.

It’s perfectly possible to see a party lose vote share without losing a single voter. For Labour, that could mean the difference between polling 35 and 37.

The reason is that the current rates of abstention are different for past voters of different parties.

Typically, 2010 Liberal Democrats are almost twice as likely as their Labour friends to say they currently Don’t know or Won’t Vote. Tory voters are also slightly more likely than Labour voters to say they currently don’t know how they’ll vote.

I’ve been taking the recent YouGov polls, and working out what happens if you assume 2010 LibDems and Tories decide to vote in the same proportion Labour voters do now ((Polling types: You will already have noticed that this is simply a crude version of past vote weighting. Indeed, if you assume people will return to their ‘old’ party, the big winners are the Lib Dems)).

This expands the total voter pool, and as a result, (assuming they don’t decide to vote Labour) the Labour share of vote drops by an average of one point. A Labour poll result of 38 typically becomes a headline share of 37, without a single Labour voter having changed their minds.

In some polls, this change knock up to two points off Labour’s headline vote share and lead1.

This probably won’t happen, but is a useful reminder that the flow between the voting and non-voting is going to be as significant to a close election as the flow between parties.

Further, doing the numbers meant I paid a lot more attention to the share of voters telling YouGov they don’t know or won’t vote. To me, this is surprisingly low, at only around a fifth of their respondents2. That would mark a huge turnout increase. If that isn’t the case, voters who decide to sit on their hands over the next year could be a crucial battle ground.

In other words, the choice whether to vote could easily be more important to deciding the next election than change minds between parties. That emphasises the role of field organisation, local organisers, voter mobilisation, quality data and voter profiling.

It also means understanding what your more doubtful and sceptical voters need to know to keep them on board, which is why these two charts niggle at me so much.

The need to focus on mobilisiation is especially true for Labour, as while you can see a pool of potential Tory ‘converts’ currently in the UKIP voters, it’s much harder to see where Labour’s next three or four points of poll share would come from among existing voters.

In other words, absent a shift in strategy, Labour task for the next year will be to hold on to what we’ve got.




  1. the maths is dull, but if anyone would like a copy of my spreadsheet, happy to share it []
  2. Perhaps this is one reason YG is a little more favourable on  Labour vote share than some other pollsters? []

Leave a Reply