Featured, Features | 12th September 2014

The incumbency effect is hurting the Tories not helping them

In a recent piece, the Times columnist Danny Finkelstein suggested that: “Although everyone thinks […]

Photo: Getty


In a recent piece, the Times columnist Danny Finkelstein suggested that:

“Although everyone thinks the Tories need a lead to eliminate their disadvantage from constituency boundaries, this fails to take account of the advantage that new incumbents have in their first run for re-election. The best estimate is that these two effects cancel each other out.”

Conservatives shouldn’t be so reassured.

We have created something that allows you to translate the latest polling numbers into how many seats each party will win on election day. It shows quite how disadvantaged the Tories are under the boundaries.

In a system where no one party was disadvantaged, two parties would get roughly the same number of seats if they got the same number of votes.

But because of the way Britain’s constituencies work, Labour would win between 35 to 40 more seats if it tied with the Tories. In the graph below, figures have been held constant for the Lib Dems and UKIP at 8 per cent, and those for other parties adjusted to make 100 per cent.

Not changing the boundaries in this parliament was a catastrophic error for the Conservative Party. It has to win nearly 40 per cent of the vote to win a majority – but no party has done that since 2001.

The Tories haven’t won a majority for 22 years and it’s hard to see that changing now that Ukip commands 10-15 per cent of the vote.

But Finkelstein suggests this is counteracted by “incumbency”, which argues voters are more likely to vote for their local MP than their answers to national polls imply. How much more likely? Will this hand the Tories the 40 seats they lose under the boundaries?

The Tories haven’t won a majority for 22 years, and it’s hard to see that changing.

There is little to suggest as much. Lord Ashcroft’s polling of the 20 closest Tory-Labour marginals doesn’t.

He asks voters two different questions. One asks which national party they would vote for in a notional election tomorrow. A second asks them to take into account “your own constituency and the candidates who are likely to stand there”. This allows us to test the incumbency effect.

There is no good news for the Tories. If anything, there appears to be a negative effect. On average Tory MPs poll about 1 per cent worse than the national party. A positive, blue number means the Tories are gaining from an incumbency effect in that seat. A negative, red one means they poll further behind Labour when the question is localised.

Without any strong incumbency effect, the Tories are far behind in most of the seats they narrowly won in 2010, and are trying to hold on to in 2015. Labour leads by at least 8 points in 17 of the 23 closest Tory-Labour marginals.

Ashcroft’s polls are picking up an incumbency effect – but not one that helps the Tories.

When voters in the ten closest Tory-Lib Dem marginals are asked which party they would notionally vote for in an election tomorrow, the Tories lead in every seat.

But when these voters are asked to consider their local candidates (as they will do on election day), there is, on average, a 15-point swing towards the local Liberal Democrat MP.

That is making many theoretically easy Tory wins into very close races. The Tories lead by more than five points in just five of the the 11 seats they are trying to take from their coalition partners .

Not only does any incumbency effect seem incapable of overcoming the 40-seat disadvantage the Tories face under these boundaries, but it is actually hurting the party. It is not helping against Labour, and it is hurting against the Lib Dems.