Fantasy election data, 2020 edition

30 September 2015 »

Tags: data politics


Let's look at the numbers behind the election speculation.


There’s been lots of speculation recently about the Labour party’s future electoral prospects. I enjoy a late-night evidence-free debate as much as the next person, but for good measure (and to satisfy my own curiosity) I thought it might be fun to test a few of the more frequently mooted scenarios against some actual data.

So in this post we’ll use maths to figure out what sort of switching between parties would be consistent with five different electoral outcomes. In goes the constituency-by-constituency data for 7 May 2015 along with a big switching matrix (X% of voters for Party A decide to vote for Party B instead, et cetera), and out comes a revised result once all the votes have been reallocated. There are also a few simplifying assumptions to keep the model tractable, which you can check out in the spreadsheet if you’re curious.

Remember: this exercise is about exploring what sorts of underlying numbers would be consistent with each of the five scenarios. You can make your own mind up about which are more or less likely to play out at the ballot box.

In case you need a quick refresher: there are 650 seats in the House of Commons, and at the last election the Conservative party won 331 of them.

All set? Ok, let’s do this.

1. Progressive majority

Proposition

“By adopting a more explicit left-wing stance, the Labour party can attract enough defections from other parties on the left to carry an election.”

What would it take?

At least 75% of all of the people who voted for the Lib Dems, the SNP, Plaid Cymru and the Greens would have to switch to Labour. In this scenario Labour would win 328 seats, which would be a majority of 6.

2. Progressive majority, sans Scotland

Proposition

“As above, but Labour can acheive a majority without any voters switching from the SNP.”

What would it take?

This doesn’t look possible. Even if every Lib Dem, Plaid Cymru and Green voter switched, Labour would still only win 295 seats (well short of the 326 needed for a majority).

3. #JezWeCan

Proposition

“By tapping into a new, down-to-earth way of doing politics, the Labour party can attract enough non-voters to carry an election.”

What would it take?

Over a third of non-voters would have to turn up at the ballot box and vote Labour. And that’s assuming no increase in turnout for any of the other parties. In this scenario Labour would win 328 seats, which would be a majority of 6. Overall turnout would go up by more than 3 million, from 66% to 78% (a rate last seen in the early 1990s).

4. The error of our ways

Proposition

“It’s just a matter of time before the people of Britain come to their senses and surge left from all across the political spectrum. It’ll be like 1997 all over again (but without the Red Tory imposters).”

What would it take?

At least 20% of all of the people who voted Conservative, Lib Dem, SNP, Plaid Cymru, Green and UKIP would have to switch to the Labour party. On top of this, 5% of non-voters would also have to turn out for Labour. In this scenario Labour would win 412 seats, which would be a majority of 174 (cf. 179 under Tony Blair in 1997).

5. Total wipeout

Proposition

“Any new support that Labour wins by shifting left will be massively outweighted by losses in the centre ground, resulting in a Conservative landslide.”

What would it take?

There are a few ways this could happen. Here’s one - if:

  • 20% of Lib Dem, SNP, Plaid Cymru and Green voters switch to Labour
  • 10% of Labour voters switch to the Lib Dems
  • 10% of Labour voters, and 20% of UKIP voters, switch to the Conservatives
  • 5% of non-voters turn out for Labour, and 5% turn out for the Conservatives

Then the Conservative party would win 402 seats, and a majority of 156. Oops.

Share this page

Like Tweet +1 Share