Saturday, April 18, 2015

An election model - part 3 - inputs

Having settled on a methodology for modelling the 2015 election (see last post) it was time to see whether the model would produce anything that looked interesting. First, a reminder of the 2010 "baseline":

PartyVotes 2010% votes 2010SeatsPoll standing 16/4/15

Source for 2010 data


  • This was not the same source as the one I obtained the figures for my model for.
  • Although I have excluded the Northern Ireland parties and "others", this covers 97.2% of the votes and 632 seats. Thus, this is a substantially complete and comprehensive model of the parliamentary election for Great Britain
  • Percentages for SNP and PC aren't specified, and in any case as they only relate to a small number of seats, are not meaningful. The BNP will not be a factor at a national level in this election. "Others" at that date was 4%.

My model permits me to apply changes to voting patterns across the country. I can use this to determine what proportion of the vote each party would get as a result, and what that would mean in terms of seats obtained. It suggests the political party of the MP returned

Next, what changes to votes would be applied? One change that has been anticipated pretty much since the Lib-Dems became part of the coalition with the Conservative party has been that their support would fall drastically. However, on the basis that incumbent MPs might be better regarded by their constituents, I applied a different and significantly smaller correction to those constituencies which had an incumbent Lib-Dem MP. I then assumed that these 2010 Lib-Dem voters would divide their votes between other parties - Labour, Green, UKIP and the nationalist parties.

There seems also to have been a strengthening of nationalist opinion, particularly in Scotland. So I assumed that a proportion of all Con/Lab/Lib votes would transfer to nationalist votes. This would have a most significant effect on Labour.

I assumed that a proportion of the Conservative vote would be lost to UKIP, but also that a smaller proportion of the Labour vote would be lost to them as well. I assumed that half the BNP votes would be transferred to UKIP. This was before I had realised just how much weaker BNP had become in this election. And finally, as the Green party is now standing in 90% of constituencies, I assumed that in any constituency in which they didn't have a candidate in 2010, they would pick up a percentage of Labour votes in addition to the Liberal defection.

Here, then, is what I came up with after a little fiddling around with the proportions:

Defection Con-UKIP - 0.17
Defection Lab-UKIP - 0.08
Loss of votes for Lib when not incumbent - 0.7
Loss of votes for Lib when incumbent - 0.2
Lib transfer votes to UKIP - 0.1
Lib transfer votes to Labour - 0.35
Lib transfer votes to nationalist - 0.15
BNP transfer votes to UKIP - 0.8
Lib transfer to nationalist that go to labour if no nationalist candidate - 0.5
Swing Lab/Con/Lib-SNP - 0.25
Swing Lab/Con/Lib-PC - 0.1
Greens standing in 90% - assume all. Defection Lab-Green for new candidate - 0.06

PartyVotes 2015% votes 2015SeatsPoll standing 16/4/15

I will discuss these results in the next post.

No comments: