Update 11/06/17
This forecast turned out to be widely incorrect. In a follow-up post I try to explore why it was incorrect. The major mistake I believe was to bend my forecast to a consensus of opinion. My follow-up post explores this further.
In the past few weeks there have been dramatic swings in the opinion polls. As a result of which I felt unable to update my forecasts. Although the Conservatives have fallen in the polls, and Labour have risen, it is a very inconsistent picture. One Yougov poll had Labour only one point behind the Conservatives. Yet Atul Hatwal at Labour Uncut believes from the canvassing returns, the Labour Party is in for a nuclear winter outside of London.
However, I will go Mike Smithson, the punters and most of the polls in predicting in showing a Conservative lead approaching double figures. Up from about 6% at the last General election, it is quite a small change considering the dramatic changes in the political landscape of the last couple of years. To use uniform swings based on the General Election 2015 would be absurd. My method has been
- Start with the absurd. I use uniform swings – but by region – updating the post of 22nd May, with the forecast for a Tory majority of 28.
- Kept the assumption of a Conservative revival at the expense of the SNP in Scotland, but with Labour now only slightly down on 2015. Assume that Labour are doing very well in London.
- Assumed that in Labour constituencies where UKIP were strong in 2015, half the UKIP vote would go Conservative, and some of the Labour voters would stay at home. That is across the North of England and in Wales I am assuming a drop in turnout, which will be at the expense of Labour.
- I have not made a general allowance for a drop in turnout. This could be significant given the appalling campaigns, and the weather forecast.
In Figure 1 is the General Election Result 2015
Figure 2 is my best guess, by region, of the result of tomorrow’s election.
This would give the Conservatives a majority of 76. This is much smaller than the forecasts in April, but much larger than that of even a week ago.
Figure 3 is the summary of seat gains
The Conservatives gain seats in Scotland, the North, Wales and the West Midlands. Labour still manages over 200 seats. Whilst this is down on 2015, it is better than Micheal Foot did in 1983, and the share of the popular vote is still around 35%, which could be higher than the share Tony Blair achieved when he won the General Election in 2005.
The part of the forecast where I will be closest is the seats with no change.
A quick look at other forecasts. Lord Ashcrofts’ constituency estimate is for majorities between 48 and 78. These are probabilistic estimates, so will not identify actual seats.
The Telegraph have published Chris Hanretty’s estimate of a 100 seat majority. Hanretty did some excellent work on the EU referendum, which I have utilized in a number of posts.
Who knows what tomorrow will bring. But having crude figures by constituency I can compare forecast with actual.
Update 8th June 18.30
Lord Ashcroft’s final estimates were published at 15.45 at Conservative Home. His central “combined probabilistic estimate” of 363 Conservative seats, giving a majority of 76, is exactly the same as mine. We differ on the other parties, as I am assuming Labour will do worse in their Northern heartlands, where disillusioned voters in the Brexit areas will stay at home. The Lib-Dems and Plaid Cymru I assume will do better due to strong local support.
ManicBeancounter 