Seats projections from GB and Scottish polls combined

by Stephen Fisher.

We’ve had eight Britain wide opinion polls since the election was called. Roughly in order, with the most recent first (and with thanks to Anthony Wells at for the figures and changes since the previous poll) they are:

ICM/ITV, CON 48%(+2), LAB 26%(+1), LDEM 10%(-1), UKIP 8%(nc), GRN 3%(-1)

Norstat/S Express, CON 42%, LAB 26%, LDEM 10%, UKIP 8%, GRN 6%

ComRes/S Mirror, CON 50%(+4), LAB 25%(nc), LDEM 11%(nc), UKIP 7%(-2), GRN 3%(-1)

YouGov/S Times, CON 48%(nc), LAB 25%(+1), LDEM 12%(nc), UKIP 5%(-2)

Survation/Mail on S, CON 40%, LAB 29%, LDEM 11%, UKIP 11%

Opinium/Observer, CON 45%(+7), LAB 26%(-3), LDEM 11%(+4), UKIP 9%(-5)

YouGov/Times, CON 48%(+4), LAB 24%(+1), LDEM 12%(nc), UKIP 7%(-3).

ICM/Guardian, CON 46%, LAB 25%, LDEM 11%, UKIP 8%, GRN 4%

The average of the most recent polls per pollster (the top six above) is

Con 45.5, Lab 26.2, LD 10.8, UKIP 8.0.

The GB share of the vote in 2015 was:

Con 37.8, Lab 31.2, LD 8.1, UKIP 12.9

So the implied changes since 2015 are:

Con +7.7, Lab -5.0, LD +2.7, UKIP -4.9

A uniform change projection from these figures, assuming other parties unchanged, gives

Con 386, Lab 179, LD 7, SNP 55, Others 23. A Conservative majority of 122.

But we have also had two polls from Scotland. Again with thanks to Anthony Wells, the headline figures with changes since 2015, are as follows:

Panelbase/S Times – SNP 44%(-6), CON 33%(+18), LAB 13%(-11), LDEM 5%(-3)
Survation/S Post – SNP 43%(-7), CON 28%(+13), LAB 18%(-6), LDEM 9%(+1)

The average of these is

Con 30.5, Lab 15.5, LD 7.0, SNP 43.5.

The Scottish share of the vote last time was

Con 14.9, Lab 24.3, LD 7.5, SNP 50.0

So the changes are:

Con +15.6, Lab -8.8, LD -0.5, SNP -6.5

Not only do these numbers suggest that the Conservatives would take more seats off the SNP in Scotland than in the uniform GB wide projection, but also since the Conservatives are doing much better in Scotland they must be doing a little worse in England and Wales than the GB changes above indicate. Taking both the Scottish and GB polls together implies that the swing since 2015 in England and Wales is, at 5.8, slightly less than the 6.35 figure from the GB polls.

Similarly, since UKIP only got 1.6 per cent of the vote in Scotland in 2015 they cannot have fallen by 4.9 points there. So the GB polls in effect indicate that UKIP are down in England and Wales by more than the GB average. Assuming UKIP have fallen by just one point in Scotland and that Plaid Cymru and Green shares are unchanged, making the necessary adjustments to have uniform change in Scotland separately from England and Wales, yields the following seats projection.

Con 390, Lab 181, LD 9, SNP 47, Others 23. A Conservative majority of 130.

While the projection from the GB only polls suggested the Conservatives would make just one gain in Scotland, the Scottish polls point to eight. The slightly lower swing in England and Wales would appear to have, in effect, cost them only three seats there.

If this differential between Scottish and GB polls persists it suggests that GB wide uniform change projections will slightly understate the implications for the Conservative majority.

Of course much could change in the polls, change may not be uniform across constituencies in many other ways and the polls overall may be out by a substantial margin again. More comment from me on all of these things in due course. In the meantime, Chris Hanretty has an important piece here about how the latest British Election Study data tell us that the Tories are doing particularly well in Labour seats.

Another Labour Meltdown?

By Stephen Fisher

The polls in Scotland just before the last election showed a 21-point lead for SNP over Labour. The SNP went on to take all but one of Labour’s 41 Scottish seats.

This week Theresa May called a general election in the wake of polls showing her Conservative party 21 points ahead of Labour. Could Labour now be headed for a Britain wide meltdown of the kind that they suffered in Scotland two years ago?

Intriguingly, the distribution of the 2015 Labour share of the vote across the seats they are defending now is very similar to the distribution of their 2010 share of the vote in the Scottish seats that they were defending in 2015. In both sets the vast majority of seats involve Labour defending vote shares of between 40% and 65% with an average of 50%. In Scotland two years ago, the largest share that they were defending was 68%. This time Labour have a dozen ultra-safe seats with shares bigger than 68%. That will be little comfort for them.

But it is not the share of the vote that matters so much as the lead over the Conservatives that are defending. In Scotland last time they lost seats where there had been majorities all the way up to 54%. Now Labour have just 13 seats with leads over the Conservatives greater than this. Not much better.

The SNP might have taken even “safer” Labour seats had there been any. Some of the largest swings from Labour that the SNP achieved in 2015 are big enough to unseat any Labour MP if the Tories achieve the same at this election.

So is it reasonable to expect a major meltdown for Labour this time?

Not really. The Tories may be going into this election with a similar lead over Labour that the SNP had at the last election, but it is not the lead alone that matters. It has to be understood in the context of the distribution of votes across seats.

The SNP managed to win nearly all the seats in Scotland with just under half the votes because the votes against them were spread out between several parties, at the local as well as the national level. This led to a very efficient distribution of the vote for the SNP. In the seats the SNP won the average margin of victory was 20 points, less than the 25-point average for the Conservatives. Not only did the SNP not waste so many votes building up bigger majorities than necessary to win, but by only losing three seats they barely wasted any that way either.  By contrast, 26% of Conservative votes in 2015 were in seats that the party lost.

This difference in efficiency was anticipated in advance and largely the product of the distribution of votes in 2010. It was clear from uniform change projections from the Scottish polls in 2015 that Labour were going to lose nearly all their seats.

At the moment GB polls are pointing towards a 6-point swing, or so, from Labour to the Conservatives. That translates to Labour losing about 50 seats with uniform change projections. They would still retain about 180 because of the distribution of votes. That would amount to their worst performance since 1935 but still Labour are in a far more secure position in England and Wales than they were in Scotland two years ago.

Since swing is never actually uniform a key question at this election is whether there will be any pattern in the variation in the swing across constituencies that might either limit or extend the scale of Labour losses. Here there are important lessons from both Scotland last time and Labour’s experience at the 2010 election. I am hoping to write about them in further blog posts soon. Some are comforting and some worrying for Labour. None are as important for Labour as trying to reduce or even reverse the apparent swing against them before election day.

Why was Trump elected?

by Stephen Fisher.

This question can be addressed from various angles and at different levels. I will start with the basic features of the result, then discuss the patterns of change since 2012, all before trying to address the bewilderment of many who find it hard to understand how anyone could vote for him.

Trump won the presidency by holding on to states that Romney won and winning some states that Obama previously won, including Iowa, Florida, Ohio, Pennsylvania and Wisconsin. He did this by a combination of a national vote swing of around 1.5 points since 2012, plus above average swings in some key battleground states, such as Pennsylvania 3 points, Michigan 5 points and Iowa 10 points.*

At the same time, Trump lost the national popular vote but won the presidency by narrowly winning Michigan, Wisconsin, Pennsylvania and Florida with leads of less that 1.5 points each. So, had the swing to Trump in these states been smaller by just 1 point then Hilary Clinton would have won outright.  This was a close election. Continue reading Why was Trump elected?

What to make of the US election forecasts?

By Stephen Fisher

The majority of forecasts point to Hilary Clinton winning tomorrow’s US presidential election. Several of the poll, market and expert forecasts with probabilities for who will win are helpfully summarised by the New York Times here.  The polls-based predictions are all, apart from one, pretty confident that Clinton will win. At the time of writing, Drew Linzer’s model at Daily Kos puts the probability of a Clinton win at 87%, HuffPost has 98% and Sam Wang at the Princeton Election Consortium estimate is as high as 99%. The New York Times’ own model is slightly less confident, on 84%. The exception is Nate Silver’s FiveThirtyEight model which puts Clinton’s chances at just 67%.

The lower probability for Clinton in the FiveThirtyEight model is partly (by comparison with most but not all models) due to 538 having a smaller estimate of her likely lead in votes. By comparison with HuffPost, for instance, this is because FiveThirtyEight put more weight on more recent polls that haven’t been so great for Clinton in wake of the FBI announcement last week of further investigations into her emails.

But the discrepancy is mostly due to the FiveThirtyEight model allowing for more uncertainty. So 538 have a higher probability of a Clinton landslide than other models as well as a higher probability of a triumphant Trump. Continue reading What to make of the US election forecasts?

Why did the UK vote to leave the European Union?

by Stephen Fisher.

It is not hard to see why Leave won. Evidence from numerous opinion polls showed that there was a clear majority for Leave on the basis of concerns about immigration and beliefs that leaving would reduce immigration. Moreover the same opinion polls showed us that there was no compensating majority who believed that the UK would be worse off if we left. Still less did people feel that they personally would be financially worse off. For further details see here.

Although it should not have come as a surprise that Leave won, the result was close enough that it could easily have gone the other way. There will be much debate as to whether the Remain side could have made their arguments more persuasively and campaigned more effectively. With such a close result a host of relatively small differences could all have made a difference. It will be hard for survey and other academic research to identify what, if anything was decisive.

Turnout, at 72%, was the highest for any UK-wide vote since 1992. Whilst this might constitute a relatively high level of participation by recent standards, it does still mean that 28% of registered voters did not bother to have their say in what, we were frequently told, was the most important national decision in our lifetimes. If opinion poll questions on likelihood of voting are at least roughly right, it is likely to transpire that non-voters more often favoured remaining. Moreover it is probably the case that if everyone eligible had voted then the UK would have chosen to Remain.

The local authority results provide some evidence that the divisions in society suggested by the opinion polls as important for referendum voting were born out. Places with more pensioners were more likely to vote for Leave.


The Remain vote was higher where there were more graduates.


These are geographical manifestations of primarily individual-level differences. There are also others related to ethnic diversity and economic prosperity.

The national divides are the most notable aspect of the political geography. Both Northern Ireland and Scotland voted Remain (at 56% and 62% respectively). Both England and Wales voted 53% Leave.

The only region within England to vote Remain was London, and it did so strongly with 60% of the vote. Setting aside London, the rest of England voted solidly, 55%, for Leave. While it is true that the Leave vote was strongest in the English countryside, many of the big cities voted Leave, including Birmingham, Sheffield, Bradford, Sunderland, Bolton, Newcastle, and Stoke. While the North and Midlands were stronger for Leave than the South of England, there were cities in the South, such as Plymouth, Southampton and Swindon, that voted Leave.

In the end only 52 out of 293 local authorities in England outside London voted to Remain. Similarly only 5 out of 22 Welsh council areas voted to Remain. Outside London, the preference for leave was widespread in England and Wales.


Final forecast from the Historical Referendums and Polls based method

by Stephen Fisher and Alan Renwick.

The polls this week have been better for Remain than they were last week. Since this is our final forecast it makes sense for us to restrict our sample of polls to include in our polling average just the most recent poll from each company (or company-mode combination) over the last week. If we do this then our polling average finds Remain at 51 per cent after setting aside Don’t Knows. This is up two points from our polling average on Sunday. The two-point difference is partly due to restricting the sample from two weeks to just one, partly rounding error and partly to the fact that more of the polls than previously include Northern Ireland. So it is not clear whether the apparent movement towards Remain is real or not.

Our forecast share of the vote is 52 per cent for Remain, 48 per cent for Leave. This reflects an expectation of a 1.5-point rise in support for the status quo, based on the change that is visible on average between the final polls and the actual result in previous referendums in Britain or on the EU elsewhere. While this reflects the average historical experience we have explained here and here why the average may not be a very reliable guide.

The unreliability means there is a lot of uncertainty in our forecast. The 95 per cent prediction interval is considerably narrower than it was at the beginning of the week. But at ±10 points it is still very wide. So wide that Remain could reasonably be expected to get anywhere between 42 per cent and 62 per cent of the vote. Neither a comfortable Remain victory nor a comfortable Leave victory can be ruled out.

That said not all the possible outcomes in this range are equally likely. Our forecast probability that Remain will win the referendum is 64 per cent.

The methods behind our forecast

Continue reading Final forecast from the Historical Referendums and Polls based method

Final combined EU Referendum forecast

by Stephen Fisher and Rosalind Shorrocks.

This forecast is based on data collected late on 22nd June. Yet again this method suggests Remain is most likely to win but there is a considerable chance that Leave may win. The probability that Remain will win is up from 62.3% on Sunday to 66.5% now. This reflects both better opinion polls in recent days and so polling models more favourable for Remain, and greater market confidence in a status quo victory.
The only component forecast that is less favourable to Remain is the citizen forecasts, and that is because we have restricted our sample to citizen forecasts over the last week. They are noticeably more equivocal.
The forecast share of the vote is once again little changed, with Remain predicted to get 53.3% and Leave 46.7% per cent of the vote.
Remain % share Leave % share Probability Remain wins
Betting markets 53.5 46.5 76.7
Prediction markets 73.4
Citizen forecasts  52.0  48.0 55.2
Expert forecasts 55.1 44.9  62.0
Volunteer forecasts 54.0 46.0 74.0
Polls 50.6 49.4 55.6
Poll based models 52.5 47.5 68.5
Non-poll based models 55.6 44.4
Combined forecast (mean) 53.3 46.7 66.5

(Individual forecasts collected on the evening of 22nd June 2016.)

There is another poll due during polling day but that would be unlikely to change our polling average. The markets may change, possibly dramatically during the day if people on the betting and prediction markets are following trends financial markets.

METHODOLOGY Continue reading Final combined EU Referendum forecast

Election forecasting and more