Top Ten Predictions for the Election

This article first appeared in the Daily Telegraph online on 6 June 2017

Will the Conservatives win, or can Labour pull off a shock result? We will know all by Friday morning, but if you can't wait until then, here are the current best predictions from the Electoral Calculus' crystal ball.

Prediction #1 Predictions are only as good as the polls

The predictions from Electoral Calculus and most other pundits are based on the opinion polls. Even if our models are good, they are vulnerable to polling error. If the input polls to the model are wrong, then the output predictions will be wrong too. The tech guys call this "garbage in, garbage out", and that's exactly what happened in 2015 when the polls underestimated David Cameron's lead over Ed Miliband by 6pc.

So this prediction is the caution that if the polls are wrong, then the other predictions below will be wrong too.

Prediction #2 Some of the polls will be wrong

Polls can never be completely accurate and always have some error. But this year, things are stranger than that. Following the problems in 2015, the pollsters have been working hard to make their polls more accurate. Their methodologies have grown more complex and now have their own models to predict turnout. (They used just to ask people if they would actually vote, but too many non-voters over-optimistically think they will.) This complexity has led to a clear divergence between pollsters. Some think the Conservative lead over Labour is fairly small but others think it is much larger.

And there are reputable and established pollsters on each side. Amongst the "low-leaders" who see a small lead are YouGov, Ipsos-MORI and Survation. On the other side, the "big-leaders" are ICM, ComRes and Kantar (formerly TNS-BMRB). And some other firms like Panelbase and Opinium are less clearly in one camp or another.

On an average of recent polls, the low-leaders have a Conservative lead over Labour of 4.5pc, whereas the big-leaders think it is 12pc. As the saying goes, they can't both be right. At least one group, or maybe both, is wrong.

So a pretty definite prediction is that some pollsters are going to be wrong.

Prediction #3 Labour's vote share will be up compared with 2015

Jeremy Corbyn entered the campaign with negative expectations. Many observers were sceptical that he could appeal to the public, and last year three-quarters of his own MPs voted no-confidence in him. But he has confounded them, and Labour's ratings (from all the pollsters) are now higher than 2015. In surveys of who would make the best PM, Corbyn also does better than Ed Miliband.

If Labour's vote share goes up, even if it loses some seats, Corbyn may try to stay on as leader. This would dismay his internal critics who had hoped the election would prove the undoing of the hard left and allow them to retake control of Labour. That is now only a possibility rather than a likelihood.

The prediction is that Labour's vote share will be up on 2015.

Prediction #4 The Lib Dems won't make progress

The Liberal Democrats had hoped to capitalise on the unelectability of Corbyn and on the 48pc of people who voted "Remain" and who want a softer or cancelled Brexit. But it hasn't worked out for them. Polls show that on average they are basically unchanged on their 2015 performance of 8pc, when they dropped from 57 to 8 seats. And with the decline of UKIP and the Greens, they should have expected to pick up some more votes.

Since both Labour and the Conservatives have gained vote share, standing still is the same as going backwards. My current central forecast is that Lib Dems might only keep three of their seats, with Nick Clegg having a struggle to retain his own seat of Sheffield Hallam. Many of their seats are now marginal, and could go either way.

It looks fairly likely that the Lib Dems will get between one and eight seats, and pretty likely they will not get more than ten seats.

Prediction #5 UKIP will do badly

Despite being on the winning side of last year's EU Referendum this election has gone badly for UKIP. Beset by internal squabbles, and lacking an established leader and message, they have floundered. They have been unable to take votes from Labour, and have been losing many of their existing supporters to the Conservatives. Their only MP has left them.

They are standing aside in hundreds of seats, which gives a further boost for the Conservatives as UKIP-inclined voters are more likely to support Theresa May than Jeremy Corbyn.

This prediction is that UKIP will win no seats. Whether UKIP will survive as a party is an open question.

Prediction #6 Caroline, Yes

Despite losing some of their national support, the Green party will keep their seat of Brighton Pavilion. Caroline Lucas, their co-leader and incumbent MP, is locally popular and local constituency polling shows her enjoying a strong lead over the Conservatives and Labour.

The prediction is that the Greens will keep Brighton Pavilion, but not gain any other seats.

Prediction #7 The SNP will win Scotland but with fewer seats

In Scotland, polls show that the Conservatives are in clear second place ahead of Labour, and that the SNP have lost some ground. In 2015, the SNP won a colossal 50pc of the vote and all but three of the 59 seats. So there is nowhere for them to go but down. There are many tight contests in Scotland, which could go either way. Even deputy SNP leader Angus Robertson faces a strong Conservative challenge in Moray, and Edinburgh contains four closely fought seats.

The prediction is that the Conservatives will gain seats in Scotland and could take between five and fifteen seats. The SNP will lose some seats, but will still win most of the Scottish seats.

Prediction #8 Low lead means a close result

If YouGov, Ipsos-MORI and the other low-lead pollsters are correct and the Conservative lead is in the low single digits, then the election result will be uncertain. There is still a fair chance that the Conservatives will have a majority, albeit a modest one of around twenty-five seats and probably less than 100 seats. But there is also a one-third chance that they will fall short of the required 326 seats and not gain a majority. There is also a 15pc chance (odds of six-to-one against) that Labour and the SNP would together win more seats than the Conservatives. That could be a route into number ten for Jeremy Corbyn at the head of a minority Labour government with SNP support.

This prediction is that if the the low-lead pollsters are right (which they might not be), then the election becomes unpredictable and although the Conservatives have a good chance of winning, a minority Labour government is possible.

Prediction #9 High lead means a solid majority

On the other hand, if ICM, ComRes and the other high-lead pollsters are correct, then the Conservatives will enjoy a greatly increased majority. In this scenario, the Conservative majority can fairly confidently be predicted at somewhere between thirty and 180. Indeed there is a fifty-fifty chance of a landslide majority of more than one hundred seats. Theresa May would be returned to Downing Street and the ups and downs of the campaign trail would be likely forgotten.

If the high-lead pollsters are right (and they might not be), then the Conservatives will have a substantial majority.

Prediction #10 Theresa May is more likely to be Prime Minister than Jeremy Corbyn

The divergence between the polling firms makes this election less certain. If the low-lead pollsters are right, then Labour could (just) win. Suppose both sets of pollster are wrong and the actual result is somewhere in between. Then the Conservatives have a 90pc chance of an overall majority and a 96pc chance (odds of 27-to-1 on) of having more seats than Labour and the SNP combined.

Taking it all together, the Conservatives have a large change of winning the election, and Theresa May is the most likely person to end up as Prime Minister. But it is not certain.


Return to Blogs and Comment page, or home page.