Posts Tagged ‘election forecasting’

Screen Shot 2017-06-05 at 2.19.00 pm

This blog has a long and honourable history of predicting election outcomes, and usually getting them right.

We have, however, recently made a right pig’s arse of the process.

We got the Australian election almost spot on, in predicting a narrow Conservative majority. We got the shellacking handed out to the Liberal Democrats in the 2015 election spot on – we may have been the only predictor expecting them to hold under 10 of their seats. But in both those cases it might be argued that it was our personal closeness to the outcomes that led to their accuracy.

In other instances recently – David Cameron winning an unexpected majority in 2015, the Brexit vote narrowly backing Leave, and most infamously Donald Trump winning the electoral college (not, note, the popular vote) we were plain wrong, much to the inordinate glee of some of our correspondents who accused us of everything from not understanding opinion polling (unlikely considering our profession), to not seeing a fundamental shift to the populist nationalist right worldwide (which was always a nonsense), or of under-estimating electorate’s sense of angst and desire to give anyone – everyone – a kicking, (which was perfectly correct, and we did indeed under-estimate it).

Bang on time for this week’s UK election, this super article by Nate Silver of fivethirtyeight.com explains in great detail and copious references just why it has become so difficult to predict elections today, and the efforts gone to by polling organisations to correct any likely mistakes – which may even cause further mistakes.

If you have any interest in politics or political forecasting at all, we cannot recommend highly enough that you click the link and read it.

Anyhow, the old line out forth by politicians anytime a poll is unfavourable to them – “There’s only one poll that counts. The one on election day.” – is proving to be more and more true.

Screen Shot 2017-06-05 at 2.23.45 pm

So will we make a prediction, or is our headline just a pathetic attempt at click bait?

Nope. We’ll give it a go, but with the loudly proclaimed proviso that we could be wrong by more than the apparent margin of error either way, as Silver so kindly points out.

The average poll lead for the Tories over the Labour Party currently sits at around 7%. In our view, this is likely to be a winning margin for the following reasons:

Labour haven’t fixed their Scottish problem. Although they will do better than their wipeout north of the border last time, we need to remember that Scotland used to be rock-solid Labour territory. They could count on hatfuls of seats from the big industrial conurbations. They are doing slightly better now, so they will win some seats back, but in our view, not enough. And essentially, with a thoroughly “SNP-ised” Scotland, Labour need to do historically better in England and Wales than at any time in their history in order to overhaul the Conservatives. And that’s not going to happen.

The Liberal Democrats have stayed stubbornly limited to around 8% in the polls, sometimes up to 10-11%, sometimes down to 6%. Essentially, they are not seen as serious contenders in this election, and there is a still a strong “on the nose” element hanging over from their disastrous conduct of the Coalition agreement. The Lib Dem result is notoriously hard to predict because what they do bring to the table, undoubtedly, is superb on the ground campaigning – the so-called “street game”. Which is why we think they will pick up some seats, but they are also in danger of losing a couple of obvious wins where boundary changes or demographic changes are running against them, so their net effect via-a-vis the Tories (their main target) is likely to be negligible. Certainly not dramatic enough to rob May of her majority.

The Corbyn factor is especially interesting. Rarely has a leader of a major political party been more demonised by the media and commentators. But along with this demonisation has come a stubborn determination by some Britons – especially the young and first time voters, who have registered in historically large numbers – to back him, to give the establishment a kicking if nothing else. He has also appeared more impressive while the campaign has worn on, in stark contrast to Theresa May, who has frequently looked like a rabbit caught in headlights.

So we see the growth in the polls for Labour as real, but we simply can’t bring ourselves to believe that a character as polarising as Corbyn can beat an incumbent government. It’s just too counter-cultural – the continuing drumbeat about his past views on the IRA, his beard – how successful many bearded politicians do you know? There’s a reason for that. – the fact he looks uncomfortable in a suit, the way he is repeatedly castigated as returning Labour to the bad old days of the seventies, let alone his recent wholesale rejection by his own MPs, etc. etc.  It’s just too much of a tsunami of criticism to push back against.

Our last reason for plumping for the Tories is the three recent terror attacks, and especially the last two, in Manchester at the Ariana Grande concert, and over the weekend on London Bridge. In times of deep anxiety, people tend to plump for “the devil they know”, and in this case that is definitely the incumbent government. It shouldn’t be so: in a perfect world, people should make a careful and cautious examination of all the reasons for voting for one or other of the options on offer, and not be swayed by the actions of a few vicious lunatics, but the fact is people simply don’t behave like that. Enough people will say, we believe, that this is no time remove backing from the Government, and that factor alone will be enough to see them over the line. Also, the Prime Minister and Government of the day receive a great deal of “free” – and mostly positive – airtime, simply by saying the sort of thing that is expected of political leaders at such times.

As became clear after Manchester, however, the advantage does not necessarily flow all one way. A spate of attacks, so close to an election, cannot but challenge the Government’s line in some people’s minds that it has successfully defended the country against terrorism. After Manchester, too, the Labour leader offered a rather different view of root causes. While Corbyn denied blaming the Government for the Manchester attack in any way – and indeed did not do so – he did ask about the possible impact on domestic security of the UK’s involvement in foreign wars. Nor – surprisingly perhaps, given the context – did he attract much public opprobrium for doing so.

So the advantage to the Conservatives of a national security election – as opposed to a Brexit or migration or austerity election – may not be as decisive as initially thought. But it will not be the same election as it would have been, even if going ahead with the vote as normal shows the necessary determination not to be cowed.

So in the final wash up, our guess is a Tory majority of 30-40 seats. Way down on initial suggestions that May might win a majority of 100 or even 150 seats, and one unintended consequence of such a result would be that May’s own leadership credentials would be severely wounded. How seriously, only time would tell. But there are certainly those around her cabinet table with a lean and hungry look, even if that soubriquet could never reasonably be applied to Boris Johnson.

But our certainty, also, is that we could be completely – even wildly – wrong. Things used to be quite easy in the world of psephology. Not so much nowadays.

For once, we will not be making a substantial bet on the outcome. That should tell you something, Dear Reader.