With YouGov’s MRP model pointing to a 68-seat Tory majority, Conservative voters might think they have pretty solid ground for optimism. But as I say in my Daily Telegraph column today, things can still go wrong for the Tories. Here are my reasons to be fearful:-
The Labour vote has grown since the election was called. Anti-Semitism charges are nowhere near as harmful for Labour as people in Westminster think. Yes, Jeremy Corbyn is having a torrid time of it on the campaign trail – the Chief Rabbi is now out against him – but this is not translating into any decline in Labour support. I had a letter from a reader this morning saying he’s cancelling his subscription because ‘I have become an anti-Semite in my views on Israel. Until recently, I would have denied it.’ It was a disturbing but honest letter, a reminder that there are a lot of people who are not appalled by Corbyn’s views. Yes, about a third of the public think Corbyn is anti-Semetic. But just as many say that Boris Johnson is racist.
Almost 4m have registered to vote since the election was called, two-thirds of them are under 35 and more likely to vote Labour. DeltaPoll’s latest poll for the Mail on Sunday finds that 59 per cent of 18-24-year-olds are planning to back Jeremy Corbyn.
Just because a pollster says there’s <1% chance of anything, doesn’t mean it won’t happen. On the day that Trump was elected, Princeton University put his chances at 1 per cent and advised progressives to campaign for congressional candidates rather than waste their time on a presidential campaign that had already been won. In 2015 Populus unveiled a sophisticated forecasting model called the ‘Predictor’ which drew not just polling data but socioeconomic data and gave odds to a fraction of a percentage point. Just before Cameron won his majority, Populus gave him a 0.5 per cent chance of doing so. It’s common for pollsters to claim they have invented a Clever New Machine that then flops.
Pollsters badly underestimated the Labour vote last time around. Labour won 41 per cent of the vote; pollsters’ final estimates varied from 33 per cent to 40 per cent. How sure can we be that they have corrected for this bias?
The Labour vote rallied late last time, and might do so this time. In fact, there is quite an eerie similarity between vote share patterns in the two campaigns. Below shows polling 15 weeks away from election day, the 2017 figures in dotted lines. Right now, one in six voters are still undecided. If they break for Labour, then that projected majority might turn into yet another hung parliament.