When Britain votes at general elections, minutes after voting ends at 10pm, the broadcasters reveal the results of their exit poll – a massive survey of around 150 seats in the country, paid for the BBC, Sky and ITV – which, mostly, gives us a fairly clear idea of how the country has voted.
But on the night of the referendum, there will be no such poll. There will be two on-the-day polls from YouGov and Survation that will come out after 10pm – but they are very different from the usual exit poll, and shouldn’t be treated as the same thing.
Why? Well, British exit polls aren’t measuring voting intention – they don’t give us much of a sense of what the percentage of the vote will be, for instance – but change. Although there are many more Labour voters in Hackney than there are in Harrogate, for instance, for the most part, if there is a five per cent increase in the Labour vote at the expense of the Conservatives in Hackney, there will be a five per cent increase in the Labour vote at the expense of the Conservatives in Harrogate – and, more importantly, in Harlow, a marginal seat.
Although almost every election is accompanied – and followed – by commentators predicting the “death of national swing”, this model of predicting elections has worked so far – and indeed has outlived some of the people who have pronounced it dead. In the last election, for example, Bridget Phillipson’s Sunderland South and Houghton seat was a routine hold for the Labour party – but the vote share showed the Conservatives doing even better than the exit poll – which put the Conservatives on 316 seats, higher than any of the polls hitherto but lower than the 330 seats they would end up with.
The last referendum on Britain’s membership of the European Union – then the European Economic Community – was in 1975. It is difficult in the extreme to measure change from that contest, not least as most of the people who voted in that election are dead. (Although, just as we were able to see a Conservative victory in a Labour landslide in Sunderland South at the general election, we will be able to have a good idea of the overall result as soon as Sunderland votes, even if it does as the polls suggest and backs a Brexit vote by a significant margin.)
To fill the time between close of polls and the first result – likely from Sunderland – broadcasters and pundits will discuss the public polls by YouGov and Survation. A number of hedge funds will be conducting their own private polls, so some people will try to guess the result based on how the market behaves the hours after polls close. However, referendum polling is as hard as it gets – a lot of polling is based on assuming that people are forgetful, lying or both. But with the last European referendum so far back, it is very difficult to readjust people’s voting intention. (To take the last election as a guide, the Conservative lead on the economy and leadership mirrored the real result much better than the headline voting intention did. We have no idea whether economic prosperity will prove a more persuasive message than closing the borders on this occasion, as there is insufficient data.)
So I would hold off any gloating/stockpiling of tinned goods/selling off your share portfolio until the results come in, instead of worrying about the polls from Survation and YouGov.