Forecasting elections is a precarious business. In recent years, professional pollsters have got most elections wrong most of the time. They failed to call the result of the Brexit referendum, David Cameron’s victory in 2015, and the hung parliament of 2017, and lost public trust at a time when, according to some, the country had already “had enough of experts.”
The media will always speculate on how people plan to vote during an election campaign, but relying on polls alone risks overstating their importance and exaggerating their influence. Markets have a better record when it comes to predicting election outcomes, especially the closer to the result you look at them. That was the case when Boris Johnson won the Tory leadership, for example. In simple terms, the more money that’s placed on an election, the more likely the market is to reflect the final result. That’s because the data set is larger; punters and traders are drawing their information from a wider variety of sources and there is more of it. The longer a campaign goes on, the more accurate the picture becomes.
That information might well include the latest national polls, but it could also include the situation on the ground, the number of canvassers on the streets or conversations at the local pub. Markets are adept at predicting probabilities – in individual seats as well as nationally – precisely because they aggregate and assess all the available information. They are the most sophisticated tool we have to measure the collective wisdom of many thousands or hundreds of thousands of individuals.
A week ahead of the vote, the Conservative poll lead was narrowing and the chances of a hung parliament were increasing, according to the pollsters. However, our markets are relatively unmoved and show the Tories are likely headed for a majority, mainly as a result of one key trend: Labour’s loss of support in seats that voted strongly for Brexit. A slew of historically safe Labour seats, from Bassetlaw in Nottinghamshire, to Don Valley in Yorkshire, through Stockton-on-Tees in the North East, are expected to turn from red to blue. Labour’s “red wall” in the North, the Midlands and Wales might not collapse but, according to our prices, it is crumbling. Even in cosmopolitan London, the smart money suggests Labour will lose rather than gain seats – including some of those it won two years ago and others it has held for years, such as Dagenham & Rainham. Scotland also looks bad for Labour. None of this is certain to happen, but the market is saying there’s a high probability it will.
Nationally, on a market-by-market (or seat-by-seat) basis, our election markets are suggesting the Tories will win a majority of 42 next week. That outcome is calculated by taking the leading candidate in each individual seat according to the bets being placed on all 649 contested constituencies. As things stand, our markets are implying that in many of the seats that used to mine coal, produce steel or bash metal, Labour is losing ground to the Conservatives, and Boris Johnson is on course to win as a result. In fact, the chances of Labour winning a majority was just 2.5%. John Curtice, the polling guru who is professor of politics at the University of Strathclyde, said at the start of the campaign that Labour’s chances of an outright majority were “as close to zero as one can safely say”: so on that, at least, the experts and markets are aligned. But with the possibility of a hung parliament becoming more statistically likely by the day, according to most polling companies, it’s interesting to note that our market is taking a far more robust view of the election outcome – it has been expecting a narrow but comfortable Tory win for more than a fortnight. It could be the market knows something the pollsters don’t and we will discover if the wisdom of crowds is more accurate than the analysis advanced by the experts. No one can say for certain if the market will prevail – but I wouldn’t bet against it.
Sarbjit Bakhshi is head of political markets at Smarkets