WREXHAM, A FORMER mining town in north Wales, had returned a Labour MP in every election since 1935. Yet in the early hours of December 13th it picked a Conservative for the first time in its history. The Tories won 45% of the vote, almost exactly as predicted in a constituency poll by Survation for The Economist in the run up to the vote (see chart).
Wrexham was one of five key constituencies we polled during the campaign. Constituency-level polling comes with a health warning: the sample size is small, so the margin of error is large. Not all of our results were accurate. In Wrexham we underestimated Labour’s performance, though we got the Tories about right. Yet overall it worked well, calling the result correctly in four of the five seats.
A poll in Great Grimsby provided a stark indication of the trouble Labour faced. It gave the Conservatives a 13-point lead in a seat that was once as safe as it got for Labour. (One former MP claimed the party could stand a “raving alcoholic sex paedophile” in the fishing town and still win.) Brexit changed all that. About 70% of Grimsby’s voters backed Leave. Melanie Onn, the Labour MP, rebelled against the party to support Boris Johnson’s Brexit deal. It did her little good. The result was an astonishing 22-point victory for the Tories.
Warwick and Leamington, a collection of handsome towns in the Midlands filled with students and Remainers, should have proven a happier hunting ground for Labour. Yet The Economist’s poll put both parties neck and neck in the marginal seat. In the end, Labour won, but barely. The Liberal Democrats squeezed Labour, but not fatally. The party scraped past the Tories by 789 votes.
In normal times, Reading West is the type of seat that Labour might have hoped to gain from a party that had spent nine years in office. The commuter town outside London is filled with the type of people who tend to vote Labour: young families, ethnic minorities and those who rent their homes. If Labour were a party readying itself for government, Reading West’s 2,876 majority would have been gobbled up with ease. Instead, our poll suggested that Labour’s vote would crash. Labour did better than we expected, winning 40% of the vote, but still well behind the Conservative Party’s share of 48%.
Where the polls did slip up was in Gedling, a constituency of suburban towns and villages east of Nottingham. The poll, taken at the start of the campaign, gave Labour a small lead of 42% to the Tories’ 37%. In the end the Conservatives snatched it by just 679 votes. It capped a miserable night for Labour. In Vernon Coaker, Labour had the benefit of a popular incumbent. Although the constituency voted Leave, it did so by only slightly more than the English average. Gedling was a tipping point for the Tories, we argued: if the Conservatives could take Gedling, they were on track for a big majority. So it proved.
Our latest coverage of Britain’s election
This article appeared in the Britain section of the print edition under the headline “Constituency polls foretold the election result”