Conservatives defy forecasts to secure victory in UK election

Who knew? Stefan Rousseau/PA

The Conservative Party has won the 2015 UK election with considerably more seats than anyone predicted. After weeks of polling putting it neck-and-neck with Labour, the incumbents have secured 331 seats in the House of Commons – a comfortable majority and many more than Labour’s 232 seats.

Labour had a disastrous night, while the SNP did even better than predicted, winning the all but three seats in Scotland. The SNP, led by Nicola Sturgeon and her colleagues now find themselves with a powerful block to deal with with the Conservative government.

The Liberal Democrats have been massacred, landing just eight seats. While party leader Nick Clegg survived in Sheffield Hallam, big names including Vince Cable and Danny Alexander lost their seats.

A night for losing

Overall, this election has produced a long list of losers. Labour’s performance is very disappointing for a party that thought it would be close enough to the Conservatives to make a credible claim on power.

For the Lib Dems party luminaries such as Vince Cable, Danny Alexander and Simon Hughes fell to the huge swing against the party. Although UKIP won a sizable share of the vote, only one of its candidates was elected – Douglas Carswell. Leader Nigel Farage failed to take South Thanet.

Finally, the election forecasters are clearly losers too. Broadly similar forecasts became commonplace during the election campaign. The Conservatives and Labour ran neck-and-neck, with each party having approximately one third of the vote.

At no time did any of the forecasts (and there were many of them) give either of the major parties a majority in parliament.

Most forecasts underestimated the Conservative seat total by a wide margin. Across seven such efforts published the day before the election, the point forecasts for the average number of seats allocated to the Conservatives was 279, with a range from 273 to 284. And as has become clear, the Lib Dem losses and SNP gains were also both significantly underestimated.

The usual health warnings were issued in the form of statistical uncertainty estimates, but these invitations to prudence were given less attention than they deserved by most consumers of the numbers.

In the end, the problem for forecasters is as familiar as it is vexing – translating votes into seats. Even with high-quality survey data with huge sample sizes, predicting hundreds of constituency-level results in a first-past-the-post electoral system with varying patterns of interparty competition remains a risky business. The 2015 election result forcefully illustrates the point.