by Stephen Fisher.

Extrapolating the results of yesterday’s local elections to those parts of Britain that did not have local elections produces a GB share of the vote of Con 30, Lab 31, LD 15, UKIP 12, and Others 12.

One way to think about these figures is that they represent what would have happened had the whole of Britain held local elections yesterday and if the pattern of candidature had been similar to that in a general election (i.e. if the Conservatives, Labour, Liberal Democrats and UKIP were standing in nearly all wards, and similarly for nationalists in Scotland and Wales). Although the election results from this year that are used for the projection are solely from England, the baseline figures for previous years from which we project the pattern of changing shares do take all parts of Britain into account. The 2015 baseline, on which we rely particularly heavily, is calculated by comparing local and general election results last year, and so reflects the decline of Labour in Scotland. For a more discussion of methodology involved see here and here.

The graph below shows the 2016 figures by comparison with previous values. A full list for previous years is here.

When it comes to thinking about the implications for a future general election, the most important aspect of the PNS is the difference between the Conservative and Labour shares. The two main parties nearly always have a lead in the PNS when they are in opposition. Blair’s leads in the PNS of 21 points in 1995 and 16 points in 1996 clearly heralded the 1997 Labour landslide. Excepting John Smith, other Labour opposition leaders since 1982 (when the PNS series began) failed to achieve such big leads and also failed to win the subsequent general election.

Labour’s PNS lead of just one point over the Conservative’s this year is the same as the party scored in 2011 and 1991, but not as bad as the tie in 1980 or the big Conservative lead in 1982 (when the local elections were during the Falkland’s war). So while this year’s PNS does not represent a record bad result for Labour, it is a long way from the kind of result that suggests a change of government at the next election.

Based on the 2016 PNS, Chris Prosser’s local elections based forecasting model estimates a 92% probability of the Conservatives having the largest share of the vote at the next general election.

The Liberal Democrats may be pleased by a modest recovery, but they are still a long way down on their average performance before they joined the coalition government.

The decline in the UKIP PNS is surprising in the context of recent general election vote intention polls showing higher shares of the vote for them than they achieved at the general election. It may be that UKIP party activists did not put so much effort into the local elections this time because they are much more focused on next month’s EU referendum.

Thanks to BBC, John Curtice, Patrick English, Rob Ford and Jon Mellon for help with data.