12February

Time to wheel out the Swingometer for the first time this year, to analyse the changes and transfers between the three leading parties at local council elections.

In order to feature in the list, the seat must have been contested by two of the Conservatives, Labour, and Liberal Democrats at both the original election and the by-election. It must have been won by one of these parties, or the Scottish or Welsh nationalists. There are too few seats held by Greens and other minor parties to make them normal in these terms. In multi-member wards, the vote of record is the highest individual vote attained by any candidate of that party.

In order to determine the transfer in an individual seat, we work out the transfers between each party individually. We then take the second largest absolute swing, halve it, and put that down as a transfer of record. We then reduce the first largest swing by the magnitude of the second, and put that down as a transfer of record.

An example should muddy the waters further:

The individual swings are as follows:

The second-largest of these by magnitude is the Lab from LD swing, which we halve to -2.8125%. We then reduce the magnitude of the largest swing by this halved-second-largest swing, leaving Con from Lab as +4.6875%. The final transfers are therefore:

Enough theory, what of the actual results?

Averaging out these swings during the by-elections over the last 13 weeks, we obtain the following:

When we last performed this exercise, in late November, Labour was losing about 7% to each of the other parties. It's become worse. Applied to the Commons, we get the following range of results:

Of the three projections we have, each using a slightly different method of calculation, the central figure we have is for a Conservative overall majority of 10.

| Permanent link