A ‘uniform national swing’ won’t happen and three factors may turn many more races than expected into tactical exercises
At FiveThirtyEight.com, we have been working on several electoral projection models intended to improve on the main deficiencies of the “uniform national swing” (UNS) technique.
One key area where UNS falls down is on the subject of tactical voting – an issue that is likely to determine the winner of this election, and set the stage for elections down the road.
In order to properly calibrate the projection, we have worked through a significant amount of demographic data, running regression analysis of key constituency characteristics against the results of the elections from 1992 to 2005.
Tactical voting can be easily seen in this context when you compare the actual results of an election against the predicted values that come from the demographic model. For example, in the Labour-Liberal Democrat marginals of Liverpool Wavertree and Rochdale, demographic modelling would have expected a Conservative share of the vote 15 to 20 points higher in 2005, while the Liberal Democrat share should have been about 13 points less in each case.
In this election, three factors are likely to turn many more races than expected into tactical voting exercises. First, the emerging national strength of the Liberal Democrats will provide anti-incumbent and opposition voters a “plausible” alternative, and second, the significant number of constituencies in England with boundary changes will complicate matters for voters. Lastly, there are about 150 MPs retiring, more than any election in recent history, at least 100 of them from Labour, which will make defections from incumbent party candidates even easier in tactical voting situations.
New tactical voting is likely to centre in a set of 66 marginal three-way seats, 39 of them held by Labour and 27 held by the Conservatives. In each of these seats, all three major parties won at least 20% of the vote in 2005.
At the margins there are seats like Aldershot and Devon Southwest, both held by the Conservatives, or Bristol South or Preston, held by Labour, where tactical voting is likely to increase but not make much of a dent in the incumbent party’s lead.
However, in places where the Liberal Democrats have a stronger base to build from, like Somerset North or Bournemouth West, a decisive movement towards the Lib Dem candidate by Labour voters could put it over the top. Similarly, a significant tactical move by Tory voters in Norwich South, Lewisham West and Penge, or Holborn and St Pancras, etc., would bring these in the Lib Dem column.
Increased tactical voting from the Tories is likely to be seen in seats like Edinburgh South, but this will only serve to extend the Liberal lead into the safe zone.
In the truly three-way seats, such as Watford, Ealing Central and Acton, Hampstead and Kilburn, and Filton and Bradley Stoke, the decision of whether to vote tactically may not be made until just before election day, depending on the polls at that point. If the Liberal Democrats are looking strong, a number of Labour supporters – particularly younger voters – will move to the Lib Dems. Tory voters are likely to stick with their candidate, even retreating from previous Lib Dem tactical votes, meaning that the Lib Dem share may not rise as much as expected.
All told, seats with incumbents retiring and particularly compelling Lib Dem candidates will see more pro-Lib Dem tactical voting. In addition, increased turnout among young people, which has been falling each election since 1997, may add strength to a tactical turn towards the Liberals.
Nonetheless, it will be a tough row to hoe for the Liberals. Of the 66 three-way competitive seats, our current estimates have the Lib Dems picking up just 10 seats: 5 from Labour and 5 from the Conservatives.