The majority of seats in England and Wales (but none in Scotland) will be fought upon new boundaries at the next general election. For each seat affected UK Polling Report has calculated a notional election result – a guess at what the results would have looked like if the votes have been counted on the basis of the new boundaries. These are the votes cast at the last election, shared out between the new constituency boundaries. Hence, for each new seat we have a notional “winner” and a notional “majority”.
It’s important to note what notional results aren’t. Firstly, notional results are NOT a prediction of what will happen at the next election. They are what the results would have been if the votes cast at the 2005 election were counted under the new electoral boundaries. Secondly, they are NOT a estimate of how people would have actually voted if the new boundaries had been used in 2005 – if the new boundaries had actually been in force, people who used their vote tactically may have chosen to vote differently. For example, some people who voted Lib Dem in a Conservative/Lib Dem marginal might have voted Labour instead had they been in a Conservative/Labour marginal. The notional results are based on how people actually voted in 2005, but divided up between the new boundaries – not how they would have voted had the new boundaries really been in use.
Nearly all boundary reviews favour the Conservative party. This is because the trend in population movement in the UK is for people to move out of inner city areas into the suburbs. The result is that Labour strongholds in run-down inner city areas see their populations decline and therefore recieve fewer seats at boundary reviews, while Tory-voting suburban commuter towns see their populations rise, and therefore recieve more seats (obviously this an extreme simplification).
The new House of Commons will have 650 seats, meaning that a party requires 326 MPs for an overall majority (although since Sinn Fein MPs do not take their seats, in practical terms an overall majority is slightly lower).
These boundary changes will see the creation of thirteen new seats (10 Conservative seats and 3 Liberal Democrat seat) and the abolition of nine seats (6 Labour, 2 Conservative and 1 Liberal Democrat). Twenty seats would notionally have been won by different parties under the new boundaries. The net effect of these changes is to increase the number of Conservative seats by 16, increase the number of Liberal Democrat seats by 1, decrease the number of Labour seats by 12 and decrease the number of Plaid Cymru seats by 1.
The means that, had the votes at the last election been counted on the new boundaries the result would have been Labour 343, Conservative 214, Liberal Democrat 63, SNP 6, Plaid 2, Others 3, Speaker 1, Northern Ireland Parties 18. This would have resulted in a Labour majority of 36, rather than they 64 they actually won.. An outline map of the new boundaries, courtesy of Andrew Leitch, can be downloaded here.
These changes are not uniform whatever the swing, as seats are not uniformly distributed in terms of marginality. If the levels of support at the next election were exactly the same as in 2005 then the Conservatives would in theory be better off by 16 seats, but it doesn’t follow that if there was a swing of 4% to the Conservatives they would gain 16 more seats than they would have under the old boundaries – in fact they’d only be 9 seats better off.
|Swing||New Boundaries||Old Boundaries||Difference|
While the new boundaries favour the Conservative party, the effect is marginal. Under the old boundaries the Conservatives would have required a swing of 2.2% to rob Labour of a majority, 4.8% to become the largest party and 7.4% to gain an overall majority. On the new boundaries they require a swing of 1.5% to deprive Labour of a majority, 4.5% to become the largest party, and 6.9% to gain a majority. The effect is there, but it isn’t huge.
How are they calculated?
Parliamentary constituencies are based on local government wards – every new constituency is based upon a number of council wards (sometimes from more than one council) grouped together. Unfortunately general election votes are not counted at a ward level, so calculations for notional results are therefore forced to rely upon other elections which are – primarily local authority elections. Clearly there are limits to the accuracy of such an approach, voting patterns at local elections are not necessarily the same as at national elections, but until such a time that ward-by-ward breakdowns of general election results are available it is the best method we have.
Notional results are calculated in a similar way to that used by Rallings & Thrasher for the “official” notional figures. Votes in local authority elections were summed for the wards making up each of the existing constituencies (in Greater London the ward breakdown for constituency votes in the London Assembly elections were used, since it eliminated the effect of councillors personal votes and the effect of different parties standing in different wards.) Each party’s vote at the 2005 general election was then divided by the sum of the local election votes in that constituency to obtain a multiplier for each party.
This multiplier was then applied to the party’s vote in each ward making up the old constituency to obtain a notional 2005 general election vote in that ward. Wards were then reallocated to the new constituency boundaries and the votes in each new constituency’s wards summed to obtain a notional result for each seat.
In wards that were split between two old constituencies, it is assumed that, in the absence of evidence to the contrary, that the local government vote is homogenous throughout the ward. In multi-member council wards the highest vote for each party was taken. In uncontested wards the winning party was assumed to have obtained the vote of the average of the percentage of the electorate who voted for winning candidates of that party in the same local authority. Parties not contesting a ward were assumed to have the support of the percentage of the electorate who voted for losing candidates of that party in the same local authority.
In a few cases manual adjustments to the figures were made to take into account of feedback I received in response to earlier estimates. These mostly effect very large split wards (such as in Birmingham, where I am indebted to Iain Bowen in particular for his input) and wards in the South-West where independent councillors skewed the result in significant wards.
People do vote differently at local and national elections, and therefore notional results like this will never be perfect. If you know one of these seats well and think the notional results I have here are way off the mark, the chances are that you, with your direct knowledge of the ward, are right and I’m wrong. Here are some particular things that can go wrong.
1) The majority of these figures take no account of particular local factors – they are purely a mathematical calculation based on the parties’s relative strengths in wards at local elections. There are many ways that these calculations could be less than accurate – such as local councillors with strong cross-party personal votes – and no attempt has been made to account for any strong local factors.
2) Some of the more isolated rural areas of England and Wales tend to have a large number of independent candidates and seats uncontested by the major parties. This makes it particularly difficult to calculate notional figures and, for that reason, seats in areas such as Cornwall, North Wales and Cumbria are particularly tentative. The same applies to areas with very strong Residents associations, such as Elmbridge.
3) When only part of a ward moves constituencies, it is assumed that party support within that ward is uniform. This is probably untrue, particularly within metropolitan councils where wards often have electorates of over 10,000.