The UKPollingReport election guide for 2010 has now been archived and all comments closed. The new Election Guide for the 2015 election is now online at The old site is archived at the UK Web Archive.

Electoral Bias

Despite the boundary changes you will probably have noticed that the electoral boundaries continue to display a “bias” towards the Labour party. It is far easier for Labour to secure a majority in the House of Commons than it is for the Conservatives. If Labour lead in the vote they will secure an overall majority, if the parties are neck and neck then Labour will be by far the largest party. In contrast, depending on how well the Liberal Democrats do, the Conservatives need to be in the region of 9 or 10 percent ahead in the polls to secure an overall majorty.

There are several reasons for this. Some are to do with the way the electoral boundaries are drawn up, most are to do with people’s voting behaviour.

1) Out of date boundaries. Boundary changes are not a case of swings and roundabouts, sometimes favouring Labour, sometimes favouring the Conservatives. The nature of population movement in the UK is such that updating the boundaries will always favour the Conservative party. In general, population trends are that people are moving out of run-down inner city areas (which tend to vote Labour) and moving into more rural and suburban areas (which tend to vote Conservative). Obviously there are exceptions to the rule – some of the fastest growing seats are places like Poplar and Canning Town and East Ham, but generally it is seats in places like Norfolk, Cambridgeshire, Lincolnshire and so on. Hence the majority of boundary changes will see a reduction in the number of seats in Labour areas, and an increase in the number of seats in Conservative areas.

Because the effect of boundary changes is one way, any delay in keeping the boundaries up to date with population movements tends to be to the advantage of the Labour party and the disadvantage of the Conservatives.

Currently, Parliamentary boundary reviews are based on the electorates at the time the boundary review commences (unlike local authorities boundaries, which are based on projections of the future electorate). In the case of the boundaries which will be used for the next election, the review began in 2000, so by the time the boundaries are first used in 2009/10 they will already be a decade out of date. By the time they are replaced by the next boundary review, due to report between 2014 and 2018, they will be close to 20 years out of date.

The effect of this time-lag is to produce Labour seats that are smaller than average, and Conservative seats that are larger than average. There are well known extremes in size in seats – notably the huge Isle of Wight seat and the tiny Na h-Eileanan an lar – but these are for specific geographical reasons. This is a more systematic problem. Leaving aside Wales (see above), Scotland (whose boundaries are relatively up to date) and known geographical oddities, the smallest seats in England are mostly Labour strongholds – places like Sheffield Brightside (50,801), Salford (53,294), Bootle (53,700). Out of the ten smallest seats, 9 are held by Labour and 1 by the Liberal Democrats. The largest seats tend to be Conservative – places like Daventry (88,758), Northampton South (89,722) and Banbury (87,168). Leaving aside the Isle of Wight, out of the 10 largest English seats, 7 are Conservative, 3 are Lib Dem. At the last election, the fact that the boundaries were 14 years out of date benefitted Labour to the tune of 10 seats, costing the Conservatives around 8 seats and the Lib Dems around 2 seats.

The new boundaries will address some of this distortion – with some minor discrepancies the electorates of the new seats would have been the same in the year 2000. However, with 9 years of population movement since then, even the new boundaries will tend to produce smaller seats in Labour held inner cities than in Conservative rural areas.

2) Over-representation of Wales. Prior the the 2005 election both Scotland and Wales had more MPs per head than England did. This was for two reasons – firstly Scotland and Wales were guaranteed a minimum number of seats at Westminster (71 for Scotland, 35 for Wales). Secondly there are separate electoral quotas for Scotland, England and Wales that are not pegged against one another – hence if the population in one of the UK’s nations grows the electoral quota will also grow, if another nation isn’t experiencing the same population growth, their electoral quotas will slowly diverge.

After the creation of the Scottish Parliament the over-representation of Scotland was removed and their boundaries were redrawn using England’s electoral quota. Wales however continues to enjoy over-representation at Westminister – using England’s electoral quota they would have 32 seats, but thanks to the lower electoral quota in Wales they currently have 40 seats. Since Wales contains large numbers of Labour strongholds, this benefits Labour. How many seats Labour would lose if Wales only had 32 seats would depend on the exact boundaries, but it would probably be about 5.

3) Diffential Turnout. A number of seats a county gets in boundary reviews depends upon the electorate, not how many of those people actually vote. Imagine if there were two seats, one solid Labour and one solid Conservative. In the Labour seat turnout is only 10% but in the Conservative seat it is 90%. Both seats would return one MP, but if you totalled up the vote for both seats the Conservatives would have 90% of the vote. This is a simpfied version of what actually happens in most British elections – social groups who vote Labour also tend be the social groups with a low turnout, hence Labour seats tend to have lower turnouts than Conservative seats.

If every seat in the country had experienced the same turnout in 2005 (but votes within seats had been cast in the same proportions), then the shares of the vote would have been CON 32%, LAB 37%, LDEM 22% – Labour would have had a five point lead rather than a three point lead. If those had been the shares of the vote, Labour would have had an extra 12 seats.

4) Support distribution the nature of the First Past the Post electoral system is that lots of votes are ‘wasted. Votes for party that isn’t in contention to win a seat are not going towards electing an MP – equally, if a party already has 50% of the vote in a seat then any votes over and above that are ‘wasted’. In terms of winning seats it is far more efficent for a party to have lots of seats with about 45% of the vote than fewer seats with 70% or 80% of the vote. In actual fact, analysis by Martin Baxter here shows the distribution of Labour and Conservative support isn’t hugely different – though Labour do have slightly more seats than the Conservatives where they have 40-50% support, the ideal distribution of a party’s vote.

5) Tactical Voting. Tactical voting is when supporters of a party that is unlikely to win in a constituency instead vote for a party with a better chance of winning in order to prevent a third party which they dislike of winning. In practical terms, for the last few elections Liberal Democrat voters in seats where the Lib Dems cannot win have voted tactically for Labour to stop the Tories, while Labour votes in Conservative/Lib Dems marginals have voted tactically for the Lib Dems to stop the Tories winning. This has further increased the number of seats that Labour get at the expense of the Conservatives.

It is fantastically illustrated by two tables on Martin Baxter’s site here which show the votes in those seats where the Conservatives have between 35% and 40% of the vote, and those seats where Labour have between 35% and 40% of the vote. There are 80 seats where the Conservatives have between 35% and 40% of the vote, but only 14 were won by the Tories. In contrast there are 88 seats were Labour have between 35% and 40% of the vote, and Labour won 42 of them. The difference is that Labour often get in because the opposition vote is split between the Conservatives and Lib Dems, while the Conservatives normally don’t get in because their opponents vote tactically for the party best placed to stop them.

This is not ‘bias’ of course, it is people making voting decisions based on the electoral system on how to use their vote in the most useful way, but it is a further explanation of why the system appears to be biased towards Labour. Obviously, were Labour to become more unpopular, the Conservatives less unpopular and Liberal Democrat voters more inclined to vote tactically against Labour, then the effect could easily be reversed and work against Labour.

NB - The constituency guide is now archived and is no longer being updated. It will be replaced by a new guide in 2013, once the fate of the boundary review is finally settled.