Why Labour is ahead of the Conservatives in seats, despite the polls being neck and neck

Opinion polls across the UK have shown the gap between the Labour and Conservative parties have narrowed to a couple of percentage points, with the consensus being a 2-point lead by Labour over the Tories.  Based on this, you would expect that the election will be close between the two parties with respect to the number of seats they are expected to win in the House of Commons.

When you extrapolate the polls to get a seat figure, however, Labour has a significant lead.  The lead is very similar to that which the Conservatives actually won in the 2010 general election, despite a seven-point victory in the popular vote.

Conservative supporters say that the allocation of seats across the regions of the UK is biased towards Labour, whereas Labour supporters say that the Conservative vote is simply inefficient.  So who's right?  Let's take a look at the two theories.

Are the constituencies biased towards Labour?

To answer this question, we need to start by looking at the number of votes cast in each constituency.  The Boundary Commission makes it a policy to, if at all possible, keep each constituency entirely within one of the 12 regions of the UK.  Ideally, each constituency should have approximately the same number of voters, and this should be reflected in the regional statistics.

This is the summary looks like from the 2010 general election:

The nation-wide average is 45,673 votes per seat, and each region should see an average close to that.  However, there is a wide variance up and down Great Britain, ranging from 41,032 in North East England to 51,122 in South East England.  This means that the vote of someone in the North East counts for about 25% more, in terms of contributing to the election of an MP, than the vote of someone in the South East.

The Parliamentary Constituencies Act 1986 (enacted by a Conservative government) sets out the rules for the drawing of constituency boundaries.  It sets minimum numbers of seats for Wales, Scotland and Norther Ireland, and that the number of voters in each constituency shall be as close as practicable to the national average.  However, it gives the four Boundary Commissions flexibility to deviate from this for a number of reasons, including the geographical size of the constituency.

The question remains, though, what effect this has on the electoral calculus.  Let's move some seats from the regions that are over-represented to those which are under-represented, so that the new geography looks something like this:

Losers: Wales (8), Scotland (5), Northern Ireland (4), North East (3), North West (2), West Midlands (1), Yorkshire and the Humber (1)
Gainers: South East England (10), South West England (6), East of England (5), East Midlands (2), London (1)

When you adjust the number of seats within each region proportionally, we would still see Labour ahead in seats, but by a margin that is more consistent with a 2-point lead.

It's worth noting that applying a similar algorithm to the actual 2010 election result would give the Conservatives 321 seats, which meant that they would not have needed the Liberal Democrats to govern (only DUP support would have been necessary).

There doesn't appear to be any malicious intent behind this disparity of seats between the regions, as the party that was in power when they were brought in is the party that's losing.  But a move to enforce more equal-sized constituencies can reduce the gap between vote lead and seat lead.

Verdict: True

Are Conservative votes wasted?

There isn't a commonly-accepted definition of what constitutes a "wasted vote", but the basis of Labour's argument is that there are a large number of votes in safe Tory constituencies, and that moving them to marginal constituencies will have a bigger impact on seats. So for this analysis, we will look at the percentage of votes above what is necessary to actually take the seat.  For example, in my home constituency, the elected Labour candidate received 22,132 votes, and the second-place LibDem candidate received 9,633 votes.  This effectively counts as 12,498 wasted Labour votes (22,132-9,634), a 56.5% waste rate.

Based on the current projections, here are the waste rates for each of the three major parties:

At a high level, the share of the vote for both Labour and the Conservatives that is wasted is not statistically different.  In fact, if we were to look at the raw number of wasted votes, Labour has about 16,000 more than the Conservatives.  So if anything, it's Labour that stands to benefit by moving around their wasted votes, not the Conservatives.

Verdict: False