Should Gordon go for it? (24 September 2007)
Labour 07: The polls look good for Labour, but thanks to the peculiar workings of the British electoral system, that is not necesssarily good enough.
Labour meets at Bournemouth in a slightly giddy state of optimism, inspired by a renewed increase in the party’s opinion poll lead to 6-8 points and other evidence including sensational local authority by-election gains in Worcester and Birmingham last week that the party is in an excellent position. Supposing the polls are right, this would hold out the hope of earning a pro-Labour swing since 2005 and an increase in the Labour majority. However, the one does not necessarily imply the other. One could have both a pro-Labour swing and a reduced majority thanks to the peculiar workings of the British electoral system.
The trouble is that not all votes have the same value. The overall result depends on the votes cast in the marginal seats. Whether Labour has 18,000 or 28,000 votes in Liverpool Riverside is immaterial to the result – the seat elects only one Labour MP no matter how many votes pile up. However, whether Labour has 15,000 or 16,000 votes in Portsmouth North is highly material, as it makes the difference between that seat electing a Labour MP and a Conservative MP. Under Blair, Labour’s share of the vote suffered a severe slump (down from 43 per cent to 35 per cent) but while thousands of votes disappeared in the safe seats, support held up better in the marginals.
The risk Brown faces is that the pattern will reverse. Given that Labour’s majority allowing for boundary changes is 48, there is not much room for slippage if Brown is going to enjoy a manageable full-term parliament. If electors in safe Labour seats who stopped voting between 1997 and 2005 come back to the polls, it will boost Labour’s national share of the vote but not win any extra seats.
There is a strong possibility that Labour could do worse in the key marginals than national opinion trends might suggest. One reason is regional variation. Polls and local elections have seen the Conservatives adding votes in the south of England while doing poorly further north. It so happens that there are a lot of marginal Labour seats in the south and a 3 per cent swing from Labour to Conservative in the region would see 15 seats change hands. A swing of the same size to Labour in the north and midlands would switch only 9 from Conservative to Labour.
Another reason is party organisation and preparation. The Conservatives, in particular Lord Ashcroft, have poured resources into the marginals they want to win and worked hard – they may well now be considerably better than Labour at the campaigning on the ground and this could pay off in winning seats. In the seats Labour should hope to take off the Conservatives, most of them are constituencies Labour lost in 2005.
Newly-established incumbents tend to do better than the national swing in their first election (hence Labour’s nearly undamaged majority in 2001) and a small national or regional swing to Labour would not manage to counteract the incumbents’ bonus. While Kettering is highly marginal, requiring a tiny 0.2 per cent pro-Labour swing on the face of it, in reality it would probably take a national swing of about 2 per cent to fall. Given incumbency and regional variation, it would be quite a risk to go early without a solid poll lead of 8 points or better.
The incumbency factor also applies to Lib Dem MPs – while in principle a post-Blair party should recover ground among the liberal metropolitan electors who deserted in 2005, it may be difficult to dislodge MPs in areas such as Hornsey & Wood Green and Cambridge and Labour will also be exposed to further possible losses for instance in Oxford East and Watford. Labour also has some cause to worry about Wales and Scotland (from the Conservatives and SNP respectively) although the SNP danger has been overstated. There are only three seats which would change hands on a 10 per cent swing from Labour to SNP since 2005, and one is not comparing like with like if one starts from the 2007 Scottish Parliament results.
Before Jim Callaghan decided against an October 1978 election, he took a copy of the Times guide to the House of Commons on holiday with him and tried, seat by seat, to work out what an election result would look like. The best he could do was a hung parliament with Labour narrowly the largest party. Perhaps, somewhere in Downing Street, there is a heavily-thumbed and annotated copy of the last edition of the same volume, pointing to a Labour win by about 30 seats. Whether that is enough, and whether the risk of going now is greater than the risk of leaving the election for another year, is a dilemma Brown must face this week.