Invisible wall separating Left and Right

YouGov President, Peter Kellner, discusses the mid-term by-elections

Beware analyses of mid-term by-elections (including this one, if you like). They are prone either to exaggerate or understate their meaning. It is “this shows the government is on its way out” versus “they tell us nothing at all about the next election”. The truth lies somewhere between: read with care, they illuminate the public’s present mood, but they are deeply fallible guides to the future.

Last week Labour gained Corby from the Conservatives on a 13% swing. Is this big, small or par for the course? The best thing to do is compare like-with like – how Labour has performed in past by-elections in its target seats when the Tories have been in government. On this basis, Labour has done better than in any equivalent contest than its first ten years in opposition after 1979. Then it either failed to take target seats (such as Croydon North West in 1981 and Hillhead in 1982) or gained them on smaller swings (Birmingham Northfield in 1982 and Fulham in 1986). Those by-elections confirmed that Labour was not doing well enough to win the following general elections of 1983 and 1987 – which the party duly went on to lose badly.

Want to receive Peter Kellner's commentaries by email? Subscribe here

The nearest post-1979 equivalent to Corby was the Vale of Glamorgan in May 1989, which Labour gained on an almost identical swing. By then, Labour appeared to be back in the game nationally. Ten months later, in March 1990, Labour went on to capture Mid Staffordshire with a swing of 21% – far larger than in Corby. But that was the month when the poll tax was about to come into force and Margaret Thatcher was a deeply unpopular Prime Minister. Had both survived, Labour might well have won the following general election. But the Tories rid themselves of both and duly prospered. In November 1991 Labour captured Langbaurgh with a swing of only 3%: a clear by-election clue that the Tories were likely to win the next general election five months later, which they duly did.

The final five years of Conservative rule, under John Major, produced huge Labour gains with by-election swings: 29% in Dudley West, 22% in Staffordshire South East and 17% in Wirral South. These results told us, correctly, that Labour was on course for a landslide national victory over the Tories in 1997.

Seen in this context, the Corby result fits with the view that the next general election is wide open: Labour did better than in its nightmare years in the 1980s, but not as well as under Tony Blair in the mid-1990s.

There is, however, something else that all three of last week’s by-elections showed, and it’s virtually identical to the message from YouGov’s recent polls, including our latest for the Sunday Times. What all show is that the Conservatives and Liberal Democrats have lost popularity, while Labour and UKIP are more popular than in 2010. Indeed, if we add together the votes for the ‘centre-left’ parties (Labour, Lib Dem and, in Wales, Plaid Cymru), and then add together the votes for the right-of-centre Conservatives and UKIP, then we find a clear pattern:

In each case, the current ‘centre-left’ total is virtually identical to the last election, while the ‘right-of-centre’ vote is much the same (YouGov and Corby) or slightly down (Manchester and Cardiff).

These figures suggest that we are currently living through an unusual mid-term, in that there is little evidence of any direct switching between the two main parties. There is some, of course: there is always a degree of churn. There are some people who will have switched from Labour to Tory since 2010, just as there were those who voted for Thatcher in 1979 and wanted Michael Foot to replace her in 1983 – just not very many of them. The two big movements since 2010 have been from Lib Dem to Labour, and Tory to UKIP. We are seeing a rearrangement of votes within Left and Right, not migration between them.

This is very different from what happened ahead of the last three changes of government. Labour voters switched straight to Conservative in 1979 and 2010, while millions of Tories shifted direct to Labour in 1997.

That’s the past record: what of future prospects? The very existence of the Conservative-Lib Dem coalition means that we are in uncharted territory. For example, none of us can be sure whether the Lib Dems will achieve their usual campaign gains during the 2015 elections, or whether they will be unable to escape the odium of their record in power. It probably depends in part on whether Nick Clegg fights the next election as their leader.

Likewise, will UKIP hold on to their gains (buoyed by their likely success in the June 2014 elections to the European Parliament)? Or will the party cede them back to the Tories when voters are choosing who should govern Britain rather than using polls and by-elections to record their unhappiness with David Cameron’s performance?

My own best guess – no money-back guarantees here – is that UKIP support will slip back to some degree to the Tories and Labour won’t be able to hold on to all their mid-term gains from the Lib Dems. That is why, as things now stand, 2015 looks like providing us with another closely-fought contest.

However, all this could change if the wall between Left and Right starts to crumble, and voters start to move between the two sides in significant numbers. Then all bets would be off. But if the wall holds, Cameron will find it extremely hard to achieve his ambition of an outright victory for the Tories at the next election. Far more likely is another hung parliament or, at most, a very small majority for Ed Miliband.

Want to receive Peter Kellner's commentaries by email? Subscribe here