Labour's got a swing in its step

Tactical voting is on the increase and this time it is likely to be used to kick 'them' out. Philip Cowley reports

The polls still fail to move. No poll this campaign has given the Conservatives more than 34 per cent of the vote. None - other than those by ICM - has given Labour less than 48 per cent. Even the lowest ICM Labour figure, published in last Wednesday's Guardian, was 45 per cent, and gave Labour a 14-point lead over the Conservatives.

But on 2 May no one will care how many votes the parties have gained. What will matter will be the number of seats. After each poll, there is usually a "projected" figure, detailing the number of seats that such a result would give. It goes something like this: "If everyone votes as they say, the effect of the latest poll, given a uniform swing, would be to give Labour A seats, Conservatives B, and Liberal Democrats C, an overall Labour majority of D seats".

The trouble with such predictions lies not with the opinion polls (although it is wise to remember that sampling error alone is worth plus or minus 3 per cent on each figure). Rather, it is that little phrase "given a uniform swing". These are national opinion polls, but the electoral system is based on 659 separate local constituencies. The problem for predictions based on the assumption of a "uniform swing" is that it simply doesn't happen.

In 1987, for example, the Conservatives managed to hang on to Hyndburn (where they had a previous majority over Labour of only 21 votes), Sherwood (1.3 per cent), and Batley and Spen (1.6 per cent), yet they lost Edinburgh Central (6.9 per cent) and Aberdeen South (9.1 per cent).

Similarly, in 1992, the Conservatives kept hold of Ayr (previous majority 0.3 per cent), Bolton North East (1.7 per cent) and Battersea (1.8 per cent), yet managed to lose Croydon North West, Pembroke (both 10 per cent), Nuneaton (10.3 per cent), Ilford South, Lewisham East (both 10.9 per cent) and even Southampton Itchen (12.2 per cent). Labour won a total of 40 seats from the Conservatives, but of these just 13 were from its top 20 target seats.

The effect of such uneven swing can be dramatic. In 1992 it was Labour that benefited, seeing a larger swing in those seats where they were challenging.

Had all the constituencies in the 1992 election moved in the same way, as projected results assume they do, the Conservatives would have been returned with a majority of 71, 50 more than they actually received. In the last five years John Major has had cause to curse those missing 50 Tories. The same sort of curses must have come from Clement Attlee who polled more votes than the Conservatives yet lost the 1951 election, or from Edward Heath, for whom something similar happened in February 1974.

So attempts to "predict" the overall result from the expected share of the vote are likely to be misleading. When we dig deeper, though, we find things remain very much in Labour's favour.

Of least importance are the candidates themselves. Most recognise this, appreciating that the voters return them to Parliament because of their party label and not because of their sparkling wit or powerful intellect. A few, though, retain a rather touching faith in their ability to win over the voters. More than one Conservative MP thinks that though the party collectively is heading for a real humdinger of a defeat, they personally will hang on.

There is evidence that the characteristics of the candidate - such as incumbency, gender or ethnicity - do shift some votes (although things are a far cry from a 1959 Gallup survey where 55 per cent of the population said that they would not vote for an atheist). But it is very much at the margins. Indeed, the authoritative Nuffield study of the last election could not identify a single Conservative MP who managed to hang on to his or her seat because of any incumbency effect.

Incumbency is likely to have less impact this time. In the last two elections all but a few of the most marginal Conservative seats were being defended by incumbent MPs. This time more than half the Conservative marginals are being fought by new candidates.

There is evidence that constituency-level campaigning might swing slightly more voters. The parties certainly think so, and pile resources - and big names - into their key target seats. One study after the last election showed that in seats where parties were more than 20 per cent behind they spent less than half the amount allowed. By contrast, Labour spent over 95 per cent of the quota in its target seats; with the Conservatives spending a similar amount to fight them off.

Whether because of this or not, there are signs that Labour is doing noticeably better in its target seats, just as in 1992. John Curtice of Strathclyde University has calculated the swing in Conservative-Labour marginals to be 15 per cent at present, four points higher than the national swing shown by ICM.

In addition, the regional variations appear to favour Labour. The biggest regional swings so far appear to be occurring in the south of England - where Labour most needs to win seats - with smaller swings in the North.

A third possible variant is tactical voting. Here again things seem to be in Labour's favour. First, because the extent of tactical voting has been growing. In 1983 around 4 per cent of the electorate voted tactically. By 1987 this had risen to 6 per cent. And at the last election tactical voting amounted to 9 per cent of voters, and cost the Conservatives around 10 seats. But second, and more importantly, because it looks as if the direction of tactical voting might be shifting. At each of the last two elections, according to the British Election Study, Lib Dem supporters who voted tactically were more likely to end up voting for the Conservatives rather than for Labour. Rather than being used to Get Rid Of Them - the title of a group trying to organise anti-Conservative tactical voting this election - it has just as often been used to keep "them" in.

The pattern might well be different this time. ICM found that 56 per cent of Lib Dem supporters would switch to Labour if they were convinced that the Lib Dems could not win in their constituency. Similarly, 66 per cent of Labour supporters would vote Lib Dem if the choice in their seat was between the Lib Dems or Conservatives.

The traffic is not all one way. About a quarter of Lib Dems and 12 per cent of Labour supporters say they would switch to the Conservatives. And around 10 per cent of each party's supporters say that they would stick with their party come what may.

Overall, however, the picture remains a good one for Labour. Their poll lead appears to be holding up. Regional and constituency factors appear to be working in their favour. And should they manage to coax them, it appears as if a large chunk of Lib Dem voters are ready to fall into their arms in order to eject the Conservatives.

They remain on course for an impressive victory. Just how impressive remains unsure.

The writer teaches politics at the University of Hull.