IT IS often reported that the Prime Minister alone among Cabinet colleagues believes that the next general election may be a much closer race than the polls now suggest. Work just completed for me by statisticians in the House of Commons Library suggests that, as always, Tony Blair’s political instincts could be surprisingly accurate.

Late in the 1987 Parliament, depressed about Labour’s prospects, I asked the Commons Library to analyse a series of local election results. Was it true, as I felt, with an unpopular Conservative Government, and Labour unreformed, that voters were prepared to move around to support whichever candidate looked most likely to unseat an incumbent Tory? The analysis showed that this trend was becoming more pronounced and I christened it “tactical voting”. The phrase caught on and highlighted just how ahead voters were of the politicians.

While Labour voters have generally been more willing to vote for a Liberal candidate, Liberal voters have been less prepared to return the compliment. Even so, up until recently, tactical voting has still worked in Labour’s interest. And this growing mobility of tactical voters has made it increasingly difficult for the Tories to win.

Labour’s landslide victories in 1997 and four years later were, in part, a judgment on the unacceptable face of modern Conservatism as well as of the extraordinary appeal of Mr Blair. Recent local government and by-election results suggest, however, that for the first time tactical voting holds out a real danger for the Labour Government. Voters again are adapting quickly to the new circumstances.

I asked the Commons Library to compute the size of Labour’s majority under the following assumptions. I assumed that no Labour voters moved directly to the Tories or from the Tories to my party — an improbable scenario, I know — and that the Tory vote stayed at the same low level as at the 2001 general election. I asked what would be the size of Labour’s overall majority if 5, 10 or 15 out of every 100 electors who voted Labour last time decided not to vote Labour at the next election. I further assumed that a quarter of these former Labour voters would simply abstain and that three-quarters of them would transfer to the Liberal Democrats, largely on the basis of being the anti-war party, as well as in response to the appeal the Liberal Democrats have been making to Labour’s traditional vote.

If 5 per cent of Labour voters desert in the way I have described, Labour still ends up with a majority of 104. However, from this point Labour’s overall majority quickly tumbles. If 10 per cent of Labour voters desert, the overall majority is slashed to 26. At a little over 11 per cent the majority, astonishingly, vanishes.

It is important to emphasise again just how limited the assumptions are for the analysis. But it illustrates how vulnerable Labour’s Commons majority is to quite a small strategic fall in its vote — small certainly compared with the seismic changes in recent by-elections. With a little over 11 per cent of votes lost overall, the number of seats Labour holds falls from 402 to 323.

Most of the talk in the media is of big Liberal Democrat gains. Not so. Despite the assumption that three quarters of the Labour vote moves to the Liberal Democrats, and not a single Labour vote moves to the Tories, the number of Liberal Democrat seats rises from 51 to only 77. The main beneficiaries are the Conservatives, whose total rises to 214 — up from 166 in 2001. A 15 per cent drop in the Labour vote leaves Labour 78 seats short of an overall majority in the Commons.

Labour strategists are already aware of this danger. Hence the trial runs in the Leicester South, Birmingham Hodge Hill and Hartlepool by-elections of “vote Liberal and get a Tory”. That campaign did not stop a very large proportion of Labour voters deserting and there are big question-marks over whether such a simple campaign will be effective enough during the general election. Two factors above all are at work here — trust and the war. When they are combined the impact is potentially lethal.

Some Labour voters simply want to punish the Government. For some of them, not all, the threat of letting the Tories in is less of a deterrent than it once was. Many of them do not believe that there will be a big enough swing to defeat the Government outright, but that enough will vote to tame the Government. For some of those Labour voters who believe the Prime Minister misled the country over the war, voting for anti-war candidates is fast becoming a duty. I write as someone who supports the war in Iraq and knows that many of my constituents will punish me at the election.

It is remarkable how quickly voters now take Labour’s successes, such as the state of the economy, for granted. Their agenda has moved on. With possibly only six months to go before facing the country, the Prime Minister’s room for manoeuvre gets ever smaller. Even if every day from now brings good news from Iraq, attitudes to the war are too entrenched to change very much. The question revolves on whether the Prime Minister can do anything about the issue of trust. He last reshuffle was a lost opportunity — half the Cabinet should have been replaced by younger faces who are not associated in the public mind with spin.

If the Prime Minister does not act, Labour could be in difficulties. When commentators come to interpret the election result, I wonder how many of them will explain what has happened as being pretty obvious. As Bill Clinton might have said: “It was trust, stupid.”

Frank Field is the Labour MP for Birkenhead

Join the Debate at comment@thetimes.co.uk

Times Recommends

• Fragility always gets my sympathy vote
• Double it and you get the real jobless total
• Has al-Megrahi paid his council tax?