There has been a sense in recent weeks among Labour optimists that the coming election need not be a lost cause after all. Are they right?

Their argument goes like this: "The Conservatives' poll lead has been shrinking. They would need an 11-point lead to secure an overall majority; some recent polls have shown them falling short of that. Governments normally gain ground as general elections approach; if Britain's economy is seen to start growing again, this could help Labour further. So will the TV debates between the three main party leaders. If we can get the Tory lead below 6 points then Labour could end up the largest party in the new House of Commons. Even if we don't manage that, we could harass the Tories, should they fall short of an overall majority; David Cameron's could be the shortest-lived government since 1924."

I have no special wish to spoil anyone's New Year but, if Labour is to make a fist of the coming election, it must start with a hard-headed view of what is really going on - and the true picture is not as rosy as that assessment suggests. Let's take each assertion in turn.

Marginal swing

“The Conservatives' poll lead has been shrinking." Well, up to a point. Between May and July 2009, YouGov recorded 13-to-19-point leads. In recent weeks, the range has been between 9 and 13 points. One much-reported MORI poll, in November, put the lead at just 6 points, but this now looks like an outlier. Apart from a brief rise in Labour's support in the days following the pre-Budget report, most polls have shown the Tory lead remaining in double figures.

“They need an 11-point lead to secure an overall majority." That figure assumes that the national swing is reflected in the marginal seats the Tories are targeting. There is some evidence - such as a recent YouGov/Telegraph poll in northern Labour marginals - that the Tories will achieve a higher swing in these seats. Why? Tactical anti-Tory voting, which has benefited Labour in key seats over the past three elections, may start to unwind. If the Tories secure a 9-point lead, they will probably achieve an overall majority. All recent polls bar two have shown them achieving this, so it is stretching things to say: "Some recent polls have shown them falling short of this goal."

“Governments normally gain ground as general elections approach." This used to be true, but not for the past 20 years. The picture is complicated by the way all the polls overstated Labour's support in 1992, and most did in 1997 and 2001. If we correct the data to allow for the polls' errors, we find that there was no signi­ficant government recovery ahead of the last four general elections. As those personal finance ads say, past performance is not a reliable guide to future behaviour. Maybe Labour will gain ground this time. My point is that it cannot be assumed this will happen automatically. Something must occur to make it happen.

“If Britain's economy is seen to start growing again, this could help Labour further." This must be one of the party's great hopes. The next few weeks should bring the release of official data showing growth in the fourth quarter of 2009; and unemployment may not now rise as much as was feared. But we should be cautious. The economy did remarkably well between 1993 and 1997, yet this did little to help the reputation of John Major's troubled government, which had been shot to pieces by the events of Black Wednesday, 16 September 1992. Labour, likewise, will be burdened at the coming election by the vast Budget deficit and memories of last winter's sharp recession. I expect economic recovery to help Labour a little, but not much.

“So will the TV debates." Perhaps. It will be Gordon Brown's gravitas and mastery of detail against Cameron's youth and freshness. In the first televised US debates, the contest was between Richard Nixon's experience and John F Kennedy's charisma. Kennedy won. All one can say at this stage is that Britain's first TV debates could go either way.

Minority government

“If we can get the Tory lead below 6 points, Labour could end up the largest party in the new House of Commons." Again, this assumes that the marginals behave like Britain as a whole. I believe the Tory target is lower: a 3-to-4-point lead could be enough to make them the largest party. In other words, although there will still be a pro-Labour bias in the way votes translate into seats (that is, Labour could still win fewer votes but more seats than the Tories), the bias will be less than it has been for the past 15 years.

A minority Cameron administration could be "the shortest-lived government since 1924". Don't bet on it. Ramsay MacDonald's government survived just ten months. If the Tories end up the largest party but short of an overall majority, I would expect Cameron to form a minority government (rather than do a deal with the Liberal Democrats) and hold - and win - a second election within a year. If the Tories win more seats than Labour in the coming election, they should be safe until at least 2014, unless inept economic policies detonate another recession, and/or tax rises provoke howls of pain. Otherwise, the only people with the power to cut short a competent Cameron premiership will be Europhobic Tory backbenchers behaving as they did under Major.

If Labour loses, its MPs and activists will be little more than bystanders to a drama they cannot control. To regain power, their main task will be to stop the party descending into the kind of civil war that caused so much damage 30 years ago, after the fall of the Callaghan government. And that will be quite demanding enough to keep them fully occupied.

Peter Kellner is president of YouGov