Could Gordon Brown still be Prime Minister at the end of this year? Six months ago, the mere suggestion seemed fanciful, as the Conservatives enjoyed a steady lead large enough to give David Cameron a clear victory. But since the New Year, the gap has halved from an average of 10 points to one of just 5. A further move of the same magnitude between now and 6 May - the assumed date for the coming election - would leave Labour as the largest party in the new House of Commons, and probably not far short of an overall majority.
In order to have a stab at the answer, we must first understand why the Tories' lead has shrunk in recent weeks. Three months ago in the New Statesman (The Politics Column, 4 January 2010), I argued that there is no automatic tendency for governments to recover lost ground as elections approach. Something has to make it happen. What is that "something" this time?
The short answer is the economy. In late January the Office for National Statistics reported that Britain's economy grew fractionally in the last quarter of 2009. On their own, ONS press releases have little direct effect on voters. But since that particular announcement, the climate of debate about the economy has changed; and, consciously or subconsciously, voters have noticed. Indeed, two interconnected things are happening: some of last year's recession-driven anger against Brown's government has softened, and more voters fear for the economy under the Conservatives. Taken together, they amount to a growing feeling that this is not the time to change horses.
We should not overstate the case. The Conservatives are down just 2-3 percentage points since the beginning of the year. And three recent surveys in Labour marginals (by YouGov, Populus and Ipsos MORI) all suggest that the Conservatives are doing slightly better in these than in the rest of Britain.
When we convert votes into seats using a uniform national swing, we find that the Tories need a lead in the nationwide popular vote of around 6 points to become the largest party, and as much as 11 points to secure an overall majority. But if they manage to achieve an extra swing of 2 points in Labour-Conservative marginals, then - even if Liberal Democrat MPs, as I expect, prove hard to dislodge - the Tories' target national voting figures will be lower. Namely, they will need around a 3-point lead to become the largest party, and 8 points to win outright. That would mean the recent 5-point average Tory lead would leave Cameron commanding a larger contingent of MPs than Gordon Brown.
What could change that prospect? We can rule out some of the things that have intermittently excited the media - shock stories about individuals that confirm the public's low opinion of politicians generally, but which seem to have little or no impact on voting intention. Since YouGov started daily polling in mid-February for the Sun and the Sunday Times, we have been able to observe the day-by-day impact of three such stories. They have been negligible. Out there, voters shrug their shoulders and get on with their lives. The Prime Minister loses his temper - so what? A Tory billionaire minimises his tax bill by choosing to be "non-domicile" - so what? Former cabinet ministers try to cash in on their contacts - so what?
For Labour to win the coming election, one or more of three things must happen. First, Brown and Alistair Darling must win the battle with Cameron and George Osborne over economic competence. Last year the Tory duo held a clear lead. Labour's pair has wrestled that back to near-evens (though the Budget on 24 March may have hurt Labour more than helped it). The first-quarter GDP figures are published two weeks before election day. Ministers must be praying that January's rotten weather and VAT rise haven't driven the economy back towards recession.
Oddly, the opposite - rapid growth - might also harm Labour, because the impression could gain ground that the recovery is secure and it is safe to let the Tories take over. Figures suggesting a continuing, but fragile, recovery could suit Labour best.
Second, Brown must improve his personal ratings to close the gap further with Cameron. Already he has made considerable progress. Last summer the Prime Minister's personal rating was -50 (23 per cent thought he was doing well, 73 per cent thought he was doing badly); Cameron's rating was +25 (58 per cent well, 33 per cent badly). That gave Cameron a huge, 75-point advantage. YouGov's latest figures put Brown on -28 and Cameron on +10: the gap has halved to 38. It is still far too wide. The three televised leaders' debates provide Brown with his best opportunity to close it further - and, of course, Cameron's best opportunity to re-establish last summer's huge lead.
A third factor could be, to borrow Donald Rumsfeld's words, an "unknown unknown". Some readers will recall how floods in Germany in August 2002 saved Gerhard Schröder from seemingly certain defeat in the September elections, because of his decisive and empathetic response; and how, 18 months later, Spain's centre-right government threw away a clear lead by responding ham-fistedly to the Madrid bombing, three days before polling day (ministers blamed the Basque separatists for what quickly turned out to be the evil handiwork of al-Qaeda-inspired Islamist terrorists).
However, surely only the most rabid anti-Tory would take any pleasure in Labour clinging on to power as the result of either a natural or man-made calamity.
Peter Kellner is president of YouGov