Do you think Gordon Brown has ever bought a fizzy drink from a vending machine? It doesn't strike me as a very Gordon Brownish thing to do, somehow. There's too much finding small change involved, and fiddling around with the ring pull is no joke when you've been biting your nails for more than a decade.

So I hope the Prime Minister doesn't mind if I use a vending machine story to explain why he needs to call a general election as soon as possible. I am sure he won't. He may not be good with machines, but he'd be at home with a good analogy, I'm confident.

About ten years ago Coca-Cola developed a vending machine that monitored the weather. The idea was to be able to sell cans for more when it was a hot day. The history of the venture is recounted by Steven Landsburg in his book More Sex is Safer Sex (a volume that sounds racy but is in fact about the economics of externalities).

Professor Landsburg explains that “economists almost universally applauded” Coke for its new scheme. The weather monitors ensured that on days when the machines were most likely to run out of drinks, beverages were sold to those who placed a higher value upon them. The Professor notes, however, that “the general public didn't see it that way”, even though the public are perfectly willing to wear other schemes that were, in economic terms, similar - offering different prices to people booking airline seats at different times, for instance.

Why the apparently inconsistent reaction? It was because of the way Coke allowed its proposition to be framed. If told that drinks would be cheaper on days when it was cold outside, the scheme would have produced a very different reaction.

The idea of loss aversion has become an important one in economics. People's attitude to an offer depends strongly on how they view the status quo. And most commercial outfits realise this instinctively. A retailer, for instance, understands that they are better off marketing a product as “£5 off, if you bring in this coupon”, than they are telling customers without a coupon that they will be fined £5. It is all about framing the question.

Which is why Mr Brown should hold a general election. Now. Or as soon as Chanukkah is out of the way. (It's eight nights incidentally, but it will be over before the new year starts)

Because people are loss averse, they are least inclined to take risks when a loss is in prospect. When a loss has already been incurred, they are more willing to take risks in the hope of making a gain. So if you want to run an election campaign arguing that it is no time for a novice, no time to take risks, you want to hold it while losses are still in prospect.

In other words, the suggestion that Gordon Brown should wait until he has “steered” the economy out of a crisis is dead wrong. He needs the crisis. It is his best card.

In making this point I am not relying solely on analogy. I am also drawing on experience. In 1992, when the economy was still in poor shape and further losses were in prospect, the Conservatives won against Neil Kinnock. It was one of the few elections in modern history in which “it's no time for a novice” has triumphed.

In 1997, by contrast, economic prospects were excellent. Michael Heseltine held a famous press conference suggesting that conditions were such that the Tories were bound to win. Instead, the absence of loss aversion led to a feeling that risking Labour would be fine. The Tory slogan - “You can only be sure with the Conservatives” - had no force. The election was won by the party that proclaimed (and just look at this framing) that “things can only get better”.

Look at the current crop of polls. In the summer the Tories were miles ahead, coasting on the negative attitude of former Labour voters towards their own party. Now the polls suggest that Labour is gaining greatly from an increased willingness of their own voters to go out to vote. And that the crisis has intensified a feeling that the Tories might be on the side of the rich in a downturn, while Labour will be on their side. These are classic loss-aversion sentiments and will last as long as loss aversion lasts.

So, go now, Gordon, go.

There is, of course, a prosaic reason why Mr Brown may ignore this advice. In all the excitement about the narrowing of the Conservative lead it seems almost impolite to point out that it, ahem, remains a lead. And that the Tories have been ahead in every poll for almost a year. So if Labour call an election, the chance that they will be returned with a majority is small. The temptation, therefore, will be to keep hanging on and see what happens. But it is a temptation they should resist. Before the summer I took to describing Mr Brown's political strategy as Travolta Micawber - staying alive and hoping something will turn up. Well, something did turn up. But it's been and gone and turned up now, so what are they waiting for? Something else to turn up?

That attitude reminds me very much of the Jewish joke about the man who falls over the edge of a cliff and finds himself clinging to a branch with his fingertips. “Is there anyone up there?” he shouts to the heavens. “Oh Lord, please tell me what to do.” And from out of the sky, magically, a celestial voice booms: “Son, let go of the branch.” The man thinks for a second and shouts back: “Is there anyone else up there?”

The probability of a clear Labour triumph at the next election may not be very good, but it is far higher now than it will be at any time in the future. This is a Government that has presided over an economic catastrophe. How long does it think the current position - in which voters value experience and authority - will last? And when it comes to an end, boy will it come to an end.

daniel.finkelstein@thetimes.co.uk

Times Recommends

• Farwell to the stiff upper lip
• It’s about who runs the US, not health
• The Iron Lady and her blue felt-tip sympathy