"Perhaps Labour's day is done, and it is time for something new" says Nick Cohen, a sentiment echoed by Janan Ganesh.

I disagree. Labour has two massive things going for it, which are immense obstacles to the formation of a new party.

One is class. Britain is a class-divided society, and strong political parties must have a class base (which is of course not to say their support should be confined to a particular class.) Labour's ties to trades unions give it this, and any new party will lack it.

It's often said that our FPTP system militates against new parties, which is true. But this is often a reflection of the fact that the class divide is partly a geographical split, with post-industrial areas divided from posher ones.

The second obstacle - which is not unrelated to the first - is path dependency. Political parties, like successful brands, are products of history. People support Man Utd not because of Marouane Fellaini but because they were attracted to the team by the likes of Bobby Charlton and Paul Scholes and got into a habit which they can't break. In the same way, some will vote for a pig if it wears a Labour rosette.

Take, for example, the Greens. In many ways, these are more attractive than Labour - and certainly more so than the intellectual and political vacuum that is the Labour right. They are sounder on immigration and, I suspect, civil liberties and their joint leadership betokens some freedom from the leadershipitis that disfigures Labour. Nevertheless, the lack of a class base and benefit of path dependency makes me doubt they can supplant Labour.

A glance at the historical record shows what I mean. One thing this tells us is that it is rare for big incumbent parties to collapse. This alone suggests that Labour's obituarists might be guilty of the fallacy of base rate neglect - as well as of wishful thinking. Where parties have collapsed it's been because they lack a class base, such as the Liberals in the 1920s, or have betrayed the perceived interests of much of that base as the Tories did in abolishing the Corn Laws in 1846. (Scottish Labour might also be an example of the latter).

Equally, new parties succeed where they have a class base - as Labour did in the early 20^th century - but fail when they are the ego trip of a narcissist; think of Mosley's BUF or David Owen's SDP. It's too soon to say into which category Ukip falls.

In saying all this, I am actually rejigging a point I made in a different context the other day. Judging political parties should be like assessing businesses. We must consider not merely or even mainly the calibre of the people running them: to do so is to commit the fundamental attribution error. We must also consider their structural advantages that come from history and market position. Very often, it is the latter that matter most. Warren Buffett has said that if these are strong enough an organization can survive being run by one's "idiot nephew".

I fear, however, that the Labour party is testing this hypothesis somewhat more rigorously than I would like.