How long can we expect the polling bounce to last?

The above table, adapted from UK Polling Report, shows what happened to the polls when John Major succeeded Margaret Thatcher in November 1990 – the last time a Prime Minister was replaced mid-term.

With all the talk of a general election in the air a critical issue will, surely, be the expected length of Gordon’s poll bounce. Clearly new or different faces and a new approach had a very marked impact then and it really is no surprise that the same is happening now.

Of course there were differences between the end of the Thatcher era and the end Blair era but should we really be expecting the change-over boost to be lasting, perhaps, six months. And if that is the case does it reinforce the view that there should be an earlier rather than later general election?

Whenever this is raised people point to Brown’s great caution – but would not he get massive kudos from having the bravery to put his government to the test so early.

In making historical polling comparisons we should remember that all the firms operated in a different fashion seventeen years ago and when it got the the general election there was an average understatement of the Tory lead of 7-8%.

I remain of the view that an early general election is probably more likely than the betting markets are currently suggesting. The latest CantorSpreadfair spread in the number of weeks to the election starting from Gordon’s arrival on June 27th 2007 is 86.75-89.75 weeks. I sold at 86 weeks.

Mike Smithson