Tuesday, May 16, 2017

If this election were a boxing match (which it isn't) ...

Earlier today, after learning of a particularly egregious shambles following the launch of Labour's manifesto (Jeremy Corbyn made a double U-turn within hours of the launch), I tweeted

"If this General Election were a boxing match the referee would have stopped it and awarded victory to Theresa May ages ago."

Lest anyone misunderstand, this was a reference to the relevant level of competence of the two sides, which in a boxing match would have made it necessary to stop the fight to protect the weaker player from injury. It was not a prediction about the result.

An election is not the same as a boxing match and as we have all repeatedly discovered over the last two years, opinions can change and opinion polls are not always right.


As Stephen Bush of the New Statesman points out, three weeks before the 1997 General Election, opinion polls were published projecting a Labour vote share of 50%, 51% and 53%.

In the actual 1997 General Election Labour polled 43%.

Three weeks before the 2001 General Election, opinion polls were published projecting a Labour vote share of 48%, 54% and 53%.

In the actual 2001 General Election Labour polled 41%.

Three weeks before the 2005 General Election, opinion polls were published projecting a Labour vote share of 40%, 39% and 40%.

In the actual 2005 General Election Labour polled 35%.

So things can change. Everyone remembers how badly the 1992  and 2015 general election opinion polls and the Brexit predictions were wrong but all polls are subject to a margin of error.

In the 1997, 2001 and 2005 general elections the polls correctly suggested a Labour win but anyone who took them as a precise prediction would have greatly overstated the winning margin in share of the vote. In 1992, 2010, 2015 and 2017 most people were expecting a different outcome to what actually happened and, particularly in 1992 and 2015, polling error was a major part of that.

Of course, after 2015 most of the polling companies have adjusted their methodology to try to correct for a consistent pattern of overstating the relative position of the Labour party and understating that of the Conservatives.

This year's general election will be the first test of how well those corrections work. Until the votes are counted we will have no idea whether the polls are still overstating Labour support, are now about right, or - and this is the key point - have gone too far in the other direction and started to overstate Conservative support.

Do I think that Jeremy Corbyn is likely to win the election? No. My experience on the doorstep suggest that the polls are about right and that Labour is heading for a heavy defeat.

Do I think Conservative victory is certain? Absolutely not. You can never be certain of the result of an election until the actual votes are counted and declared.

Is the Conservative campaign scaremongering when they suggest that a Corbyn win is possible?

No, they really are not.

The electorate have the right to do whatever they wish when they get into the polling booth. And ironically the worst danger for the Conservatives is if our supporters assume that a Tory win is "nailed on" and don't bother to vote.

We have to campaign as if the polls were neck and neck as they were in 2015. We have every right to point to how disastrous a Corbyn government would be - because on June 8th the voters do have the ability to elect one. They deserve an explanation from the Conservatives as to why they shouldn't.

Theresa May has started to spell out positive reasons to vote Conservative and I am confident that she will do much more of this. But it is entirely fair to point out that the only realistic alternative candidate for Prime Minister - Jeremy Corbyn - simply is not up to the job.

No comments: