Monday 5 November 2012

'Calling' the Presidential election


So after all the billions of dollars; all the debating; all the screaming and shouting and hating and fearing: it's nearly all over.

And it's time to make calls.Who's going to win?

Now, let me be absolutely clear: this column, were it to have a vote, would cast it enthusiastically for President Obama, a leader who has hauled the United States and the world back from the brink of an economic precipice. If you have a vote, and you're reading this in the United States, I would urge you in the strongest possible terms to go out and use it - and to back to incumbent.

But I think it'll be tight. Really, really tight. A squeaker. Choose whatever term you want - you get my drift.

Right now, President Obama has a 0.4 per cent lead in the Real Clear Politics average, a 1.2 per cent lead from the website Pollster, and a 0.7 per cent lead on the Talking Points Memo website. Those numbers are unlikely to change much as the last polls trickle in. How did Real Clear Politics do in 2004 and 2008? Well, they were 1.2 per cent adrift from the final gap in 2004, and they were 0.3 per cent 'wrong' in 2008. Pollster, by the way, was also 0.3 per cent off the final result in 2008. Really close; this time, that close to the gap between the candidates means that either could win the popular vote. Statistical guru Nate Silver gives the President an 86 per cent chance of winning that (statistician that I am at heart) I just can't square with this sort of uncertainty. Intrade punters, betting their own money, are saying 67 per cent right now, and somewhere between those figures might be about right.

It'll be so tight that Americans will be up all night - and possibly biting their nails for many days to come (or at least until Ohio counts its provisional ballots on November 17). And arguing about Republican governors' vote suppression in Florida and Ohio: in the latter state, its Secretary of State is trying to chance election rules at the last moment in order to make it harder to understand and comply with ID rules. There is some possibility of an electoral college draw - in which case Governor Romney will almost certainly be chosen to sit in the White House by a Republican House of Representatives - or yet another Supreme Court tie-breaker will be required.

You don't have to take my word for it - you can look at the early voting data, which shows President Obama narrowly ahead (insofar as one can tell) in Iowa and Coloroda - states he won at a canter last time. The early voting figures tell us exactly what the polls do: it's basically neck-and-neck at a national level, with perhaps some key swing states leaning over a little to Mr Obama. Fewer Democrats have voted early this year when compared to 2008 - though enough have turned up to give the President most of the swing states. If, that is, Democrats haven't been cannibalising their election-day turnout. Because if they have, they're toast.

And all this, of course, assumes that the polls are right. Now don't listen to any crazy talk here: their record is actually pretty good, on the whole, as the numbers above imply. There is a small, but non-negligible chance that they are wrong. 'Correcting for house effects', a process conducted by most aggregating websites, assumes that right in the middle of the polling range, there's a 'zero' target that's right. Take that assumption away, and anything could happen. British voters, of course, have their own memories of this: in the 1992 General Election, most polls predicted a hung or undecided Parliament in which Labour was the biggest party. The Conservatives won by eight (count them - eight) percentage points. Mr Romney could do really well with conservatives who don't like talking to pollsters, upsetting the odds: Mr Obama might bring tens of thousands of Spanish-speaking voters to the polls, or appeal particularly to young cellphone users who are hard to reach. Your guess is as good as mine on this one, readers.

In the end, it'll all come down to turnout. If Democrats turn up at the polls, the President will be very, very narrowly re-elected. Many swing state polls are assuming electorates that look a bit like the Presidential election of 2008 - younger, less white and more female than others we've got used to. So if the electorate looks like the Republican wave year of 2012, Mr Romney will win quite easily. There's plenty of evidence this year that Democrats are a bit disheartened by what they perceive as 'their' man's so-so performance in office. Republicans are passionate, fired-up, desperate to get rid of a man they perceive as a European-style socialist at best, and possibly a 'foreign-born' threat to national security at worst. You can see that from tonight's Gallup Poll, which has Mr Obama up three amongst all registered voters - but down one against likely voters. Balance that against the fact that the demographics are changing, and states such as Nevada, New Mexico, Colorado and even Arizona are inexorably moving away from the Republican Party as their electorates become more hispanic... And you have a story of anger and pride against the shifting tectonic plates of America's racial balance.

So the bottom line: expect Obama. But really, really don't be surprised if you get Romney, a candidate who gets top marks only for obfuscation and mendacity: but who might become the most powerful man in the world tomorrow night.

That is all. Enjoy!