I don’t see it.
I don’t know if there are enough of us yet, but I know more and more people are just not going to vote.  McCain was the first obviously unacceptable candidate- the first lesser evil who was just too evil to vote for regardless of the evilness of Obama.  Romney is unacceptable too, and he shut down the Ron Paul people at the convention.  The GOP changed rules and strong armed people- there is another group of people who are simply not going to vote Republican, and probably not at all. 
I have a great personal dislike for the man, which I do not have for Obama.  I just think Obama is an idiot socialist; but I think the Republicans have been lying to me for a very long time.  The pretend to believe the stuff I believe, but even Paul Ryan voted for the bailouts.
The GOP deserves to go out of existence.
If Romney were remotely competent he’d be ahead in the polls since 9/11/2012, but he’s got some sort of personality disorder and he appears to have screwed up the political edge the Libyans handed him. Maybe he ekes out a win- no one is particularly enthused with Obama, but it seems to me traditional racial/class warfare tactics the democrats tend to employ will get more dumb people to the polls than Romney can pull- though the old folks still tend to vote religiously.

The polls are likely to be false because of the following mechanism: one of the questions asked is “do you consider yourself a Republican or a Democrat” [and variations of this]. Also questions about gender, occupation, race and religion. Then, the pollster takes the demographic data of how the various genders, ethnic groups and income levels voted *IN 2008* and guesses how different turnout will be this time.

Frankly, that’s not easy to do. But when you have a landslide election in 2008, estimating how different it will be is likely to be wrong.

The best way to poll would be 100 random people from each of the 435 congressional districts of the USA plus another 100 from Washington DC. Apart from checking that they are local I would ask three questions:
1) “Who do you think will win in your state?”
2) “Who do you think will win nationally?”
and
3) “Who would you like to win?”

I don’t believe anyone has ever tried this. Because the sample would be a) large and b) asks a non-partisan question, it wiouldn’t need demographic weighting (Wisdom of Crowds would come into play).

The election is based on an electoral college with 56 districts (50 states, DC, and the separate congressional votes in Maine and Nebraska). The local question would help forecast the actual result. The national one would help gauge the popular vote overall. And the “who do you want to win” especially compared with how people predict the outcome would make a good measure for turnout and enthusiasm.

While sympathetic to the “Let it bleed” argument that says it would be better for Obama to win so his socialistic policies would “own” the coming crisis, I’d still like Romney to win. Not so much because of anything to do with him, but because I am so sick of the smug dishonesty of the mainstream media.

In favour of the idea that Romney will win: what you said. Basically a repeat of the “shy Tory” effect from the UK 1992 election (which some US commenters are aware of and discussing). Possibly, also, the MSM attempts to depress Republican voters might instead fill them with vengeful vigour. It would be fun if the R’s won precisely because of this!

Sites like http://battlegroundwatch.com/ have plausible sounding reasons to suppose that the pollsters are using a turnout model too close to the historically unusual 2008 election.

However, against that -

Both that “47%” video as a gaffe and the on-the-ground fact that what he said there is, unfortunately, largely true. There is a vast class of dependants who will vote for “free” goodies to continue. It works. Gordon Brown did surprisingly well in our last election in part because he had created and expanded his client class.

I know you posted some good reasons to suppose that not all the dependants want to stay that way, but I can’t help thinking that these admirable people may be a small minority.

My bet: Romney by a squeak. But I would not bet much.

Just to back up your comment on voters in ‘08 thinking it was about time for a black president, the Guardian writer, Gary Younge, who is quite observant despite his daft politics, said something along the lines of “You can only vote for the first African American president once.”

Meant to say in my last comment “Just to back up your comment on voters in ‘08 thinking it was about time for a black president but this effect being a one-off”

Antoine, can you think of anyone whose interests would be well enough served by your “wisdom of crowds” poll to pay for it?

Are the Romney campaign sure enough in their claim that the turnout estimates are wrong to want to pay for a poll such as you describe?

When I saw all the emails for these comments, I just assumed it was all spam.  But it was all genuine!

Natalie: strongly agree about the reason for Romney winning being to kick the mainstream media in the gonads, rather than expecting anything of Romney.

Also: I actually said, on the afternoon of the 1992 election, when I breezed into the ASI to announce my finding, that Major would win with an absolute majority of 50 seats.  The consensus then being: hung Parliament or Labour win.  Final count: nearer 30, but not bad.  My thinking was as you describe.

The moment one stops looking for an African American with a PhD living in Idaho or a Baptist gun-owning single mother of five in San Francisco, opinion polling should be a lot cheaper. Perhaps if the polls get it badly wrong this sort of approach might get tried out.

One of the curious problems pollsters have is the system for charging people for mobile phones: you pay to receive calls (only happens in the EU if you’re “roaming"). As a result, pollsters aren’t allowed to call cell phones therefore making it harder to get a demographic spread: younger voters are less likely to have a land line.

So they “adjust” the results, or as Michael Crichton put in in State of Fear “make a guess.”

I think Perry is right ("let it bleed” as Natalie says). But I struggle to actually *want* Obama to win. I honestly have no idea. Isn’t there some sort of betting market that would make better predictions than polls?

Rob: the Iowa Electronic Market is often cited. It’s essentially a futures market in which you can buy or sell election outcomes. It is small sums only market principally for research purposes, but its record is reasonably good. It’s presently showing Obama slightly ahead.

You can just go to a betting exchange such as Betfair. This is presently showing Obama miles ahead. However, it is a British market, on which which it is illegal for people resident in the US to bet on, so this might be showing foreigners’ biases more than prospective outcomes. (I wonder if it is possible to arbitrage between the two markets. I must try it. Even if you can’t - a return of better than £5 for a £1 bet on Romney is surely a good bet).

On the post, though, I see two factors slightly helping Romney compared to the last election. The first is the “This is my historic opportunity to right historic wrongs by electing the first black president” factor that existed last time and does not exist this time. Quite a few people with politics nowhere near as left as Obama and who should have known better decided to believe that Obama was in fact something more moderate than the Chicago machine politician with far left roots that he is. That was a factor that would only help Obama once, and many of these people have thought better about this since and will not vote for him again. 

The second is possibly my imagination, but the main US media seems so obviously and preposterously biased this time to the point that it is laughable, and to an extent I have not seen before. It could be that I have changed rather than the media, though, so I am not sure how much credence I should give this. However, it feels that at least some people are going to respond to the media bias with “Don’t insult me.” rather than believing it. But as I said, I may be deluded on this one.

In Obama’s favour though, you have the advantages of incumbency, which are considerable. There is no better way to look presidential than to be president, as the trappings of office are designed to make the holder look presidential. Any challenger has to fight this, which is why most American presidents do in fact get elected to a second terms.

The media is presently reporting that Romney is running a terrible campaign. Most of the media would report this whether or not he actually was running a terrible campaign, but I am seeing it in places like Fox News and various Republican sympathetic blogs too, which makes me thing he might actually be running a terrible campaign. And let’s face it, he is an unpleasant man and a fairly bad candidate.

So on those general feelings, I see Romney with a chance but with various factors against him. And one really must go state by state, look at the local and vested interests, and do the electoral math (as the Americans say). Do that, and I can’t find a Romney landslide. In fact, a Romney landslide is preposterous. He might win, although I am less confident of this than I was when I last spoke to you about this (as you quoted at the top of the post). I am thinking Obama by a small to smallish margin at the moment.

On the other hand, you called the debate right.

Is it online somewhere? I have a couple of long train rides today.

It’s still listening to politicians talk, and I can’t generally bear that any more.

I could get something like 5.05 for Romney on Betfair last night, and it’s now down to 3.85, so certainly that debate had an impact. I could have made good money just by betting on Romney last night and against him this morning.

Sometimes I think that Obama doesn’t actually want to win.
It is very possible to get Romney angry, flustered, and have him say things he doesn’t really want to say. 
From what little I’ve heard, (because I can’t watch them lying anymore) it sounds like Obama didn’t even try.

It wouldn’t seem to much of a stretch to me to think Wall Street is running the whole show now.  Obama got in because he’s a complaint tool, but now he’s up against one of the finance world’s own.  They’ll lock down the private profit, public risk/losses model and keep making us pay for their mistakes until there isn’t anything left.

Michael - shame about those odds. The only reason I didn’t bet yesterday after reading your comment is that I’ve been having trouble logging in to my account.

You can only get 3.05 on Romney now. Buying at 3.85 after the debate would have been a good buy, too. (And obviously, so will 3.05 if he actually wins). I think the debate was the moment when the mainstream on this side of the point discovered that Romney might actually be a chance of winning the election). The Iowa market has hardly moved, though, so it might still be that it is perception outside the US that has moved more than the actual election. I am still personally saying it will be close by which I mean an electoral college margin probably of less than 50 votes and certainly less than 100 - for reference Obama had a 208 vote margin in 2008, and George Bush a margin of 35 in 2004.

And lord. Did we really get this close to President John Kerry. It’s frightening.