I have been doing politics full time for over 30 years now, and in most of those elections, even the full-time pols with great insider polling and analysis to look at had some big surprises when the results came in. Sometimes these surprises were on individual races, but much of the time, they were on the overall election pattern.
In 1980, my first election year working politics, most of the experts said the Presidential race would be tight, and no one predicted the kind of tide that swept in a Republican Senate. In 1984, everyone rightly predicted the Reagan landslide, but few people thought that two of the most openly progressive Senators elected in the last half century (Tom Harkin and Paul Simon) would buck the trend and upset Republican incumbents in the Senate. In 1986, few experts predicted the Democrats would win most of the close races and retake the Senate. In 1990, pundits were surprised by the level of backlash against (the first) Bush and the number of seats Democrats picked up. In 1992, most politicos predicted Perot would fade in the end- few thought he would get 19 percent of the vote. In 1994, the experts knew Democrats would have a bad year, but almost no one predicted the Republicans would win both houses of Congress. In 1998, with the Lewinsky scandal looming over the political landscape, most predictions were for Republican pick-ups of 20 to 30 House seats; when the Democrats instead picked up 5, embarrassed Republicans stripped Newt Gingrich of his Speakership. In 2000, The Bush team was so confident going into election weekend that they were having Bush campaigning for other Republicans -- they never expected the high levels of African-American turnout that caused the close election in FL and meant that most other swing states went for Gore. In 2004, everyone was blown away by the high levels of Republican base turnout, in swing states and nationwide, that not only allowed Bush to win but swept in several new Republican Senators. In 2006 and 2008 both, people generally knew it was going to be a Democratic year, but almost no one predicted that Democrats would win almost every close Senate race and go from 45 to 51 in 2006, and 51 to 60 in 2008. And in 2010, everyone knew there would be a strong Republican tide, but few predicted that it would be so historically big that it would mean a 63 seat pick-up in the House.
If you look at the patterns in all these elections, it isn't that the polling is necessarily so far off or that the predictions get all the big things wrong. But in politics, a little change here and a little change there can mean a big difference. It's why no matter how expert you are that no one in this business is ever 100 percent on the mark. Very modest changes in who turns out and how different demographic groups break really matter, and even a last-minute change of 1 percent can swing a whole bunch of elections. Remember too that polling is always 24-48 hours behind what is actually happening in the real world, which if something is starting to move really matters. And polling, of course, doesn't do a very good job of measuring on-the-ground field ops effectiveness.
One other factor which has an impact on the surprise thing is an intangible which is hard to measure or account for in advance, or even figure out why after the election in the post-mortem: in many election cycles, one party or another wins most of the close races. And it doesn't necessarily have all that much to do with which party is having the better year overall. To add to the complication factor, in several recent cycles, one party has won most of the close Senate races and the other most of the close House races: in 2006, for example, which was a very Democratic year overall, we Dems won every close Senate seat but one, yet lost over 60 percent of the closest House races- even though we picked up 31 seats and won back the House, we still left more seats than we should have on the table. If they are the ones to win, party committees will sometimes argue that that their superior field and communications operations were the reasons they won the close ones, but I have never seen any hard evidence of that one way or another. I am sure there is a lot of random luck to the whole phenomena. But I will say this: in most of the elections of the last few cycles, the party having a better year won at least most of the close Senate races. 2010, where Ds and Rs split the closest races, was actually the big exception. In 2002 and 2004, the Republicans won most of the closest races; in 2000, 2006, and 2008, Democrats won most of the closest.
So will 2012 give us some surprises? Almost certainly. The thing about surprises is that they are hard to predict, but the biggest surprise for me would be an election result that looked exactly like the current polling and prognostication predicts.
My great fear, naturally, is that Gallup and Rasmussen could put all the other pollsters to shame, the undecideds could all break against the incumbent, and Romney could end up winning by several percentage points and sweeping in a Republican Senate. But I am feeling optimistic today, so my guess is that the surprises could very well be on our side of the ledger.
One gut feeling I have is that we may end up a little further ahead nationally and in most of the swing states than it now looks. I tend to agree with Greenberg's argument on cell phones being under-polled in a lot of the national polling; I tend to think that with so many of the undecided voters being young unmarried women, we have a good shot at getting more of them in the end than Romney does; I continue to think our field operation is out-performing Romney's; and the great job that Obama is doing re disaster relief may help him in the end as well. If even 2 or 3 of those hunches turn out to be true, we could win the national popular vote by 4 points, and most of the swing states by at least that much; if all 4 of them are true, we could be looking at a final margin of 5 points.
When it comes to the Senate, the biggest question in my mind, the thing that could turn into a surprise, is if one party or another, for whatever mysterious set of reasons, ends up winning most of the closest races. Given that there are now 15 races considered by both parties to be competitive, a strong trend by either party in winning the close ones will be a big deal. One other thing to add here: if Obama does end up with a 4-5 point victory, that will almost certainly help some Dem Senate candidates win close races.
The biggest surprise of all according to Conventional Wisdom would be the Democrats winning the House. I am not going to predict that one -- I am not that much of an optimist. The money that Rove and Co. are spending on House races makes that scenario very tough. But this scenario is not as impossible as the CW would have it. If my optimistic hunch about Obama turns out to be right and he is winning by 4 or 5, that will definitely give Dem House candidates in close races a boost. And a superior field operation for Obama in FL, NC, VA, NH, PA, OH, MI, MN, IA, CO, and NV could well bring several House seats home. In the end, Democratic House candidates will have to win over 60 percent of the closest races to pull this off, but that isn't out of the question.
One final note that gives a Democratic surprise scenario a little more hope: we continue to win the basic argument in this race. Romney has had to move our way rhetorically, the tea partiers have been forced on their heels. On the issues and on values, we are winning this year's debate.
Wouldn't that be amazing if after all this talk of the closest election ever, Democrats ended up with a major trifecta sweep? Yeah, I know, I'm being more than a little optimistic here. But a boy can dream, can't he? And I guarantee you, having lived through big surprise elections like 1980, 1994, 1998, 2000: stranger things have definitely happened.