In the final weeks of the presidential campaign, there have been two major schools of thought about who is going to win. One school points to President Obama getting reelected for a very simple reason: polls show Mitt Romney losing Ohio and all of the states that he would need to win to make up for a loss in Ohio. The other school argues that state polls have been systematically biased against Romney by assuming that Democratic turnout will rival – or even exceed – elevated 2008 levels, when Obamamania was at its peak. Those who believe that Romney will win have pointed to polls showing him ahead among independents, predicted a late break toward the challenger, or pointed to economic fundamentals. To read the conventional wisdom pointing to an Obama win, check out Nate Silver. To read predictions of a Romney win, check out our own Michael Barone, as well as Dan McLaughlin, Ben Domenech and Jay Cost. Also read Ted Frank, who still thinks Obama will likely win, though he makes a strong devil’s advocate argument for how Romney could pull it out.
I believe the arguments about polls understating Romney’s position have some merit, but only up to a point. I also believe that by and large, despite some high profile errors, polling is generally accurate when results from multiple pollsters overwhelmingly point in one direction. So, I’ve decided to split the difference in my prediction. That is, I’ve given Romney the states that are essentially tied, in which he’s led in at least some recent polls. But in states where Romney has trailed in nearly all polls, and in some cases by a comfortable margin, I’m giving them to Obama. My thinking is that even if Romney over-performs the polls somewhat, he still is unlikely to over-perform by a wide enough margin to win these states.
Applying this philosophy, I give Romney Florida, North Carolina, Virginia and Colorado. But I assume that Obama takes Nevada, Pennsylvania, Michigan, Wisconsin and Ohio.
That already gives Obama sufficient electoral votes to win reelection, but still leaves out Iowa and New Hampshire. Obama has been ahead in most recent polls of the states, and has generally polled better in them than he has in Colorado or Virginia, but not as well as in Ohio or Wisconsin. Once again, I’ve decided to split the difference. I give Romney New Hampshire while giving Iowa to Obama. There are three reasons why I think Romney has a better chance of winning New Hampshire: 1) He spent much more time in the Granite State during the primary season and built a strong organization there. 2) New Hampshire is a more affluent state than Iowa, where populist-style attacks like those waged by the Obama campaign are less likely to succeed. 3) There is no early vote in New Hampshire, so Romney can win there on Election Day, while Democrats have already built up an edge in the early vote in Iowa. Even if the Democrats’ early vote lead in Iowa is down from 2008, it still should be sufficient given that Obama won the state by 9.5 points last time.
Add it all up and the final tally is Obama 277, Romney 261.
For what it’s worth, in 2008, I predicted a 338 to 200 Obama win. I missed his victories in North Carolina and Indiana.