Predicting the Election: Will It Be As Close As People Think?
As many of you know, I have compiled an exceedingly good track record of predicting the outcome of presidential and congressional elections over the years. My best cycle was 2006, when I was off by just one seat when predicting how many seats the Democrats would pick up in their take over of the House. For over a year, I have been predicting that this presidential election would not ultimately be a close one, and that it would not be as close as 2008. This has been a bold prediction because I have never been able to point to any current empirical data that supports such a prediction. My judgment has been based on some data, but it's all historical in nature. For example, there's just no precedent in the last 150 years of an incumbent being reelected by a narrower margin that they were elected in the first place. It doesn't happen. The only exception you will find is FDR's third and fourth elections, but that's an anomaly that can't be constitutionally replicated. The rest of my judgment has rested on my estimation of the difference in quality between Barack Obama and the Democrats on the one hand, and Mitt Romney and the Republicans on the other hand. We've seen flawed candidates battle it out before. Nixon and Humphrey, Ford and Carter, and Gore and Bush definitely come to mind. But, when there's been a clear talent differential, as we saw between Reagan and Mondale, Clinton and Dole, and Obama and McCain, the end result has always been decisive.
Arguing against my prediction all along was polling data and a weak economy that argued persuasively in favor of this being a toss-up election, or even one in which the incumbent should be considered the underdog. In particular, the polling data has been extraordinary stubborn and stable all year long. While Obama has consistently held a lead, his lead has always been extremely narrow, particularly on the national (popular vote) level. Yet, even as the polls have been close, I have argued that Romney is losing, losing badly, and is going to get blown out. Nate Silver's analysis has largely backed this up, although his numbers have spoken much louder than his words.
The numbers have finally started to move decisively in the president's direction, beginning (roughly) with the onset of the Republican National Convention. Mr. Silver is now predicting that Obama will emerge from the conventions with an 8-10 point advantage in the national polls. Let me be honest. This is exactly what I expected to happen. What I want to know is why Mr. Silver says this:
You would figure that at some point over the past year, Mr. Romney would have pulled into the lead in the polls, given how close it has usually been. John McCain held occasional leads in 2008; John Kerry led for much of the summer in 2004; and Michael Dukakis had moments where he was well ahead of George H.W. Bush in the spring and summer of 1988. But Mr. Romney, if there have been moments when his polls were ever-so-slightly stronger or weaker, has never really had his moment in the sun.
Instead, the cases where one candidate led essentially from wire to wire have been associated with landslides: Bill Clinton in 1996, Ronald Reagan in 1984, Richard Nixon in 1972 and Dwight D. Eisenhower in 1952 and 1956.
There is almost no chance that Mr. Obama will win by those sort of margins. But this nevertheless seems like an inauspicious sign for Mr. Romney. If even at his high-water mark, he can only pull the race into a rough tie, what pitch can he come up with in October or November to suddenly put him over the top?
Due to the basic strength of the Republican Party in the South and the Plains States, I have to agree that Obama cannot win by margins reminiscent of Nixon in 1972 or Reagan in 1984, but I see no reason he can't win by margins reminiscent of Dwight Eisenhower or Bill Clinton. As Mr. Silver points out, candidates who never break into a clear lead during a whole campaign get blown out. That's what happens. We've never seen a candidate fail to break through at any point who nevertheless lost narrowly on election day.
So, how does this work? How might it work?
if Obama really opens up an eight-to-ten percentage point lead in the national polls, what will that mean for the state polling in Indiana and North Carolina and Georgia and Missouri and Arizona? If Obama is going to win by a bigger margin than he won in 2008, he's going to have to hold the states he won back then and win a couple states he lost.
What I have predicted is that this is exactly what will happen and, if it doesn't, I stand ready to eat my hat. I know that voter suppression will have an impact. I know the Democratic base is a little less enthused. I know Romney has a lot of money. All of those factors are worrisome, and they might ultimately combine to prove me wrong.
But history tells me that the people will make a decisive choice this time around. And I still believe that to be the case.