My regular correspondent nicknamed “Number-Cruncher” — not to be mistaken for NRO Corner contributor Josh Jordan, who uses the Twitter handle “Numbermuncher” — checks in with an interesting look at how swing states have compared to the national popular vote in recent cycles.
I think the chart below blows away some of this conventional wisdom that Obama can win the electoral college but lose by a significant margin in the popular vote (e.g. greater than 2 percentage points). First and foremost that doesn’t happen often, and the way the map is drawn today I seriously doubt it will happen again. When the popular vote is close, the Electoral College will be, too. I went ahead and looked at the past three elections and compared the individual state results to the national results and came away with a pretty interesting picture. I have included the raw data below, but want to point out a few bullets.
1. In the past three cycles, the Republican candidate has done better in Florida, North Carolina, and Virginia that he has nationwide. This is why Obama will gradually move on from these states.
2. In Ohio, which has proven to be the most important state in the union over the past three elections, the GOP candidate done better than the national popular vote, except for 2004, where it was almost spot on (e.g. -0.2%). Chalk that up to massive campaigning on both sides, basically making Ohio a microcosm of the United States. Any poll that tells you that Obama is winning Ohio by 5 points when the national numbers are tied or slightly with Romney is, to put it mildly, whacked. On average over the past three cycles, Ohio goes 2.13 percentage points more towards the GOP presidential candidate than the national vote total.
3. Pennsylvania is very much in play. The Keystone State has been remarkably consistent with regard to following the national result, with the Democrat on average performing four percentage points better than he does in the national vote.
4. A Democrat pollster would argue a little with me that there are demographic changes to worry about, especially in Oregon, Colorado, and Nevada. While I know Colorado has undergone some changes, I am not sure I buy it totally. I have another theory.: in these Western states, there may have been an election day bandwagon effect. Especially in Nevada and Oregon. Afternoon voters knew the election was over in 2008 well before they got to the polls (and thus we saw the large swing).
I went back and checked; Obama was not declared the winner by the networks until after the polls had closed in California and Washington, and Oregon is largely a vote-by-mail state. Having said that, there may have been a bit of a time-zone effect, as late West Coast voters may have seen the results from the earliest states. Also note last cycle may be tough to measure the phenomenon of voters not bothering to vote because they think their guy doesn’t have a chance, because a lot of people were convinced Obama was a sure winner before anyone showed up to vote on the morning of Election Day 2008.
If Pennsylvania falls into Romney’s hands, I would not be surprised if you would see the reverse of this trend in 2012, whereby the Western states will see a Republican surge based on Election Day results in other states.
Keep in mind the polls close at 8 p.m. locally in Pennsylvania, and if the state is close, a Romney win won’t be announced until considerably later in the evening.
Anyway there is a lot to look at with this data. But with the Democrats still pushing the “Obama’s win is inevitable” narrative, except now they’re basing it on the Electoral College. Looking at the real data should pour a lot of cold water on that theory.
Number-Cruncher’s chart can be found below; it’s a little wide for NRO’s pages, so I’ve broken it into two pieces for legibility below its resized version.
Here’s the portion of the chart simply calculating the margins in the swing states, and how they compare to the national margin:
Here’s the difference between the state’s result and the national popular vote in percentage points, averaged for the past three cycles. Of course, there’s no guarantee that 2012 will be an “average” year.