Electoral College Predictions and False Premises

The Electoral College doesn’t matter in the way pundits think it does.

November 7, 2011

Comments

Jonathan Bernstein makes a seemingly obvious but too-often-missed point about those trying to forecast the Electoral College:

States are not units which move independently of each other. They just aren’t. Instead, when Obama is more or less popular, or generically whenever one party becomes more popular, it produces more-or-less equal shifts across all states. If Obama wins by five points nationally, he’s going to win Pennsylvania and Michigan and Iowa and Wisconsin and, basically, the Al Gore states from 2000 plus a few others, and he’ll win easily. If he loses by five points nationally, he’s going to win what John Kerry won in 2004 minus a few states, and he’ll lose easily. He’s not going to run the same in most states but surge or drop dramatically in a handful of battleground states. It doesn’t work like that.

It is possible that the shape of the electoral college could give one party a relatively small advantage in very close races, but (1) that doesn’t appear to be true, and (2) the way to show that is to demonstrate that one party has more “wasted” votes (such as the huge Democratic majorities in New York or the huge Republican majorities in Utah). If you’re not doing that, don’t talk about the states this far out. And even if you are doing that, don’t talk about the states this far out. It’s possible we’ll have a very close election and quirks in the electoral college and distribution of votes by state will matter, but it’s far more likely that it will be just like most elections, with a clear national win one way or the other that is reflected and, for that matter, magnified by the electoral college.

Quite right. It’s true that there are “swing states” and that virtually nothing could happen to make the most hard core Red and Blue states flip. But the reason Florida was so close in 2000 and Ohio so close in 2004 wasn’t because of some unique local factors in those states but rather because 2000 and 2004 were very close elections on a national level. That naturally mirrors itself in the handful of states without a strong partisan lean.

2 responses to “Electoral College Predictions and False Premises”

PD Shaw

November 7, 2011 at 2:40 pm

I tend to see the states as constituting regional groupings that tend to move in similar directions. Chuck Todd has grouped the 11 potential swing states into three groups: five in the Midwest/Rust Belt (IA, MI, OH, PA, WI), three in the New South (FL, NC, VA), and three in the West (CO, NV, NM). So, I can see a Republican campaigning sucesfully in the Midwest with certain issues, which might be critically harmful in other regions. (E.g., illegal immigration)
anphang

November 7, 2011 at 5:06 pm

I wasn’t politically attentive for 2000, but I seem to recall Gore’s drop in the so-called Greater Appalachians being exceptionally precipitous, disproportionate to the shift against him in the rest of the country. Ditto but in the Democrat’s favor for virtually every county of Indiana in 2008 – not just outer-ring Chicago.