The months and months of campaigning, the hundreds of millions of TV advertising, the incessant travel schedules of the candidates, and the vigorous efforts of both sides to get their vote out made little or no difference in the outcome of the Election of 2012.
To begin with, only two states — North Carolina and Indiana — changed sides. And the change in Obama’s vote share from ’08 to ’12 was pretty much the same whether it was in a swing state or not. The obsessive campaigning in swing states did not seem to have much effect.
This election was notable because of the massive concentration of advertising in eight “swing” states — Florida, Virginia, Colorado, Ohio, New Hampshire, North Carolina, Nevada, and Iowa. Well over 80 percent of the television advertising at the presidential level was directed at these key states. The candidates visited them over and over, cycling back around every few days. Except for fund raisers, incidental travel, or return trips to their headquarters in Boston and Chicago, Obama and Romney rarely set foot outside these swing states.
And it didn’t work.
In these eight swing states, all of which Obama carried in 2008, his vote share dropped by 2.1 percent from ’08 to ’12. In the twenty-one other non-swing states Obama carried — where neither candidate did much campaigning — his vote share drop was almost the same: 2.4 percent.
Among states McCain carried in 2008 (plus Indiana), the drop in Obama’s vote share was more significant: 3.2 percent.
Nor would it have made a difference if Obama’s vote share fell in the swing states by the same 2.4 percent that it fell in the non-swing states that went for Obama (as opposed to the 2.1 percent decline that in fact happened in the swing states). The 0.3 percent change would not have moved a single state to Romney from Obama. Florida, the closest, went to Obama by 0.8 percent!
So if ads and candidate campaigning did not move the dial, what did? Events.
The debates, the conventions, the storm coverage, Benghazi, the state of the economy, jobs data and all other events that affected all fifty states mattered. But the paid media, the in-person campaigning in swing states, and the massive ground game deployed by both sides accomplished nothing. Obama lost all the votes he was going to lose anyway in the swing states and Romney gained of the votes he was going to gain anyway in the swing states. Otherwise, how can one account for the virtually identical change from ’08 to ’12 in all the states Obama carried, whether swing states or not?
Two implications flow from these data:
1. Television is losing its impact. Particularly in the presidential race, it is astonishing that the almost one billion dollars spent advertising in eight states did very little to move the vote share. Voters are not watching television as much these days and those that are still turning it on are fast forwarding through the ads. And negative campaign ads — in fact, all ads, — are losing their impact.
2. Demographic voting is the new norm in America. You vote based on who you are, not where you live or how well each campaign has articulated its case. 93% of blacks, 70% of Latinos, 60% of those under 30, and 62% of single people, voted for Obama. And white married couples over 30 years of age voted for Romney. Not much else matters. A president who was elected and re-elected through identity politics has brought about a state of affairs where demographic voting determines the outcome. Our votes are predictable based on our race, ethnicity, age, and marital status well before anybody does any campaigning.
Even the vaunted ground games of the two parties didn’t do much. Voter turnout was eleven million lower than in 2008 — reversing the upward trends of the past four elections — and Obama’s vote share change from ’08 to ’12 was about the same in states where vigorous get out the vote campaigns raged and in those where they did not.
It is bad enough that America is now divided into red and blue states. It is divided into red and blue people as well based, not on their opinions, but on their demographics.