The results are in and I was wrong when I predicted that Mitt Romney would win 315 electoral votes. For those of you who sent in nasty emails and for those who sent in reasoned arguments that I was wrong, please be assured that I will be on a diet of crow for some time. For those who sent in positive or thankful emails, remember that none of us wants to live in a country where one party wins every election even though we tend to wish our party would win every time, and so disappointment is a necessary attribute of living in an electoral republic.
Congratulations are in order to a reelected President Barack Obama, who has won 303 electoral votes and appears to be on his way to winning 332 when the final votes in Florida are counted. (The Fox News decision desk, in whose company I worked on election night, decided not to call the state when the Miami-Dade vote counters decided to turn in for the night; there are many precincts uncounted there and many precincts in the county vote very differently).
I will still defend my prediction as reasonable, just as other predictions that either Obama or Romney would win, or would win with more than 300 electoral votes, were reasonable. You could look at the polling data, try to assess the balance of enthusiasm and consider which factors would be more important to voters and come up with pretty different conclusions. Remember that predictions of 290 (Larry Sabato’s prediction for Obama) or 315 aren’t necessarily predictions of popular vote landslides. Barack Obama won 52.9% of the vote in 2008 and won 365 electoral votes—more than anyone predicted, so far as I know, for either candidate.
The clear difference between national and state election polls, which Sean Trende noted in a really smart Real Clear Politics election day article, has been resolved, for the moment, in favor of the state polls. The percentage of the popular vote in this election, as I write after 2:00am Eastern election night, is 50% Obama and 49% Romney. When the entire count is in—and California, with one-eighths of the nation’s popular vote, takes a long time to turn its count in—it could end up Obama 51% and Romney 48%.
An electoral vote lead of 332-206, the likely outcome, is pretty wide given the narrowness of the popular vote margin; George W. Bush led John Kerry 51%-48% in the popular vote in 2004 and won the electoral college by only 286-252. But that is how our system works, and all the players and observers knew that going in. Including me. And, by the way, if you didn’t hear me before, I was just plain wrong.
I was wrong because the outcome of the election was not determined, as I thought it would be, by fundamentals. Some fundamentals, I thought, favored Obama. Americans like to think well of their presidents (and Obama’s approval ratings rose, slightly, over the fall) and many, perhaps most, Americans believe it would be a bad thing for Americans to be seen as rejecting the first black president.
On the other hand, most voters opposed Obama’s major policies and found unsatisfactory the sluggish economic recovery that seems to them to be the result—negative factors that seem to have been confirmed by responses to exit poll questions as they were by responses to poll questions for many months now. It is true that Obama won a second term by a reduced percentage and electoral vote margin, making him the first president to do so since Woodrow Wilson in 1916, nearly a century ago. This is not, I think, a grand triumph for his ideas or ideology. It is a triumph for his campaign strategists.
What happened? I think fundamentals were trumped by mechanics and, to a lesser extent, by demographics.
The Obama campaign strategists—and congratulations to them, by the way—argued that they would win by organizing and turning out the vote in the key states that would determine the outcome of the election. They had no illusions that they could expand the president’s appeal beyond the 53% of the popular vote of the 365/359 electoral votes they won in 2008; on the contrary, they conceded Indiana’s 11 electoral votes and the single electoral vote of the Nebraska 2nd congressional district even before the campaign started. They didn’t contest the 15 electoral votes of North Carolina very much after holding the 2012 Democratic National Convention there; they concentrated in their pre-convention negative anti-Romney advertising and in their organizational efforts on a three-state firewall of the next three states in order of Obama 2008 percentage, Florida (29 electoral votes), Ohio (18) and Virginia (13).
After Obama’s disastrous performance, and Romney’s sparkling performance, in the October 3 debate, it looked like the firewall was crumbling. But it seems to have held. The popular vote, as I write relying on 1:30am Eastern Associated Press numbers, is Obama 50%-49% in Florida, 50%-48% in Ohio and 50%-49% in Virginia. You could cite various polling and other evidence and suggest those states were going for Romney. This would make a big difference in electoral votes: with them Obama holds 332 electoral votes (his likely total), without them he’s at 272 (and loses the election if he loses one more state). But he seems to have held them, narrowly.
Which is to say that mechanics and, to a lesser extent demographics, determined the outcome of the election. The fundamentals which I pointed to simply did not operate at all powerfully in the results in the states that were target states from when the electoral cycle started, when the Obama campaign began advertising heavily against Mitt Romney when he clinched the Republican nomination in April. Compare the Romney-Obama numbers in the popular vote percentages (as of circa 1:30am election night Associated Press totals) or with the exit poll percentages (as of the same time in states with less than 80% of precincts then reporting) with the McCain-Obama popular vote percentages in 2008 and the Obama percentage in both elections. First, the three firewall states:
2012 2008 O diff
Florida 49-50 48-51 - 1
Ohio 48-50 47-51 - 1
Virginia 49-50 46-53 - 3
Then look at the other states that were target states all the way from April to November. All these comparisons are necessarily imprecise because I’m rounding out percentages and we don’t have final 2012 election results.
Colorado 48-48 45-54 - 6
Iowa 47-52 44-54 - 2
Nevada 45-51 43-55 - 4
New Hampshire 47-52 45-54 - 2
North Carolina 51-48 49-50 - 2
Only in Colorado was there real slippage for Obama beyond his slippage in national polling percentages from 53% to 50% or 51%.
Those of us who thought the Obama campaign was falling short of its goals in turning out early voters seem to have been proven wrong. Or at least they got enough out to prevent sufficient slippage to deprive him of these key electoral votes.
What about the states which became target states later in the campaign? Here the results look like:
Michigan 46-52 41-57 - 5
Minnesota 47-51 44-54 - 3
Pennsylvania 47-52 44-54 - 2
Wisconsin 47-52 42-56 - 4
These states came into play, but did not produce enough weakening of Obama margins to place any of their 56 electoral votes into the Romney column.
The overall story seems to be that fundamentals drove some numbers in this election, reducing Obama’s popular vote margin from 53%-46% to 50%-49% or 51%-48%–but only or almost only in the states that didn’t matter. If you aggregate the results from the Republican states that weren’t seriously contested (AL, AK, AZ, AR, GA, ID, IN, KS, KY, LA, MS, MO, MT, NE, ND, OK, SC, TN, TX, UT, WV, WY) and the Democratic states that weren’t seriously contested (CA, CT, DE, DC, HI, IL, ME, MD, MA, NJ, NM, NY, OR, RI, VT, WA), based on current AP results as of 1:30am Eastern or, when 90% of precincts weren’t reported, exit polls (which after the more educated massage this year were pretty closely on target: another story), and weight the 2012 results by the number of congressional districts (a reasonable proxy for population and, at this stage in the returns, for voter turnout), and compare them to the actual vote totals in those states (AK and HI reporting no 2012 election returns and exit polls, and so omitted for now) you get the following:
R states 58-40 55- 43 - 3
D states 40-58 38-62 - 4
In other words, fundamentals if they had been evenly applied in all states would have drive Obama’s percentages down by 3% or 4%, enough possibly to deprive him of a majority, and enough to deprive him of pluralities quite possibly in Florida, Ohio, Virginia, Colorado and perhaps enough of the other target states to give Romney an electoral vote majority. The trend apparent in the national polls toward Romney gains over previous elections among affluent voters was real, but it was either suppressed or offset in target states by stickiness to Obama. And so in the target states mechanics—organization, turnout efforts, early voting, etc.—trumped fundamentals and Obama won.
And demographics? The Hispanic turnout is not increasing as rapidly as in some projections. But it wasn’t good news for Romney. In Florida the exit poll showed him carrying Cuban-Americans by only 50%-47% and losing non-Cuban Hispanics by 68%-32%. AP showed lost Osceola and Orange Counties—DisneyWorld areas with large numbers of Puerto Ricans by larger margins than John McCain. By way of contrast, Republican Governor Rick Scott in Florida narrowly carried Hispanics in 2010. In Colorado he lost Hispanics 74%-25%–worse than McCain 2008—and in Nevada by 69%-24% (although it showed him carrying Asians there 49%-42%, puzzling to me since I think many are heavily Democratic Filipinos: more to learn on this).
So I was wrong. I take some pleasure in finding I have been wrong, because it’s an opportunity to learn more. As I prowl through the 2012 election statistics I will have an opportunity to learn much more about America and where we are today. A nation dissatisfied with the results of a Democratic president, Democratic Senate and Republican House has decided to return a Democratic president, Democratic Senate and Republican House. Lots to learn for all of us.