Why Most Postmortems of Virginia’s Gubernatorial Race Are Wrong

by Stuart Rothenberg November 14, 2013 · 9:56 AM EST

The dust has settled (mostly) from last week’s elections, so I thought it time to present a very different assessment of what happened in Virginia than the snapshot I’ve seen from others.

For example, Democracy Corps and Women’s Voices, Women Vote Action Fund distributed a wholly self-serving and unconvincing memo titled “Unmarried Women Cast Deciding Votes in Virginia Election.” It’s unconvincing, of course, because Republicans always lose unmarried women, regardless of an election’s outcome. Unmarried women are more liberal than most voters and are not part of any winning Republican coalition.

NBC’s Domenico Montanaro and The Washington Post’s Jonathan Capehart credited African-American turnout for Democrat Terry McAuliffe’s victory, as did Jamelle Bouie of The Daily Beast. Wrong as well, I’m afraid.

Others have noted, quite incorrectly, that the partisan makeup of the 2013 electorate wasn’t very different from the makeup of the 2012 electorate in Virginia, suggesting that Democrats have found some formula for turning out key voting groups in lower turnout elections that could help them offset what most expect to be a less Democratic-inclined electorate for the 2014 midterms.

While these assessments tell a part of the story and certainly should force Republican voters and strategists to take a clear-eyed look at the long-term prospects of the current GOP coalition, they don’t explain last week’s results in Virginia, nor do they offer meaningful insights into 2014.

While African-American voters made up one-fifth of the Virginia electorate in both 2012 and 2013 and Democrats held a partisan advantage in both contests (7 points in 2012 and 5 points in 2013), McAuliffe did not win because of the makeup of the electorate or the GOP’s weaknesses with black voters.

He won because Republican Ken Cuccinelli failed to get the same level of support from the normally Republican voting groups that Mitt Romney had a year earlier. And any complete comparison of the 2012 and 2013 electorates in the commonwealth suggests turnout problems for Democrats in next year’s midterms.

The 2013 Virginia electorate was older, wealthier, more married and, surprisingly, more male than the Virginia electorate during the presidential race just a year earlier. In other words, it was a measurably more Republican-looking electorate than the one that turned out in the commonwealth for President Barack Obama’s re-election, even with the impressive black turnout.

In 2012, the Virginia electorate, according to exit poll data on CNN’s website, was 53 percent female and 47 percent male. This year, the exit poll showed an electorate that was 51 percent female and 49 percent male.

Voters age 18-29 constituted a strong 19 percent of the electorate in 2012, but made up only 13 percent of this year’s electorate. On the other hand, voters 65 and older were 18 percent of this year’s Virginia electorate, 4 points more than the 14 percent they constituted in 2012.

This year’s electorate was also much wealthier. In 2012, 34 percent of voters made $100,000 a year or more, but this year 40 percent fell into that income category.

In 2012, only 62 percent of the electorate was married. This year, 67 percent said they were married.

So, in many important demographic categories, the 2013 electorate was much different than the one that showed up to the polls last year. And the differences should have favored the Republican ticket. That wasn’t true of race and partisanship, of course, but there is no reason to pull out only two categories for examination, especially given the richness of the exit poll data.

The survey data are pretty clear on why Cuccinelli lost. He lost because he was unable to match Romney’s percentages with key demographic groups that almost always vote Republican. Those voters showed up at the polls, but too many Romney voters crossed over to cast ballots for McAuliffe or Libertarian Robert Sarvis.

The Republican nominee for governor won a plurality of male voters (48 percent), but well below the 51 percent that Romney won in the state last year. Cuccinelli would have gained an additional 48,000 votes if he had matched Romney’s percentage, much of which would have come from McAuliffe, thereby completely erasing the Democrat’s 55,100 victory margin. (See Virginia’s total vote here.)

Add in white women (Romney won 59 percent of them in the state in 2012, while Cuccinelli won only 54 percent this year) or wealthy voters (Romney won 51 percent of voters earning at least $100,000 a year in Virginia in 2012, while Cuccinelli drew just 43 percent of them and lost the category to McAuliffe) and the Republican would have had a comfortable victory last week.

And if you don’t want to focus on gender, the marital status numbers tell the same story. Romney won 55 percent of married voters in Virginia last year, while Cuccinelli won only 50 percent of them this year. That’s about 75,400 fewer married voters than a Romney-like Republican gubernatorial nominee should have drawn.

Though you hear a lot about the changing face of the electorate, both nationally and in Virginia, that’s not why Cuccinelli lost last week.

The Virginia election in 2013 was one where the Republican nominee would have won merely by attracting the votes of the same people who voted for Mitt Romney. The party’s candidate for governor did not need to improve his showing among young voters, African-Americans, Hispanics or unmarried women. He just needed to get white guys and their wives.

That conclusion, which is based on an evaluation of all of the data, not on merely cherry-picking one or two variables, ought to be little comfort for Democratic strategists worrying about the makeup of the midterm electorate.