Many people think of the off year Virginia Governor's elections as being like a "midterm" election for Congress. True, the turnout is lower than Presidential years but it isn't quite the same electorate. This is one of the biggest mistakes some national consultants make when working in Virginia in off year elections because the differences between a midterm electorate and a Governor's year electorate here are substantial both in targeting voters- but also in targeting issues.
The 2010 midterms did not have a statewide Senate race, so we need to look back at 2006 to really compare. To make that as even of a comparison as possible, we'll match it up with the 2005 Governor's election here in Virginia.
In the 2005 Governor's election there were 1,983,778 votes cast. In 2006 there were 2,370,445 votes cast. That's a difference of 383,667 votes or more than a 16.2% shift in the electorate opting to show up.
These turnout differences are not even across the Commonwealth. For example, the City of Richmond which as the Capital area always shows more interest in state politics than the rest of Virginia had 51,238 voters come out in 2005 while 53,769 voted in the 2006 midterms. Those 2,531 voters are a 4.7% turnout bump, well below the 16.2% statewide average. The same applies in Henrico County (just outside Richmond) where 92,359 voted in 2005 and 101,017 voted in 2006. That's an 8.6% turnout bump- well below the 16.2% statewide average.
But in Arlington County- home of voters more interested in federal races than state elections the turnout bump was much larger. 56,989 voted in 2005 while 73,069 voters came out in 2006. That's a bump of 22.0%, well above the statewide average of 16.2%. The bump was similar in Virginia Beach City where 96,889 voted in 2005 but 124,050 voted in 2006- a move up in turnout of 21.9%.
In short the Richmond metro area is much more important in state election cycles than in federal ones.
But there are even more differences than just regional in the turnout models.
According to the statewide voter file in 2011, there was a 4 year age difference in the median voter in Governor years turnout than in midterm elections. It's important to note this is median- not average- so a four year jump means the average age change is even higher. This median change occurs as voters opt out of the electorate. This jump causes an increase in the electorate of about 7% of voters who are no longer in the workforce (retired, although not always by choice).
Meanwhile, what voters don't fall off? Voters between 65 and 85 are the strongest Governor year voters, with over 80% that vote in Presidential elections returning to the polls and over 90% of the midterm electorate coming back in Governor year elections.
That's where we get to our most important demographic of all. Life expectancy. The difference in life expectancy between blacks and whites has been getting smaller- but are still at a very important 5.4 year difference for men and 3.7 year difference for women. Since seniors are the most reliable voters in Governor year elections this causes a huge turnout gap. Check out this chart to see why:
Virginia Seniors (65+)
Black Men (Obama 95/5) avg. live for 6 election cycles post 65
White Men (Romney 76/24) avg. live for 11 election cycles post 65
Black Women (Obama 96/4) avg. live for 12 election cycles post 65
White Women (Romney 64/36) avg. live for 16 election cycles post 65.
In a perfect 50-50 statewide split, let's say the black women's vote exactly cancelled out the white men vote with similar life expectancies. White senior women- the most conservative and GOP leaning group of women in Virginia- then live for 10 more election cycles than black men. When all the younger people are participating in a Presidential year you might not notice this- but in a Governor's election when older voters increase their share of the electorate this is a really important stat.
The Democratic victories for Warner and Kaine in 2001 and 2005 masked some of these demographic issues with good candidates and unique electoral environments at their time. Warner outspent his opponent more than 2-1 to win by five points, then Kaine had Warner at 80% approval when he was running to help him in 2005. But in 1993, 1997, and 2009 the GOP candidates for Governor not only prevailed- but won in huge landslides of double digit margins. This trend really started as the seniors who grew up in the great depression era slowly began dying off. Those older voters tended to be much more Democratic- especially white seniors- and that allowed Democrats to win low turnout Governor's elections in the 1980's. Without them the next generation of white seniors have been trending more and more Republican every cycle- and with their longer life expectancy have been able to dominate the electorate.
None of this makes it impossible for Democrats to win- just very difficult. Throw in a party structure that hasn't spent much time reaching out to lower participation voters to get them interested and a less than perfect top of the ticket and the path for 2013 becomes even harder.