The Virginia Governor’s Race Is Not A Good Political Barometer

Virginia is about to elect a new governor, and some Democratic operatives are worried. Worried that the public polls that over the past month have given Democrat Ralph Northam an average lead of 7 percentage points over Republican Ed Gillespie are off. Worried about a relative lack of small-dollar and outside donations to Northam. Worried, most of all, that Virginia won’t be the launching pad into the 2018 midterm elections that they had hoped for.

Maybe Democrats should fret. Virginia is a key swing state, after all, and Gillespie is running hard to the right on immigration and cultural issues, just as President Trump did in 2016. Republicans may try to copy the Trump-Gillespie playbook if the Virginia GOP candidate emerges victorious.

But regardless of whether Northam wins, I’d urge you not to read too much into Virginia’s governor’s race. State politics are different from national politics. Even if Northam isn’t raking in cash hand over fist, he has still outraised Gillespie and pulled in nearly three times as many donations of less than $100 as the Republican. More than that, gubernatorial elections — especially in Virginia — don’t necessarily reflect the national mood.

The best way to see that this race is not necessarily a reliable barometer for the rest of the country is to compare the result in past Virginia gubernatorial elections with the national House vote in the following midterm. Rather than use the raw Virginia result, though, we want to adjust for the state’s “partisan lean,” a measure of how the state would vote in a neutral political environment in that year.¹³

Let’s use the 1993 Virginia governor’s race as an example. Back then, Virginia had a Republican lean of +10.6 percentage points — meaning that if there had been a presidential race in 1993 and that race had ended up tied 50-50 nationally, we’d expect the Republican to win in Virginia by 10.6 points. Republican George Allen won the Virginia governor’s race by 17.4 points that year, and if we had used that race to gauge the national environment, we would have concluded that the country favored Republicans by nearly 7 points. The next year, Republicans won the national House vote by 7 points. Unfortunately, most Virginia governor elections haven’t been so predictive.

The average difference between Democrats’ over- or underperformance in Virginia and the following national House vote has been 7 percentage points. That’s a pretty big miss — and just a few points could make or break the Democrats’ shot at a House majority.

In 2009, for example, Republican Bob McDonnell’s landslide victory over Democrat Creigh Deeds accurately presaged Republicans taking over the House in 2010, though the national House margin ended up being much smaller. Four years later, however, Democrat Terry McAuliffe’s win over Republican Ken Cuccinelli gave Democrats false hope ahead of the 2014 midterms, in which they lost control of the Senate. The 2001 Virginia governor election was similarly misleading: Democrat Mark Warner outperformed the presidential lean of the state by 14 points, and the following year Republicans easily held onto control in the House.

So why isn’t the Virginia election a good national barometer?

First, one election is never a reliable barometer of much of anything — the same way watching one NBA game doesn’t tell you much about how the whole season will play out. That’s why, for special elections, I’ve been arguing that we need to look at the results as a group to get a sense where the national political environment is.

Second, while most politics have become nationalized (people vote for one party up and down the ballot), governor’s races are still an exception. How else do you explain Republican governors in the very blue states of Maryland, Massachusetts and Vermont? Or the fact that Democrats control the governorships of Louisiana and Montana?

In the 2016 elections, there was barely any relationship between the outcome in the 12 gubernatorial elections and the presidential vote in those states. In theory, you’d expect the presidential lean of a state to have its greatest impact when presidential candidates are literally on the ballot. Yet there was a 17 percentage point difference, on average, between gubernatorial and presidential results in these 12 states.

So even if Northam or Gillespie wins by 10 points, we really won’t be able to say that much about what it means for 2018.

Now, the Virginia governor’s race will matter in other ways. It could, for instance, drive turnout for state legislative races, which did a good job in 2013 of gauging the midterm environment. It could serve as a morale boost (or a downer) for Democrats as they look to recruit candidates for the 2018 cycle. The race also matters because if Democrats don’t take an interest in state politics, they won’t control the redistricting process in many states after the next U.S. Census. The way congressional lines are drawn helps explain at least some of the reason that Democrats have won fewer House seats than their vote share would indicate they should have in the past few cycles.

So the Virginia race matters. It’s just not a very good crystal ball.