Could Non-Citizen Voting Have Tipped the Virginia Senate Race?

by Andrew Biggs

On October 24, two political scientists from Old Dominion University — Jesse Richman and David Earnest — published an article in the Washington Post summarizing their research on voting by non-citizens, which is not legal. (The voting, that is; as far as I know the research is perfectly legal.) Click here for their more technical follow-up. Richman and Earnest cite data from the Cooperative Congressional Election Study (CCES) showing that 6.4 percent of non-citizens stated that they voted in the 2008 election and 2.2 percent voted in 2010. Given that 80 percent of non-citizens in the survey stated that they voted for President Obama, Richman and Earnest conclude that in certain very tight races — they cite the 2008 Minnesota Senate election — non-citizen voting could tip the balance.

By coincidence, the Virginia Senate race between Democrat Mark Warner and Republican Ed Gillespie has turned out to be extremely close. At the time I write this, with 99.2 percent of precincts reporting (which I believe is around 95 percent of the total vote), Warner leads Gillespie by around 12,000 votes out of more than 2 million cast.

This provides a potential test case for Richman and Earnest’s theory — though, to be clear, the results could vary significantly with small changes in assumptions. To start, in 2008 around 6 percent of Virginia’s population was reported to be non-citizens. If Virginia non-citizens turned out to vote at the same 2.2 percent rate as in the 2010 off-year election, that would make for 10,904 non-citizen voters, based on the current Virginia population of 8.26 million. If we assume that 80 percent voted Democratic and 20 percent Republican, that would produce a net gain in Democratic votes of 6,542, equal to about half the current gap between the candidates.

But as Richman and Earnest note, their figures are based on a national sample and voting behavior could differ between states. In some states non-citizens might be more or less likely to vote, and more or less likely to voter for either party, than in others. (Exit polls for the Virginia race didn’t have a large enough sample to break down Hispanic voters by party.) The percentage of non-citizen residents may differ from the percentage of non-citizens of voting age, if non-citizens are younger on average.

My best guess is that non-citizen voting wouldn’t have been enough to shift the results of the Virginia Senate race, but that’s only a guess. If the Gillespie campaign team was feeling dedicated, it might look at an alternative approach also recently in the news in Maryland: comparing voter rolls to jury rolls, where non-citizens have an incentive to declare their status in order to avoid jury duty. I have no idea what that would show. What’s important is people trying to address these issues in a rigorous way. 

— Andrew G. Biggs is a resident scholar at the American Enterprise Institute.

The Corner

The one and only.