For more on the June 18 primaries, see our voter guide.
After the presidential primary in March, some readers (primarily Donald Trump supporters) suggested to me that Nikki Haley’s totals were inflated by Democrats who had crossed over to cast ballots in the Republican primary.
I discounted that, but after my recent column where I listed five factors that would determine the outcome of the hotly contested 5th District Republican primary between Rep. Bob Good and state Sen. John McGuire, I heard from a reader who felt I had ignored what he called “The Trojan Elephant” in the room: the prospect of Democratic crossover voting influencing that outcome.
That reader was David Beiler of Stafford County. He was formerly with Campaigns & Elections magazine, was the national political director for Ross Perot’s 1996 presidential campaign, and is a former elected official himself — a former Stafford County supervisor who claims the distinction of being the first person in the country to be elected under the Reform Party banner.
He offered some estimates on how much crossover voting we might expect in the 5th. But before we get to that, let’s set the stage.
Virginia is one of 20 states that don’t register voters by party. That means when a political party holds a primary, any registered voter can cast a ballot. There’s nothing except a conscience to stop the chairman of one party from voting in the other side’s primary. That’s been a particular point of contention for many Republicans in Virginia; it’s one reason some have favored conventions over primaries. But a new state law makes it all but impossible to nominate candidates any other way — as some Lynchburg Republicans discovered to their chagrin earlier this year.
Virginia has one famous example of crossover voting influencing an outcome, although it happened 75 years ago. In 1949, when the Byrd Machine dominated the Democratic Party and Republicans were of little electoral consequence in Virginia, Francis Pickens Miller challenged Byrd’s candidate in the Democratic primary for governor. He almost won — and probably would have if Republicans hadn’t voted. Party leader Henry Wise warned that if Republicans didn’t intervene, the liberal Miller might win. “Alien influences,” Wise said. By all accounts, a lot of Republicans that year did vote in the Democratic battle, and some believe they made the difference on behalf of Democrat John Battle. Historian Virginius Dabney, who was around back then, writes that “it seems probable that Battle owed his nomination to the GOP.” Miller certainly believed it. He claimed to the Richmond Times-Dispatch that 50,000 Republicans voted — and Battle won by just under 23,000 votes. Some Republican leaders believed it, too. Then-state Sen. Tad Dalton, R-Radford, declared: “The Byrd organization may deny that it owes its political neck in the state government to the Republicans . . . but the county and precinct workers know otherwise.”
Without that crossover voting, the course of Virginia history might have been very different.
A more recent example is the 2014 Republican primary in the 7th Congressional District, where Dave Brat challenged incumbent Eric Cantor. Richmond Times-Dispatch columnist Jeff Schapiro recently revisited that race: “Democrats, particularly in Henrico and Chesterfield counties, went into the Cantor-Brat primary hoping for a twofer: ending Cantor’s career and Brat’s — before it began.” Brat won the primary, which those crossover Democrats cheered, but then won in November, which they did not.
Because Virginia doesn’t register voters by party, it’s hard to say how much crossover voting happens but, the point is: It does happen.
So what about Haley’s vote in Virginia? She took 34.98% of the vote — 244,586 in all. Skeptics point out that she won in the more liberal parts of the state — she carried parts of Northern Virginia, Charlottesville and Albemarle County and Richmond. Surely that’s because Democrats voted, right? After all, there wasn’t much reason for Democrats to vote in their own presidential primary this year. I’m skeptical. Trump has always been weakest in the suburbs; it seems more likely to me that the Republican voters in those areas are of a more moderate variety. Haley’s vote patterns in Virginia mirrored those of Marco Rubio in 2016 — except that Rubio ran stronger.
Without party registration, though, we can’t know for sure. Here’s what we can do, though: look at how Haley has fared in states with closed primaries and compare those results with states with open primaries.
On March 5, the same day that Virginia held its primary, so did 14 other states. In some states with closed primaries, Haley polled better than she did in states with open primaries. For instance, in Maine (closed), she took 25.4% of the vote. In Alabama (open), she took 13.0% of the vote. In primaries since she’s dropped out, Haley did better in closed Connecticut (13.9%) than in open Wisconsin (12.8%).
While we obviously can’t rule out that some Democrats voted in the Republican presidential primary, what I see are naturally occuring regional differences. Some Trump supporters, though, say that more than 20% of the voters in Virginia’s Republican presidential primary were really Democrats. If we assume (as I suspect those Trump supporters do) that those Democrats voted for Haley, then about a little more half her votes came from Democrats. If that’s so, then her “true” vote would have been about 19% of the total, not the 35% it was.
Let’s take a closer look, by zeroing in on some strongly Democratic precincts. In Lynchburg, those would be in the city’s Second Ward. In the 2020 presidential election, one of the four precincts there voted Republican and three voted Democratic, with totals ranging from 62.4% to 69.4% Democratic. If there were a lot of Democratic crossover voting, that might be one place we’d see it. However, in the March 2024 primary, Trump won all those precincts. In two of them, his share of the vote exceeded his statewide margins (63.7% and 84.4% compared to 62.99% statewide). In the third precinct, he took close to that (57.92% of the vote). If there was a lot of crossover voting, it must have been somewhere else. However, I do notice this: In those three precincts, where usually about two-thirds of the voters cast Democratic ballots, the proportions were quite different in the 2024 primary. For instance, in the Second Ward’s Third Precinct, Republicans cast 45 ballots, but Democrats only 55. Does this suggest that there was a lot of crossover voting? I’m not persuaded. In 2016, when both parties had well-contested presidential primaries, there were 58 Republican ballots in that precinct, but 170 Democratic ones. What I see is a general lack of interest in the Democratic primary this year, either because there wasn’t much of a race, or because of general lack of enthusiasm for Joe Biden in general (something last week’s Roanoke College poll highlighted).
How about Charlottesville, where Haley ran up the score? The first thing I noticed is that in Charlottesville, more people voted in the inconsequential Democratic primary (3,020) than in the Republican primary (1,634). That’s about what I’d expect in a strongly Democratic place like Charlottesville. However, here is one figure that might suggest there was some crossover voting in Charlottesville in March. Those 1,634 voters in the Republican primary represent 52.8% of the 3,094 votes that Trump took there in the 2020 general election. That seems an awfully high percentage to me, since statewide the Republican vote this year accounted for 35.6% of what Trump polled in the general election four years ago. Perhaps Haley supporters in Charlottesville were simply more motivated this March than Trump supporters were elsewhere — or maybe there were crossover votes.
The best place to look for crossover voting might be to find unusual turnout spikes such as Charlottesville, but they are mostly in Haley localities in Northern Virginia. In Arlington, another place she won, the Republican presidential primary turnout amounted to 68% of Trump’s totals four years ago. Outside of Charlottesville and Albemarle County (where the primary turnout was 44.2% of the Republican vote four years ago), I don’t see any unusual spikes. In Lynchburg, the primary turnout was 30% of the Republican general election totals four years ago. In Pittsylvania County, it was 24.8%.
Based on those numbers, I can be persuaded that there was crossover voting for Haley in some places, but not others.
Here’s one key thing to remember, though: Democrats didn’t have much of a presidential primary going on this year, so that helped enable any potential crossover voting. However, in the 5th District, Democrats do have a primary of their own going on — and seem to be voting at a pretty good clip. At last count, 6,930 people have voted in that district’s Republican primary, more than in any other Republican congressional primary in the state this year. (In second place is the 7th District between Northern Virginia and Richmond with 4,321.)
However, 6,957 people have voted in the Democratic primary in the 5th District, so we’re seeing more Democrats than Republicans voting early in that district — and they’re voting in the Democratic primary, not the Republican one. That’s a lower Democratic figure than in some other districts, but we have to remember that the 5th skews Republican, so we’d expect fewer Democrats to vote there. Democrats are also more enthusiastic about early voting than Republicans are, which helps explain why the Democratic vote in the district is outpacing the Republican one. Still, those numbers don’t make me think there’s a lot of crossover voting going on. Since you only can only vote in one, any Democrats crossing over into the Republican primary would forfeit a chance to pick their own nominee, and plainly, a lot aren’t doing that.
The other thing that makes me skeptical of crossover voting in the 5th is the nature of the Republican contest: two candidates arguing over who’s the more conservative and the more supportive of Donald Trump. I can see a rationale for a Democrat entering the Republican presidential primary to vote for Haley — they can’t stand Trump, and figure it’s better for the republic for Joe Biden to lose to Haley than Trump. A Haley administration might be a pretty normal Republican one; a second Trump term might tear the nation apart. (I can also see some Democrats thinking Trump would be easier to beat than Haley, but more likely any crossover Democrats would be voting for Haley as an insurance policy for democracy.)
Who would Democrats crossing over into the 5th District Republican primary want to vote for, though? I suppose they’d get a thrill out of seeing Good go down, but McGuire’s voting record would be the same; he just might not be quite so disruptive in Washington. Given the Republican lean of the 5th District (the mapmakers who drew the lines calculated it at 53% Republican), Democrats might be better off with Good as the Republican nominee — he offers a more inviting target than McGuire.
Beiler, though, sees things differently, and I can’t say he’s wrong. As the early voting has been underway, we’ve been corresponding about how much crossover vote he expects to see, based on the levels of voting between the two parties. His most recent estimate, as of Monday,was that 10% of the primary vote may be crossover voting. “But,” he wrote Monday, “I think Trump’s endorsement of McGuire has shuffled the deck here: Now there will probably be just as many Democrats crossing over to embarrass Trump as there are Dems trying to sink Good. That means while the earliest crossovers benefited McGuire, it will probably be a mixed bag for those cast after the endorsement.”
He’s guessing the ultimate turnout in the 5th District Republican primary will be about 45,000 voters. If so, that will put it slightly north of the highest Republican primary in 2022 (in the 2nd District in Hampton Roads, won by Jen Kiggans) but slightly south of the highest Republican primary in 2018 (in the 10th District in Northern Virginia, won by Barbara Comstock) and well south of the biggest Republican congressional primary turnout ever Virginia — the 7th District in 2014 where Dave Brat defeated incumbent Eric Cantor. That primary drew 65,021 voters.
“Now that there are two breeds of crossovers at work — those looking to a) ditch Good or b) diss Trump — the figure may turn up a click, perhaps 12% or about 5,500,” Beiler wrote, based on his estimate of 45,000 eventual votes.
If he’s right, how significant would 5,500 Democratic voters be in a Republican primary?
Ultimately, that depends on how close the race is. If we look back to 2022 at the four Republican congressional primaries across the state that year, 5,500 votes (all for the same candidate) would have made a difference in two of them. That number of votes wouldn’t have made any difference in the three Republican congressional primaries in 2018 but would have made a difference in two of the three Republican congressional primaries in 2016. If we go back to that epic Brat vs. Cantor primary in 2014, some 5,500 votes wouldn’t have made a difference — Brat won by 7,193 votes. Ultimately, the potential impact of any crossover votes in the 5th this year depends on how close the Good-McGuire primary is, and whether crossover votes tend toward one candidate or another or are split evenly for the reasons Beiler outlined.
So, yes, under certain circumstances, it is possible maybe crossover Democrats could make the difference. So what should they do?
Nobody asked me, but if any 5th District Democrats did ask for my advice, I’d tell them to go vote in their own primary, and let nature take its course on the Republican side.
In this week’s West of the Capital:
I write a free weekly political newsletter, West of the Capital, that goes out every Friday afternoon at 3 p.m. You can sign up here:
Here’s what I’ll have in this week’s edition:
- The latest early voting numbers across the state.
- More information on who’s speaking at the cannabis conference that Cardinal News is hosting Oct. 15 at Roanoke College. The conference is expected to attract participants from across Virginia, but space is limited. More information about the program, sponsorships and early bird registrations are available now. To take advantage of a $25 discount off the $150 ticket, use the promo code “early bird” before Aug. 1.
