Do political scandals still matter? Or is ticket-splitting dead?

Jay Jones’s text message scandal will test the limits of both propositions in what should be a good year for Democrats in Virginia.

Jay Jones walks with Abigail Spanberger and Ghazala Hashmi in happier times, at the Annual Buena Vista Labor Day Festival parade on Sept. 1. Getty Images.

Compared to last year — and compared to next year — this is going to be a quiet November at Silver Bulletin. But it’s Election Day next Tuesday in Virginia, New Jersey, California and New York City. The election is (obviously) important if you happen to live in one of these places. And California’s Prop 50, which would remove control of the redistricting process from a bipartisan commission and allow Democrats to aggressively redraw boundaries, has major implications in the redistricting wars.

But we have to be honest: Nate and I are more skeptical than most analysts about the forecasting power of off-year and special elections. Yes, elections are fun (at least if you win). And we’re always starved for hard data, especially if you don’t trust the polls.¹ But it’s easy to miss the forest for the trees by focusing on minutiae in every election and using them to make pronouncements about the broader environment.

But also, these races have been, to use a technical election analysis term, boring. For instance, we’d be covering Prop 50 more if its passage didn’t look like nearly a foregone conclusion. Meanwhile, Democrats had held strong leads in polls of Virginia’s three statewide races (governor, lieutenant governor, and attorney general) for most of the year. And in general, the party out of power tends to do well in these elections — more good news for Democrats in VA. Did I watch the Virginia gubernatorial debate? Yes, but only in the background while building a piece of IKEA furniture.

Then on Oct. 2nd, news broke that Jay Jones, the Democratic candidate for attorney general in Virginia, had been convicted of reckless driving in 2022 after going 116 miles per hour in a 70 mph zone. As a former Florida driver I normally wouldn’t look twice at that number.² But more controversially than the speeding ticket, Jones served 500 of his 1000 hours of court-mandated community service at his own political action committee.³

And that was just the oppo appetizer. On Oct. 3rd, National Review published texts Jones sent to a Republican member of Virginia’s state legislature in 2022 about Virginia House Speaker Todd Gilbert. Here’s an excerpt:

Three people, two bullets

Gilbert, hitler, and pol pot

Gilbert gets two bullets to the head

Spoiler: put Gilbert in the crew with the two worst people you know and he receives both bullets every time

In a subsequent phone call, Jones allegedly suggested that he wished Gilbert’s wife could watch her child die in her arms. Why? Because it might make Gilbert reconsider his position on gun control. It feels unnecessary to say, but this is an especially damaging scandal at a time when people are understandably more sensitive to political violence. Jones apologized after the texts came out and again in Virginia’s attorney general debate. Other Democrats in the state have condemned the statements, but the most prominent state-level figures like gubernatorial candidate Abigail Spanberger and VA senators Tim Kaine and Mark Warner have stopped short of calling for Jones to drop out.

Republicans have an opportunity, but face a tough tactical choice

So here’s the question that makes this race interesting: do scandals like this still matter in American politics? And do they matter more for a Democrat than for a Republican — like, say, the current president? Or will an unpopular President Trump, combined with a friendly off-year environment for Democrats, be too much for Republicans to overcome?

In the AG race, Jones has gone from 4-point leads in the DDHQ and VoteHub polling averages on Oct. 1st to trailing incumbent Republican Jason Miyares by 3 and 0.9 points, respectively. Most polls since the text messages were revealed show a small Miyares edge. The Washington Post even found that 51 percent of registered VA voters, including 26 percent of Democrats, think Jones should drop out.

These are bad numbers if you’re a Virginia Democrat. But they open two distinct options for Republicans: a) use Jones to bring down other Democratic candidates or b) go all in on the AG race. Zack Roday, a Republican strategist, hinted at the former approach, saying that he sees the texts, along with transgender policies, as an indicator that “Democrats have lost the plot on common sense.” “They did that in ‘24 in a major way, and they’re doing that again in Virginia in ‘25,” he told me.

But so far, the scandal hasn’t moved the needle upballot. Spanberger leads Republican Winsome Earle-Sears by between 8 to 10 points, depending on which polling average you use.⁴

Just to level-set here, that kind of lead is probably safe for Spanberger. Virginia, like other states, has had better and worse years for polling. In 2017, for example, the Morning Joe-type conventional wisdom was that Democratic candidate Ralph Northam was in trouble; he led Republican Ed Gillespie, a favorite of the consultant class, by a relatively modest 3 points in polls. Indeed, the polls were off — but in Northam’s favor, who won by nearly 9 points instead.

These errors don’t always favor Democrats. In 2013, you could basically flip the polling average and the end-result columns: Democrat Terry McAuliffe was ahead by about 8 points in the polling average, but he won the race by only 3 points.⁵ Still, a win is a win. The historical average polling error in Virginia is about 4 points. A mathematical extrapolation of these past polling errors would suggest that prediction market odds putting Spanberger’s chances in the mid-90s are roughly fair. You’re welcome to bet on Earle-Sears’ ~5 percent chance if you like — 5 percent chances happen. Although it’s similarly possible that Spanberger wins by double digits with even a minor polling error in her favor.

Still, there’s no sign of post-text movement toward Republicans in the top-of-the-ticket polls. Across averages from DDHQ, FiftyPlusOne, Race to the WH, and VoteHub, the gubernatorial margin has shifted by an earth-shattering 0.5 percentage points toward Democrats since Oct. 1st. Virginia politics expert Sam Shirazi told me that “in theory, could some of the stuff that has come out against Jay Jones affect [Spanberger]? Maybe, but it just seems like we’ve traditionally not seen […] someone else’s scandal stick to another candidate.”

The race for lieutenant governor — by far the lowest-profile statewide contest — is playing out similarly. Democratic candidate Ghazala Hashmi’s margin in the DDHQ average has dropped only infinitesimally, from D +4.5 to D +4.0 since Oct. 1st. Clearly, her AI likeness’s debate performance — you can’t spell “Commonwealth of Virginia” without C-R-I-N-G-E — was strong enough to keep her in the driver’s seat.

Those are the kind of numbers you’d expect in a state Kamala Harris won by five points in 2024. Did Republicans win all three of Virginia’s statewide offices in 2021? Yes. But Virginia is really only purple-ish. Even Republicans like Roday expected this to be a tough year: “Historically, the party in the White House struggles to win in Virginia. That’s been true for 50 years […] And so you’re swimming uphill as a Republican with a Republican in the White House.” Trump is unpopular in Virginia, and Jones’s main strategy has been to connect Miyares to Trump as closely as possible. That’s not exactly difficult, given that Miyares is the only statewide Republican candidate with a Trump endorsement.

The ongoing government shutdown could also be a real issue for Republicans in a state home to nearly 8 percent of all federal workers. Democratic strategist Jesse Ferguson thinks “putting the federal government and the Trump administration front and center at the close of the campaign makes it more of a referendum on him. That’s more of a problem… for the Republicans.” That might make going all-in on the potentially winnable race for attorney general more appealing. But what would that approach look like? Something like this Miyares digital ad:

The point of this strategy isn’t to use Jones to drag Spanberger down, but just the opposite: to convince Spanberger voters to split their tickets and vote for Miyares. Shirazi thinks “there’s this potential where Miyares is trying to survive and just doesn’t necessarily do a whole lot to help Earle-Sears,” he said. “And I think, frankly, a lot of Republicans are kind of encouraging Earle-Sears to spend most of her time on the Jay Jones side of things, because that also helps Miyares.” The logic is simple enough: if the gubernatorial race is a lost cause, why waste resources there when you can instead focus on your one competitive statewide race?

The AG’s race was the closest of the three even before the texts came out, which makes this approach even more understandable. Let’s use advertising spending as an example. According to Ferguson, “even months ago, the most likely Republican opportunity in Virginia was the attorney general’s race, mostly because [of their] financial advantage.” Spanberger has a massive ad spend advantage in the gubernatorial race of about $13 million. Democrats are also outspending Republicans (albeit by a more modest $2 million) on ads in the race for lieutenant governor. But Jones has spent substantially less on advertising than Miyares: a cumulative $6 million, compared to Miyares’s $16 million.

Miyares’s cash advantage has only increased as Republican donors have been energized by the text scandal, while some Democratic donors have become more reluctant to give to Jones.

How many Virginians will split their tickets — or sit the AG race out?

Ultimately, Jones’s fate will depend on which of the two T’s resonates more with voters: Texts or Trump. In most states, you might assume Trump. But Virginia is a politically engaged state that has historically prided itself on decorum. In our live SBSQ last week, Nate said his modal Virginian would find the texts unacceptable, and the polls back up that assumption.

Still, how plausible is a big upballot-vs-downballot split? Historically, Virginia’s attorney general candidate has underperformed the winning gubernatorial candidate from their party by an average of 1.3 points. There have been some bigger splits, however. Donald McEachin underperformed the Democratic gubernatorial candidate by a whopping 26 points in 2001, while in 1985 and 1989, Mary Sue Terry ran 12 and 26 points ahead of the Democratic gubernatorial candidate.⁶ But in general, most AG candidates underperform by between one and six points.

That means Jones running behind Spanberger wouldn’t have been surprising even before the texts were released. But it also means that, based on Spanberger’s current position in the polls, a historically average level of underperformance won’t be enough to sink Jones.

But this isn’t an average scandal. And Shirazi pointed out that voters don’t necessarily need to split their tickets: they could sit the AG race out instead. “I also think there’s an under-discussed issue of undervotes, where some people who vote for Spanberger will just skip the attorney general race,” he said. One hint about this possibility is that there are still a meaningful number of undecided voters: 8.5 percent on average in the AG’s race compared to only 5.3 percent in the gubernatorial race.

And Jones and Miyares won’t necessarily split those voters evenly. Especially if some of these undecideds are really just shy Jones voters: respondents who plan to vote for Jones but don’t want to say so during a survey.⁷ There’s some evidence of this already. Jones dropped from 51 percent support to 46 percent between the September and October Washington Post polls, but Miyares’s support only increased from 45 percent to 46 percent. Where did everyone else move? “Neither” or “Would not vote.”

Our research has also found that polling errors are correlated from race to race within a state. In other words, if Spanberger overperforms or matches her best polls and wins by double digits, it would take a monumental amount of split-ticket voting for Miyares to win. Indeed, you can see this in the polls already: the correlation between Spanberger’s margin and Jones’s is 0.83, so not quite a perfect match but close. For instance, the only recent survey showing Jones ahead, from State Navigate, also had Spanberger winning by her widest margin: 13 points. Trafalgar/InsiderAdvantage has shown Democrats’ worst numbers in both races, conversely. (Those firms’ polls are typically strongly Republican-leaning, and Virginia is apparently no exception.)

Scandals aren’t what they used to be

In a previous political age, you could easily imagine someone dropping out of the race over a scandal as big as Jones’s. But we haven’t lived in that world for some time, which is probably why no one I spoke to for this article had a confident prediction for the AG’s race. The data show a close race, and the career-killing scandals of yesteryear simply aren’t anymore.

On one hand, it’s hard to imagine these texts not changing the outcome. Miyares is ahead in most polls, he has more money than Jones, and bettors on Polymarket give him a slight edge (56 percent to Jones’ 44 percent as of this post). On the other hand, a moderate polling error in favor of Spanberger would likely be enough to drag Jones across the finish line. It’s easy enough to envision watching the results come in on election night and seeing that Jones has beaten his polls by a few points because the large number of Virginians who disapprove of Trump saw him as the lesser of two evils.

1. Though most non-Silver Bulletin-readers go too far with that.

2. Miami is one of the most dangerous places to drive in the US.

3. An investigation into Jones’s community service was opened just a few days ago.

4. Full disclosure, I’m automatically more sympathetic to Earle-Sears as a fellow hyphenated-last-name-haver.

5. CORRECTION: This article originally incorrectly stated that Terry McAuliffe was behind by 8 points in the 2013 polling average.

6. She was also the first woman elected to statewide office in Virginia.

7. This kind of behavior is expected when polling people about sensitive topics: they usually don’t like admitting to potentially embarrassing things like supporting someone who sent those texts.