|Dear Readers: We wish all of you a Happy Thanksgiving; the Crystal Ball will not be publishing next week.
Before we move full speed into assessing 2024, we wanted to take a quick look back at a few notable findings from last week’s elections — specifically in Virginia, Pennsylvania, and Kentucky.
— The Editors
Virginia result looked a lot like four years ago
Vote counts in Virginia are nearly final but not yet certified. There were no late changes to the topline result that seemed likeliest the morning after the election: Democrats won their barest possible majorities in both chambers: 51-49 in the state House of Delegates and 21-19 in the state Senate.
There has been a “choose your own adventure” aspect to interpreting the results, ranging from the elections being a devastating rebuke to Gov. Glenn Youngkin (R-VA) — because he invested so much political capital into winning unified control of Richmond only to come up short — to the outcome actually being a great sign for Republicans because their candidates ran well ahead of the 2020 presidential results in many places.
After looking more closely at the numbers, our own assessment is that the result was… very familiar. In fact, it was extremely similar to 2019, the last standalone state legislative election in Virginia, when Democrats won both chambers before losing the House in 2021 as part of Youngkin’s gubernatorial victory.
Before the election, we noted that presidential performance is a difficult-to-reach ceiling for Virginia Democrats in state legislative elections. For whatever reason, likely at least in part because of the smaller and different turnout mix in an odd-year state-level election, Virginia just performs as a more competitive state in these kinds of elections than in a presidential year. So Republicans running well ahead of Republican presidential margins in state legislative races is not some sort of unusually good performance for them in Virginia — rather, it’s normal. If that was not the case, the race for both chambers would have had a clear Democratic edge as opposed to being as competitive as it was. So let’s go through that recent history again before looking at the numbers for 2023.
Over the last few election cycles, only in 2017 did Virginia Democrats come close to matching presidential-level performance, a Democratic wave year in Virginia that saw Democrats win all three statewide offices and come within a tiebreaking drawing to force a tie in the House of Delegates. In that year, Democrats in contested races ran, on average, a point behind Hillary Clinton’s two-party presidential margin from 2016.
In other years, the gap between presidential performance and down-ballot, two-party Democratic margin was much wider. In 2015, when Republicans won a huge edge in the House and a narrow 21-19 advantage in the Senate, Democrats on average ran 22 points behind what Clinton would win a year later in House districts and 13 points behind Clinton in Senate districts.
In 2021, when Republicans flipped the House, Democrats ran 12 points behind Biden’s margin on average in contested districts.
Finally, in 2019, Democrats on average ran 7 points behind Biden’s eventual 2020 performance in House races and 8 points behind in Senate races. That translated into a 55-45 edge in the House and 21-19 in the Senate in the standalone state legislative elections that year (this was the most recent election comparable to 2023, when both legislative chambers but no statewide offices were on the ballot).
So, how did it go in 2023? On average, Democrats ran about 8 points behind Biden in both House and Senate races. That includes an average of the 2023 state legislative results compared to the 2020 presidential results in the 34 of 40 state Senate districts and 65 of 100 state House districts that were contested by both major parties. In other words, this was just like 2019, at least by this metric.
The Democratic majority in the House is smaller than in 2019 (51-49 versus 55-45 in 2019), while the Senate produced the same outcome in both elections. The maps are different this time, drawn by court-appointed special masters as opposed to a Democratic-drawn Senate map in the 2010s and a GOP-drawn but court-modified map in the House that was used in 2019 and 2021.
Part of the difference between 2019 and 2023, and why the Republicans are at a smaller deficit in the House than four years ago, is because they basically won all of the close races this time: Republicans ended up winning the five closest races (districts 82, 41, 89, 71, and 57), all of which were decided by 2 points or less. The Democrats’ smallest victory came in HD-21, which Del.-elect Josh Thomas (D) won by 3.6 points. Back in 2019, the parties split the four closest races (the only four that were decided by less than 3 points). Meanwhile, Republicans limited the Democratic Senate majority in both 2019 and 2023 by doing well in the closest races: Sen.-elect Danica Roem (D), whose Senate district overlaps with Thomas’s Northern Virginia seat, was the closest Democratic Senate victor, at a 3.8-point margin; the Republicans won the two closest races, districts 24 and 27, which were decided by 1.1 and 1.8 points, respectively. In 2019, Republicans won the four closest Senate races.
So while one could say that in 2023 the Republicans were just two seats away from an outright House majority and just one seat away from a tied Senate where Lt. Gov. Winsome Earle-Sears (R) could have broken at least some ties in favor of Republicans (her tiebreaking power would not be absolute in all matters), it’s also fair to say that the Democrats’ advantage could have been more robust in both chambers had the GOP not scratched out the closest victories. That could prove important if there are vacancies in the coming years, because Republicans are just one special election flip away from forcing a tie in either chamber.
One can interpret what happened in different ways. On one hand, Democrats essentially maintained their overall performance from 2019, at least on average, despite having an unpopular president in the White House, whereas in 2019, Republicans were the ones saddled with the unpopular president. That perhaps suggests the Democratic showing was more impressive than 2019, given how off-year elections can sometimes break against the presidential party, a dynamic that was likely at play when Republicans did well in Virginia in 2021.
On the other hand, remember that the 2019 elections happened before the 2020 presidential race, which in Virginia further solidified the state’s move toward Democrats. That Republicans kept up their overperformance compared to the presidential results after that event may be a mark in their favor — they also kept up a very similar performance in a post-Dobbs election, a decision that of course was years in the future in 2019.
Again — choose your own adventure. But just remember that a lot changed from 2019 to 2023 — Biden became president, Youngkin became governor, the district maps changed, and the Dobbs decision came down, among other things. Despite that, the 2019 and 2023 Virginia state legislative elections, both in terms of overall outcome and how the results compared to presidential performance, were very similar.
Pennsylvania: A more federalized race, with exceptions
We noted last week that the voting patterns in the Pennsylvania state Supreme Court race appeared to follow federal patterns a little more closely in 2023 than a similar race in 2021. This year, Democrat Daniel McCaffery won by about 7 points statewide in what are still unofficial returns. Back in 2021, Republican Kevin Brobson won by about a point statewide. McCaffery’s victory restored the 5-2 state Supreme Court advantage that Democrats enjoyed prior to the death of former Chief Justice Max Baer (D) in 2022.
In 2021, Brobson’s narrow win was powered by overperformances in the historically GOP-leaning but Democratic-trending Philadelphia suburbs — Brobson did a lot better than Donald Trump had done in those places. That was part of the reason why the 2021 Supreme Court race reflected the 2020 presidential results a little less than, say, the 2022 Senate race won by now-Sen. John Fetterman (D-PA). There was a .99 correlation between the 2020 county-level Democratic presidential margin and the 2022 Democratic Senate margin, but only a .96 correlation between the 2020 presidential and 2021 judicial Democratic margins. The 2023 race, overall, had a higher correlation with the presidential patterns than the 2021 race did — the correlation with 2020 was .98, so closer to but not quite as high as the 2022 Senate correlation.
As we were looking over the 2023 Supreme Court county-level results, though, there was a result that seemed at least a little out of place.
In our preview of the 2023 race, we noted that despite losing statewide, the Democratic candidate in 2021 carried Erie County, a closely-watched and competitive county in northwest Pennsylvania, by an impressive 15-point margin. But in 2023, the victorious McCaffery actually lost Erie by less than a point despite winning by 7 statewide. That seemed strange, particularly because none of the candidates in either 2021 or 2023 had any obvious local ties to Erie County, as far as we can tell. Erie was a reliably Democratic county at the presidential level prior to flipping to Donald Trump in 2016 and then back to Joe Biden in 2020, so it’s one of the key places to watch in the state; Fetterman gave the county a ton of attention in 2022, and he won it by 9 points, better than his 5-point statewide margin.
There is at least one possible contributor to Erie County’s change between 2021 and 2023, though: differences in TV ad spending.
Thanks to ad spending numbers supplied to us by our friends at the advertising analysis firm AdImpact, (which puts out reams of great data), we can see that in the 2021 Supreme Court race, Democrat Maria McLaughlin outspent a couple of Republican groups in the Erie media market, about $125,000 to $108,000. But in 2023, it was Republican Carolyn Carluccio and her allies outspending McCaffery $241,000 to $86,000 — close to a 3-to-1 advantage. That may have had an impact on the results, especially when one considers that the two other counties in the Erie media market — Crawford and Warren — also saw Carluccio do better than Brobson did in 2021, even though Carluccio lost statewide while Brobson won, meaning that Carluccio did worse than Brobson in much of the rest of the state. Overall, McCaffery and his allies enjoyed an ad spending edge statewide — but not in the Erie media market, which might have been a factor in why the Republicans did relatively well there compared to their showing statewide. Democrats also did well in some other Erie County races, which may suggest the importance of the ad spending edge in the specific state Supreme Court result.
The efficacy of political ad spending is much debated in political science circles, but here’s an example of ad spending differences between two similar elections in a small media market potentially shedding some light on what happened.
Disaster response may have boosted Beshear
As a Democratic governor in a Republican state, Gov. Andy Beshear (D-KY) needed to find ways to win crossover support from voters who are Republican at the federal level. Disaster relief may have been one way he did it.
Nate Cohn of the New York Times noted that Beshear’s biggest improvements from his 2019 victory came in eastern Kentucky, where devastating flooding was the worst in the state in 2022. Beshear emphasized his response to this disaster in advertising. Some of this territory is ancestrally Democratic but extremely Republican at the presidential level in the Trump era.
This reminded us of other governors who seemed to be rewarded for disaster response, like former Gov. Jay Nixon (D-MO), who improved in the area around Joplin, MO in 2012 following devastating tornadoes that struck the area in 2011 (this improvement came as Nixon did worse statewide in 2012 than he did in 2008). While it’s harder to specifically track on a county-level results map, Gov. Ron DeSantis (R-FL) may have been helped by a strong state-level response to Hurricane Ian before the 2022 election — that election was never close, but perhaps the late test of competence helped pad DeSantis’s overall margin (he outperformed the polls). We summarized some of the research on the politics of natural disasters back in 2017.
Kentucky governors don’t have a lot of power versus the state legislature, at least when it comes to vetoes, which can be overridden by a simple majority vote (and Republicans of course have gigantic majorities anyway). This lack of legislative sway might have actually helped Beshear, too, because it’s not like he can really successfully champion progressive legislation through a hostile legislature in a conservative state. Instead, a governor in his shoes needs to focus on other ways to make a difference and win approval. Being seen as a competent manager in a time of crisis can be helpful in this regard, and it likely contributed to Beshear winning by a surprisingly clear 5-point margin.
One other thing: We noted above that Virginia is a different state in its state legislative election years than presidential years. It’s probably fair to say that Kentucky is different in its gubernatorial election years than in presidential years, too. After Bill Clinton carried the state twice, the Republican presidential margin in Kentucky has been 15 points or more in each of the last six presidential elections. Meanwhile, Republicans have won only two of the six gubernatorial contests since that 2000 presidential election, and their biggest margin was 10 points, in 2003. Democrats with the last name of Beshear — Andy and his two-term Democratic governor father, Steve — won the other four contests, even as the state has otherwise become much more uniformly Republican down the ballot. As the nation’s elections have become more partisanized, quirkier results like GOP overperformance in Virginia odd-year state legislative races or Democratic overperformance in Kentucky governor races have become less common — but these results that don’t neatly conform to presidential partisanship clearly do still exist.