Midway to the Midterm: The Imperfect Bellwethers of New Jersey and Virginia

KEY POINTS FROM THIS ARTICLE

— It will be very tempting to use the New Jersey and Virginia gubernatorial races as predictors for next year’s midterms.

— Sometimes these races do provide a preview of the following year, but there are enough instances where they do not that we would urge caution against overinterpreting the results.

— One key factor is that the political situation could just be different in the midterm year than it is in the gubernatorial year, something we arguably saw as recently as 2021 and 2022, the most recent comparable cycle.

— Another confounding factor is that New Jersey and Virginia are both more Democratic than the nation at the federal level, which was not consistently the case until recently.

NJ and VA as imperfect bellwethers

There is going to be a real temptation to treat this November’s gubernatorial elections in New Jersey and Virginia as a preview of next year’s midterm. But a look at these races over the last several decades and how they compared to what happened in the following year’s midterm should provide some caution against doing so.

Sometimes the New Jersey and Virginia elections are suggestive of what will happen in the following election. But there are other examples where they are not.

Part of why these races are not really bellwethers is that they, ultimately, are just two races held in what is a barren spot on the political calendar—there are not many major statewide races regularly held in odd-numbered years, whereas even-numbered midterm years feature a plethora of statewide races and every seat in the U.S. House of Representatives. Incumbency may also skew results—Virginia is always an open seat because of the state’s unique prohibition on governors seeking a second consecutive term, but incumbents in New Jersey can seek a second consecutive term.

Let’s take a look at the history, which is shown in Table 1. We are using 1965 as a starting point because it represented the last pre-modern Virginia gubernatorial election—it was the last gasp of the old conservative Democratic dominance of Virginia, as Mills Godwin became the last of 21 consecutive Democrats to win the governorship. The election of Republican Linwood Holton in 1969 ushered in Virginia’s modern era of two-party competition. Table 1 covers 15 Virginia/New Jersey gubernatorial cycles and then the 15 midterms held in the years immediately following those elections. Obviously this is not a huge sample size, but it does give us a chance to see the commonalities (and differences) between what happened in these gubernatorial races versus what happened in the midterms.

Table 1: NJ/VA gubernatorial history compared to subsequent midterm results

Source and notes: Gubernatorial results from Dave Leip’s Atlas of US Presidential Elections. Midterm net change from Vital Statistics on Congress, The Long Red Thread, and Crystal Ball research. (I) indicates incumbent.

In at least a few of these cycles, the gubernatorial race results clearly fit in with what happened in the ensuing midterm. For instance, in 1993, Republicans won Virginia in a rout while Christine Todd Whitman (R) narrowly unseated incumbent Gov. Jim Florio (D) in New Jersey. The 1994 “Republican Revolution” followed a year later, with Republicans flipping both the House and the Senate while also netting a substantial number of governorships. The back-to-back cycles of 2005 and 2009 also show a symmetry between the gubernatorial results and the ensuing midterms—Democrats doing well in 2005 and then 2006, and Republicans doing well in 2009 and 2010. But those are the only three cycles out of the 15 on Table 1 where one party swept all five categories—meaning that they won New Jersey and Virginia and then netted seats in all three categories of midterm races (governors, House, and Senate). In all three of those cycles, the non-presidential party was the one sweeping the categories, as is familiar in midterm election cycles.

A couple of other wave years against the president’s party—1974 and 2014—were not necessarily previewed by the previous year’s gubernatorial races, as the president’s party salvaged the open Virginia race in both years. Each of those Virginia races featured somewhat unusual circumstances: Ken Cuccinelli was an outspokenly conservative Republican candidate in 2013 who also might have been hurt by a Republican-sparked federal government shutdown in his narrow loss to Democrat Terry McAuliffe (as an aside, a federal government shutdown could be on the horizon this fall, perhaps sparked by a Senate Democratic filibuster). In the 1973 race, liberal/populist firebrand independent Henry Howell supplanted the Democrats, leaving conservatives to rally around the aforementioned Godwin, who after winning in 1965 as a Democrat then narrowly won in 1973 as a Republican. Neither of those races suggested the waves that came the following year against the party that had won in Virginia. (New Jersey had a somewhat strange election in 1973 itself, as moderate Republican incumbent William Cahill, hurt by scandal in his administration, lost his primary to a more conservative challenger, aiding Democrat Brendan Byrne’s landslide victory).

The 2017 races were suggestive of 2018, as Democrats performed well in both of the gubernatorial races and also had a strong midterm, albeit losing a little ground in the Senate (which is why we didn’t include that with the 1993, 2005, and 2009 cycles). Democrats losing ground in the Senate in 2018 was due in large part to a challenging map that year (Democrats were defending the lion’s share of the seats being contested and were trying to defend incumbents living on borrowed time in several red states).

We often note how the Senate can operate differently than the House (or even governors) in a midterm cycle because of the profile of the seats being contested—that same dynamic likely contributed to some degree to the president’s party holding up relatively well in Senate elections in years like 1970, 1982, 2002, and 2022.

Those four cycles are years where the gubernatorial results didn’t necessarily map neatly on to what happened in the following year’s midterm. Republicans swept in 1969, but Democrats did OK all things considered in the 1970 midterms. In 1981, both governorships were open and flipped in opposite directions, but Democrats ended up having a decent though not necessarily great midterm in 1982.

This does lead to another observation that is worth remembering—perceptions of the president, which can be an important factor in a midterm environment, are sometimes different in the first November of his presidency compared to his second. President Reagan, under whom 1981 and 1982 were contested, provides an example. Table 2 shows how the public viewed the president on the day of each cycle’s off-year gubernatorial and midterm elections (thanks to Nate Silver’s Substack, the Silver Bulletin, for producing modeled averages of presidential approval).

Table 2: Presidential approval at time of NJ/VA elections and midterm

Source and note: Silver Bulletin. We used a president’s approval rating on the day of the odd-year gubernatorial election and even-year midterm election.

As you can see, Reagan’s approval was significantly lower in late 1982 than it was in late 1981. Another example that stands out is 1965-1966, when Lyndon Johnson’s approval dipped a lot from 1965 to 1966 and the twin Democratic victories in the gubernatorial races in 1965 suggested nothing about the GOP wave to come a year later. More recently, Trump’s approval rating in November 2018 was weak but still a little better than it was in November 2017—perhaps that, paired with other factors (like the battle over Brett Kavanaugh’s appointment to the Supreme Court in fall 2018), helped Republicans activate their base voters and salvage some key gubernatorial and Senate races that year.

George W. Bush’s approval rating was substantially lower in 2002 than it was in 2001—but his approval was still excellent overall. Remember that November 2001 was just a couple of months after the Sept. 11 attacks, and Bush was riding a wave of goodwill and national unity following that traumatic event. Bush’s numbers were still strong in 2002 and he campaigned vigorously for his party, helping Republicans avoid the usual midterm headwinds and make small gains in the House while flipping control of the Senate (Democrats had taken control of the chamber in mid-2001, when Vermont Sen. Jim Jeffords left the GOP, became an independent, and moved over to the Democratic caucus). Overall, the 2001 Democratic gubernatorial wins did not really provide a preview of 2002.

Four years earlier, Republicans swept the gubernatorial races in November 1997. That came before news of President Clinton’s affair with Monica Lewinsky emerged in January 1998, leading to the Republican impeachment push widely credited with helping Democrats perform relatively well in the 1998 midterms (note, too, that Clinton’s approval was actually better in November 1998 than it was in November 1997 despite the firestorm).

More recently, 2021 was an imperfect harbinger of 2022: Republicans won Virginia and came close to unseating Gov. Phil Murphy (D) in New Jersey. Even though Republicans did not win both races, the results on their face suggested big problems for Democrats in the midterm. Part of that was, by 2021, both New Jersey and Virginia had become more Democratic than the country at the federal level in an era where presidential voting is more predictive of down-ballot voting than it was a few decades ago.

Figure 1 shows how both New Jersey (the yellow solid line) and Virginia (the green dashed line) voted for president relative to the country—we use a metric we call “presidential deviation” for this, which compares how a party does in the national popular vote to how the party does in the state. In 2024, for instance, Donald Trump received 51% of the two-party national vote, while he got 47% in each of New Jersey and Virginia. So the presidential deviation for both states was D +4, expressed as -4 on Figure 1.

Figure 1: NJ/VA presidential deviation, 1964-2024

At the presidential level, both New Jersey and Virginia had a Republican lean from the late 1960s to the early 1990s, with the states voting basically the same in both 1976 and 1980. New Jersey acquired a clear Democratic lean starting in 1996, while Virginia’s did not really emerge until the Trump era. That’s where they both are today, which is part of the reason why Republicans could claim something of a moral victory in New Jersey four years ago despite losing there, while Republicans had been on a dry spell in Virginia for more than a decade prior to their 2021 win.

As it was, 2021 was somewhat suggestive of 2022—Republicans did flip the House, after all—but it certainly felt like things would be worse for Democrats in 2022 immediately following the 2021 elections than they ended up being. One factor was likely the Dobbs Supreme Court decision on abortion rights, which came out before the 2022 elections but after the 2021 elections. Republicans also had better-quality candidates in the 2021 gubernatorial elections than they did in many of the key Senate races in 2022.

Another two-year cycle that shares some similarities with 2021-2022 is 1977-1978. The gubernatorial races split, with Republicans turning in a strong performance in Virginia while Byrne was reelected by double digits in New Jersey, albeit with a much smaller margin than four years prior. Republicans would go on to net seats in all three categories in 1978, but the results were disappointing to some Republicans at the time.

Overall, there definitely are instances where the New Jersey and Virginia gubernatorial elections were very suggestive of the following year’s midterm. But there are enough examples when they were not that we would urge caution against overinterpreting the results in November. We’ll try to remind ourselves of that as we’re hungrily devouring the results in early November.

Kyle Kondik is a Political Analyst at the Center for Politics at the University of Virginia and the Managing Editor of Sabato's Crystal Ball.

See Other Political Commentary by Kyle Kondik.

See Other Political Commentary.

Views expressed in this column are those of the author, not those of Rasmussen Reports. Comments about this content should be directed to the author or syndicate.