A study of battleground Senate races in presidential years suggests that Democrats will struggle to separate themselves from Joe Biden.
A study of battleground Senate races in presidential years suggests that Democrats will struggle to separate themselves from Joe Biden.
D emocrats face a dilemma in the presidential race: stick with a visibly failing Joe Biden, or take a leap into the unknown by escalating the pressure for him to drop out. Any of these options put them on uncharted ground. Dropping out — which would be unprecedented this late in a national election — could mean replacing him with the deeply unpopular Kamala Harris, or it could mean embarking on a chaotic and divisive quest for a new nominee not chosen by the voters. It appears increasingly unlikely that Democrats will be able to escape Biden.
The last time a national party faced a choice even vaguely similar was in 1968, when Lyndon Johnson’s withdrawal in March and Robert F. Kennedy’s assassination in June backed the Democrats at their violence-marred Chicago convention into choosing as their standard-bearer vice president Hubert Humphrey, who had competed in just eight primaries and didn’t win any of them. Trailing badly in national polls throughout the late summer and the fall, Humphrey closed the gap at the end but still lost to Richard Nixon. The Democratic ticket dropped from 43.1 million votes and 61.1 percent of the popular vote to 31.3 million votes and 42.7 percent. Republicans gained five Senate seats, five House seats, and five governorships.
Therein lies the real driver of Democratic panic: not just fear that they could lose the presidential race, in which Biden has been trailing in the polls since last fall, but fear that if the bottom drops out, Biden could drag down the party’s fortunes across the ballot. How likely is it that a Biden implosion could spread down the ballot? I decided to look at other modern presidential election cycles to see what they might tell us.
The Scene: The 2024 Senate Map
For the moment, I’ll focus on the Senate, where the stakes are highest. These seats won’t be up again until 2030 and Democrats are defending a lot of seats. Senate races are (unlike races for governor) impossible to separate from the national environment. But unlike in House races, Senate elections are typically high-profile enough that the candidates can build their own brand: One side or the other often runs well ahead of his or her national party ticket. That has gotten harder: Across 2016 and 2020, 68 of the 69 Senate races were won by the same party that carried the state at the presidential level (Susan Collins in 2020 being the lone exception).
The two complicating factors are Donald Trump and the map. The gravitational pull of presidential-approval ratings in midterm races would have predicted significant Republican gains in 2022, but Democrats instead picked up one Senate seat while losing none. That was driven in part by the Democrats’ strategy of turning it into a two-incumbent race, where Trump was hugely unpopular, independents turned against Republican candidates, and the Democrats had better turnout in the key Senate battleground states than they did nationwide.
On the other hand, the 2024 Senate map is outrageously favorable to Republicans, who hold just eleven of the 34 seats up for reelection. The bluest state with a Republican-held seat this year is either Texas or Florida, where Ted Cruz and Rick Scott are unlikely to lose if Trump is cruising. (There are also open seats in Indiana and Utah.) The last three times this set of seats was on the ballot were all blue-wave years: 2018, 2012, and 2006. Even so, amidst the Brett Kavanaugh hearings in 2018, Republicans knocked off an unprecedented four incumbent Democratic senators, all of them in red to deep-red states (Florida, Missouri, Indiana, and North Dakota) while Democrats were gaining seats in Nevada and Arizona.
It’s generally assumed that Jim Justice will easily pick up the retiring Joe Manchin’s seat in West Virginia, where Trump will win handily. That alone gets Republicans to 50 Senators if they hold their own seats. Two embattled Democratic incumbents who survived 2018 (Sherrod Brown in Ohio and Jon Tester in Montana) are also running in states that seem safely red at the presidential level. Brown and Tester will need to run well ahead of Biden to escape again — and 2018 was the only time in Tester’s career that he won a majority of the vote. Even in blue-wave years in 2006 and 2012 he got by on the skin of his teeth.
At the other end of the spectrum, in the blue states, Larry Hogan is the best possible candidate to make a real race for the open seat in Maryland, where it would take a national catastrophe for Democrats to produce a Trump victory. Democrats have won Maryland by double digits in the presidential race in every election since George H. W. Bush won it in 1988. New Jersey, which is safely blue at the presidential level, is a longer shot for Republican Curtis Bashaw that depends on the unlikely hope that scandal-tarred incumbent Bob Menendez (currently running as an independent) disrupts Democrat Andy Kim’s path.
In between, however, there are Senate races in nine states that range from likely Trump wins to states that could conceivably go on the board if Biden’s support sags badly enough: open Democrat-held seats in Arizona and Michigan, and Democratic incumbents in Nevada, Pennsylvania, Wisconsin, Virginia, New Mexico, Minnesota, and Maine. (That includes every state Biden won by single digits in 2020, plus Virginia and New Mexico, each of which he won by under eleven points). Republicans have plausible candidates in most of those races (Minnesota being the most dubious), although as of now, none of them really seem strong enough to win without help from a good national environment.
In short, it’s at least conceivable that Republicans approach their Senate landslide in 1980, when the party gained twelve Senate seats as Jimmy Carter lost 44 states to Ronald Reagan. But a lot needs to go very, very, very wrong for Democrats in order to deliver an outcome even close to that territory. Trump is still no Reagan. And in current polling, Republicans lead the RealClearPolitics average only in West Virginia and Montana.
Then again, it’s mid July. As I’ve detailed in past studies, the historical tendency is for Senate races to shift after mid September in the direction of the national trend — although that didn’t really happen in 2022. We’re still a long way away from even seeing the mid September baseline.
How likely is it that a Democratic presidential wipeout would shift the Senate well past 51 Republican senators?
Senate Coattails
Presidential victories haven’t always produced Senate coattails. Since 1948, the party that won the presidential election has gained Senate seats ten times, lost Senate seats seven times, and played to a draw twice. In two of the losing years, 2016 and 2000, the winning party lost the national popular vote in the presidential race. If we count those for the losing side, the winner of the popular vote has picked up seats twelve times and lost them five times.
But the coattails trend looks clearer if you focus on more recent cycles. In the last five elections, the winner of the national presidential popular vote has gained seats all five times — whereas 1980 was the only election between 1976 and 1996 where the same party won the White House while gaining Senate seats. Clearly, ticket-splitting isn’t what it once was.
Democratic incumbent senators have been particularly hard targets to take down in presidential years ever since the 1980 wipeout in which the party lost eleven incumbent senators. Between 1984–92, seven incumbent Republicans and three incumbent Democrats lost reelection in a presidential cycle. Between 1996 and 2004, it was even more lopsided: six Republicans and two Democrats (granted, one of those two was Tom Daschle, then the sitting Senate minority leader). Between 2008 and 2020, twelve Republican incumbents have lost reelection in a presidential year, and only one Democrat (that being Doug Jones in Alabama in 2020) has done so.
Of course, the Senate map has always mattered. Four times since 1972, the winning presidential party has lost more Senate races than it won: Democrats in 2020 (15–20) and 1996 (13–21) and Republicans in 2000 (15–19) and 1988 (14–19). But Republicans were defending a lot of territory in 2020, such that Democrats gained three seats even while losing five more races than they won. They missed a lot of opportunities. In 1976, by contrast, Democrats won 21 of 32 races but needed 22 just to break even. And that’s before you consider where the contested Senate races were held.
Senate Battlegrounds
To drill into the “where” and “by how much” question, I studied every battleground Senate race in a presidential-election year since 1992. I picked 1992 partly because it was the start of the post–Cold War political era, and partly because David Leip’s Political Atlas had the data. I defined “battleground” by the outcomes of the races. A Senate race qualified as a battleground if it met one of two qualifications:
- The final margin of victory was under ten points; or
- Party control of the seat flipped.
Out of 273 Senate races, that yielded 100 battlegrounds. That has a clarifying effect. In the non-battleground Senate races, the party that won the presidential race won the Senate race 51.4 percent of the time, but in the battleground races, the presidency-winning party won 59 percent of the time.
The battleground races were strongly aligned with presidential outcomes in those states. In the 100 races, the winning presidential party won 64 times, and won 59 Senate races. That means that the winning presidential party would be expected, 95 percent of the time, to win battleground Senate races in the states it won at the presidential level.
That’s not an iron one-to-one rule, of course: In those 100 races, the losing presidential party won 15 Senate races in states where they lost the presidential race. But they also lost ten Senate races in states where they won the presidential race. So, in total, the winning candidate’s party won both races statewide 45 out of 100 times, lost both races statewide 31 out of 100 times, and resulted in a split ticket 25 times. That means the tendency for Senate races to follow the presidential race has been far from uniform in individual cases.
Moving from outcomes to margins, how far ahead of the national ticket have Senate candidates run when their presidential candidate was going down to defeat? Out of 100 battleground races, the losing presidential side ran three points (net) ahead of the presidential ticket in 45 races, five points in 30 races, and eight points in 23 races — almost a quarter. But there’s an important caveat. That was much more common in 1992–96 than in 2000–04, and much more common in 2000–04 than in 2008–20.
In particular, the deepest pool of voters for Senate candidates when their presidential standard-bearer was losing was the H. Ross Perot voters in 1992–96. If you assume zero ticket-splitting (which is mathematically impossible, as well as being unrealistic, but bear with me), across Senate battleground races, Republican Senate candidates won 71 percent of the Perot vote in 1992 and 115 percent of the Perot vote in 1996.
That raises a question: If Biden sags, can Democrats prop up their Senate tickets with Robert F. Kennedy Jr. voters? They would have to lay off their strident attacks on RFK. They might, at some point, need to do what Republicans did in 1996, when almost half of their battleground Senate candidates ran five points ahead of Dole: Give the green light to Senate candidates to argue openly that voters need to put a check on the guy who was actually going to win the election. Dole could do that: As a man who had led Senate Republicans until five months before the election, he was a loyal party man capable of subordinating his own ego to the team. If Biden was built that way, he wouldn’t still be running — and Democrats will be hard-pressed to start openly acknowledging a looming Trump victory so long as they are campaigning on the idea that a Trump win means there will be no more elections in America, ever.
On average, across states with battleground Senate races, the losing party in the presidential race has run 1.7 points ahead of their national ticket. But again, that has varied from year to year, and is lower if you start in 2000 after the Perot movement collapsed. Also, an individual race can throw things off. The biggest victory since 2000 in running ahead of the national ticket was pulled off by Democrats in 2004, but that’s almost entirely a factor of the Illinois Senate race, where the GOP’s top recruit pulled out due to a sex scandal of sorts, and as a result, Barack Obama won by 43 points in a state John Kerry won by nine. The Democrats only actually won two battleground Senate races that year, the other being Colorado.
By contrast, Obama actually ran almost five points behind the Democrats’ Senate candidates in battleground races in 2008 — or, one might say, John McCain ran five points ahead of Senate Republicans. Either way, McCain carried five of the twelve states with key Senate races in 2008, four of them with at least 56 percent of the vote, but Republicans won only three of those.
Post-Perot, there has usually been one side but never both that managed as many total votes in the battleground Senate races as in the presidential races in those states. But since 2000, only the Democrats in 2004 — again, due mainly to Obama’s blowout race — have actually gotten more votes across those Senate races than their presidential ticket.
How much does partisan turnout matter? If you look at turnout for the two party tickets, combined, as a share of voter-eligible population, turnout was highest in 2020, with 64.9 percent of eligible voters casting ballots for one of the two major-party candidates — and the Senate races ran very close to those. The two highest turnout years after that were 2008 and 2004. It was lowest in 1996, with just 46.5 percent of eligible voters voting for either Dole or Bill Clinton. So, if Biden is losing, Democrats may be better off pursuing a strategy that isn’t focused on maximizing their base turnout across the ticket.
