I’m going to date myself, though my intent is analytical, not autobiographical.
The first Senate campaigns for which I served as pollster were those of then-Rep. Al Gore (D-Tenn.) and the late Sen. Carl Levin (D-Mich.) in 1984.
A presidential election occurred the same November day. Gore performed 19 percentage points better than Democratic nominee Walter Mondale, while Levin exceeded Mondale’s vote in his state by 12 points.
The next cycle, I polled for three key Senate candidates who were then House members. Colorado’s Tim Wirth did 15 points better in his state than Mondale had done two years before. The late Harry Reid surpassed the presidential nominee’s vote by 18 points in Nevada, while South Dakota’s Tom Daschle exceeded the Democratic presidential vote in his state by 15 points.
Senate and presidential votes seemed largely disconnected — and it wasn’t just my clients.
In 1986, 59 percent of Senate races were won by the party that lost the 1984 presidential contest in that state. In fairness, it was the peak year for presidential/Senate mismatches, but through the late 90s, mismatches were greater than 30 percent.
By 2020, things changed dramatically: Just one Senate election that year produced a mismatch between the party of the Senate victor and the winner of the state’s presidential contest.
2022 also yielded just a single mismatch between the party of the Senate winner and the party of the 2020 presidential victor.
Today, the presidential vote seems like a near perfect indicator of the Senate vote.
My Mellman Group colleague Edward Wu and I examined this phenomenon from a different perspective.
We looked at the difference between the Senate vote and the presidential vote (or the previous presidential vote for a midterm) over the course of full Senate cycles. One third of the Senate stands for election every two years, so over three two-year cycles, all 100 senators face the voters.
During the 1984-88 Senate cycle of my polling youth, on average, Democratic Senate candidates outperformed their presidential nominee by over 9 points.
By the 2018-22 Senate cycle, that average difference plummeted to 1.6 percentage points.
Of course, averages can obscure as much as they illuminate. So, let’s dig a bit deeper into the data to capture some important nuance.
During the three latest Senate election years, no Democrat ran more than 23 points ahead of their presidential candidate.
Only 26 percent of Democratic Senate candidates ran 5 points or more ahead of the presidential nominee.
Over half (52 percent) ran within 3 points or less of their presidential candidate.
Numbers like these give Democrats heartburn about retaining control of the Senate. Fighting in solid Trump states like Montana and Ohio seems almost quixotic.
However, while winning these states will be quite difficult, the data offers some basis for hope.
Take Montana. Winning reelection will require Sen. Jon Tester to run at least 11.2 points ahead of Vice President Harris’s current poll average (as per 538.com).
That’s admittedly rare — it happened only four times since 2018. But it’s squarely in the realm of the possible.
In fact, one of the four previous cases was Tester himself, who bested Hillary Clinton’s 2016 showing by nearly 15 points.
If Sen. Sherrod Brown (D-Ohio) overperforms Harris’s current polling average in his state by the same margin he outperformed Clinton in 2016, he too will be reelected.
That said, there’s no question that presidential and Senate votes are becoming increasingly correlated. Why?
Correlation is not causation, so this does not necessarily imply that presidential voting is somehow causing Senate voting.
More likely the similarity in vote emerges from fundamental political facts of life I’ve pointed to here on several occasions — partisanship and polarization.
Most people have a partisan and/or ideological identity that puts them on one side of the political divide or the other. Increasingly, those identities are mutually reinforcing (Republicans are conservative, Democratic are liberal), rendering them even stickier.
Those political identities do far more to structure our thinking about candidates and issues than anything else.
Polarization makes it even harder for people to escape the gravitational pull of partisanship. Democrats dislike Republicans more than ever before (at least since the advent of polling), while Republicans also dislike Democrats more than ever.
Crossing the lines drawn by those identities has become more psychologically costly and therefore increasingly rare.
In response to the same partisan pressures, presidential, Senate, House and even gubernatorial votes are becoming more highly correlated.
The precise strength of that correlation across states will go far in determining which party controls the Senate in 2025.
Mark Mellman is a pollster and president of The Mellman Group, a political consultancy. He is also president of Democratic Majority for Israel.
