(CNN) Senate races across the country are converging with the trend in the presidential race in that same state, underscoring the increasingly parliamentary nature of American elections.
In 2016, for the first time since the direct election of all senators began in 1914, every Senate race finished the same way as the presidential contest in that state. It now appears likely the 2020 election will offer either a repeat or an extremely limited number of exceptions, far fewer than seemed possible even a month ago.
Detailed results provided to CNN from a wide array of public polls by different pollsters released through October show that Senate candidates in both parties are attracting vanishingly little support for voters who back the other party's presidential candidate.
A presidential race like no other
"Because the partisan makeup of the electorate is so polarized in a presidential year, all of these lines converge," says Republican pollster Gene Ulm.
This pattern affects both parties, but it creates a greater complication for Democrats, who need a net gain of three seats Tuesday to regain Senate control if former Vice President Joe Biden wins the presidency. Almost all of this year's competitive Senate races are in states that voted for Donald Trump in 2016 (Michigan, Maine and Colorado are the big exceptions). That means that to win a Senate majority Biden must either flip some states that backed Trump last time or Democratic Senate candidates must win a state that votes for Trump again. Because the latter has grown more unlikely, Democratic hopes of winning the Senate may pivot on whether Biden can win either North Carolina, Georgia or Iowa, the states that seem most likely now to decide Senate control.
"In part because of the polarization that exists in the last decade, these Senate races -- especially in battleground states where there has been substantial spending -- have basically been running in line with the presidential," says Democratic pollster Matt McDermott.
Earlier in the election cycle, Democrats hoped that several of their candidates might win even if Biden lost narrowly in their states -- a list that included Mark Kelly in Arizona, Steve Bullock in Montana, Theresa Greenfield in Iowa, Cal Cunningham in North Carolina and maybe Jaime Harrison in South Carolina. Republicans touted the potential for Sen. Susan Collins in Maine and challenger John James in Michigan to surmount a presidential loss.
Such divided results remain possible; analysts in both parties give Kelly, Bullock and Collins the best chance of surviving a presidential loss. But operatives in both parties closely following the Senate races say the chances of such split outcomes have dwindled in recent weeks as voters have increasingly sorted along the fault line of support or opposition to the presidential nominees.
"The death of split-ticket voting is real unless you are talking about a dead heat at the top of the ticket," said one Republican strategist closely involved in the party's Senate planning. "You are pretty much at the whims of what the national atmospherics are."
This pattern creates a near insurmountable challenge for two incumbents: Democrat Doug Jones in Alabama and Republican Cory Gardner in Colorado. Each faces very long odds in states that are likely to back the other party by a substantial margin in the presidential race.
Yet even beyond them, there's no candidate who faces a clear path to victory in a state likely to vote the other way for the top of the ticket.
"I think that ... we are going to see the highest rates of straight-ticket voting in the last 70 years, as far back as we've got data," says Alan Abramowitz, an Emory University political scientist who has extensively studied the relationship between voting in presidential and congressional contests.
Both parties, for instance, consider Kelly in Arizona probably the Senate candidate most likely to win even if his state votes against his party for president. But the latest surveys show even he is attracting very little support from Trump voters, while his opponent, Republican Sen. Martha McSally, draws even more negligible support from Biden voters. In a CNN poll released Sunday, Kelly was winning 6% of voters backing Trump -- not much, but double the 3% of Biden voters supporting McSally. The decline of Kelly's crossover appeal may prove academic, however, because polls have consistently shown Biden leading narrowly in the state; that leaves McSally facing extremely long odds.
Mike Noble, a nonpartisan Phoenix-based pollster, says McSally's biggest problem is that independents are abandoning the President, who she has lashed herself to unreservedly.
"Independents have always leaned center-right in Arizona, hence the strong GOP track record electorally here," says Noble, whose firm OH Predictive Insights also found minuscule split-ticket voting between the Senate and the presidential race in its final poll there. "Now, they are leaning not just left -- but significantly left."
Other expectations of big crossover support are fizzling for both sides. Republicans had high hopes that James, an African American Army veteran and business executive in Michigan, might peel away a reasonable share of Democratic voters. They say they still see some signs of that in their private polling, but it's not evident in the public surveys. A new CNN poll last weekend showed James winning just 4% of Biden voters, even less than the modest 6% of Trump voters backing Democratic Sen. Gary Peters. An EPIC/MRA survey released Sunday had similar results.
Earlier this year, Democrats had likewise anticipated that North Carolina nominee Cunningham could win even if Biden did not, because Republican Sen. Thom Tillis was caught in a crossfire between moderates who considered him too close to Trump and conservatives who thought him too disloyal. Democrats also believed that Cunningham, a small-town lawyer and military veteran, could run a little bit better than Biden with rural voters. But as Cunningham has been snared in a scandal surrounding an alleged extramarital affair, the presidential and Senate races have largely converged.
The most recent New York Times/Siena College poll showed Tillis winning just 3% of Biden voters and Cunningham just 4% of Trump supporters, according to detailed results provided by Donald Levy, director of the Siena College Research Institute. CNN's poll last weekend put the numbers at 6% of Biden voters for Tillis and just 2% of Trump voters for Cunningham. That still gives Cunningham a slim edge because most polls show Biden leading narrowly in the state -- but if Trump pulls out a victory by generating another big turnout of rural Whites, Cunningham could be swept away too, both sides believe.
In Iowa, the Democratic nominee, Greenfield, had been running ahead of Biden most of the fall, leading Republican Sen. Joni Ernst even in surveys where Biden narrowly trailed Trump there. But the latest Des Moines Register poll, released Saturday, showed Trump and Ernst leading, the latter for the first time this year. Many believe that survey exaggerates the GOP's advantage in the state, but other polls suggest it will be hard for Greenfield to win unless Biden does: the New York Times/Siena poll in the state showed her winning just 4% of those backing the President, even less than Ernst's modest 7% among Biden voters.
Democrats nationwide have contributed enormous sums to Harrison's campaign against Republican Sen. Lindsey Graham in South Carolina, but this dynamic looms as a huge headwind for the Democrat: Trump is virtually certain to carry the state, and the NYT/Siena poll in South Carolina found Harrison winning just 2% of Trump voters. Trump's strength in the state represents a likely insurmountable barrier as well for competitive Democratic Senate candidates in Kansas and Alaska.
Georgia may test the flip side of that equation. Both Republican Sen. David Perdue (at 3%) and his Democratic opponent, Jon Ossoff (at 2%) drew very little support from those backing the other party's presidential candidate in a recent NYT/Siena survey. But with Biden showing late strength in the state, Republicans are growing more concerned he might carry Ossoff past Perdue, and even potentially past the 50% support level required to avoid a January runoff under state law. A second Georgia US Senate race involving multiple candidates seems more likely to produce a runoff, but even there a strong Biden finish conceivably might lift the leading Democrat, Raphael Warnock, to a winning majority.)
Texas also appears headed for a surprisingly close finish in the presidential race. There it's less certain that a narrow Biden win would flip the Senate seat, because Democrat MJ Hegar, who hasn't raised as much money as other Democratic Senate challengers, isn't well known across the sprawling state. That might allow Republican Sen. John Cornyn to squeeze past even if Biden wins, though Democrats still believe there's an outside shot that a Biden win (which itself would be an earthquake) might lift Hegar.
With the popular governor, Bullock, running, Montana may be the Democrats' best remaining chance to win a state Biden is unlikely to carry. Bullock has remained close in most polls with GOP Sen. Steve Daines, though Trump has consistently led (albeit by much less than the 20-point margin he amassed in 2016). But the decline of split-ticket voters is complicating the pathway for Bullock too: In a recent NYT/Siena survey, he was winning just 5% of Trump voters, not much more than Daine's 3% of Biden voters.
Jason Adkins, an assistant professor of political science at Montana State University (Billings), thinks there may still be an opening for Bullock. The university's recent poll also found that each candidate was winning only 4% of voters backing the other party's presidential candidate. But, Adkins notes, while just 1% of Biden voters said they were undecided in the Senate race, 8% of Trump voters remained unsure. That differential could mean Bullock is fishing for his final votes in unfavorable waters -- or that a significant slice of Trump voters just can't quite commit to Daines.
"I'd say the latter is more plausible considering polling in the past few months has shown Daines' support has consistently been several points lower compared to Trump's numbers," Adkins told me in an email. "Trump voters are still not completely sold on Daines." (The MSU poll showed Bullock slightly ahead overall, though the NYT/Siena survey gave Daines a modest lead.)
No senator has been more renowned for attracting crossover votes than Maine's Collins. In 2008, she won easily even though Barack Obama carried the state by 17 percentage points. In 2014, exit polls found that she won about 40% of Democrats (albeit against marginal opposition).
Those days are gone. Collins' votes to confirm Supreme Court Justice Brett Kavanaugh and to acquit Trump of the charges brought against him in impeachment have stamped her as much more partisan to state voters (even though she also voted against repealing the Affordable Care Act and confirming Justice Amy Coney Barrett). Sara Gideon, the Maine House of Representatives speaker, who is running against Collins, has tied her to the Republican leadership in a year when Biden is likely to win the state by a large margin. Gideon has hammered Collins for refusing to say whether she is voting for Trump and also stressed that if reelected the incumbent's first vote would be to elect Mitch McConnell as majority leader -- a potentially powerful argument in a state likely to vote so heavily Democratic for president.
All of this has taken a toll. Collins still draws more crossover support than any other Senate candidate in the recent public polling but not nearly as much as she once did: The NYT/Siena survey showed her winning 14% of Biden voters, while a recent Colby College survey put her number at 11%. Both surveys showed Collins slightly trailing Gideon.
Operatives in both parties say Collins is keeping the race closer than many expected given the size of Trump's deficit. But even so Trump's unpopularity has left her in a deep hole. Adding to Collins' challenge: The state's unique ranked-choice voting system means Gideon could benefit when voters are reallocated from a left-leaning minor party candidate on the ballot.
While Collins is drawing more crossover support than Senate candidates in other states, "I think it's a lot less than she expected," says Nicholas Jacobs, a Colby College visiting assistant professor of government who works on the poll. "I think she thought she would be getting more. She thought she would be able to thread that needle and find that place between turning off the Republican base in Maine and then being able to shift to the general electorate and then win over that moderate voter she's successfully won over in every reelection."
But, Jacobs adds, "from what I see -- and I do not say it as a partisan, I say it for what the data suggests -- it might be impressive by comparative standards, but it's not going to be impressive enough to win. Susan Collins was able to escape Trump more than anybody else, but nobody was able to escape Trump."
The microscopic level of crossover support that the 2020 Senate candidates are attracting in polls intensifies a long-term trend. As Abramowitz has documented, polling results from the University of Michigan's American National Election Studies show a steady decline since the 1970s in the share of voters who "split their tickets" -- that is, vote for one party in presidential races and for the other in Senate or House elections.
Split-ticket voting was actually relatively rare in the early 20th century, but it rose quickly after around 1960 and peaked during the 1970s and 1980s. (Split-ticket voting rose largely because conservative Democratic voters in the South started voting for GOP presidential candidates long before they backed Republicans for Congress, and the reverse happened for moderate GOP voters along the coasts.) At that point, in the American National Election Studies between 23% and 29% of voters consistently voted for presidential candidates from one party and Senate candidates from the other. But in recent years, no more than 1-in-7 voters have split their tickets, the surveys have found. If anything, the public polls suggest that number may dip even further this year.
"It's such a dramatic contrast with what was going on 40 or 50 years ago in the era of candidate-centered electorate," Abramowitz says. "You'd see 25 or 30% crossover voting. If you had a popular incumbent in a race you could get really a huge percentage of votes to cross over for you."
Can you take selfies at voting polls in your state? Learn the rules here
The tightening correlation between presidential and Senate results is both a cause and a reflection of Washington's increased polarization. The decline of split-ticket voting has heightened polarization in Washington by reducing the number of cross-pressured senators who have an inherent political interest in finding common ground with voters who backed them but also supported a president of the other party.
Yet many analysts consider it understandable for voters to treat Senate elections, in effect, as parliamentary contests because Congress is now shaped by party-line voting much more than before. With the parties now presenting much more unified congressional blocs defined by support for or opposition to the president, Jacobs notes, it's become more difficult, almost anachronistic, for candidates to try to sell themselves on constituent service or their local ties.
"Sometimes I watch Susan Collins and it's pleasant, that she's campaigning in this bygone era of American politics that I read about in textbooks: that people care about their senator bringing home the bacon, or jobs to the Bath Iron Works," he says "Part of me wishes that was true, but I don't know how many Mainers are interpreting their political world that way."
The 2016 election, when every Senate race followed the presidential outcome in that state, marked a new peak in this process. Even as late as 2008, five Democrats won Senate seats in states that backed John McCain over Obama and Collins triumphed in Maine despite Obama's win there. In 2012, four Democrats and Republican Dean Heller in Nevada won Senate seats in states that voted the other way for president.
But each party lost almost all of those crossover seats the next time those senators faced reelection. While Collins survived in 2014, Democrats that year lost all five seats in their crossover states from 2008. Then in 2018, West Virginia's Joe Manchin was the only crossover winner from 2012 in either party who won reelection.
Send us your stories about voting issues
After those results, if all, or almost all, of the Senate races this year again follow the presidential outcomes, that will send a dramatic message about the continuing evolution of American politics into a quasi-parliamentary system in which each party is consolidating control over large swathes of the country up and down the ballot.
These dynamics are creating a political landscape that is not only more divided, but also more volatile. The ability of each party to tighten control over most of the Senate seats in the states that favor it in presidential races has left them closely balanced in the chamber overall. After this election, Democrats could control all 40 Senate seats in the 20 states that backed Hillary Clinton and are likely to also support Biden on Tuesday. In parallel, Republicans seem likely to control all but two or three of the 40 Senate seats in the 20 states Trump won last time that he is most likely to carry again. With so many seats locked up that securely, neither side in recent decades has established a Senate majority big enough to survive for any length of time.
If Democrats recapture control this year, it will mean that neither party has held the Senate for more than eight consecutive years since 1980, a 40-year span. Never before in American history has either party failed to control the Senate for more than eight years over so long a period. The paradox is that the decline of split-ticket voting is simultaneously making the outcomes of individual Senate races more regimented -- and the balance of power in the overall system more unstable, as Tuesday's results could demonstrate again.