The fight for control of the United States Senate is really close. It’s basically a coin-flip. Democrats have a 53 percent chance of winning a Senate majority according to FiveThirtyEight’s polls-plus model, or a 52 percent chance according to our polls-only forecast. There is about a 16 percent chance, in fact, that we’ll end up with a 50-50 Senate and the presidential race will decide control. (The vice-president breaks ties in the Senate.)
There are structural factors working in both parties’ favor. Republicans hold 24 of the 34 seats that are up for reelection in 2016. Democrats, therefore, have more opportunities to pick up seats than Republicans. The political geography also favors Democrats. Of the 10 seats Democrats are defending, none are in states that were won by either John McCain or Mitt Romney in 2008 or 2012. Republicans, by contrast, are defending a lot of turf that’s either blue or purple. That’s because the last time these seats were up for reelection was in 2010,1 a GOP-wave year when Republicans won a lot of seats in Democratic-leaning territory. In total, nine of the GOP-held seats up for grabs on Tuesday are in states that President Obama won in either 2008 or 2012.
On the other hand, Republicans are defending those seats — that is, they have more incumbents on the ballot. Incumbent candidates generally outperform non-incumbent candidates, all else being equal. And indeed, Republican incumbents are generally polling better than Donald Trump.
Speaking of the presidential race, one of the clearest trends this cycle has been that Democratic chances of taking back the Senate have generally run hand-in-hand with Hillary Clinton’s chances of winning the presidency.
The past few weeks have been no different. Democrats, who in the middle of October held a clear advantage in their fight for a Senate majority, have seen their fortunes fall as the presidential race has tightened. This matches a pattern seen over the previous few election cycles: More voters casting their ballot for the same party in both Senate and presidential races. There is some good news for Senate Democrats in this tightening, though: If voters think Clinton is less of a sure bet to win the presidency, they may be less likely to vote for Republicans in down-ballot races in order to act as a check on her power. That seemed like a possibility after Clinton opened up a large lead over Trump after the first debate in late September.
OK, so that’s the broad view of where things stand. Republicans currently hold 54 Senate seats. Democrats have 46 (including independents Angus King of Maine and Bernie Sanders of Vermont, who caucus with them). So Democrats need to pick up a net gain of four seats for a tie, or five seats to get an outright majority.
The FiveThirtyEight Senate forecast has 11 seats to watch. Let’s go through those 11 seats, from most likely to go Democratic to least likely, according to the polls-plus model. I’ve also added two extra wild-card races to keep an eye on as the returns come in Tuesday.
Likely Democratic pickup
Illinois (97 percent chance of a Democratic win in polls-plus forecast)
Republican Sen. Mark Kirk was always facing an uphill battle in solidly blue Illinois. Clinton is favored to win the state by more than 10 percentage points, and no Republican has won a Senate race in Illinois in a presidential election cycle since 1972. Kirk wasn’t helped when Democrats were able to recruit Rep. Tammy Duckworth, an Iraq-war veteran, to challenge him. Kirk has the voting profile (moderate) and geographic base (suburbs of Chicago) that a Republican needs to win a statewide race in Illinois, but even so he would have had to run an almost perfect campaign to have a chance. He hasn’t. He also had to fight off the belief among some voters that a stroke he suffered while in office has limited his effectiveness. If Kirk is to upset Duckworth, he’ll have to hold her margin in Cook County (Chicago) under 35 percentage points and lose pretty much no place else.
Probable Democratic pickup
Wisconsin (86 percent)
Wisconsin may be closer than our forecast suggests: Our model may not have fully picked up a late shift in the polls toward incumbent Republican Sen. Ron Johnson. Democrats were able to recruit former Sen. Russ Feingold to take on Johnson, and polls showed Feingold as the favorite for most of the year. Wisconsin hasn’t voted for a Republican for Senate in a presidential year since the Republican wave of 1980, and Johnson has a very conservative record. Johnson, though, appears to be mounting a comeback. Both Democratic and Republican groups have invested heavily in the race in the past few weeks. And Marquette University, perhaps the most reliable pollster in Wisconsin, put Johnson down just 1 percentage point in its final survey of the race. Other recent polls also point to a close contest. Feingold is still a favorite, and he will probably be helped by Clinton’s coattails — she holds a healthy lead over Trump in the state. But keep an eye on eye on the traditionally Republican-leaning counties outside of Milwaukee to see by how much Johnson is outperforming Trump. If it’s by a lot, as some polls suggest, Johnson has a shot.
Pennsylvania (67 percent)
Yet another seat won by a Republican in a midterm election that will be harder to hold in presidential year. (Are you noticing a trend?) Republican Sen. Pat Toomey is one of the most conservative senators in Congress, despite representing a state that tends to vote Democratic in presidential elections. Beltway Democrats were able to get their preferred candidate, Katie McGinty, through a competitive primary, but the race between McGinty and Toomey was neck-and-neck through early October. Since then, however, McGinty has opened a small lead in almost every poll. Toomey has had his hands full with Trump, trying to keep his distance without outright disavowing the GOP nominee. McGinty, on the other hand, is probably going to be helped by Clinton, who is a favorite to win the state. Watch Chester County in the southeast part of the state. Chester County traditionally votes Republican, but Clinton is expected to overperform there. Toomey may need some ticket-splitting in that county to win the state.
Missouri (44 percent), Nevada (56 percent), New Hampshire (52 percent)
These are the three seats where the Senate is most likely going to be decided. If Democrats win two of these three, it’s difficult to see how they don’t end up with at least 50 Senate seats. If they don’t, they’ll have to pick up at least one seat that appears to be trending away from them.
Missouri is perhaps the most interesting of the three because it’s the state where we have the best chance of seeing one party win the Senate race and the other win the presidential race. Although Trump should win easily in Missouri, Democrat Jason Kander has made the Senate race there quite competitive. Running on his biography as an Army veteran and outsider, Kander has fought his way to a near tie against Republican Sen. Roy Blunt. Blunt seems to have been caught flat footed, tagged as a Washington insider while his party’s presidential nominee runs against Washington. Still, Kander seems to have hit a wall in recent polls. Blunt has the slightest of edges in our forecast, but the race is really too close to call. One county to pay extra attention to is Jefferson County, near St. Louis, which has a history of matching the statewide vote. In 2012, the final margins in the gubernatorial, presidential and Senate races in Jefferson looked a lot like the state as a whole.
If turnout for Trump might help Blunt in Missouri, he’s likely to hurt Republican Joe Heck in Nevada. Heck is running against Democrat Catherine Cortez Masto to fill the seat of retiring Democratic Sen. Harry Reid. Heck, like Toomey in Pennsylvania, has struggled to figure out how exactly to deal with the Republican presidential nominee. While the polls point to a close race, the early vote in Nevada has been decidedly negative for Republicans. That doesn’t mean Heck is toast. Four years ago, Republican Sen. Dean Heller was able to win by 1 percentage point even as President Obama carried Nevada by 7 points. Heller, though, benefitted from a Democratic opponent who had ethics problems. Heck doesn’t have that luxury. Heck’s best chance is to win a larger share of Latinos than Trump in Clark County, where they are turning out in large numbers.
Unlike in Nevada, the recent trends have been good for Republicans in New Hampshire. Trump has closed the gap with Clinton there, and Republican Sen. Kelly Ayotte has made her race with Democratic Gov. Maggie Hassan a tossup again. There’s a touch of irony in Ayotte’s fate being tied to Trump’s; she has said she won’t vote him after facing backlash for calling him a good role model during a debate. The tight race shouldn’t come as a surprise: New Hampshire is the ultimate swing state. On the presidential level, it has been within a few percentage points of the national vote in every election since 1996. This year looks to be no different. With two well-known candidates running, the race looks like it will come down to the wire. The candidate who carries the independent vote in the state (which has one of the largest independent votes in the nation) will probably win the race.
Plausible Democratic pickups
Indiana (40 percent), North Carolina (30 percent)
Democrats thought they hit the jackpot when they were able to recruit former Indiana Sen. Evan Bayh into the race to replace retiring Republican Sen. Dan Coats. Bayh had been a two-term governor and senator. He also had a giant fundraising war-chest and opened the race with a large lead over Republican Todd Young. But that advantage has evaporated in the past few months. That’s not totally surprising given Trump is expected to easily carry the state. Bayh, though, hasn’t helped his cause. He’s been tied to Washington lobbyists and has been questioned on how much time he really spends in Indiana. Young, for his part, has improved his name recognition. The result is that Young took his first lead in a public poll in the last two weeks. Still, the race is close and could be won by either candidate. To win, Bayh will need strong turnout in heavily black Lake County (Gary) and Marion County (Indianapolis).
Democrats have a different problem in North Carolina. Republican Sen. Richard Burr had probably hoped the slight Republican tilt of the Tar Heel State would ensure his re-election, especially after the Democrats’ strongest potential candidate, Attorney General Roy Cooper, decided to run for governor instead. That left Democrats with Deborah Ross, a lawyer and former member of the North Carolina House of Representatives. Ross has kept it close with Burr the entire campaign, but she’s never been able to overtake him. That makes this race the only one examined so far in which the Democrat has never led in the FiveThirtyEight polls-plus forecast. There are very few swing voters in North Carolina, so the race is all about turnout: If Clinton outperforms expectations in the presidential race by just a few points, she might carry Ross across the finish line.
Longshot Democratic pickup
Florida (13 percent)
Democrats were a favorite to win the Senate seat in Florida until Republican Sen. Marco Rubio changed his mind and decided to run for reelection after all. Rubio has held a consistent lead over Democrat Patrick Murphy. Rubio has strong ties to Cuban communities in Florida, and he’s expected to vastly exceed Trump’s poll numbers among Hispanics in the Sunshine State. Keep an eye on heavily Cuban Miami-Dade County to see how much Rubio is outperforming Trump’s margins with Cubans. Murphy, for his part, has been called out for exaggerations about his record and has attacked Rubio for supporting Trump, even after Rubio called Trump a con artist in the presidential primary. For whatever reason, the attacks don’t seem to be sticking. Rubio’s advantage is so large at this point that it’s difficult to imagine Murphy winning even if Clinton wins the state.
Really long shot Democratic pickups
Arizona (3 percent), Ohio (2 percent)
Early polls indicated that incumbent Republican Sens. John McCain of Arizona and Rob Portman of Ohio might be in trouble. And Democrats recruited strong candidates: Ann Kirkpatrick in Arizona and former Gov. Ted Strickland in Ohio. But McCain and Portman have been perhaps more successful than any other Republican Senate candidates in distancing themselves from Trump. Portman was also helped by a massive fundraising advantage over Strickland.
Colorado (95 percent), Louisiana (14 percent)
No one has talked about Colorado as a potential Republican pickup. Democratic Sen. Michael Bennet has only trailed in one poll all year against Republican Darryl Glenn. Colorado, though, is still a swing state in presidential elections, even if Clinton is favored to win it this year. It also elected a Republican to the Senate in 2014. And Bennet’s lead in the polls has shrunk to within single digits over the past month. Bennet will almost certainly still win, but don’t be surprised if the race ends up closer than many once thought it would be.
Louisiana is something entirely different. The state will hold on a primary on Tuesday in which all the Democratic candidates and Republican candidates run against each other. As long as no candidate gets 50 percent (and there’s basically no chance of that happening), the two top vote getters regardless of party affiliation will advance to a runoff. It’s expected that one of those candidates will be Republican John Kennedy who leads in the polls. But Democrats Foster Campbell and Caroline Fayard are not too far behind Kennedy in second and third respectively. With five Republicans projected to get at least 4 percent of the vote, it’s possible that Republicans will split their vote enough that Kennedy drops to third and both Democrats end up advancing. Keep in mind that polls are much more likely to be wrong in primaries than in general elections.
It is possible that one party will win most of the battleground races. The polls are sometimes off in the same direction in multiple races. That happened in 2014 when Republicans beat their polls across the board. If that happens again, then some of the long-shot candidates may end up as surprise winners.
Still, the polls show a close overall race — about as close as you can get. In nearly half of our simulations, the Senate ends up as either tied or 51-49 in one direction or the other. So as you watch the presidential race unfold, remember that it’s not just determining control of one branch of government but potentially half of another branch as well.