I'm going to contradict that a teensy bit right now-most elections probably don't entirely come out to turnout, or at least not in the way that that cliche is meant (unless taken literally). The reality is that most elections are not that close, even what we conventionally think of as close elections. While we see probably 8-10 truly competitive Senate races this year, for example, my gut says that on November 6th, only 2-3 of those races will have been so close a last-minute push (experts disagree on how much a strong ground game is worth, but it's no more than 2-points, and some argue even less) would have made the difference. Most of the races in Arizona, Missouri, Indiana, Florida, and the like this year will be decided by 4-5 points, maybe even into the double digits, because the polls underestimated a candidate or a wave or any number of other factors. In other words, your vote, or even a number of votes, probably won't ultimately make the difference in the battle for the Senate.
Before I start getting irate comments, I want to point out three things here, as you'll see throughout this article. One, voting always matters-enough people thinking that their vote doesn't matter and therefore they don't show up certainly decided every election on this list, and undoubtedly more of them than I can mention in one article. Two, you don't know which Senate races are going to be the ones that are decided by a slim margin. We'll profile several races on this list that were on virtually no one's radar that just got decided by a few handful of voters, either protecting an incumbent party or throwing them out. And third, and most importantly, you have no idea what that one senator will be voting upon during his or her tenure in office. I point this out a lot on Twitter, but the reality is that four current Republican senators won their last elections by less than two-points. Two of them have their Democratic opponents win instead, and Brett Kavanaugh is Robert Bork. Neil Gorsuch is Merrick Garland. And names like Scott Pruitt and Betsy DeVos are political asterisks rather than cabinet secretaries. Every election matters, particularly for the Senate which is like playing a game of chess. So, you need to vote, even if only 2-3 states will ultimately be close enough where your efforts are close to making the difference. That fact, that only a few states are going to make the difference, should make you want to vote more, not less, in my opinion, as you genuinely don't know if you're one of the few people that could make the difference. And for the record, I've been "one of those people" on this list and it was a ballot that I was truly apathetic about casting until I realized how much it mattered on Election Night.
But how often does this happen, and what do those races look like? I thought it would be fun to profile each Senate Cycle from this century below and give you an idea of the kinds of races that tend to be close, and what races ultimately were slim wins. I also want to give you an indication of some of the surprising senators that might not be in office anymore were it not for these elections, proving that not only is this, say, a good chance to oust someone like a Marsha Blackburn or Ted Cruz, it's possibly the only opportunity Democrats in those states have.
|(Almost) Sen. Katie McGinty (D-PA)|
Ultimate Net for Cycle: Democrats +2
Won By Less Than Two Points: New Hampshire (Gov. Maggie Hassan (D) won by 0.2 points over incumbent Sen. Kelly Ayotte (R)) and Pennsylvania (Sen. Pat Toomey (R) won by 1.5 points over Chief of Staff Katie McGinty (D))
What the Polls Were Saying: I'm going to try and use the Real Clear Politics average as long as their data will allow during this article since it's arguably the best way to track poll performance vs. actual results, and looking at their polling averages, both of these races were something of an upset. Ayotte led by +1.5 points in the polling averages, overtaking Hassan in the last week of the campaign, while McGinty had led for most of the fall in Pennsylvania, and was at +2.0 points. Ultimately what mattered the most was the presidential race was far more important than anyone thought in straight-ticket voting. While I don't have the precise data on ticket-splitting, the margin for Hillary Clinton in New Hampshire (+0.4-points) and Donald Trump in Pennsylvania (+0.7-points) was less than a percentage point difference between the two candidates. 2016 was the only year since the passage of the 17th Amendment where there were no splits between the POTUS/Senate ultimate winners (though the margins in some states saw dramatic differences between the two offices' candidates like in Missouri), so a stronger push by Clinton could have elected McGinty in Pennsylvania and vice versa for Trump in New Hampshire for Ayotte, though in the latter's case that may not have been possible as Ayotte didn't endorse Trump and was actively distancing herself from him at the time (a problem that Clinton/McGinty didn't have).
|(Almost) Two-Term Sen. Kay Hagan (D-NC)|
Ultimate Net for Cycle: Republicans +9
Won By Less Than Two Points: Colorado (Rep. Cory Gardner (R) won by 1.9 points over incumbent Sen. Mark Udall (D)), North Carolina (House Speaker Thom Tillis (R) won by 1.5 points over incumbent Sen. Kay Hagan (D)), and Virginia (Sen. Mark Warner (D) won by 0.8 points over RNC Chair Ed Gillespie (R))
What the Polls Were Saying: 2014 was a bloodbath for Democrats, losing nine seats, including five incumbents (in addition to Hagan & Udall, there was also Mark Begich (AK), Mark Pryor (AR), and Mary Landrieu (LA)). That said, arguably the biggest shock on Election Night, and perhaps the best reason to always vote, was Mark Warner. Warner showed a very healthy lead in the polls in 2014, up by +9.7-points in the RCP average, and yet won by less than a point thanks to a depressed Democratic electorate and Virginia still being a slightly swingier state (even since then, the state has shifted further left and may move further still in 2018 if the Democrats can nab a majority of the congressional seats). The other two seats were expected to be close, with Gardner slightly underperforming his RCP average of +2.5 and Tillis well overperforming his (they had Hagan winning by 0.7 points). I'll say this several times, but it's impossible not to see that had Democrats been remotely excited about voting in 2014 in the way they normally are in a presidential election, Hagan & Udall being in the Senate right now would stop Brett Kavanaugh. It's worth noting that Gillespie tried to propel himself to further victory in 2017 with a nasty governor's race he ended up being clobbered in, and Udall has been rumored to want a rematch in 2020.
|(Almost) Sen. Shelley Berkley (D-NV)|
Ultimate Net for Cycle: Democrats +2
Won By Less Than Two Points: Nevada (Sen. Dean Heller (R) won by 1.2 points over Rep. Shelley Berkley (D)) and North Dakota (Attorney General Heidi Heitkamp (D) won by 0.9 points over Rep. Rick Berg (R))
What the Polls Were Saying: 2012, unlike 2014, was a year where Democrats saw a bit of an over-performance of the polls, something that would surely help them if that were the case in 2018. Both Heitkamp & Berkley vastly out-performed the polls, with RCP's final averages in these races being Berg +5.7 and Heller +4.0 (it's worth noting that Jon Tester was also projected to lose using the RCP average, though he ultimately won by +3.7 points, so this is a sure case where you can't just count on the polls). Heitkamp was a notable upset, as virtually no one saw that coming, including boy wonder that year Nate Silver; she may well have to overcome similar odds in 2018 if polls are to be believed, though Cramer isn't doing nearly as well as Berg was at this point. Berkley has publicly complained in the years since that the Democratic Party abandoned a winnable race, and it's hard not to see the validity there. Berkley greatly out-performed the polls, and lost by less than 2-points in a state that Barack Obama won by seven. Considering Obama ultimately could have had more use for Berkley in the long run than, say, padding his electoral college count slightly (for all of the talk about 2012 being a particularly close election, it really wasn't compared to 2000, 2004, or 2016), were he to have leant more support to dragging her on his coattails, it surely would have helped his second term in office.
|(Almost) Sen. Alexi Giannoulias (D-IL)|
Ultimate Net for Cycle: Republicans +6
Won By Less Than Two Points: Colorado (Sen. Michael Bennet (D) won by 1.7 points over District Attorney Ken Buck (R)), Illinois (Rep. Mark Kirk won by 1.6 points over State Treasurer Alexi Giannoulias (D)), and Pennsylvania (Rep. Pat Toomey (R) won by 2.0 points over Rep. Joe Sestak (D))
What the Polls Were Saying: I oftentimes think of 2010 as the year the Democrats just skated by a true disaster, and in some ways that's true. Bennet, for example, winning by such a slim margin was surely a result of the Republicans fouling up their primary, same as in Delaware and Nevada (the latter of which, famously, saw one of the biggest over-performances of a candidate ever in the RCP as Harry Reid beat Sharron Angle by 7 points over the RCP average, which had predicted a narrow victory for Angle). But it's also worth noting that the cycle that once looked like the Republicans would take back the Senate could have been even more disappointing for the Republicans, as Giannoulias & Sestak both ran much better than the polls (Kirk was at +3.3 and Toomey at +4.5), and wins there also would have given the Democrats two more votes for Merrick Garland (this is the earliest class that would have been guaranteed to vote on that critical judicial nomination). It's worth noting that 2010 is the first of these years that we've already seen the sequel, and that these races all swung to the Democrats by a stronger margin in 2016, though in Pennsylvania's case not by enough (Toomey being the only person on this list to have won both his initial election and reelection's by less than two points). Kirk got crushed in his 2016 bid and Bennet was fine when he ran for a second full-term.
|(Barely) Sen. Mark Begich (D-AK)|
Ultimate Net for Cycle: Democrats +8
Won By Less Than Two Points: Alaska (Mayor Mark Begich (D) won by 1.5 points over incumbent Sen. Ted Stevens (R)) and Minnesota (Comedian Al Franken (D) won by just 312 votes, or roughly .01-points over incumbent Sen. Norm Coleman (R))
What the Polls Were Saying: With a tanking economy, an unpopular war, and a mishandled hurricane relief effort, George W. Bush was certainly going to pad the Democrats' majority, though even he had no way of knowing that 60 votes would soon be in store for the Democrats with Arlen Specter switching parties, giving Barack Obama seven months with a filibuster-proof majority, in the process passing one of the most comprehensive healthcare bills in American history (they needed all 60 of those votes to do it). The RCP average starts getting a little bit wonky here (Alaska, for example, they only list one poll), so I'm going to have to give you a bit of a gut reaction based on my own polling research for that race. As it looks, most polls showed Begich doing considerably better in Alaska, something that was a telling sign of his future in the state as he would lose his reelection despite solid personal numbers and may well lose the governor's race this year despite ample opportunity. Minnesota does have an average, which had Coleman at +6.3. In hindsight, we should have seen a narrower victory coming considering Bush's unpopularity and the fact that Obama was going to win, but Minnesota (at the time) genuinely didn't like either of these guys that much which is why former Sen. Dean Barkley would win a crucial third party bloc of votes. Coleman has not run for major office since (despite a lot of people assuming he'd seek the governor's mansion or the RNC Chairmanship a few years back), and Franken would go from career highs (a solid reelection, rumors of a 2020 presidential bid) to a major career defeat when he was accused of groping multiple women and was forced to resign. But it has to be said-313 more people vote for Coleman instead of Franken (roughly the population of my old 3-story apartment building), and 11 million people don't get healthcare.
|(Barely) Sen. Jim Webb (D-VA)|
Ultimate Net for Cycle: Democrats +5 (it's +6 if you don't count the Joe Lieberman going from Democrat to Independent as a loss for the D's since he still conferenced with them)
Won By Less Than Two Points: Montana (State Sen. Jon Tester (D) won by 0.9 points over incumbent Sen. Conrad Burns (R)) and Virginia (Sec. of Navy Jim Webb (D) won by 0.4 points over incumbent Sen. George Allen (R))
What the Polls Were Saying: I have seen bigger Senate victories for the Democrats, and bigger victories in general, but there will never be an election night I enjoyed more than watching the 2006 elections (if the D's win both houses in 2018, it will probably top this night, however). The entire cycle had been a conversation about how the Democrats needed to run a perfect race to win the Senate, and even then it wasn't likely to happen. But the Democrats quickly cleaned house and managed to pick up these two races, odd twins of themselves if you look in hindsight. RCP, which called every single one of the Senate races right in 2006 (basically unheard of since-at least one state gets off-kilter), had Tester at +3 and Webb at +1.5, both of them ultimately under-performing as the Democrats didn't close as well as they could have in the last few days of the election (likely not enough to cost them any Senate seats, but may well have plucked a House seat or two off of the map). Still, if you look at the races they had very different trajectories. In Montana, former "accidental" Sen. Conrad Burns (he was a shock win in 1988, defeating longtime Sen. John Melcher that cycle as part of the Bush landslide), had been trailing the entire race until the last minute when his state's natural red tendencies caught up with the polls, but the Republicans who had largely abandoned him didn't have a strong enough operation to take advantage of such a race (have the election two weeks later, and Burns probably wins). George Allen, on the other hand, had an easy shot at reelection and potentially even the White House in 2008 until he was caught on camera using a racial slur, making the contest a tied race and giving the Democrats a shock victory in what in hindsight was a harbinger to a Blue Virginia.
|(Almost) Sen. Betty Castor (D-FL)|
Ultimate Net for Cycle: Republicans +4
Won By Less Than Two Points: Florida (HUD Sec. Mel Martinez (R) won by 1.1 points over former Education Commissioner Betty Castor (D)), Kentucky (Sen. Jim Bunning (R) won by 1.4 points over State Sen. Dan Mongiardo), and South Dakota (Rep. John Thune (R) won by 1.2 points over Sen. Tom Daschle (D))
What the Polls Were Saying: The polls only accurately called one of these races in what appears to be our final RCP year with full data. John Thune's average headed into the night was 1.3 points, but I still think there were Democrats who were stunned that the Republicans would attack their leader and actually successfully take him out. Coming two years after the demoralizing loss of the Senate in 2002, the Democrats endured for four years their longest period of GOP control of the White House/Senate/House since 1929-33 (if the Republicans win both the House and Senate this cycle, this will be duplicated and make Donald Trump the first GOP president since Herbert Hoover to have a Republican majority in both houses of Congress for a full-term). Castor was up +0.5 points in Florida that year, and her loss can be attributed to Bush doing quite well in Florida four years after the ultimate "every vote matters" election in 2000, while Bunning's under-performance was a huge surprise as he was up by +7.5 points. The race featuring the former Detroit Tigers Hall of Famer became so close because Bunning repeatedly used racial and ethnic slurs against Mongiardo, comparing him to one of Saddam Hussein's sons. Had Bush not won the state by twenty points, Bunning surely would have sunk in a surprise loss that year. Two odd things worth noting in 2004, however. One, with the exception of Daschle, none of the Republican pickups that year were in states where they defeated incumbents, as five retiring Southern senators (John Edwards, Fritz Hollings, Zell Miller, Bob Graham, and John Breaux) left huge openings that the Democrats couldn't sustain, thus making this one of the only years in modern history to have only one defeated incumbent in a general election. And secondly, despite a plethora of incoming Republicans who were elected this year, it would be a Democrat who would become the indisputable star of this freshman class: a young state senator from Illinois named Barack Obama.
|(Barely) Sen. Tim Johnson (D-SD)|
Ultimate Net for Cycle: Republicans +2
Won By Less Than Two Points: Missouri (Rep. Jim Talent (R) won by 1.1 points over incumbent Sen. Jean Carnahan (D)) and South Dakota (Sen. Tim Johnson (D) won by just 524 votes, or roughly 0.1 points over Rep. John Thune (R))
What the Polls Were Saying: Okay, so I lied above (RCP does have an average for South Dakota with Thune winning by 1.3 points and had Talent up by 5.3 points in Missouri, so obviously a bit of an under-polling of Democrats that year). The 2002 Senate races remain some of the odder ones on this list, and a weird companion to 2004. While in 2004 it was a bunch of Southern Democratic seats that swung to the right, giving them a number of pickups, in 2002 it was supposed to be a rough ride for the Republicans with open Senate seats throughout the South (and New Hampshire) that were vulnerable, and would have gone their way in a more traditional midterm. But in 2002 the country was still reeling from 9/11 and the Iraq War had just started, so President Bush wasn't as toxic as he would become later in his presidency, and the Republicans were able to stave off top-flight Senate challengers like Gov. Jeanne Shaheen, White House Chief of Staff Erskine Bowles, and Judge Alex Sanders in open Republican seats that might have been friendlier to them in other circumstances. The two closest races were in Missouri, where Jean Carnahan's rough first year in office (chronicled a bit more here) nearly pulled off an upset against Rep. Jim Talent when the ballots were counted (a telling sign, perhaps, that Talent wasn't living up to his name & would in fact lose in a tougher race four years later), and in South Dakota, where John Thune nearly upset Johnson, but record strength in Sioux Falls managed to keep Johnson in office for another term (though Thune would of course gain his revenge two years later and is still in office). It's worth noting that 2002 was the last year where both Republican (Tim Hutchinson) and Democratic (Carnahan, Max Cleland) incumbents lost general elections in the same cycle, though 2018 could end that streak.
|(Barely) Sen. Debbie Stabenow (D-MI)|
Ultimate Net for Cycle: Democrats +4
Won By Less Than Two Points: Michigan (Rep. Debbie Stabenow (D) won by 1.6 points over incumbent Sen. Spencer Abraham (R) and Washington (Rep. Maria Cantwell (D) won by 2,229 votes, or roughly 0.1 points over incumbent Sen. Slade Gorton (R))
What the Polls Were Saying: 2000 was the first year that I remember distinctly following the elections at the time of the release (I also remember about this time that my English class got LexisNexis for a week for a project, and instead of researching for said project, I spent the entire time reading Roll Call articles you needed a subscription to read about old Senate contests). As a result, my relatively photographic memory remembers both Stabenow & Cantwell being in incredibly close races, though public polling is very difficult to come across from this time period. It's worth noting in taking out incumbent senators they've had very fruitful careers, and are still in the Senate now (and will likely win reelection for a fourth term in 45 days). The main scuttlebutt at the time was that the Democrats managed to win so many close Senate races despite the Republicans doing well in getting their incumbents to run. Republicans also lost Sens. John Ashcroft, Rod Grams, and Bill Roth, and as a result the Senate was tied, and in fact the Democrats had the majority briefly from January 3rd-January 20th of 2001 before Dick Cheney became the tiebreaker. The closest a Democrat came to taking that 51st seat was Brian Schweitzer in Montana (he'd go on to parlay his near-successful campaign against Conrad Burns into a two-term stint in the governor's mansion). It's worth noting that Democrats criticized Sen. Joe Lieberman, the running mate of Al Gore, for not giving up his Senate seat due to his replacement being guaranteed to be a Republican (Lieberman was going to be replaced by then-Gov. John Rowland, a member of the GOP). Lieberman, perhaps selfishly assuming A) that he needed a backup plan in case he wouldn't be VP or B) that the Senate couldn't possibly tie as the Democrats didn't seem likely to run the boards with close races, didn't care about this, and though he lost, I recall reading in US News & World Report at the time that Lieberman may have felt pressure not to give up the Senate seat, but the vice presidency, to ensure the Democrats controlled the upper chamber under a President Gore, being replaced by former President Jimmy Carter. It was definitely in US News & World Report, but it was just speculation at the time so who knows if anything would have come from it had 538 more people voted for Al Gore.