Could a third-party candidate cost Republicans a Senate seat -- or even the Senate, period -- this year?
Theoretically, the answer to both questions is yes. But the impact of third-party candidates is often oversold, and just as notably, these candidates tend to fade at the end of the campaign.
There's already a good bit of chatter about third-party candidates potentially swinging some key Senate races this year -- including in North Carolina, Alaska and South Dakota -- as well as governor's races in Hawaii and Maine.
The logic goes like this: A third-party/independent/libertarian candidates takes 5-6 percent of the vote, or more, mostly at the expense of one candidate (generally thought to be the Republican), which can swing a very close race (often for the Democrat).
If this storyline picks up this fall, though, just remember what history shows. And that is that these candidates almost always under-perform the late polls on Election Day.
Below is pretty much every recent example of a nominal third-party candidate who was included in late polls of a Senate or governor's race.
Here's independent Tom Horner in the 2010 Minnesota governor's race:
This is Tim Cahill in the 2010 Massachusetts governor's race:
And Chris Daggett in the 2009 New Jersey governor's race:
Not only did all three men fade at the end; they all also under-performed the late polling of their races. In Daggett's case, for example, polls showed him as high as 20 percent two weeks out, his polling average on Election Day was just over 10 percent, and he finished at 5.8 percent.
Examples of semi-viable third-party candidates running in top Senate races are a little more scant. And to the extent third-party candidates in these races have won significant votes, they often haven't been included in polling heading into Election Day. So we don't really know whether/how much they faded at the end.
But here's one example, of former appointed senator Dean Barkley (I) running in the 2008 Minnesota Senate race with Al Franken and Norm Coleman:
Barkley had a similar arc as the three men listed above. He polled as high as 19 percent in October and was getting the spoiler treatment -- big time. He finished with 15 percent after tailing off at the end. Given that Franken won by just 312 votes, though, the spoiler talk was valid.
Going back a little further in time to some long-ago governor's races, here's the 2006 Texas governor's race, which featured two semi-viable independents: Kinky Friedman and Carole Keeton Strayhorn (a.k.a. the mother of a certain former White House press secretary):
Both had respectable showings (18 percent for Keeton Strayhorn and 12 percent for Friedman), but you'll also note that voters seemed to come home to the Democratic nominee, Chris Bell, at the end of the campaign. Both independents also finished slightly lower than the late polls showed.
And finally, two more examples. These two, unlike the charts above, don't show the third party candidates fading at the very end.
Here's the 2006 Illinois governor's race, featuring a pre-federal-inmate Rod Blagojevich (D) and Green Party candidate Rich Whitney:
And here's independent Robert Sarvis in the 2013 Virginia governor's race:
But while neither of these polls showed a fade at the end, these candidates certainly faded on Election Day. While the polling average had Sarvis in double digits, he took just 6.6 percent of the vote on Election Day. And while Whitney was in the mid-teens for much of the final two weeks of the campaign, he finished at just more than 10 percent of the vote.
In all of these cases, it seems, people might have wanted to support the independent/third-party candidate but decided in the end that it was a waste of their vote and opted to go with the major parties. Either that or they stayed home because their guy didn't have a shot.
So when some of these 2014 third-party candidates rise in the polls in the months ahead, just remember that their best day on the campaign trail is likely to be a couple weeks before Election Day.
That doesn't mean they won't matter, but it does mean they'll probably matter less than you think in mid-October.