That question might not be immediately apparent or totally accurate, so please allow me to explain. The literal answer will probably be "somewhere in between," if the yardstick used is total percentage of presidential votes cast. But what I'm really asking is whether this year will follow the model of third-party support collapsing in the voting booth or actually holding firm on Election Day.
Both Ralph Nader and H. Ross Perot influenced the elections they ran in -- Nader in 2000, and Perot in both 1992 and 1996. Nader won 2.7 percent of the national popular vote. Perot won a whopping 18.9 percent in 1992 and a respectable 8.4 percent in 1996. Without their candidacies, the races might have elected different presidents (George H. W. Bush's second term, Bob Dole, or Al Gore). Or perhaps Bill Clinton would have won with more than 50 percent of the vote both times. You can endlessly speculate about what might have been, but the fact is that in three successive elections, third-party candidates had a measurable impact (whether good, bad, or indifferent depends on your political outlook, really).
Nader's effect on the 2000 race was the most keenly felt (by Democrats, at least), for a number of reasons. The Supreme Court case, the Florida "recount," the fact that Gore had won a half-million more popular votes, and the thinness of Bush's Electoral College victory (if Gore had managed to win his home state -- or even New Hampshire -- he wouldn't have had to care about Florida, one way or the other) -- these all contributed to the depth of feelings some Democrats had (and even now, still harbor) towards Ralph Nader. But Nader only got less than three percent overall -- nothing like the influence Perot had in either one of his runs.
Pollsters generally scoff at any third parties included in the pre-election polling, and not without reason. The pollsters have found that in a normal election year, many voters may say they're going to vote for third-party candidates, but when the votes are actually counted these candidates usually pull in a lot less than their polling percentage had indicated. Voters like to vote for a candidate with a chance of actually winning, perhaps. Or perhaps third-party voters just aren't all that motivated to get to the polls -- they may answer a pollster's phone survey but then fail to turn out to vote. For whatever reason, the safe bet for pollsters is to discount any third-party responses, since they almost never equate to actual votes.
That's "in a normal election year," mind you. This is, to put it mildly, quite likely not a normal election year. This means conventional wisdom may not be a reliable indicator of what might happen.
There are two third-party candidates this year who are polling far above the usual levels seen by minor political parties: Libertarian Gary Johnson, and the Green Party's Jill Stein. As of right now, Johnson is polling at a national average of 8.9 percent, and Stein is at 3.4 percent. Both are polling much better in some individual states, including states that might be big battlegrounds in November. The big question that not enough pundits are even considering right now is whether this support in the polls will translate into actual votes in November.
For the most part, and rather astonishingly, both Johnson and Stein taken together do not seem to be changing the dynamics of the race between Hillary Clinton and Donald Trump all that much. This is astonishing because normally one party is more affected by third-party candidates than the other. Although it's not really true, a strong third-party candidate is usually said to be "stealing votes" from one or the other of the major party candidates. It's not really true because every vote must be earned -- if a third-party candidate wins it, it means there are some serious deficiencies in the two major party candidates that might mean that voter would just stay home on Election Day, if given the choice of only two names on the ballot.
Perhaps because Stein balances Johnson out ideologically (Stein might be drawing disaffected Bernie Sanders supporters, while Johnson pulls in "Never Trump" Republicans), their inclusion in pollsters' surveys doesn't seem to change the margin between Clinton and Trump all that much. Currently, Hillary is up 5.5 percent in polling that doesn't include third-party candidate names to choose from, while she's up 4.3 in polls which mention all four candidates. This stability has been fairly consistent for the past few months -- the third parties may lower both Clinton and Trump's absolute numbers, but they seem to be drawing fairly equally from both.
Of course, as any regular reader of this column knows, national polling is almost worthless when attempting to predict the race's outcome. Given our Electoral College system, state-level polling is necessary to see the true picture of the race. This is where third-party candidates could actually cause upsets, to put it another way. In Utah, for instance, disaffected Republicans won't just have Johnson on the ballot (if they can't bring themselves to vote for Trump), there will also be a local (and Mormon) name on the ballot to choose, as well. The two combined may shrink Trump's vote to the point where Hillary Clinton could actually win this deep-red state. The opposite may be true elsewhere, if Stein draws enough support from Clinton in a key battleground state (such as Ohio, perhaps).
National numbers are meaningless for another reason, as well. Voters in states that are comfortably in either Clinton's or Trump's column may feel a lot freer to choose a third-party candidate. "Hillary's going to win Vermont," they tell themselves, "so I can vote for Jill Stein and not worry about handing the election to Trump." Voters in states that are going to be close won't have that freedom, because they know full well that their state might be decided by a few hundred votes. Voters do indeed make these sorts of calculations, which the polling often entirely misses.
The outcome of the 2016 election may instill a certain amount of caution among those who watch polls for a living (or even as a serious hobby). Conventional wisdom about the possible turnout for third parties may shift, as the old model ("Let's just ignore them altogether -- their voters never actually make it to the polls") proves wrong. Having Donald Trump on a major party ticket sets up a rather unique situation (a "black swan event," to number-crunchers), but then every election has its own unique dynamics of one sort or another.
If Gary Johnson manages to even match H. Ross Perot's 1996 percentage, it will change the outcome of the race. It might not change the overall winner, and it may not even change any particular state outcomes, but it will be seen in the total percent of the nationwide vote the winner claims. Perot got almost 20 million votes in 1992, meaning Bill Clinton won the race with only 43 percent of the national popular vote. Clinton then went on to face the charge that he "hadn't won a mandate" because of his low vote total.
If 2016 turns out to be a Perot year, the total third-party vote percent may influence the "political capital" the winner can count on in his or her first term. If it turns out to be a Nader year, then this won't be as true, but if you define a "Nader year" as being one where certain states flip between the two major party candidates, then the third-party influence could be even greater than the percentage would have you believe. It'd have to be a pretty close race (as it was in 2000) for that to be true, of course. But one way or another, the conventional wisdom of just ignoring the third parties may turn out to be wildly wrong this November.
-- Chris Weigant
Follow Chris on Twitter: @ChrisWeigant