Perhaps the most surprising result in presidential polls since the conventions is that more than one in 10 voters say they would support candidates not named Donald Trump or Hillary Clinton: Libertarian Gary Johnson and Jill Stein of the Green Party. In the New York Times/CBS poll released September 15, those candidates' combined 12 percent share was boosted by support for Johnson from more than a quarter of voters ages 18 to 29.
In real elections, neither of those two parties has ever come close to the level of support these polls indicate: Libertarians peaked at about 1 percent in 1980 (with David Koch on the ticket) and again in Johnson's previous run in 2012. Running as a Green, Ralph Nader — one of the country's "most admired" people for decades — captured 2.74 percent in 2000, just enough to swing that election to George W. Bush.
Reminded of this history, many voters who now say they favor Johnson or Stein will ultimately conclude that voting for either one risks the election of the candidate they like least, whether that's Trump or Clinton. (Others might change their minds when they recognize that Johnson and Stein are almost as ill-prepared to be president as Donald Trump.)
As Clay Shirky recently wrote at Medium, "There's no such thing as a protest vote." Elections are not a way to "send a message," Shirky pointed out, because the political system isn't set up to hear those individual self-expressions. Elections offer one binary digit, not a whole alphabet to describe your feelings. There are other opportunities, outside of elections, to express or advocate for your more specific views, as people like Shirky and I will maturely explain.
But why should voters who actually think Stein or Johnson would be the best president have to forgo the opportunity to vote for them? Is it a given, in the American political system, that we can't vote for a candidate or party that better reflects our views, or the candidate we like most, without risking the worst outcome?
The answer is that while all electoral systems with single winners naturally tend toward two parties (a rule of thumb known as Duverger's law, although it is far from a law) there are particular features of the US system, which are not written in stone or in the Constitution, that reinforce this dynamic. And those features could be changed.
Having more than just two viable parties would do more than provide an outlet for frustrated libertarians, socialists, or anti-Trump conservatives. Even if you don't agree with any of the parties that are likely to emerge in a multi-party system, having more of them could improve our politics in subtle ways that go well beyond our choices in a presidential election.
The basic failure of American politics is its rigidity — we're locked into two distinct camps, divided not just by ideology but by region, race, cultural identity, and to some extent, gender. For the most part, each conflict of policy and identity breaks along exactly the same lines. Each camp is armed with more than enough money and unshakable strongholds in certain states and branches of government. There's no room for compromise, as we've seen most sharply in Congress, but, even worse, there's little room for creativity or new solutions that build new coalitions.
The problem is not partisanship. Parties aren't just random sides in an artificial fight. There's no democratic politics without parties. Parties define clear choices and help organize people's participation in democracy. Because parties intend to stick around for the long term, we expect them to be more interested in winning over average voters, and cementing their loyalty over time, than in scoring one big win, taking the presidency, Congress, and the Supreme Court and playing each one for maximum advantage.
That's the theory, anyway. But it may be that when there are just two parties, aligned precisely along lines of race, class, level of urbanization, and education, they don't have the same incentives. When there aren't many voters floating in between, available to be won over, the best strategy might be to just go for the biggest win possible and lock in the gains, in the way that Republicans in Wisconsin and North Carolina have tried to do. That's also a strategy that's more likely to be adopted by a party whose demographic share of the electorate is on the decline.
In 1963, the historian and political scientist James MacGregor Burns identified "four parties" in American democracy — liberal and conservative factions in both the Democratic and Republican parties — whose members and voters combined and formed alliances along many different lines, sometimes furthering progress and sometimes impeding it.
These quasi-parties were even semi-officially recognized at the time: Congressional Quarterly's records of House and Senate votes were labeled not just by party but by whether members were part of the "Conservative Coalition" of Southern Democrats and conservative Republicans, a coalition now fully coterminous with the GOP. Even in the 1990s, you could see the outlines of the four parties and surprising alliances among them. But that was long ago.
Today there are alternative coalitions that one could imagine, in theory, but that never come together. There could be a left-right cross-party coalition forming around, say, economic development in areas hit hard by the decline of the coal industry or one advocating stronger supports for Social Security.
But the structures of the electoral system, and a Republican Party that seems to be in search of a final showdown, make those alliances impossible to achieve in real life. Additional parties — even parties that you or I might not favor — could open up Congress and create new opportunities for bargaining and coalition building on unusual lines.
Critics of the two-party "duopoly" tend to make three mistakes from the start. One is to imagine that a party would emerge first at the presidential, or national, level. One reason the Libertarians and Greens have the modest presence they have is that over decades, they at least built a fragment of the structure of parties at the state and very local level.
Neither party has won statewide elections (Johnson and his running mate, William Weld, both won governorships as Republicans), but each boasts more than 100 elected officials in municipal and regional office. When new parties have emerged in the US, such as the Populists and Progressives of the late 19th century, or even the Republican Party in the 1850s, it is from the bottom up, not the top down. New parties could have more influence in Congress or legislatures than simply as losing names on the presidential ballot.
A second mistake is to reject parties and partisanship altogether, imagining that a single candidate, such as Michael Bloomberg, might emerge who will rescue the country from fixed ideological combat and "get things done."
Usually this imaginary candidate will be a socially liberal, economically conservative centrist, with experience outside of politics. This was the dream of Americans Elect, the lavishly funded 2012 effort that Thomas Friedman promised would "do what Amazon.com did to books ... remove the barriers to real competition, flatten the incumbents and let the people in." Americans Elect spent $35 million in 2012 but couldn't even identify a candidate.
The third and most consistent mistake is to ignore the structural conditions that make it so difficult for a party to take hold. In most states, for example, a party can't endorse candidates from another party — it has to run all its own candidates or none. In the eight states that allow fusion voting, parties can pick some of their own candidates and some from another party.
The Libertarian Party, for example, could embrace some Republicans and not others, and libertarians could send a message by voting on their party's line. In New York, voters who want to nudge the Democratic Party to the left can vote for Clinton on the line of the Working Families Party, a group that endorsed Bernie Sanders in the primaries. Such parties can often serve as a safety valve if one party or its elected officials grow too complacent, corrupt, or out of touch ideologically.
Innovations such as instant runoff voting or ranked-choice voting would give new parties an even stronger role. Voters could cast a first-choice vote, but also indicate which party or candidate they prefer if their first choice doesn't get enough votes.
Candidates would then have incentives to seek the second- or third-choice votes from supporters of other candidates. Instead of writing off Jill Stein supporters, for example, Clinton would want to ensure that she got their second-choice support. Several municipalities, including Minneapolis, have adopted ranked-choice voting for mayoral and council elections, with the additional benefit of lowering the barrier to entry for candidates who start without much money or establishment support. The process can be complicated — many voters omit their second choices — but these local experiments will show whether voters can get used to the system.
Additional parties created through fusion or ranked-choice voting would not only give citizens a chance to express their views in the voting booth but would also open up the system and create new alliances and bargains, breaking down the sense of two well-fortified, unmoving camps.
Gary Johnson and Jill Stein might triple their parties' previous levels of support but will nonetheless be quickly relegated to political trivia quizzes after the election is over. It is possible to build a true multi-party democracy in the US, and it could have some real benefits — but it will require profound patience and strategy, not a single-election breakthrough.