Why don’t third parties win US presidential elections?

third-party-win

As always, some of the best questions come from readers of the blog.  “Why don’t third parties win US presidential elections?” came to us via email.

It is also timely with the recent discussion from Kristol and polling information from Data Targeting.

Polling Third Parties

In a traditional poll, pollsters may ask a question like, “Would you consider supporting a third party candidate? Yes or No?”  Traditionally, because the question is asked like this, support for third parties is overstated, and the question has proven to be a poor predictor of voting behavior.

Another way to ask the question is in a horse race question like, “If the election were held today, who would vote for?”  Because the answer has no ramifications, again support for third parties is overstated in polling.

So, if people tell pollster they support the idea of an independent candidate then why don’t they actually vote for them?  Why do most people vote for their party’s nominee?

History of Third Parties in the US

YEAR PARTY CANDIDATE VOTE% ELECTORAL VOTE OUTCOME in Next Election
1832 Anti-Masonic William Wirt 7.8% 7 Endorsed Whig Candidate
1848 Free Soil Martin Van Buren 10.1 0 5% of the vote, absorbed by Republican Party
1856 Whig-American Millard Fillmore 21.5 8 Dissolved
1860 Southern Democrat John C. Breckinridge 18.1 72 Dissolved
1860 Constitutional Union John Bell 12.6 39 Dissolved
1892 Populist James B. Weaver 8.5 22 Absorbed by Democratic Party
1912 Progressive Teddy Roosevelt 27.5 88 Returned to Republican Party
1912 Socialist Eugene V. Debbs 6.0 0 Won 3% of the vote
1924 Progressive Robert M. LaFollette 16.6 13 Returned to Republican Party
1948 States’ Rights Strom Thurmond 2.4 39 Dissolved
1948 Progressive Henry Wallace 2.4 0 Won 1.4% of the vote
1968 American Independent George Wallace 13.5 46 Won 1.4% of the vote
1980 Independent John Anderson 6.6 0 Dissolved
1992 Reform H. Ross Perot 18.9 0 Won 8.4% of the vote
1996 Reform H. Ross Perot 8.4 0 Did not run
2000 Reform Ralph Nader 2.7 0 Ran Next election
2004 Green Ralph Nader 1.0 0

Source: http://www.thisnation.com/question/042.html

Research

The authoritative scholar on the issue is Steven Rosenstone at the University of Minnesota.

While there are structural barriers in place for third parties: ballot access, campaign finance laws, polarization of the party elites preventing cross-over third party validators, sparse media coverage, debate exclusion, and the lack of belief that a third party can win, there is a more fundamental way to look at this:  the psychology of the voter.

Our political system is dominated by the two party system.  Voters operate in primarily in this two party system and many voters take short cuts or cues provided by the parties to simplify their decision making.  To step outside the two party system takes considerable effort.

Rosenstone developed a “Four Part Test for Third Party Support” and I believe it provides a framework to answer the question.

On average, for a voter to consider voting for the third party candidate, they have to reject 4 things.  Not 2 of 4, but rather all 4.  They have to reject both parties’ candidates and both parties, then and only then will they consider a third party.   Then the voter has to find a third party candidate they consider to be legitimate and finally, vote for them.  A tall task.

rosen

 

Additional Polling

We can see Rosenstone’s framework at work in a CNN poll in May 2016.  When respondents were asked if their support for a candidate was an vote of support of opposition, we observe the following:  (editor note: quite a sad commentary on our current choices.)

polling

Conclusion

In part, this explains why Donald Trumps’ numbers have improved in the past 4 weeks.  His support is growing from Republicans who rejected his primary campaign, but are “coming home” to support the party and/or reject Hillary Clinton/democratic nominee.

In part, this also explains, in part HRC’s soft numbers.  The Democrats are still split between HRC and Sanders and have yet to coalesce around a candidate.  We should expect the Democratic nominee’s numbers to improve after the voters move through Rosenstone’s framework.

As you can see, third parties have a difficult road ahead of them and the most likely path is not to win, but rather to prevent someone else from winning or the “spoiler” role.

Gainesville Votes! and its ramifications

Gainesville Votes! is an organization that is proposing “following the model successfully used by the Alachua County School Board: Non-partisan elections occurring in August (alongside county, state and federal primaries) of every even year. When a run-off is necessary, it would be held on the first Tuesday in November along with county, state and federal elections.”  Harvey Ward. This is being proposed in the vein of promoting higher turnout in Gainesville City Elections. This goal and […]

Read more »

How to read a Political Poll!

The sheer number of polls this political cycle is amazing.  The sheer number of bad polls this political cycle is stunning. Regardless, the manner the press reports on polling is just God-awful. Biggest Polling Complaint My largest complaint is that the amount of uncertainty in a poll is not reported, is under-reported or misunderstood. A poll is only a sample (hopefully of a random one that is well constructed) of a population.  When any pollster moves […]

Read more »

Quick Reaction: The Hidden Lesson from the GOP Debate for Candidates

The morning after the GOP Debate, the Washington Post has an article “Marco Rubio had a rough night on Google“. It is an interesting story, but I think it misses the MOST important story from any candidates’ point of view. If you look at the Google search peaks, BOTH peaks were when candidates related a personal story to the current political context. THAT’s the ‘story’, if you are candidate.  THAT’s the takeaway if you are […]

Read more »

How productive is Congress?

Over the Thanksgiving break, I figured NOTHING could be any more frustrating than watching my Gators, so I decided to begin to tackle learning R. Part of teaching myself R is playing around with an interesting data set – Congressional Bills.  “This public resource provides information about more than 400,000 bills introduced in the U.S. Congress, currently 1947-2008, along with extensive information about each bill’s progress and sponsor.” I just started with a simple question, […]

Read more »