here are many lessons to be found in studying the American presidential election record.
We’re currently moving through a period of profound frustration with the Republican Party. Many were looking and hoping for a Reagan-esque candidate to arise during the 2012 primary debates. Alas, there wasn’t one anywhere to be found.
Obama and Romney took 98.17% of the popular vote. Obama won by 3.85%, and with only 1.83% of the popular vote going to candidates outside the top two parties, the non-two party vote was effectively irrelevant.
But it hasn’t always been this way, and — if history is any guidance — the tide will turn again, and soon.
The following figures show the percentage of the popular vote outside the top two candidates and the percentage of the popular vote for the third-place candidate since 1824.
What we see is a cyclical rise-and-fall of two-party dominance. Prior to the current 2000-present period, the average number of elections in a row without more than 5% of the vote going to candidates outside the top two parties was 2.9. We are already at four consecutive elections (2000 [3.7%], 2004 [1.0%], 2008 [1.5%], and 2012 [1.8%]) where the top twocandidates garnered more than 95% of the popular vote.
Similarly, the average number of elections in a row without more than 5% of the vote going to a single third-party candidate is 3.4. We are currently at four, with Ross Perot (8.4%) in 1996 being the last third-party candidate at this level.
Thus, from a historical perspective, the rise of both a significant non-top-two-party vote and a prominent third-party candidate is overdue.
This becomes even more clear if we look at the five-election running mean for the popular vote going to candidates outside the top two parties and the popular vote going to the third-place candidate.
Keep in mind that the latest data point on this graph is 2004, which is the average of 1996 through 2012. Consequently (and assuming that what is past is prologue), come 2016 — or 2020 at the latest — we should begin to see the share of the vote outside the top two parties rise again, and significantly so. Also note that the amplitude of the non-top-two-party and third-party peaks are declining over time, suggesting a progressive consolidation of the two-party dominance.
The question remains: does the vote outside the top-two candidates matter? Occasionally, the answer is yes.
In 1992, Ross Perot ran as an independent and took 18.9% of the popular vote. Bush 41 lost to Clinton in the popular vote by only 5.6%. If we give Perot’s vote to Bush 41, then Bush 41 would have taken California (51 electoral college votes), Colorado (8), Connecticut (8), Delaware (3), Georgia (13), Hawaii (4), Illinois (22), Iowa (7), Kentucky (8), Louisiana (9), Maine (4), Massachusetts (12), Michigan (18), Minnesota (10), Missouri (11), Montana (3), Nevada (4), New Hampshire (4), New Jersey (15), New Mexico (5), Ohio (21), Oregon (7), Pennsylvania (23), Rhode Island (4), Tennessee (11), Vermont (3), Washington (11), West Virginia (5), and Wisconsin (11) — and very narrowly lost Maryland (10) by <0.01% and New York (33) by <0.1%. Conceivably, had Perot’s vote all gone for Bush 41, Bush would have won the popular vote 56.4% to 43.0%, taken 47 of 50 states, and won 483 of 538 electoral college votes.
We’ll never know how Perot’s voters would have gone had he not been in the race. Exit polling data is unreliable on this front, since the campaign would have been different with Perot not in it — making it effectively impossible to run a reliable counterfactual. That said, given Perot’s policies, it’s reasonable to assume the large majority of his vote would have gone to Bush 41. Perot almost certainly took a landslide victory away from Bush 41.
Some argue that Perot’s exit polling data after the 1992 election showed that an equal share of Perot’s voters (38% vs. 38%) would have voted for Bush and for Clinton had Perot not been in the race (the remaining 24% claim they wouldn’t have voted at all if they couldn’t have voted for Perot). This seems like junk data. Perot’s running mate was United States Navy vice admiral James Stockdale, and Perot’s political platform was based on reducing national debt and government bureaucracy and balancing the budget — hardly liberal positions all around, and instead being more consistent with the current conservative Tea Party non-establishment platform. It seems bizarre that Perot’s voters would comprise a perfect liberal-conservative split.
We need to reflect on the implications of this election. If Bush 41 had adopted fiscally responsible conservative policies that would have negated the need for Perot’s entry, Bush 41 would have undoubtedly crushed Clinton, thereby ending Clinton’s political career at the federal level, likely precluding his wife’s future political successes, and possibly preventing an Obama 2012 victory (where voters were swayed by Clinton’s campaign efforts on his behalf). Post-1992 American political history would have been very different without Perot’s 1992 election run.
Along comes 2000. Bush 43 wins Florida’s 25 electoral votes by only 0.01%. The third-place candidate — Ralph Nader of the Green Party — takes 1.63% of Florida’s popular vote. It would be absurd to suggest that Ralph Nader and the Green Party had a dominantly conservative following. Ergo, absent Nader in the race, Gore easily takes Florida and the 2000 election.
Third party candidates rarely rise up to significant levels because of voter dissatisfaction with both of the two major parties at any point in time. Rather, they typically represent aggrieved voters from one of the major parties. Analysts must keep this in mind when running counterfactuals, regardless of what dubious exit polling data tells them.
The following figures show the difference between the popular-vote margin from first to second place and the share of the popular vote going outside the top two candidates (left plot), or the share of the popular vote going to the third-place candidate (right plot). In short, if the value is negative, it means that the popular vote share going either outside the top two candidates or to the third-place candidate was greater than the margin of victory for the winner over the second-place candidate. If the value is positive, the vote share for the third-place and lower finishers could not overcome the first- to second-place margin.
Again, we see that — other than the period between 1920 and 1956 — there are generally no more than three or four consecutive elections without a victory margin that could have been overcome either by the third party candidate on his/her own or via the combined vote of all non-top-two candidates.
Currently at three consecutive elections with no possibly determinative vote share outside the top two parties, we appear due for the rise of a potential electorally decisive third-party candidate (or, at least, non-top-two party vote share) in 2016 or 2020. With the agitation in Tea Party circles and widespread displeasure at the GOP establishment’s performance over the past few years, the pressure is clearly building in the conservative ranks. For the Republican Party establishment to ignore this angst and assume that no serious third-party right-of-center candidate threat will develop in the next decade is to ignore history, and to do so at its own peril.