If you look in the Daily Kos Elections oh-so-handy polling database, you may notice about a third of the races polled have a third-party candidate polling at 5 percent or above.
As David Nir noted the other day, this seems like a lot. Is it?
Indeed, it is. Below is the percent of Senate elections with three or more candidates getting more than 4 percent of the vote (roughly equivalent to polling >5%), going back to 1914, the first year with direct election of senators by popular vote.
If the polling holds up, we're on track to have the most three-way Senate races since
1920. Incidentally, that means we're on track to have some pretty big polling misses as well, since polling with third-party candidates performing this well is
much less accurate.
More below the fold.
The Governors
Below is the graph for the Senate with the races for governor added in, back to 1966, which is approximately when the current pattern (most races in midterm years) began.
The prevalence of third parties in both types of races follows the same general pattern. Races for governor, however, have a clear presidential/midterm pattern. In midterm years, the yellow squares tend to be above the green circles. This could be due to the fact that states that are more likely to support third-party candidates—say, Maine or Minnesota—always have elections during the midterms.
The numbers
Here's a table to compare the Senate and governor races:
In the aggregate, races for governor are much more likely to feature a third party with more than 4 percent than races for Senate. Even if you restrict the Senate races to the same time period, you get the same result.
Third parties are much less likely to be found in presidential year races for governor.
When looking at races for Senate and governor together, this means midterm years are likely to have a greater number and proportion of races with third-party candidates >4 percent than presidential years.
What is going on?
Looking at the graph, there are some potential explanations for the blips, bumps and peaks on it. The peak from 1992-1996 could arguably be related to the Ross Perot phenomenon. The bump in the 1960s and 1970s seems like the reasonable outcome of the turbulent politics of those days. And the initial peak in 1914 is clearly the result of the popularity of both the Progressive and Socialist parties at that time.
But what about now? Are we witnessing the birthing pangs of a new third-party movement? Or will this, too, pass? Is it just the irascible expression of general discontent, eventually to be folded into the main parties as Democrats embrace an economic populist style and Republicans become entirely engulfed by their unruly creation?
Perhaps this is related to the dismal view of the Republican Party, as many of the candidates this year are Libertarian. The latest polling numbers give Republicans a net favorability around -20, except for Fox. Democrats are also in negative territory, although well-loved compared to Republicans.
There's some evidence to back up this idea. If you look at this table of the incumbents with the worst winter approval ratings in the last few election cycles, you see that of the 17 who survived their primary, 11 (65%!) were in three-way contests (three candidates >4%).
A few words about the fascinating election of 1914
Popular elections for Senate started with a bang in 1914. It was kind of a free-for-all compared to the next 96 years.
Four candidates won with less than 40 percent. Although all the winners were Democrats or Republicans, the Socialist candidate took 21 percent in Oklahoma and 25 percent in Nevada, and the Prohibition candidate took 15 percent in Arizona. Progressive Party candidates had more than 20 percent of the vote in Georgia, California, Washington, Pennsylvania and Kansas. Progressive candidate Francis J. Heney took 29 percent in California, where Progressive vice-presidential candidate Hiram Johnson won the governor's seat.
Overall, Progressive candidates took 11.7 percent of the votes for Senate that year; Socialists, 4.2 percent; and Prohibition candidates, 1.5 percent.
For context, recall that in the 1912 presidential election, the incumbent, Taft, placed third with 23.2 percent and 8 electoral votes. Eugene V. Debs of the Socialist Party had 6.0 percent, his best showing ever. Theodore Roosevelt, having lost the Republican nomination, founded the Progressive Party and ended up with 27.4 percent of the vote.
Data sources and other comments
One pertinent question to ask is, do third-party candidates tend to fade as the campaign begins in the fall? A cursory look through the 2010 polling doesn't show that happening. I did not look into the matter very closely, though.
The 4-percent threshold is based on the 5-percent threshold for polling. The 5-percent threshold for polling is roughly the amount a candidate needs to be getting before the candidate will be routinely included in polls.
Data sources: David Leip's atlas, for all info back to 1990; the Clerk of the House, Our Campaigns and wikipedia pages. Off-year special elections for Senate were not included.