For much of the past half-year, Johnson’s support was about equally comprised of Democrats, Republicans, and independents. Given his consistently impressive showing among Millennials, a key part of the Obama coalition, it’s no wonder that Democrats were freaking out about and throwing millions at their perceived Johnson problem. “Vote for Johnson, elect Trump” warned lefty sourpuss Harold Meyerson. Well, that turned out to be almost exactly wrong.
If you look at a dozen national polls with detailed Johnson numbers over the past three weeks, a striking new pattern, which I first flagged here, has emerged: No longer is the Johnson coalition anything like a 33-33-33 split among Dems, Reps, and indies. Now, amidst the Libertarian losing around half his support over the past two months, his remaining base is consistently 10 percent Democrat, 25 percent Republican, and 65 percent independent. This ratio has been stable not just throughout national polls, but in the half-dozen or so detailed battleground state polls I’ve checked out.
What does that mean in, say, Nevada? Well, Johnson is at 5.3 percent there, according to FiveThirtyEight‘s poll average, while Trump and Clinton are tied at 46.7. If you add 10 percent of Johnson’s totals to Clinton, and 25 percent to Trump, the Republican would have a lead of 48.0 percent to 47.2. The potential Johnson Effect here is worth about 0.8 percentage points in Clinton’s direction.
Yes, yes, there is no such thing as a spoiler, votes do not inherently belong to anybody, and so on. It is plausible, I suppose, that 100 percent of Johnson’s voters in Nevada would have just stayed home if his name was not on the ballot. But it’s also possible that consumers change behavior when presented with more options (especially palatable ones), altering the shape of the overall market.
As mentioned before, the 10/25/65 D/R/I ratio of Johnson’s support is stable across battleground states where polling is sophisticated enough to check. So where else could there be a Johnson Effect, if you choose to believe in such things? Basically, wherever it’s close, or possibly in some Johnson-friendly states that don’t look close now but might end up that way.