Ranked-Choice Voting doesn’t fix the spoiler effect
In seemingly every discussion thread about American politics, the same conversation happens again and again:
“Man, both political parties suck.”
“Yeah, we really need a third party.”
“But any third party entering the race acts as a spoiler, so nobody votes for them.”
“That’s why we need Ranked Choice Voting! It eliminates the spoiler effect and makes it safe to vote honestly, ending the two-party system! Check out this CGP Grey video!”
Others then watch the video, get excited, and pass it on, perpetuating the cycle, yet everyone seems to have missed what the video says at 2:58, about First Past the Post vs Alternative Vote (which is another name for RCV, or Instant-Runoff Voting):
“Over time, they both trend towards two parties.”
The video itself says that RCV is not a solution to the two-party system…
But how could this be, if it eliminates the spoiler effect? Wouldn’t that open the door for third parties?
Well, no, because it doesn’t actually eliminate the spoiler effect.
What it really does is protect the two-party system from spoiling by weak third parties, but without giving them a path to victory. If a third party enters the race, but doesn’t have a lot of support, they can’t affect the outcome of the election, for better or for worse. Any votes for the weak third party are essentially symbolic, as they are discarded and transferred to one of the main two parties.
But it allows the third party to exist in the race and get votes, and therefore they can become more popular over time?
Not so much, because when they become more competitive, they become a spoiler again, so it ends up having the same problem as FPTP.
For example, imagine a race on a one-dimensional political spectrum:
Blue is preferred by a 57% majority, so they win. But then a weaker third party, Green, enters the race:
Green only gets 27% of the vote, but under FPTP this means 27% fewer votes for Blue. This vote-splitting of the left side of the spectrum causes Red to win.
IRV corrects this problem by eliminating Green and transferring their votes to Blue, so that Blue is the winner. This is good because the voters preferred Blue over both Green and Red. The Red and Blue parties are protected from spoiling by the Greens.
That sounds like it fixes the spoiler effect to me!
Well, what happens if the population then shifts position, so that Green becomes more popular?
Green’s presence in the race takes away even more votes from Blue, causing Blue to be eliminated first. Then, enough of Blue’s votes transfer to Red that Red beats Green, which is the worst possible outcome for the Greens. If the Greens had instead voted dishonestly for Blue > Green > Red, or if the Green candidate had not run at all, then Blue would have won, just like under our current FPTP system.
So the spoiler effect still exists under RCV. It still punishes third party voters for voting honestly, breeding resentment against them for participating in the race, and encouraging them to vote dishonestly, in a way that perpetuates a two-party Red/Blue system.
(And if you don’t like the above graphs, here’s a video illustrating the same issues using cute little blobs:)
