Follow GOVERNING on:

AddThis Social Bookmark Button

Ballot Box

May 2010

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          

« You Know You've Lived in DC Too Long... | Main | Obama and Clinton -- Stay Out »

February 07, 2008

A Viable Third Party, Instantly?

posted by Josh Goodman

Ballot Access News had quite an intriguing scoop a few weeks ago: The Vermont legislature is likely to vote to make the state the first in American history to use instant-runoff voting (IRV) for congressional general elections.

For those not familiar with IRV, it's a system where voters rank the candidates, rather than simply voting for one. If one candidate has an outright majority of first-choice votes, he or she wins. Otherwise, whoever receives the fewest first-place votes is eliminated and the second choices of that candidate's voters are then added to the others' totals.

This process continues until one candidate has more than half the votes. IRV is currently used in municipal elections in a handful of places from San Francisco, California to Cary, North Carolina.

Yesterday, I spoke with Vermont Rep. David Zuckerman, a longtime IRV supporter. He confirmed that the bill's prospects look good, although a gubernatorial veto is possible. Zuckerman said the measure doesn't have a veto-proof majority.

Debates over instant-runoff voting are always interesting because they tease out competing conceptions of democracy. Supporters argue that democracy demands we elect the person who is acceptable to the largest number of people, which IRV produces. Opponents of IRV argue that first choices are all that should matter and that there's nothing wrong with plurality winners.

IRV is also sometimes opposed on the grounds that it might cause voter confusion or lead to behind-the-scenes deals between candidates. Such a deal could look like the West Virginia Republican presidential convention on Tuesday.

I thought Zuckerman made a pretty good case, however, that these types of deals are unlikely in a general election. It's much easier to boss around a few delegates than to control tens of thousands of voters. If a deal seemed ideologically incoherent  -- if, say, the Republicans and the Socialist Workers Party encouraged their supporters to list each other as second choices -- the press would not be kind.

The biggest impact of IRV might be to make third parties viable. With IRV, liberals in 2000 wouldn't have been tormented over the thought of a vote for Nader electing Bush. They could have voted for Nader as a first choice and Gore second.

In this way, if IRV were implemented nationally, it would eliminate "spoiler" fears, one of the key factors that perpetuates the dominance of the two major parties (more so than their popularity with the public). Zuckerman, a member of the Vermont Progressive Party, is one of only 21 state legislators nationwide who is neither a Democrat nor a Republican (not counting Nebraska's non-partisan legislature), out of nearly 7,400 total.

For now, legislators in Vermont are focusing only on congressional races because there are constitutional concerns about applying IRV to state races. In particular, there's a possible conflict with Vermont's unusual constitutional rule where elections for governor and lieutenant governor are thrown into the legislature if no one receives a majority of the vote. Still, it's not hard to imagine that, if voters like IRV on the federal level, they would demand that all elections use the instant-runoff system.

So would IRV help the Progressives in Vermont? Not unless the voters liked Progressive ideas, Zuckerman notes. "What IRV would allow," he says, "is voters to vote with their true beliefs."


Rob Richie

Fascinating debate underway in Vermont. Just to add a few parts of it:

- The biggest practical impact would be upholding of the principle of majority rule. If a candidate is strongly opposed and wins only because the second-place finisher lost votes to a third candidate, that really could be understood to be as much of a violation of majority rule as the electoral college allowing someone to win who loses the national popular vote -- yet we use voting methods that allow that all the time.

- The West Virginia results did show how the principle of majority voting is a part of our politics when it's feasible. IRV-type balloting also is at work when parties elect their leaders in Congress and most state legislators, say --- with repeated balloting until you get a majority winner.

- The Republican primaries more generally have been a great example (as were the Democrats until John Edwards dropped out) of the weirdness of our system where a candidate is allowed to get all the credit for winning (and all the delegates in many GOP contests) even if that win is founded on the majority splitting its vote.

- That makes John McCain's backing for instant runoff voting all the more appropriate to point out. You can hear him talk about IRV at:

- Barack Obama has also backed IRV as the lead sponsor of IRV legislation in Illinois See:

- DNC chair and former Vermont governor Howard Dean has spoken out strongly in favor of IRV. See:

Current Vermont governor Jim Douglas signed the bill to have Burlington go to IRV, and as a backer of Sen. McCain, he may well be ready to uphold the principle of majority rule in Vermont -- or at least not block it. We may have a chance to see.

ivan swift

IRV sounds like a big improvement over how we go about electing people now, but who's going to keep track of totals? who's going to set up the computer programs? how do you ensure honest counts? Dieboldt?

Warren D Smith

This post unfortunately is rife with errors.

**First of all IRV does NOT "make third parties viable." In every IRV country in all the history of the world, all IRV seats have soon become massively 2-party dominated. For example, Australia, the most experienced IRV-using country, has 564 IRV seats counting both federal and statehouse. We tallied them and found of these 564, exactly ONE seat was occupied by a 3d-party member. For this reason, Australia's third parties want to get rid of IRV. They want PR, which is also used in Australia (for other seats) and which there does work to give seats to 3d parties.


**Second you said, after an IRV-rules-description, "IRV is currently used.. San Francisco." Actually, the IRV rules you described are NOT used in San Francisco.
E.g. for details see

**Third, you say
"Supporters argue that democracy demands we elect the person who is acceptable to the largest number of people, which IRV produces."

That statement is false. You meant to say
"approval voting," not IRV. E.g. see

And for a counterexample to your statement,
one may take any election whatever in which the Approval-voting winner differs from the IRV winner (there are an infinite
number of examples).

**Fourth, you say
"With IRV, liberals in 2000 wouldn't have been tormented over the thought of a vote for Nader electing Bush. They could have voted for Nader as a first choice and Gore second.
In this way, if IRV were implemented nationally, it would eliminate "spoiler" fears,..."

That again is false.
Spoiler phenomenon still happens in IRV
and is very common too.
For a simple example, see

**Fifth, you quote Zuckerman:
"What IRV would allow," he says, "is voters to vote with their true beliefs."

That again is false.
There are many many scenarios in IRV where
voters are stupid to vote their true beliefs,
and wiser voters, would vote in ways
not representing their true beliefs because in that way, and only in that way,
could they improve the winner from their point of view.

I already mentioned an example of that

More examples (there are an infinite number) include
and those in

For examples of IRV pathologies in real life
elections consider these

and you will see that it is very common for IRV elections to exhibit pathologies.

Why is it that IRV-proponents are constantly either unaware of the facts, and/or tell
lies? You might ask yourself...
where do those errors and/or lies trace to?
There is one place on the internet that seems to be responsible for promulgating a great number of them. They call themselves "FairVote." I hope you will not join that team.

Dave Ketchum

Your claim that Instant Runoff (IRV) elects "the person who is acceptable to the largest number of people" is false.

Here is a counterexample:

#voters their vote
------- ---------
35: H>W - they would like to live where it is HOT
33: C>W - they would like to live where it is COLD
32: W - they like WARM, neither HOT nor COLD

A majority dislike HOT (worse than WARM).
A majority dislike COLD (worse than WARM).
All find WARM acceptable.
IRV will discard the WARM voters and declare:
A MAJORITY of those we are still counting ike HOT, so HOT wins.

So IRV in this example elects HOT, even though the largest number of people (100% in fact) find WARM acceptable, while only a minority find HOT acceptable.

Rob Richie

Dave -- your example would "work" the same if the numbers who wanted warm were 3%, as in:

49: H>W - they would like to live where it is HOT
48: C>W - they would like to live where it is COLD
3: W - they like WARM, neither HOT nor COLD

I know you think that warm should win even when it's luke-warm, with that low a degree of first choice support, but I think a lot of folks would disagree with you.

K Venzke

IRV has its own spoiler issues, though apparently Zuckerman didn't mention them to you.

But did he actually say this:
"elect the person who is acceptable to the largest number of people, which IRV produces" ?

That would be the worst summary of IRV I've ever read. IRV doesn't even attempt to do that. What it's doing is assembling a majority using the highest preferences it can. In that respect it isn't really that different from plurality philosophically.

I don't think plurality is better than IRV, but among all possible reforms I don't think IRV is very good.

I think IRV would cause third parties to receive more votes, but I'm skeptical that more than two candidates would become viable. As soon as it isn't obvious which two candidates will be left in the end, there are spoiler concerns.

Josh Goodman

That was some sloppy writing on my part. Here's what I was trying to get at: If your democratic ideal is "the person who is acceptable to the most possible people should win," you're probably going to prefer IRV over our present system because, in my opinion anyway, it gets closer to that goal more of the time.
One interesting thing about IRV is that it rules out any candidate that is most unacceptable to a majority of voters. It's impossible for a candidate that is ranked last by more than 50% of the voters to win (assuming everyone votes for every candidate). In that way, IRV favors candidates that are at least somewhat acceptable to a large number of voters.

Dave Ketchum

As to IRV, it does offer the voter ability to vote for more than one candidate and to indicate relative liking among them - NEEDED improvement over Plurality.

And IRV does, usually, correctly select the winner.

But, we need to go for a method, such as Condorcet, that does better at the whole job rather than suffer IRV's occasional failures.

In the example I presented, and in Rob Richie's response, a majority preferred W over H, as did a majority prefer W over C - but IRV could not see this. IRV satisfied the minority who voted H.

Josh talks of all voters voting for both those they like AND those they hate. This might not be too bad for 3 candidates, but it is an ugly pain when there are a dozen or more, especially if the voters are expected to reasonably rank those they hate.


I'm not a fan of IRV and I agree with Dave's criticism. (For what it is worth, I am a fan of Approval voting)

You can read some criticisms of IRV here:

You can read about approval voting here:

That said, I think that Rob's 'luke-warm' retort is very funny and clever.

Bart Ingles

I am also a fan of approval voting. IRV and traditional runoffs are unsuitable for use in partisan elections because of the center-squeeze effect. In a three-way race where the candidates fall into a left-center-right arrangement, where the left and right candidates aren't too close to the fringe, then all else being equal the centrist will be eliminated in the first round.
Approval voting doesn't have this problem. It's also much simpler than IRV, and doesn't require IRV's expensive and complicated ranked ballots.

Clay Shentrup

Approval Voting is great for simplicity, but the advantages of a larger range, like 0-9, are large in every other regard, and almost certainly worth it.

The comments to this entry are closed.