Strengths and Weaknesses of Instant-Runoff Voting

I had some more thoughts on instant-runoff voting (IRV). Specifically I want to talk about some of the implications of it.

IRV makes tactical voting extremely difficult.

Voting, ostensibly, should be a voter entering the booth and casting a vote for the candidate they think will be most likely to enact the policies they support. However, in a plurality voting system like we have currently, that can be detrimental to getting your supported policies enacted. Take an electorate where the majority of the electorate believes that we must enact laws mandating that dinner plates must be between 10-12 inches in diameter and a sizable minority believe that dinner plates should only be 5-7 inches in diameter. In a two-candidate election with each candidate espousing a different side of the argument, the winning candidate would likely be the one that supports the larger dinner plate size (yes, this is a weird, single-issue electorate).

Adding a third candidate to that mix can cause the majority of the electorate to lose, if two large dinner plate candidates split the large vote, allowing a candidate who supports a minority opinion to get through. Suddenly, the majority of people are trying to squeeze their dinners onto bread plates, against their wishes. In subsequent elections, voters will attempt to support the candidate they think will most closely match their views and is also most likely to get elected.

This is bad for democracy, since it narrows our choices. However, IRV stops that by offering little incentive for voting tactically. Since voting for a candidate will not “spoil” another candidate’s chances of victory. In the first round of counting, the winners will be the candidates with the strongest amount of support, and then will pick up votes based on waning levels of support, just like in a runoff election.

IRV eliminates the need for separate runoff elections.

Let’s talk about Central Falls in 2012. Central Falls did something weird in 2012 — they held a nonpartisan primary on the regular November election day, and then held the official “election” between the two highest vote-getters on December 11th. This was a hilariously stupid way of doing it. In a sensible system, this would’ve been a runoff election. That would’ve meant that when James Diossa won over 2000 votes and close to 60% in November, he would’ve immediately been declared the mayor-elect, since he had won a majority of voters.

However, since Central Falls committed to it being a “primary” Diossa had to run in a second, later election against an opponent (Joseph Moran) he had already convincingly defeated. In that election, turnout fell, and around half of Diossa’s voters failed to show up while Moran’s held steady. If Diossa had lost (say because more of his voters failed to turn out), the legitimacy of the Central Falls government would’ve been called into question, since a majority of voters had already noted their preference for Diossa in November. Ironically, Moran had already noted the cost of holding elections in Central Falls, which was exiting bankruptcy at the time. Of course, that concern over costs did not lead to Moran dropping out, even though the whole December election was a frivolous waste of money.

IRV negates the need for separate runoffs like Central Falls’ de facto runoff. It’s cost saving, because it pretty much holds a bunch of elections in one instance. One complaint directed against IRV is that it violates the “one man, one vote” principle because a losing candidate’s supporters get to vote more than once. However, this only works if your concept of IRV ignores the “runoff” part of the system. IRV holds multiple sequential runoffs. A second round of voting is essentially the same as a second election, with a winning candidate’s voters casting another vote for that candidate in that second election.

The other complaint is that it lacks the distance between elections that’s a feature of regular runoffs. But personally, I find this complaint unconvincing. It’s usually “who knows what could happen in that time?” But this is a feature of the current system. After November, a winning candidate could be outed as a pervert or otherwise deeply flawed. Waiting for whatever potential disaster might strike unnecessarily stalls action. There comes a time where you need to stop the campaign and vote.

IRV marginally increases the chances of third parties.

I want to be clear that is not a great reason to institute IRV, but it is a feature that’s worth talking about. One of the structural issues that prevents third parties from is the spoiler effect talked about earlier. Since IRV eliminates that, it makes it easier for third parties to be successful.

However, just because one impediment is taken away doesn’t mean others are lifted. In RI for example, to qualify for party status, a political party has to collect signatures equal to 5% of either the presidential or gubernatorial vote, and then win above 5% in the subsequent election for either president or governor (depending on when they put in the signatures). That requirement doesn’t disappear, and thus IRV would be most likely to benefit political parties with a credible chance of getting 5%. I imagine the laws would have to be rewritten so that it was 5% in the first round of voting, but it’s workable.

The other major impediments to third parties are finding enough candidates, and the fact that the lack of a national organization and profile limits the ability of voters to identify them. It is still possible in Rhode Island to have completely ignored the existence of the Moderate Party, given it has had zero impact on government.

It would probably require new equipment and thus a new manner of voting.

This is a more practical concern, but the issue of whether Rhode Island’s current set of voting machines are up to a change in system has to be dealt with. The connected arrow system has been around for a long time, and voters are familiar with it. It would be ideal for the State to acquire voting equipment that could handle a range of different voting methods, with the intent that any changes to the voting system would have machinery capable of handling it.

The worse-case scenario would be mandating a change in the voting method, while not making an allowance for new machines, which could ultimately just throw the future of the whole system into doubt.