The case for Instant Runoff Voting

An edited version of this piece was published in the March 8, 2006, issue of Urban Tulsa Weekly. The archived version is no longer online. Posted on the web July 1, 2010.

The case for Instant Runoff Voting
By Michael D. Bates

The polls haven't even opened as I write this, but by the time you read this, the votes will have been counted.

Let me make this one prediction: Whoever wins the Republican nomination for Mayor will win with less than a 50% of the vote. Whoever the winner will be wasn't the top choice of a majority of Republican voters.

As I write this, I don't know if my favorite candidate won or lost, so this isn't sour grapes. I believe in majority rule, and there's something wrong about a candidate winning when most of the voters preferred someone else. Instead of the office going to the candidate who could put together a majority coalition of voters, it goes to the candidate with the most motivated, cohesive minority bloc of voters.

For state offices, Oklahoma deals with this problem by holding a primary runoff election if the primary doesn't produce a candidate with at least 50% of the vote. Tulsa's city charter doesn't provide for a runoff, and in special elections, Tulsa doesn't even have party primaries.

When a primary has three or more viable candidates, and there is no runoff, the voter is faced with a difficult choice. He may have a favorite candidate, other candidates that might be marginally acceptable, and a candidate he doesn't want to win. Does his favorite have a chance to win? If not, it might be more strategic to vote for a marginally acceptable candidate with a chance of winning in order to defeat the utterly unacceptable candidate.

At that point, the problem is knowing who really has a chance to win. You could look to polls, but if the numbers between second and third place are close, that doesn't help your decision. You could try to measure candidate viability by the number of dollars raised and spent.

Wouldn't it be nice to have a system of voting that allowed you to vote your conscience, a system that didn't require you to guess how your neighbors are likely to vote?

Hold that thought.

Adding a runoff for city elections would be a good start. A single runoff works well when you have two candidates neck-and-neck in the 40% range, and a group of other candidates splitting the rest of the vote. The primary reveals who the two most popular candidates are; the runoff settles the question, for whom would you vote if these were the only candidates in the race.

But a single runoff breaks down when you have three or more candidates who are clustered around 30%, with other candidates splitting the remainder.

The legendary example of this situation was the 1991 Governor's race in Louisiana. Incumbent governor Buddy Roemer, a moderate and respectable political figure, finished a very close third, giving voters a runoff choice between "The Crook" (Edwin Edwards) and "The Klansman" (David Duke). The also-ran candidates had enough votes between them that, had they not been in the race, would have been enough to put Roemer in first or second place and into the runoff.

There are plenty of Oklahoma examples of close three-or-more-way primaries where the single runoff system broke down. In 1990 both parties had gubernatorial primaries with three closely bunched candidates and a lot of also-rans. Burns Hargis might well have placed second in the GOP primary had it been a three man race. For the Democrats, Steve Lewis easily could have finished ahead of David Walters and Wes Watkins, if a couple of minor candidates had not been in the race.

Wouldn't it be nice to have a system of voting where there is no such thing as a spoiler candidate or splitting the vote? Where the winner would have won even in a one-on-one race with each other candidate?

Of course it would.

In the ideal system, you'd have a series of runoffs, and in each round, you'd eliminate the lowest vote-getter and vote on the rest, until someone gets 50% of the vote in a round.

Some civic organizations handle their elections in this way, but that's because they can count their votes in a matter of minutes and hold another round of voting right away. It would be burdensome enough to add a runoff election date to the city election calendar, much less multiple runoffs.

There is a way to get the effect of multiple runoff rounds without having the expense of multiple elections. Historically it's been called the "alternative vote," but recently it's been given the more descriptive name of "instant runoff voting" (IRV).

I first encountered this voting technique in my college fraternity. When voting in an election with multiple candidates, you'd write the candidates' names on the ballot in order of preference. As a voter, you'd ask yourself, if my favorite candidate weren't in the race, for whom would I vote? And if my two favorites were out, who would be my choice? And so on through the list of candidates.

The vote counters would take all the ballots to the table in the chapter room, and for each candidate there'd be a stack of the ballots that had his name marked as first choice. If no candidate had a majority, they would take the smallest stack of ballots - eliminating the candidate with the fewest first choice votes - and sort them into the other stacks based on the second choice listed. The process would repeat until one candidate had a majority of the ballots in his stack.

Xi Chapter of Zeta Beta Tau, I later learned, isn't the only place that instant runoff voting is used. Ireland uses IRV to elect its president. Australia uses it to elect its House of Representatives. London uses it to elect a mayor.

In 2002, the Utah Republican Convention used IRV to pick the top two candidates to compete in their congressional primary election in two districts and to pick a nominee for the third district.

Right here in Tulsa, the 1st District Republican Convention used instant runoff voting to choose delegates and alternates to the 2000 and 2004 Republican National Conventions.

San Francisco has adopted IRV. San Diego is looking at it - their last mayor was elected with only 35% of the vote.

Although our fraternity hand-counted our IRV ballots, modern voting machine technology makes it possible to scan preferential ballots optically and then conduct the sorting, elimination, and resorting by computer.

Some sort of runoff will become even more important if Tulsa ever switches to non-partisan elections. As it is now, primaries present general election voters with two candidates with a credible chance to win, and most officials are elected with over 50% of the vote.

We don't have to guess about the dynamics of a non-partisan winner-take-all election, because our city special elections are structured that way.

Last May's District 5 special election had seven candidates, four of whom had a significant base of support. The winner, Bill Martinson, only managed 29% of the vote, just 11 votes ahead of second-place finisher Andy Phillips. Based on the results, you could make the case that Martinson would not have won head-to-head races against any of the other major candidates; the split vote gave him the election.

Without a runoff, non-partisan city elections would give us a set of elected officials who lack the mandate of the majority.

Adopting IRV would require some changes to the City Charter, but since we're already taking another look at the charter and at the role of partisanship in city elections, let's not overlook the voting system we use.