I haven't come to a final opinion about the instant runoff voting procedure NC has used to fill Judge Wynn's former seat on the Court of Appeals, but I have some interim thoughts. And I also want to share some thoughts about comments received from a genuine expert on voting matters: Joyce McCloy.
The "instant" part of the procedure clearly refers to the voting, not the counting. The "runoff" part of the name is misleading. We don't have runoff primaries in judicial elections. IRV in this context provides an accelerated procedure to avoid the expense of a special election.
An ordinary judicial election with more than two candidates (thirteen sought this Court of Appeals seat) is held in two stages. The first stage is the primary, the purpose of which is to reduce the number of candidates to two for the general election. The two candidates then meet head to head. In this case, by definition, the winner will take more than fifty percent of the votes cast at the general election. But at the primary election, there is no requirement for any candidate to receive a majority of the votes cast in order to advance to the general election. It could therefore be said that the appearance of winning a majority of the votes is an illusion.
The same can be said of non judicial elections with many candidates from two or more parties for a single office. That is, the winner may actually enjoy the support of only a minority of the voters.
Ms. McCloy comments quite rightly that McCullough's vote amounts to only 27.99% of the total votes cast for this office, and Thigpen received only 27.65% of the votes. A true comparison with other judicial elections, though, would require counting total votes received at the general and dividing that by the total votes cast both at the primary and the general elections. I expect that would reveal that judicial candidates often are elected with a plurality rather than a majority of the vote.
Another way to compare would be to divide the votes received by each candidate by the total ballots cast in the state (2,003,130) instead of ballots cast for the office (1,943,771). That results in a vote percentage for each candidate of about 20%, give or take.
Counting the votes accurately in an IRV election presents special challenges because our machines are not programmed for such a count. The state board did a marvelous job of developing "work around" procedures. I'm just glad that Pamlico County uses the iVotronics machines. Our paper ballots (absentee by mail, curbside and provisional) had many more errors than the direct record equipment would allow. I'm confident in our count. Next week's recount is occasioned by the closeness of the result, not by the fact that it was an instant runoff.
If this procedure is to continue in the future, the state should make sure that replacement equipment is programmed to count IRV elections.
Another way to avoid the expense of runoff primaries, as I previously suggested here is to abolish the runoff. We don't need it. Most states don't use it. Despite its name, though, the purpose of IRV in judicial elections is not to avoid a runoff, but to avoid a special election. There are other ways to do this. For example, if the judicial opening comes too late to hold a primary at the regular time, just let the gubernatorial appointment to last until the next election cycle.
A more pointed question might be, "why do we elect judges, anyhow?"
That's the subject of another post.