December 21, 2010
- Rob Richie, rr [at] fairvote.org, (301) 270-4616
- Toby Rowe, jrowe [at] fairvote.org, (301) 270-4616
NC Court of Appeals Instant Runoff Election Results
Gary Bartlett: "Whether you like it or not, it worked"
In an election that saw the first-ever use of instant runoff voting (IRV) in a statewide election in North Carolina, Doug McCullough has defeated Cressie Thigpen for a seat on the NC Court of Appeals by 50.3% to 49.7% in the instant runoff. With 13 candidates running for the position, the election would have posed a challenge for any electoral system. Using IRV, voters expressed their preferences among the available candidates and the candidate with the most popular support was elected.
[North Carolina State Board of Elections director] Gary Bartlett said there were no major problems with the count and voters got to participate in picking the winner . "Whether you like it or not, it worked," he said.
-- Associated Press story circulated on December 20
The election was one of four instant runoff voting elections in North Carolina this year. Superior Court race 12A in Cumberland County also was decided in a second round of counting, as we reported earlier this month.
Why instant runoff voting was used
In 2006, North Carolina adopted a law establishing that instant runoff voting would be used to fill judicial vacancies created between the May primary and early September of election years where a traditional two-round election system would be costly and result in low turnout in one round. Before 2006, such vacancies were filled in a single election where voters cast only one vote. In 2004, the winner in an eight-candidate statewide race for the North Carolina Supreme Court earned 23%.
This year, four judicial vacancy elections were created in the proscribed time period: one for the statewide Court of Appeals, for which 13 candidates ran, and three for Superior Court positions, each with three candidates.
What instant runoff voting is
On November 2, voters in these elections had the opportunity to rank their 1st, 2nd and 3rd choices for the position. IRV simulates a traditional runoff. If a candidate wins a majority of 1st choices, that candidate is elected. If not, all but the top two candidates are eliminated, and each ballot is counted once for whichever runoff candidate is ranked higher on that ballot. The candidate who is ranked ahead of the other candidate on more ballots is elected, just as the candidate with more votes in a traditional runoff is declared the winner.
No candidate came close to winning a majority of the 1st choice vote in either the statewide race or the Superior Court election in Cumberland County. The races proceeded to an instant runoff, with the top two candidates advancing and all others eliminated. The instant runoff started on November 29, after the statewide canvass of ballots was complete
Note about the ballot-counting: If North Carolina had instant runoff software that is used in other jurisdictions such as Oakland, California, the instant runoff results would have been available on election night. The cost of obtaining and implementing such software is estimated to be less than a quarter of the cost of a single statewide runoff.
After 1st choices were counted, no candidate came close to a majority of ballots cast. The top candidates were Doug McCullough with 15.21% of 1st choice rankings and Cressie Thigpen with 20.33%.
Since no candidate received a majority in the first round, the race proceeded to an instant runoff, with McCullough and Thigpen advancing and all others eliminated. Ballots ranking eliminated candidates 1st were then counted instead for whichever remaining candidate (McCullough or Thigpen) was ranked higher on each ballot.
Once these ballots had been added to the candidates’ totals, McCullough was the winner with 50.31% of the final-round vote. This means that among voters who expressed a preference between the top two candidates, 50.31% preferred McCullough. The vote-count was close enough to trigger a statewide recount; the margin between the candidates changed by fewer than 100 votes, underscoring the quality of the original count.
Key facts and their implications
Such a large field of candidates, combined with a relatively short time frame for conducting the election, would have placed strain on any electoral system. But IRV provided a better means of conducting the vacancy election than other alternatives – and clearly performed better than the previous method of electing judges.
As a tool for conducting vacancy elections, IRV showed that it provides a more accurate expression of voters’ preferences than the previous system. (The old system, in which the candidate with a simple plurality of votes won, became a target of public scrutiny when it allowed candidates with a very small percentage of the vote to win). Under the old system, Thigpen’s 20% of the first-round total would have been enough for an outright victory. This year, however, the use of IRV allowed more voters to express a preference between the two front-runners. The result was that 50.31% of voters who expressed a preference between the top two candidates preferred McCullough to Thigpen. This demonstrates that McCullough narrowly had more support than Thigpen, but that some of McCullough's backers also chose other candidates as 1st or 2nd choices. The current IRV method of filling judicial vacancies allows this support to be taken into account, rather than simply awarding the election to the candidate with the highest plurality of votes cast.
Voter participation in the race was higher than it would have been had a traditional runoff been used to fill the Court of Appeals vacancy. Traditional runoff elections are typically poorly attended, both in North Carolina and elsewhere – particularly in late November or December after a major November general election. Statewide, turnout in primary election runoffs typically declines sharply in North Carolina. For example, in 2008, just 63,910 voters participated in the Democratic primary runoff for Commissioner of Labor, as opposed to 1,200,407 voters in the first round in May. Overall, statewide turnout for the November 2010 election was 44% of voters, but runoff turnout almost certainly would have been in the single digits.
An illustrative recent example of the pitfalls of separate runoff elections comes from the state of Georgia, which held a statewide runoff for a judicial position in December 2010. Turnout in that race was well under 10%, just a fraction of turnout for the general election that preceded it. By contrast, IRV allowed North Carolina to fill the vacancies in in a single, higher-turnout general election.
Runoffs are also costly to taxpayers: In Georgia, the estimated cost of the statewide runoff was $4 million. Costs in North Carolina would have been comparable, perhaps as high as $5 million. Runoffs also exacerbate campaign finance demands, forcing candidates to raise and spend lots of money quickly.
The Court of Appeals election demonstrated the advantages of IRV over plurality voting and potential December runoff elections as a method of conducting judicial vacancy elections. As an alternative to separate runoff elections, IRV has the benefit of allowing for much higher voter turnout and participation than is typical of December runoffs. As an alternative to the former plurality system, IRV reduces the occurrence of low-plurality winners by providing a more comprehensive reflection of voters’ preferences between candidates. Taken together, these benefits suggest that IRV is the best available method of filling judicial vacancies that occur too late for the normal primary schedule.
FairVote tracks these and other IRV elections around the country as part of its work on election analysis and reform. For more information, contact FairVote’s Toby Rowe or Rob Richie at (301) 270-4616 or firstname.lastname@example.org.