Don't call Pierce County's ranked choice voting weird just because it's different

Peter Callaghan // Published June 11, 2009 in Tacoma News Tribune

Boil down the criticism of Pierce County’s voter-approved experiment with ranked choice voting, and you get this: It’s weird.

By holding just one election in November – instead of separate primary and general elections – Pierce is different than every political jurisdiction in the state.

Pierce asks voters to rank candidates as their first choice, second choice and third choice. If their favored candidate finishes last in the first count, the election computer looks for their second choice and gives it to one of the surviving candidates.

Contributing to the confusion, county voters had to take part in a traditional election for federal and state offices and an RCV election for county offices at the same time.

See, weird. But ranked choice isn’t just weird, critics say, it influenced which candidates won. For example, Pat McCarthy didn’t get the most votes in the first round of RCV voting but somehow ended up winning the county executive race.

And assessor-treasurer? Critics – mostly partisan Democrats and Republicans – claim the only way frequent candidate Dale Washam could have won is because voters were confused by the new system.

Expect to hear all this again in the fall as voters consider whether to repeal RCV.

But is it really weird, or just different? A trio of political scientists – led by University of Washington doctoral candidate Loren Collingwood and with help from professors Todd Donovan at Western Washington University and Matt Barreto at the UW – think Pierce’s RCV experience wasn’t weird at all.

“Our conclusion is that ranked choice voting mirrors the dynamics of a traditional election,” Collingwood said. That is, it effectively simulates both a primary and general election.

For example, candidates with the most votes in a traditional primary don’t always win the election. In 2008, 22 of 26 state Senate candidates who won the most votes in the primary were elected in November. Four, however, were not.

And 14 state House candidates who did not win a majority or plurality in the primary won the general election. This happens mostly in very competitive races.

“The fact that (Shawn) Bunney was ahead initially but failed to capitalize in the final round of voting speaks to the competitiveness of the race, not the iniquities of RCV as an electoral system, per se,” the authors wrote.

RCV also tracked traditional elections in another measure – the transfer of votes from a losing candidate in the primary to one of the two finalists. Usually, party labels are the dominant cue – that is, voters who voted for a losing Democrat in the primary are very likely to support another Democrat in the general.

That’s exactly what happened in the RCV election. Democratic executive candidate Calvin Goings’ supporters went heavily to fellow Democrat McCarthy, but independent candidate Mike Longeran’s second-choice votes were distributed fairly evenly among the surviving three candidates.

Washam’s victory followed a different pattern because it was a nonpartisan race and voters couldn’t rely on party labels. Each time a candidate was eliminated, their votes were divided relatively equally among the surviving candidates. Washam won the most votes in the first round and built on that in each successive round.

One promise of RCV appeared to come true – that candidates with less money have a better chance. In 2004, five or six county candidates who had the most money won. In 2008, only three of the six biggest spenders won

Collingwood said the experience of other jurisdictions suggests RCV also can lead to less-negative politics.

The author’s one concern is undervoting – voters taking part in federal and state elections but not voting for county races.

That’s always happens but was higher – nearly 25 percent – last November. Collingwood cited the fact that RCV was on a separate ballot from other races as a major cause.

While worrisome, “in coming elections, as voters become used to ranked choice voting, undervoting is quite likely to match levels reported in traditional elections.”

That is, if they don’t kill RCV in November.