Navigation

Memo: The Many Flaws of Ranked Choice Voting

Denver City Council is considering changing city elections to ranked choice voting. This member thinks that’s a bad idea.
Image: Denver City Hall, aka the City and County Buidling.
Denver City Council will consider a proposal to change city electios. Brandon Marshall

Today is the last day of our summer campaign, and we’re just $250 away from our goal!

We’re ready to deliver—but we need the resources to do it right. If Westword matters to you, please take action and contribute today to help us expand our current events coverage when it’s needed most.

Contribute Now

Progress to goal
$17,000
$16,750
Share this:
Carbonatix Pre-Player Loader

Audio By Carbonatix

There will be a public hearing on changing city official elections to ranked choice voting at the August 11 Denver City Council meeting. Denver City Councilmember Kevin Flynn prepared this memo on the proposal:

The big picture

Before diving into this lengthy memo summarizing seven years of research and data-mining (or in lieu of, if you don’t want to read all of it), consider these main points. Unlike our runoff system, Ranked Choice Voting:

• Cannot guarantee majority winners, forcing us to repeal the charter requirement for it

• Disenfranchises voters by not including their ballots in the final results if they fail to rank the right candidates – those not eliminated during rounds of counting – which is what most often prevents majority winners

• Does not save money from runoffs because it has to be spent instead on comprehensive and continual voter education campaigns, eating up those savings

• Uses confusing ballot configuration that contributes to voter error, disqualifying their ballots

I’ve collected data on 72 RCV-decided elections between 2004 and 2024. Eleven of them were regular mayoral elections in peer cities San Francisco, Oakland, Portland and Minneapolis. Not a single one of them has yet elected a candidate by a majority of the vote, even incumbents seeking re-election. Based on the data, it’s statistically more likely than not that if we were to switch to RCV, Denver’s mayor would not be elected by a majority of our voters. We deserve better than this.


The data analysis

Ranked choice voting is a step backward for Denver because, by repealing the runoff’s guarantee of a majority winner, it effectively takes us back to plurality voting, which we used to have, in which the person with the most votes wins no matter the percentage.

How so? Doesn’t RCV or “instant runoff” result in majority winners? No, most often it fails to do so. That’s why the draft charter amendment includes a repeal of our requirement for majority winners. RCV can’t guarantee that result, although its lobbyists and literature keep claiming so.

Data I’ve compiled from 72 RCV-decided elections going back 20 years in peer cities and elsewhere show that nearly 3 out of 4 winners failed to achieve a majority of the vote.

In the worst case, a San Francisco winner was declared despite receiving only 24 percent of the vote on the 20th round of transferring votes. Yet 76 percent of voters did not vote for that candidate. RCV reported that this candidate won with a “majority” of 53 percent. How did it figure that? By tossing out more than half of the voters from the final count because they failed to “rank” either the first or second-place finishers. The winner received only a slim majority of fewer than half the total ballots cast. Of all the voters who turned in their ballots, only 46 percent of them had their votes counted in the final result, and only 24 percent  chose this candidate. The data:

• 17,808 voters cast ballots in that race.

• Only 8,200 voters’ ballots were included in the final 20th round.

• Only 4,321 voters chose the “winning” candidate.

• 9,503 voters were removed from the final count, more than twice as many as voted for the “winner.” If those disenfranchised voters instead were given a runoff ballot, would the result be different?

Claiming that RCV results in more representative governing bodies flies in the face of the reality in Denver, where the gold-standard two-round runoff system has produced the most diverse city council Denver’s ever had: nine women, with six Latinas; two African-Americans. Nine of the eleven current district members first won their seats through the runoff, including six of the seven women district members and six of the seven members of color.

A candidate of color has been in six of the eight Denver mayoral runoff elections since 1983. The candidate of color won five of those six runoffs, Mares vs. Hickenlooper in 2003 being the only exception.

Since 1983, a candidate of color competed in 26 runoff elections for mayor, council, auditor and clerk, and won 23 times. More than half defeated white opponents.

RCV is not needed to provide diversity in Denver elections; we already have it.


RCV history in Denver

Denver had a ranked voting system from 1913 to 1935, in which voters could rank their first, second and third choices among all candidates. unlike the current proposal, Denver’s earlier version did not throw out people’s votes. Instead, all voters’ second and third choices were counted without eliminating any candidates. No one’s ballot was left out of the count, unlike current RCV.

Here’s an example of how it worked: In the 1923 election, second- and third-choice rankings pushed challenger Ben Stapleton over incumbent Mayor Dewey Bailey by fewer than 2000 votes. Bailey received the most 1st-choice votes but not a majority. So the 2nd- and 3rd-ranked choices of all voters were added to the total, giving Stapleton the win. (Rocky Mountain News, May 16, 1923)

Then from 1935 to 1952, Denver voters used plurality voting, which didn’t guarantee majority winners for mayor or council. Only the top vote-getters won. In 1951, the last election before the runoff system, the winning District 1 council candidate got only 28 percent of the vote among 18 candidates. This could also happen with RCV.

Finally in 1952, partly because of that result, voters adopted the runoff system, guaranteeing all Denver’s elected leaders would win with majority voter support.

The current RCV proposal

The RCV proposal doesn’t solve any problems, but it would create an entirely new problem – ballot “exhaustion” and disenfranchisement of voters who turned in legitimate ballots that gradually are eliminated from the counting – which can take up to ten days. Here’s how:

Voters who fail to rank any of the continuing candidates are not included in the outcome of the election. It is as though they stayed home and didn’t vote at all. In some RCV elections, between one-quarter to more than half of the ballots have been removed from the count. RCV calls these “exhausted ballots.”

The winning round shows vote totals only for the winner and runner-up. The totals are always less than the number who actually voted, and the votes for all eliminated candidates are shown as “zero.”

In fact, RCV always reduces turnout as the “instant runoff” counting process continues, because ballots that rank only candidates who already have been eliminated from further contention are no longer included in the total (2010 San Francisco Board of Supervisors District 10).

Unlike in real runoffs, these voters are denied a second opportunity to vote on the two finalists. Their voice no longer matters. The only voters who matter in the final results are those who include the last two remaining candidates among their rankings.

That can be a substantial percentage of the total who voted. In 72 RCV elections in peer cities and elsewhere, an average of 12.9 percent of voters’ ballots are thrown out of the count. This is disenfranchisement.

While the sponsors argue that it’s no different than choosing not to vote in a runoff, that is wrong. It is very different. In a runoff, voters who didn’t pick the two finalists get a second ballot and the opportunity to vote and take part in deciding the outcome, a second vote that RCV denies to them.

Illustrating how RCV reduces turnout while only runoffs can increase it, the average “exhaustion rate” of RCV ballots was 12.9 percent, as noted above. That’s the functional equivalent of a 12.9 percent drop in voter turnout in a runoff election compared to first election. Yet the claim that runoffs mean lower voter turnout is incorrect. Data show that since at least 1983, most Denver runoffs have had higher turnout.

In the 44 Denver runoff elections since 1983, an average of 5.8 percent more voters have turned out in June compared with the first election, and that’s combining runoffs with lower turnout as well as higher. The largest increase was in a District 7 runoff in 1995 that had an amazing 24.6 percent  increase in voter turnout over the first election.

Twenty-four Denver runoffs with higher turnout averaged 16.5 percent more voters than the first election while 20 runoffs with lower turnout averaged 6. percent  fewer voters. Even that is substantially better than RCV’s average “exhaustion” rate of 12.9% in every election.
click to enlarge A man looks.
Mike Johnston was one of seventeen mayoral candidates in 2023.
Bennito L. Kelty

The 2023 election

If RCV had been used in the 2023 Denver election, RCV election data analysis indicates that every voter who failed to include Mike Johnston or Kelly Brough among their choices would have been disenfranchised, their ballots “exhausted” and left out of the denominator when calculating “majority.”

With seventeen candidates in the race, it was extraordinarily difficult for voters to make well-informed decisions on their choices. “Debates” featured each candidate giving short answers to few questions. How could voters rank more than two or three? The proposal in front of us would allow six choices. Could the average voter authentically rank six of those seventeen candidates in order of their preference? Runoffs are better.

Even if they ranked only Calderon, Rougeot, Herod, Ortega or any of the other candidates, but not Johnston or Brough, no voter was denied another ballot to decide the winner, unlike in RCV.

Some RCV advocates have speculated that the mayoral outcome might have been different using ranked choice voting. The data suggest just the opposite – Johnston would have won, but would have fallen significantly short of 50 percent  of the vote.

No candidate among those 72 RCV elections ever topped a true majority from as far behind as Johnston was in the first election (he had 24.54%). The average increase in true winning percentage from first round to final was 9.7 percent. The data showed that the two candidates with the most first- or second-place votes in the opening round won 70 times out of 72. In 64 of those elections, the first-round leader ended up as the winner, two-thirds of them still falling short of a true majority because of “exhausted” ballots. In 6 of those elections, the second-place candidate in the first round won in the final.

In 72 elections, only twice did the third-place candidate in the initial round win. In both cases, they began close to the first-place opponent and relied on “exhausted ballots” tossed from the count. One was less than 1 percent away from the lead, the other was 3.5 percent away. In Denver, the third-place candidate in 2023 was 6.2 percent  off the lead, and no third-place candidate in these 72 RCV elections ever came close to making up that much ground to take first-place.

Still think “ballot exhaustion” is an acceptable compromise in order to eliminate our runoff elections? Consider this: New York City’s Democratic mayoral primary is decided by RCV.

Eric Adams won the 2021 primary by a mere 7,197 votes, out of 942,031 ballots cast, a margin of 0.076 percent among thirteen candidates. But 140,202 of those ballots – 15 percent of the total – were thrown out of the count because of “exhaustion.”

Consider: If only 5 percent more of those disenfranchised NYC voters, a good number of whom were Maya Wiley voters, had been given a runoff ballot and voted for Kathryn Garcia, Eric Adams would be a footnote right now.


Our runoff system is superior

The runoff system is superior to RCV in outcomes. Runoffs always guarantee a majority winner whereas RCV cannot.

Analysis of those 72 RCV elections showed only 20 produced a majority winner, fewer than 28 percent. That’s just not good enough. Only runoffs can increase participation in deciding the winner.

RCV always decreases “turnout.” But sometimes, RCV decreases total turnout. Portland instituted RCV last year. Instead of the higher turnout that was anticipated, turnout “cratered,” to quote the city’s newspaper. Total vote in the
mayor’s race fell 15 percent from the 2020 mayoral election pre-RCV.

In Portland, 50,010 voters’ ballots, or 16.1 percent of the total, were tossed out of the final count, resulting in the winner not achieving a majority of the vote. And 11 percent of Portland voters left their choices for mayor blank, nearly twice the percentage in the 2020 election.

In Portland’s District 1, where communities of color and lower-income households are over-represented, 22.2 percent of voters made no selection at all for any council candidates to fill three district seats. Of the 42,516 votes cast in the first round, only 31,186 were counted in the final 17th round.

Bottom line, there is no problem being solved by switching to RCV. Our current system is better.

The claim that RCV saves money by eliminating a potential second election is contradicted by the request that the money be spent instead on a permanent program of comprehensive voter education, because some studies show RCV negatively affects participation among communities of color, low-income voters and seniors.

Even as a financial break-even, RCV comes at the non-monetary cost of having to repeal our long-standing charter requirement that officials be elected by a majority. While RCV might arguably be an improvement over strict plurality voting, it makes no sense for Denver to move backward from our runoff model.

The claim that RCV campaigns are more civil is merely aspirational. Most Denver campaigns are civil, and at the same time many RCV campaigns have been nasty. In addition to New York’s Adams-Garcia RCV race, look at June’s Mamdani-Cuomo primary for a reality check.

The fact is, nearly all negative campaigning comes from independent expenditure committees, not the candidates, and that will be true in all election systems. RCV cannot change this, unfortunately, and it will continue either way.

The claim that RCV incentivizes candidates to ask voters to make them their second choice if not their first is also true in Denver runoffs. Those who’ve had potential runoffs know this because they’ve asked voters who supported opponents to make them their second choice if they made the runoff. RCV offers nothing different.

There simply is no reason for this change.

Kevin Flynn is a career journalist who was elected to represent District 2 on Denver City Council in 2015, after nearly three decades of covering the mayor, council and other Denver area stories for the
Rocky Mountain News. He was re-elected in 2019 and 2023; this will be his last term. Learn more on his Denver City Council page.

On weekends, westword.com publishes commentaries on matters of interest to the Denver community; the opinions are those of the authors, not Westword. Have a piece you'd like to share? Email [email protected], where you can also comment on this essay.