How ranked-choice voting worked in District 1

See how the ranked-choice votes were counted in Oakland's 2012 District 1 City Council race.

It took seven rounds of eliminations and transfer votes to determine the winner in Oakland’s 2012 District 1 City Council race.

What exactly does that mean? Oaklanders may be used to voting in ranked choice elections, but the system by which votes are counted is a little more complicated. OaklandNorth breaks down the distribution of ranked-choice votes in the District 1 race in this video to make the process a little easier to understand.

Correction: The the meaning of the term “Under-vote,” which has more than one common use, has now been clarified in the video.

29 Comments

  1. This is an excellent video. I believe some would-be voters have avoided voting because they do not understand how ranked choice works.

    I’m proud of my neighbors for what was a well-informed, well-exercised turnout with few problems and all-in-all pretty decent democracy.

    In the absence of an incumbent, a large number of candidates made this race more demanding for the voter.

    As a District 1 voter, I spent extra energy to understand who the candidates – and their support base – were before making my decision. I was tactical with my ballot and could not be more pleased with the outcome.

    I support Ranked Choice Voting for Oakland.

    Thanks,
    MTK

    • Max Allstadt

      Ranked Choice Voting doesn’t work.

      Read this: http://rangevoting.org/Burlington.html

      It’s a great explanation of the failure of this form of voting, and it’s flaws, as expressed by an election in Burlington, VT that ultimately led the public to vote to abolish instant runoff voting.

      • Rob Richie

        Nonsense about Burlington. The repeal was a highly partisan move led entirely by backers of the candidate who led in first choices , but was in fact less popular both than the IRV winner and the third-place candidate.

        In Oakland, 16 of the 18 RCV winners have won with more votes than the previous winner of those offices without RCV. See the summary here:
        http://www.fairvote.org/assets/Oakland-RCV-Analysis-2012.pdf

  2. Bruce Nye

    This is really well done. I’d like to see it get lots of play in the time leading up to the 2014 election.

  3. Max Allstadt

    Um, this actually has one major error, and a total lack of any explanation of what exhausted ballots mean.

    Quiz: There were 34180 votes cast. Dan Kalb won 12293 votes. What percentage of the electorate voted for Dan Kalb?

    Oakland North’s analysis does not answer this question.

    Also, if I’m reading the ROV data and math correctly, the 5511 undervotes might actually be people who lived in district one who didn’t vote for ANY candidate in the council race. Then we have 149 over votes (people who voted twice or more within the 1st, 2nd, or 3rd choice column on their ballot). Add those together and you get 5660 voters in D1 who’s votes didn’t count in the council race due to non-voting or error. (and there’s even another unknown: if you leave your 1st choice blank and vote in the 2nd and 3rd choice columns, does your vote get counted?)

    So, if we do count over and undervotes in the final tally, Dan Kalb’s 12293 total votes out of 34180 total cast means that after the Ranked Choice algorithm ran, he received 35.97% of the total votes cast.

    If we exclude the undervotes and overvotes, 34180-5660 leaves 28520 total votes cast. Kalb’s 12293 total out of 28520 would mean that he got 43.1% of the total VALID votes cast by citizens who lived in D1, chose to vote in the D1 Council Race, and who did not mark their ballots incorrectly.

    And that my friends, is why Ranked Choice Voting is crazy.

    • Max Allstadt

      Here’s the ROV spreadsheet for anybody who wants to check my math, and cross reference all the numbers I’m mentioning in my previous comment: http://www.acgov.org/rov/rcv/results/rcvresults_6464.htm

    • Rob Richie

      Max — I just sent a note to Sam making the same point about undervotes being people who skipped the race entirely. So yes, Kalb won more than 40% of those who had a first choice preference and more than half of those who had a preference for either him or Lemley.

      Here’s what you’re missing, though. LOTS Of people voted in the November election. In Oakland’s corresponding elections in 2008, every city council district race was won in June and every race was won with fewer votes than the winners earned last month. If you want to get a sense of something being crazy. see this spreadsheet comparing turnout in the often decisive June primaries with turnout in November elections:
      http://www.fairvote.org/assets/OaklandJunePrimariesFactSheet-1.pdf

      So in 2008, of example, without RCV Disrict 3 was won with 4,503. votes. This year, in a highly competitive RCV race, District 3 was won with 9,397 votes — more than twice as many.

      And if you want to get into REALLY crazy stuff, check out some of the weird outcomes in top two primary races this year, where big field were reduced to two in June. In Congressional District 31, say, Gary Miller won as a Republican because he had only a GOP opponent. His district is heavily Democratic and majority-minority, but Democratic candidates split the vote in June.

      Finally, we have found that fewer people at the polls usually skip RCV races than they did in the same kind of races in pre-RCV races — probably because more candidates are out there engaging with voters. It was GOOD that Oakland voters had a chance to hear about and participate in city elections this November and hear from more candidates rather than have most everything end in June.

      • Max Allstadt

        Here is the true math on what percentage of the vote the winning and runner-up candidates got in elections where ranked choice mattered:

        Votes/(Valid Last Round Votes + Exhausted Ballots) = percentage of electorate won.

        Gallo: 5716/(10460+768) = 0.50908443177 = 50.90%
        Juarez: 4744/(10460+768)= 0.42251514072 = 42.25%
        Exhausted Ballots = 6.84%

        McElhaney: 9397/(17427+4516)=0.42824590985 = 42.82%
        Sullivan: 8030/(17427+4516)=0.36594813835 = 36.59%
        Exhausted Ballots = 20.58%

        Kalb: 12293/(23741+4779)= 0.43103085554 = 43.10%
        Lemley: 11448/(23741+4779)= 0.40140252454 = 40.14%
        Exhausted Ballots = 16.76%

        The bigger the race, the higher the likelihood of more exhausted ballots.

        The current system only shows the electorate the percentage that the winner and runner up won out of the votes in the final round. It does not account for exhausted ballots. This is a fundamentally untrue way of expressing the results, because it does not show the true mandate of the winner.

  4. Mollie Cohen-Rosenthal

    Sam, I think this is an absolutely excellent video! Well done!

    Max, while the discrepancy you found might be valid, and I don’t have time to double check your math, would it effect the outcome differently?

    • Max Allstadt

      We don’t know if it would affect the outcome.

      We do know that Kalb’s win with 43.1% of the valid votes cast means a very different mandate that a true 52%/48% victory would mean.

      Put simply, as of January, Oakland will have 3 officeholders who did not get over 50% of the votes from their respective electorates.

      A more extreme example is Malia Cohen in SF District 10. At the end of the Ranked Choice algorithm, she won, with 52% of the votes in the final round. But because there were 22 rounds, Cohen actually only won about 24% of the valid votes cast.

  5. Terry Christian

    Nobody is required to vote for City Council (or any office), so you cant necessarily clump nonvoters into the “RCV is confusing” bucket, can you?

    I was a Lemley supporter and did talk to one voter who refused to vote in any RCV election, but that was because it produced Jean Quan. Ask any candidate/volunteer and I think you’ll find people dont like RCV because of that, not because of the system itself (although granted its hard to distinguish the two)

    RCV voters should always mark all three choices, saving their third for a nose-holder between who they likely think the last two candidates will be (Len Raphael even did this). Most voters appeared to participate in the Lemley-Kalb run off, but those who exhausted or undervoted were by definition unlike the rest of the electorate. Their voices should have been heard.

    Id bet dollars to donuts Lemley would have beat Kalb in a straight head-to-head–the guy’s platform spoke not at all to North Oakland issues and he relied heavily on non Oakland money and volunteers from other organizations (as opposed to people who actually liked him). But he did win, which shows the power of the Democratic endorsement in a Presidential year and perhaps the impact of free publicity for getting mugged.

    RCV’s never going to get changed, though. This video is very helpful in sorting it out, particularly among multiple candidates. Most videos simplify to only three candidates, but that doesnt tell the whole story.

  6. Allan Feinstein

    As best I can tell, we were fortunate to have excellent candidates, and I am happy with the winner. However, I still prefer the old runoff method.

    With runoff we would have had debates between Lemley and Kalb, rather than large debates, which really illuminated little. Candidates did not have the opportunity to go back and fourth on an issue.

    • Rob Richie

      Allan — The great majority of elections in Oakland’s old system didn’t got to runoffs and were won in low turnout elections in June. For instance, Ron Dellums won in a multi-candidate race in June, garnering far fewer votes than Quan did in 2010.

      See this factsheet:
      http://www.fairvote.org/assets/OaklandJunePrimariesFactSheet-1.pdf

      • Max Allstadt

        Ok, so you are a “FairVote” advocate.

        Answer me these:

        1: Have you ever tried to survey voters during a ranked choice election to see how many of them can name all the candidates? What about how many can tell you anything about every candidate’s platform and qualifications? Did anybody try that in SF District 10 in 2010?

        Voter confusion as FairVote tries to define it means mismarked ballots. In reality, the voter confusion problem is too many candidates to keep track of.

        2: Have you ever compared a November primary, December runoff structure to RCV and also to a June-November primary/runoff structure? A November primary solves your turnout problem.

        3: Have you ever compared the amount of vetting that the media and watchdogs are able to do in a one-on-one election or runoff to the amount they can do in RCV? In Oakland we had 31 candidates on the ballot, and that made it impossible for the press to do background checks on all of them.

        I want direct answers to these, because every time I bring them up with RCV advocates, what I get is evasion and changing the subject.

        Also, Rob, please disclose if you are a paid advocate for RCV.

        • Rob Richie

          Max. Briefly.

          1. I’m director of FairVote. No secrets there. You can see the range of things we do on a range of election analysis and reform ideas at fairvote.org. But support for RCV in Oakland is grounded in Oakland, which helps explain its support on the city council and by key civic players like the League of Women Voters.

          2. As a general matter, it’s easy to criticize ANY system as less than ideal. But one has to keep contrasting it with alternatives. I believe, for example, that you seer traditional runoffs through a rosy halo that overlooks big downsides,.

          For example, as linked in my earlier comment, most Oakland races never went to a runoff, but were won in low turnout primaries; And going to a December runoff is going to an even bigger problem — the number of “exhausted voters” who wouldn’t have returned to a December runoff this yer would have dwarfed the “exhausted ballots” that seem to trouble you — and led to a less representative electorate as well.

          Furthermore, two-rounds of voting (having to survive a low turnout primary, then a one-on-one runoff) takes a lot more campaign money to raise — especially as runoffs easily devolve into a virtual shouting match punctuated by personal attacks.. In contrast, RCV gives less well-funded candidates a better chance by rewarding grassroots campaigning.

          3. I don’t have a lot of patience with arguments based on bad media and bad voter education. If the media does a bad job covering candidates in the decisive general election, they surely are going to do a MUCH worse job in a primary that will eliminate all but two candidates, and often elect one of them. Limiting the field to two is a big deal, and can effectively determine outcomes.

          For me, a better answer is to keep working on ways to get information to voters — good voter guides, media coverage, more creative debates and so on..

          In a related observation, a common pattern in RCV elections is a decline in undervotes — that is, more candidates at the polls for a national or state race decide to vote in a city’s RCV race. That suggests to me that RCV elections are engaging more voters — resulting in higher turnout than old rules and an electorate more aware of city issues. That’s a good thing.

          As a final word (and I suspect it will be), RCV may not be perfect, but nothing is. I believe it’s clearly working better than Oakland’s old system and is better than other alternatives that rely on a vote-for-one-only system

          • Max Allstadt

            You didn’t answer my questions on the following issues:

            1. Have their been surveys to assess voter knowledge and/or ignorance in large ranked choice races vs. one on one races: no answer from you.

            2. I asked for data on December runoffs. you gave me conjecture.

            3. “Who cares if the media can’t keep up?” is not a valid counterargument. To carefully check on a candidate, you need to search county records, city records, bankruptcy records, state elections commission records. When there are 31 candidates in a cycle and there’s no second stage, a small city like Oakland’s press can’t keep up.

            Now,

            A couple of other questions:

            1. Why did Aspen, CO, and Pierce County, WA adopt and then ditch ranked choice voting?

            2. Why are ranked choice results expressed only as the percentages of the final round votes, and not, as I showed above, more accurately, with exhausted ballots included?

          • Max Allstadt

            Two more questions:

            1. In San Francisco District 10 in 2010, The MAJORITY of the vote was exhausted ballots? Is that OK?

            2. In January, when the new Oakland City Council is sworn in, 4 out of 11 of Oakland’s elected officials will have won elections in which the majority of the voters did not vote for them. Is that OK?

  7. Chris J.

    Great job with the video! Very professional and really well done.

  8. Apart from the important discussion as to whether RCV increases the accuracy of an election to what the voters really wanted, there is the question as to what see is the negative impact of RCV on voter education and civic participation in local politics.

    RCV is a neat easy way for voters to even less time thinking about local politicians and issues.

    I’m not sure that traditional run-off elections are the cost benefit answer, but we need to do something to increase voter awareness and understanding of local government issues both so they can elect competent officials and so that politicians can make count on the support of residents for painful decisions.

    • Terry Christian

      Len’s right here. One solution, of course, is term limits. One thing that seemed readily apparent from the D1 result is that people had really switched off any truly critical and informed consideration of local politics given 16 years of one-Councilmember rule. It really was just “Democratic Party endorsement? Um, OK, sounds good to me.” If people actually participated in elections more often they would, y’know, participate in elections more often.

      And its two terms, not three. If its good enough for the President, its good enough for an Oakland City Councilmember.

    • Rob Richie

      Len — that analysis of the impact of RCV just doesn’t hold up. RCV winners are earning FAR more votes on average than pre-RCV races, and voters tend to be LESS likely to undervote (skip the race) than in RCV races in Oakland than in pre-RCV days.

      There’s a lot we need to do everywhere to engage more people in local elections, but blaming RCV is not a sensible place to start. See a lot of things to look for at another FairVote project, http://www.promoteourvote.com

  9. John Adler

    Here’s another very informative video about Ranked Choice Voting – an example of a Mayoral Election. But the “winner” looses?

    http://youtu.be/5ZGTnp3cgFY

  10. David Cohen

    Good video. Makes eminent sense. The fervor of the anti-RCV’ers reminds me of birthers. They’ll never be convinced. By the way, Mr. Allstadt, what do you think about Obama’s birth certificate?

Comments are closed.