Home » Columns, Headline, Island Talkback

Island Talkback: For your consideration … ranked choice voting

Submitted by on 1, October 22, 2010 – 4:50 am11 Comments

Photo from Wikimedia Commons

Editor’s note: The Island has a new comment policy. Check it out here.

By Mark Irons

In a single-seat race with a large field of candidates like our current mayor’s race, it is possible to have a candidate elected by an extreme minority of voters. With four viable candidates, a winner could be declared with just 26 percent of the vote, leaving nearly three-quarters of voters without representation.

The plurality ballot employed for City of Alameda elections happens to be the crudest ballot method available. Plurality sidesteps the expensive and lengthy runoff process, but the outcomes are less democratic.

The very simple solution to the problems of plurality and majority runoff is a Ranked Choice Voting (RCV) system, and specifically, Instant Runoff Voting (IRV). These ranked voting methods are not new and have been successfully employed in places like Australia for decades. The Alameda County Registrar of Voters recently adopted this ballot method, which means the county’s voting machines have the capacity to accommodate RCV. The basic IRV will be employed for the first time in Oakland, Berkeley and San Leandro next month.

By ranking candidate choices in order of preference, if none of the candidates receives a clear majority on the first ballot count, the ranked choices are employed to render a winning candidate with a majority mandate. The mechanism being employed is very simple and the process logical, but to many the concept may seem counter intuitive. The County Registrar has a number of video tutorials available on their website. I discovered these videos on the Oakland blog A Better Oakland, where blogger V. Smoothe gives a comprehensive explanation in addition to other links.

Our presidential races are also by plurality, which in the recent past has allowed both Bill Clinton and George W. Bush to be elected without a clear majority vote. The electoral college adds another layer of convolution to that race, but in simplest terms, if IRV had been employed in the 2000 presidential race in Florida, without a 50 percent plus one majority the first-choice votes for Nader would have been discarded as non-viable and the second choice of Nader voters would have been distributed to the two remaining viable candidates, Bush and Gore. The result would have been an “instant” majority and the protracted and controversial Supreme Court ruling avoided.

Filling multiple seats simultaneously, as in our two-seat council race, presents a slightly more complex equation, but RCV again addresses the issue rather neatly. A preferred ranking system in this type of race is called Single Transferable Vote. In this case one’s intuitive resistance is the desire to fill two seats without giving one candidate preference over another, which STV requires, as with all ranking methods.

STV defies simple explanation, but it is a very compelling choice, particularly in races such as this November’s when not just one, but two seated councilpersons vying for mayor present the possibility that a third council seat will need to be filled if either candidate succeeds in being elected mayor.

No matter which candidates are seated in the upcoming election, we should prevail on them to implement more democratic ranked choice voting systems.

Got a burning issue? Then talk back! Opinion pieces can be sent for consideration to michele@theislandofalameda.com.


  • Richard Bangert says:

    Thanks for putting the issue before the public at a timely moment. Perhaps with the memory of the November 2nd election fresh in our minds, an election task force could be convened to explore how Alameda’s voting system might be improved for the future.

    The hardest problem to deal with would be the fact that we do not have district elections, and I hope we keep it that way. But along with that comes the problem of ranking candidates for two open seats. In Oakland, for example, they have district elections and the ranking system is simple. But as the author points out, the system for ranking multiple candidates, the single transferable vote, “defies simple explanation.”

    After reading the link several times, I would concur.

    I would also like make a correction regarding the example of the 2000 presidential election regarding instant runoff voting. If instant runoff voting had been in place, the 2000 presidential election would have been settled by voting, rather than the Supreme Court, long before the elimination process got to Ralph Nader.

    The elimination process starts by instantly eliminating the lowest vote getter and re-tallying. I find it hard to believe that the election would not have been settled in Gore’s favor by the time the thousands of votes of the Socialist Party, Socialist Workers Party, the Workers World Party, and the Reform Party had been eliminated and re-assigned. Gore only lost by a little under 600 votes.

    It is not inaccurate to say that if Nader had dropped out, Gore would have won. But it is much more important to note that had just 300 of the Florida Democrats who voted for Bush switched their votes, Gore would have had the necessary 600 votes to edge out Bush. One million registered Democrats in Florida voted for Bush. Another million didn’t bother to vote at all. The Democratic Party, and its candidate who couldn’t carry his own state (which would have made Florida votes irrelevant), bares the brunt of the responsibility for the outcome of that election.

  • Dan Wood says:

    This is a great time to start thinking and talking about this issue. Now that Alameda is surrounded by cities that have RCV — and they have taken the “early adopter” hit, it makes sense for Alameda to pursue this for our subsequent elections.

    Chances are very good that there will be 3 city council seats filled come November, since it is likely that whoever wins as mayor will vacate his/her council seat. The current rule is that if the third-place vote getter has gotten at least 10% of the vote, he or she will be appointed to fill that position.

    It would be nice to have an RCV system in place so that we could all express our choices to fill the two or three seats available. The way it is now, I can vote for the two openings, but I’m leaving it to other people’s votes as far as how that third seat will be filled. That’s kind of disturbing.

    Does anybody know what the procedure would be to move Alameda to a RCV system? Does it involve changing the city charter? A ballot initiative? Council?

  • John E. Palmer says:

    Excellent piece, Mark. I believe that the success of IRV/RCV in city and county elections will give it credibility to be used in state and federal elections at some pint in the future. Even if we keep the electoral college (pretty darn likely), IRV could be used in each state to remove the spoiler effect in determining a winner of its electoral votes. I appreciate your handling of STV in the piece as well. It’s a great method, but complicated to explain–we have to walk before we can run. JP

  • Mark Irons says:

    I have watched a 13 minute power point on DVD which is about the most simplistic explanation of STV imaginable. It convinced me that it doesn’t matter that one doesn’t cast two votes with exactly equal weight in a two seat election (or even with three seats). “It all comes out in the wash”, as they say. The power point also asserts that the de facto outcome of STV is to increase the likelihood of minority constituencies having representation. Again, it’s too complex for me to elaborate here. It further claimed that the result lends to an effect similar to district elections where various pockets end up with better representation.

    I think Alameda is too small for district elections, but people often argue that various constituencies like the West End are inadequately represented. STV would appear to address multiple inequities all at once.

    I’m fairly certain San Leandro, with population of about 85,000, uses district at large election where each district has a set of candidates but everybody in the city gets to cast a vote on each seat. Imagine how that would go over in Alameda!

    There are many anecdotes about people being confused by IRV ballots with results not unlike the “butterfly ballot” scenario in Miami, and though it is a serious consideration to ponder, I think the benefits ultimately out weight these problems. Switching from Outlook to Gmail was initially aggravating, but I’ve adapted.

    The current Oakland ballot repeats the entire list of candidates three times and voters fill it out consecutive ballots three times. This being the first IRV ballot, I think the Registrar wanted to maintain some graphic familiarity. However, imagine a ballot which lists the candidates only once vertically but following the names to the right there would be a series of vertical columns with numbered bubbles similar to multiple choices tests with which we’re all familiar. One would simply fill in the bubbles corresponding to the candidates according to rank, 1, 2, 3,etc..

  • anne spanier says:

    Hooray! After working on this issue for 6 years and finally seeing it realized in Oakland, I am anxiously awaiting the outcome. Most of the politicians (not all) were opposed to it. The Oakland Tribune, which supported out efforts has been busy wriing about the role ranked choice will play in the current mayor’s race. And it appears that this election isn’t going to be a slam dunk to the biggest spender. It also make money less important in this race.Yep, I am all for Alameda taking this on.

  • anne spanier says:

    I mean, the biggest spender is not the shoo-in nor is the one with the biggest name recognition. There are at least 4 candidates with a chance in this race. And no one is a spoiler.

  • Mark Irons says:

    Richard, “The author” does not disagree with any of your observations about the 2000 presidential elections, but he doesn’t see any of it as a correction, rather just an elaboration. Greens may bristle at Nader being labeled a “spoiler”, but that is the popular perception. Ross Perot played a similar roll. My point had nothing to do with assigning blame, only that ranked choice ameliorates the spoiler issue and allows third parties a foot hold without forcing them to be relegated to spoilers status or otherwise marginalizing them as “not serious”. In doing so, RCV also opens the door for moving toward proportional representation and breaking the strangle hold of the “duopoly”

  • RCV does not eliminate the spoiler issue; it just kicks it down the road a ways. Consider an election, A vs. B, and B is wining 55% to 45%; but then we add a third candidate, C, who is preferred by most (but not all) B voters:

    45%: A > B > C
    10%: B > A > C
    15%: B > C > A
    30%: C > B > A

    Under RCV, who wins? It’s not B, and it’s not C: it’s A. C’s presence spoiled the election of B. RCV has spoilers.

    To be a spoiler, the 3rd parties need to get rather more votes than they do now, but since it’s being branded a spoiler that kills a party, we can expect that they WILL grow… until they are large enough to, once again, spoil elections. And spoilers means that two-party domination will continue; just as it has in Australia.

    There ARE voting methods which *actually* don’t have spoilers: approval voting is one of them. It’s also cheaper to implement and easier to understand than RCV: Rather than vote for one, vote for any number you like; most votes wins. That’s it.

    If you like RCV because you think it will eliminate spoilers, or you think it will help third parties, then you’re backing the wrong reform. Back approval voting.

  • Jon Spangler says:

    Thanks for this timely piece, Mark. I hope our memories of November 2, 2010 will not be too bitter after we get the results….

  • Mark Irons says:


    Your post does just enough to cast a shadow on RCV without really shedding serious real light, because even your A,B,C, example is difficult to follow.

    My understanding of IRV is that on the second ballot it isn’t all the second choices which are counted, it is only the second choice from voters whose first choice candidate is not viable on the initial count. Does your example above reflect that?

    I’m not invested in being the smartest guy in the room, just advocating that we improve on plurality at a minimum and do anything we can to break two party domination. If you can explain “approval voting” as superior so that we can all understand it, by all means go for it. All in all, it’s hard to imagine a cruder system than plurality and I can’t see how RCV is not an improvement.

    The current East Bay Express describes accounts of candidates collaborating in strategies to unseat incumbents. I’m not even certain if that constitutes gaming the system.

    Was it Rumsfeld who said “democracy is messy”?

  • Mark,

    Yes, your understanding of how the IRV process works is correct; ballots are initially assigned to the candidate listed as the 1st preference, then in each step the candidate with the fewest ballots assigned to them is eliminated and all of their ballots are reassigned to the next (non-eliminated) candidate listed on the ballot.

    My example takes that into account, and in fact relies on it. The first-preference votes go 45 for A, 30 for C, and 25 for B. So B is eliminated, and when B’s ballots are redistributed (10% to A and 15% to C), A is the winner, 55:45. Note that B is preferred 55:45 over A, and 70:30 over C, so B would win any one-on-one contest. IRV fails in this instance because the “beats all” winner is eliminated instantly, leaving the run off to the two lesser candidates.

    The C voters can get a better outcome–their second favorite, B, rather than their least favorite, A–by some of them choosing to disingenuously rank B above C. In other words, by voting for the lesser of two evils; and it’s the fact that this strategy works which keeps IRV, like plurality, stuck with a two party system.

    The short-form for why approval is better is as I’ve already said: that it ACTUALLY has no spoilers. You can always vote for your true favorite.

    The long-form is a bit more involved. I strongly recommend William Poundstone’s “Gaming the Vote”, which covers the subject delightfully. Alternately, this image is taken from the book’s conclusion. You’ll note that even though you find it “hard to believe”, in the worst case, IRV is precisely zero improvement over plurality.

    Finally, my own blog on this subject is The Least of All Evils, where I go into a lot more detail on the various nuances involved in this seemingly-pedestrian but secretly-fascinating subject.

Leave a comment!

Add your comment below, or trackback from your own site. You can also subscribe to these comments via RSS.

Be nice. Keep it clean. Stay on topic. No spam.

You can use these tags:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

This is a Gravatar-enabled weblog. To get your own globally-recognized-avatar, please register at Gravatar.