What Can Portland Learn from America’s Oldest Proportional Election System?

Thanks to the charter change ballot measure voters approved in November 2022, Portlanders will see a few changes to the city’s election method and governance in the coming years. One of the first will be the November 2024 election, when voters will choose 12 councilors in 4 districts using proportional ranked choice voting.

Under the new system, Portlanders will be able to rank multiple candidates on their ballot, bubbling in their first choice, second choice, and so on.¹ Instead of a single candidate winning one seat with more than 50 percent of the vote, 3 candidates in each district will win seats with more than 25 percent of the vote each.² Ballots are counted in rounds until these 3 candidates come out on top. Losing candidates in each round are eliminated, and their votes are transferred to the voters’ next-choice candidates. Candidates who pass that 25 percent threshold during counting will win a seat, and any votes above that threshold will be transferred to the voters’ next-choice candidates.³

To get some ideas about what voters can expect from the new system, I talked to Cambridge, Massachusetts, City Councilor Burhan Azeem. Azeem is serving his first term on the council and is the youngest councilor in the history of Cambridge, home to the United States’ longest-running proportional election system. Since 1941, Cambridge voters have selected nine citywide councilors using proportional ranked choice voting. While Portland’s experience won’t look exactly the same as Cambridge’s, Rose City residents can learn a lot from the practices and results they’ve had over the last 80 years. Here are excerpts, edited for brevity and clarity, from my conversation with Councilor Azeem.

Why did you run for office?

There’s all sorts of political causes and stuff I really care a lot about: housing and transportation policy, as well as other things. But my first real call was showing up to City Hall and thinking, “This is not the city I see right outside the doors.” Cambridge is a very young city, even without counting kids.⁴ And especially when I first got involved in politics, everyone was fairly old. Besides our mayor, all my other colleagues currently on the city council are between 50 and 80. They’re all great people, but it was just very different than the city that was represented. Some of my main issues as a candidate were topics related to the universities we have in town, and I ended up getting lots of votes in areas where lots of students live. So I think the message you generate resonates with people.

How do Cambridge voters understand the proportional ranked choice system? Do you find yourself doing lots of voter education?

As a baseline, we have a good political culture, there’s really good explainers and other materials, Cambridge residents are extremely well educated, and enough people get it to explain it to other people who don’t. The government and council actually do a lot of voter education on top of that, because the median length of residency of a voter in Cambridge is so low that for a lot of people it’s their first time doing ranked choice voting.

One trade-off of our system of doing everything at-large is that we have almost 20 candidates run, and getting voters to understand the comprehensive differences between 20 candidates is really, really difficult. Cambridge voters tend to have a good understanding of who’s likely to win a seat and who the marginal candidates are that don’t really have a shot, but even the short list can still be a lot of candidates to sort through. So what the benefit might be of Portland having smaller districts is that you might have more like six candidates running. Six is a much more understandable number for most people; they’ll be able to get a sense of what’s going on in each district.

How do voters decide between all those candidates?

In Cambridge we tend to see slates, groups of candidates you’re aligned with based on your issues. The overlap between slates is not one-to-one, so there’s some people who I might be with on a pro-housing slate but aren’t part of my pro-bikes slate. Most of these slates are lists of candidates endorsed by advocacy groups, rather than a group of candidates explicitly allying themselves. It’s nonpartisan, it’s not Democrats and Republicans, but we have a pro-bike lobby that will independently say, “If you want to support bikes, go support these candidates and rank them however you want.” Voters tend to get a lot of voter guides from different advocacy groups.

There are some candidate-selected slates, it’s just that they’re less common because Massachusetts campaign finance law puts strict limitations on us. I was on a candidate slate where it was just three candidates, and we chose to work together and then made it clear to the voters that we were allied. The slate was two incumbents plus me, and one of them lost re-election while I won a seat. So the overall number of councilors from the slate stayed the same, even though who those specific councilors are changed.

How do these candidate slates impact campaign dynamics?

Here Councilor Azeem talks about the “surplus vote transfer” that happens during ballot counting for races using proportional ranked choice voting. If a candidate finishes a round of counting with more votes than the threshold needed for election, the votes beyond that threshold are transferred to voters’ later choices (similar to if a voter’s earlier choice was eliminated in a round). This surplus transfer helps create the proportional effect of the system, letting a big enough voting bloc elect multiple preferred candidates regardless of how individual voters ordered those candidates.

Those different overlaps between slates actually give you an incentive to be pretty nice to everyone, to get along and get things done. If we didn’t have ranked choice voting, if your goal was just to be the number-one vote-getter or be in the top nine, you’re worried about the person who’s closest to you in political positions just stealing your votes. But with proportional ranked choice voting, if you have even a slightly different position, it actually greatly reduces the stress level. I can afford to get along with someone who has the same positions and views as me, because we’re not directly competing. If there’s enough voters out there, it doesn’t matter if 60 percent of that bloc vote for him and 40 percent vote for me, because it’ll get rebalanced out during the surplus vote transfer.

If I have a campaign kickoff event, people who are nominally my competitors will show up because they might get a number-two or -three vote from my supporters. And when you’re in a debate, you don’t necessarily want to go after any particular person—well, maybe if there’s someone who’s on the complete opposite side of the ideological spectrum from you. But debates are much more constructive and moderated; people are gentle with their differences.

Do you think campaigning in Cambridge’s system is more civil than in other cities?

The way it plays out is that you want to keep other candidates close, but not too close; you want to be seen as someone who if a nearby candidate loses, their voters feel comfortable coming to you as their second choice. So if I’m pro-housing and pro-transit, the people who support candidates that are only pro-housing or only pro-transit have an option whether to support me or an anti-housing, anti-transit candidate with their second choice. So you’re in direct competition, to some extent, with the candidate who’s exactly opposite you. With those other two candidates, you want to try to be close to them so that their voters will give you a number-two vote, but you don’t want to be so close to one that the other’s voters won’t put you number two.

And that’s just the ranked choice voting part, but having multiple winners plays into this, too. Take the mayor, who is also pro-housing and pro-bikes. She’s extremely popular, she always gets 20 percent of the number one votes. So since those votes over the quota (10 percent) will get redistributed, everyone actually tries to be very close to her because they know that she’s going to win.⁵ So if they know that you’re close to her, if they’ve seen you at events together, her voters will rank you number two. So you definitely want number-two votes from other candidates, but you also need number-one votes of your own, a distinct identity and a base of people who will vote for you no matter what goes on.

How do these candidate slates impact governing?

“The slates help you naturally get a sense of what issues we can move on, because people campaign on them, and we’re pretty clear on where they stand.”

Once we’re elected, slates help align the council’s policy ideas. Because we have two-year terms and new people join every term, it can be hard to figure out where you have a supermajority and where you don’t. The slates help you naturally get a sense of what issues we can move on, because people campaign on them, and we’re pretty clear on where they stand.

I think that we function mostly like a normal government. We’re fairly clear and transparent as far as governments go, and Cambridge voters hold us to a pretty high bar. The slates kind of say, “Here are the issues that we expect movement on, because these issues got a majority elected.” That tends to be fairly responsive; we’ve moved a lot on transit issues so far, and hopefully in the second half of the year we’ll move pretty aggressively on some housing issues.

One thing that changes with slates and ranked choice voting is that councilors can have a hard time with deal-making, giving up something of lesser value to get your main issues done, which I think is mostly good. People like that after the election they have a fairly good sense of how the next two years will turn out. Say my number-one issue is pro-housing and my number-two issue is pro-bikes, and there’s another councilor whose number-one issue is against bikes and number-two is against housing. So if they’re willing to give me the vote I need to get my pro-housing issue done, I might give them a vote on their anti-bike issue. That ends up being not ideal for me to do in a ranked choice voting system, because my number-one voters are people who are pro-housing and pro-bike. So if I end up being pro-housing but not pro-bike, that’s getting into a different lane, and those voters will find another candidate to support.

Who do you consider to be your constituency? Is it the entire city, the ten percent of voters who voted for you, or the communities and issue groups you’ve focused on?

It’s very clear to me who my base of core voters is, the people who I will get every time regardless of what happens. For me, this is people who are pro-housing and pro-transit, and people who tend to be affiliated with the universities. If they have something come up, I’m 100 percent on top of it. I’m leading the charge to make sure we get those issues resolved as quickly and effectively as we can.

But the vast majority of voters barely pay attention to local politics, and they like a lot of candidates, because with ranked choice voting candidates have an incentive to get along with everybody else. So if somebody comes to me and has an issue around gun violence, which is a signature topic for one of the other councilors, I won’t ignore that person and say, “Just go to the other councilor.” I’ll try to help them as much as I can, because it helps me to become their number-two vote; it’s hard to win with just number-ones.

But there’s also a chance I could become their number-one vote, and they put that other councilor number two; there’s a lot of fluidity. Cambridge has 22,000 people voting in every municipal election, which means 2,000 ballots helped get me elected. But there’s 8,000 people who put me on their ballots at all, maybe number two or three or four or later. If they’re impressed enough with me, they might move me to a number one.

So you do see that mobility historically: if somebody has a really good term, they could do a lot better than before, and if they ended up having a worse one, they go down in the rankings. So it forces you to be on top of your game even with people who are not your core base. If somebody voted for Burhan number two, they still consider themselves to be a Burhan voter. And most people even forget their exact ordering afterwards, so it’s very possible for them to vote for me number one in the next election.

So in between elections, how do you interact with constituents?

I’d put them into some different levels. Say 22,000 people vote in a municipal election. Fifteen thousand of them you never see after election day. They don’t really talk to city council, they don’t care what’s going on in City Hall, they did their job by voting. So for those people, you do a campaign event from time to time, go out door-knocking to say hi, show up to block parties and neighborhood events, give people a sense of who you are.

Then you might have 5,000 people who kind of pay attention, who are lightly interested. They vaguely know what’s going on, they read the newspaper, maybe they’re subscribed to your newsletter, maybe they’re part of a neighborhood group, they talk to you when they see you around in the neighborhood. They’ll also stay out of politics, unless there’s something that very deeply touches them or something that’s a hot-button issue. I think that those people end up being the group leaders for the first 15,000, where most of those 15,000 will go to one of the 5,000 that they know to give them their first level of advice on the election, then mix that with their own information.

Then you have maybe 2,000 people who are super involved, who will write to city council and comment on stuff, show up to meetings, follow you on Twitter, and really pay attention. Those people end up being the people who you most interact with in terms of writing policy or getting people to show up to protests or rallies.

And this might be the negative part of politics, but those different groups have different impacts on getting policy made. If I’m working on something that I think is extremely popular, but no one or even if only five people are going to show up to City Hall to speak in support of it, no one’s going to vote for it. I think that you don’t get anything done in City Hall unless you can have 100 people email or call in or show up to community meetings. So the reason we’ve had such transformative change in bicycle infrastructure is because we’ve had maybe 50 people who are dedicated, who will show up to a weekly community meeting every week for five, six years. And it sucks that you can’t just vote in an election for bike stuff and then it’ll appear, but it’s the system we have.

Since voters don’t have a single councilor they can claim as theirs, how do constituent services shake out?

Constituent services are casework that legislative offices take on for their residents outside of the legislative body, particularly interfacing with other parts of the government. Examples include accelerating the repair of a specific street, personally advocating for a property tax appeal, checking on a building permit that should have been issued sooner, or anything else a resident might call their legislator to get help with.

We do get constituent services; it just happens a little bit differently. Right now, we’re talking about the Massachusetts Institute of Technology (MIT), the biggest landowner in Cambridge. They’ve had an open campus, you can go through any of their buildings, and a lot of Cambridge residents use MIT’s library. And right now, they’re continuing shutdowns and not letting the public use it, which is a big constituent issue.⁶ But everyone knows to just go to Burhan for that, so people reached out to me, and I was able to reach out to MIT. There was an issue around homelessness in Central Square, and there’s one city councilor whose brand is homelessness issues, so people knew to reach out to that councilor specifically. So councilors have pretty distinct identities, and for the vast majority of cases constituents know who to go to.

How does Cambridge’s nine-seat proportional system affect incumbents’ chances at reelection?

I think Somerville is a really great city to compare us to. They have seven ward seats and four at-large city councilors as well.⁷ They never get much competition for any of their ward councilors. And then at large, they get some competitors, but I can’t remember the last election where an incumbent lost. Cambridge is very different. We have nine at-large city councilors, and typically at least one incumbent loses every election cycle. So it’s very dramatic turnover. I will say as an elected official, it’s very stressful because you can lose every election.⁸

But even though city councilors change over time, it’s very proportional in that you will always have seven out of nine city councilors who are pro-other forms of mobility, like transit and bikes and stuff. That has been fairly consistent for a long time. And similarly, you see a fairly stable pro-housing majority, which has even increased over time. I think it’s because ranked choice voting to some extent negates the spoiler effect, so even with so many candidates to choose from, voters still get good representation at the end of the day.

Does Cambridge get a broad geographic dispersal of councilors with its citywide elections?

Absolutely. I think this is actually the best part about proportional ranked choice voting. We have had city councilors who were basically just representing East Cambridge, which is a very heavy immigrant, Portuguese-speaking community, and they would win solely by getting lots of votes in that district and not really competing anywhere else.⁹ So kind of like a normal district system. And then you look at the Black community in Cambridge, which I think is very strong and integrated, and you have at least one Black candidate win every election cycle by getting votes spread throughout the city.¹⁰ Sometimes you get pockets in other ways; I have strong support from students so wherever there’s a university, there’s clusters of my voters around there.¹¹

“The most important identities to people are really what people end up voting around, whether that’s geographic or race-based or age or other sorts of occupationally based things.”

So you see geographic clustering around different places, not necessarily just uniformly throughout the city or in a specific ward and district. I think that allows for really interesting combinations; the most important identities to people are really what people end up voting around, whether that’s geographic or race-based or age or other sorts of occupationally based things.

People imagine that the number one identity is geographic, that distribution really matters to people, and if you have an at-large proportional ranked choice voting system you won’t get much geographic distribution. I think that’s just false. If people care about it, it still happens. For example, a decent portion of the city is affiliated with the universities, which play a huge role in Cambridge. But there’s no district where those voters make up a majority, so you wouldn’t get that voice on council with just single-member districts. Even having just one person speaking to that issue now really helps. I think geography can be what people go to because it’s easy, but especially in a highly mobile city like Cambridge, geography isn’t the number one issue. I’ve lived in Cambridge for almost a decade now, and I’ve moved every year. My neighborhood wasn’t the most important identity that I carried to the council.

See the full interview on YouTube here.