Practical post on 4 #ProRep methods (Proportional Representation)

post by Jameson Quinn (jameson-quinn) · 2021-01-09T14:56:29.431Z · LW · GW · 12 comments


      Note: there is a short summary of this article at the end.
    and Cons of #ProRep in General
  Four #ProRep methods
      Interlude: Allocated Score
    (in 3-5 seat districts)
    (Modified-Bavarian style)
    ("Too Long, Didn't Read"; Summary)


I have written a series of longer posts about voting theory: a general primer [LW · GW], single-winner pathologies [LW · GW], multi-winner method building blocks [LW · GW], and a longer, harder-to-read piece where I struggled through to a metric for multi-winner methods [LW · GW].

I believe that I should write a simpler, (relatively) shorter piece on proportional multi-winner methods. This is partly because my own thinking has evolved and sharpened, and partly because I think that current events in the US, Canada, and UK make these ideas more timely:

In all three of these cases, the situation favors grassroots calls for reform, but it seems to me that a substantial number of existing political incumbents will have to be on board for this reform to succeed. This means that reform that is especially hostile to such incumbents would probably be vetoed.

It is utterly clear to me that election method reform, based on some form of proportional representation, could be a key step towards healthier politics in all three places. It would not be a panacea; politics is, and always will be, a battleground of competing interests, and whatever your sympathies are, I suspect you see some of those interests as less valid than others. But it would help.

This article only mentions the best methods. As a co-organizer of the 2018 British Columbia Symposium on Proportional Representation; a committee-member of the Equal.Vote committee on proportional representation methods; and the primary designer of the proportional voting method used for Hugo Award nominations; I could explain dozens of different (and interesting) #ProRep methods if I wanted. But this is focusing on those I think are worth knowing from a practical standpoint.

Note: there is a short summary of this article at the end.

Pros and Cons of #ProRep in General

Any form of proportional representation would have several huge advantages over the current "first past the post" (FPTP, aka single-winner plurality) method used to elect the lower house of the legislature in all three countries (US House of Representatives, Canadian and UK National Parliaments). These advantages include:

Note that while some of these are particularly desirable from certain partisan standpoints, at least the first three are desirable on meta-ideological grounds. In other words, I believe there's reason to want this kind of reform almost no matter where you are on the ideological spectrum.

This is not to say that there are no potential downsides to proportional representation. For instance, some proportional methods, such as Israel's closed-list one, seem to lead to excessive party fragmentation, which in my opinion can result in myopic single-issue parties getting disproportionate power over their issue of interest. But even at its worst, I think this downside is small relative to the upsides above; and well-designed voting methods can minimize this downside even further.

Four #ProRep methods

Here are the four reasonably-good proportional methods that I think it's worth knowing about. For each, I'll briefly explain how it works, then mention a few pros and cons relative to the other three. Note again: the "downsides" are only relative to other methods here; all these methods are better than the current FPTP in almost every way. All of them share all the advantages listed above, and I've left out #ProRep methods like Israel's which are strictly worse than the other options. The cons I will list are only relative to the other options listed here, not relative to the status quo.


This base method is called "Single Transferable Vote", and has also been rebranded in some cases as "multi-seat Ranked Choice Voting". (It's not the same as the single-winner version of RCV, though; single-winner reform is out of the scope of this article.) At the bottom of this section, I'll quickly sketch a proposed improvement on this method, known as Sequential Monroe, which is slightly better but shares most of the primary pros and cons. But for now, here's how the basic STV works:

  1. The state/province/country is divided up into subregions, each with enough population to elect 3-5 legislators. (Annoyingly, there are different names for these regions: "districts" in the US, "ridings" in Canada, and "constituencies" in the UK. I'll use "districts" from here on. The number of seats in a district is known as its "magnitude" M.)
    • Note that states/provinces with fewer than 3 seats total cannot really use this or other proportional methods.
  2. Within each district, voters rank the local candidates in order of preference.
  3. Votes are tallied for their top choice.
  4. Any candidate with over a "quota" — 1/(M+1) of all the district votes — is seated. If so, a quota of their votes are "used up", and the rest of them are transferred to the next candidate in that ballot's preference order.
  5. The lowest-tallying candidate is eliminated, and those ballots are transferred.
  6. The above two steps (4 & 5) are repeated until all seats have been filled.

Here are the upsides of this method:

Here are the downsides:

Interlude: Allocated Score

Allocated Score is an STV-like method which uses 0-5 star ballots (that is, voters can rate candidates independently on a 0-5 star scale, with tied and skipped ratings allowed), and without the need for bottom-up eliminations. This would remove the first downside above and slightly mitigate the complexity of voting; but it would also remove the last upside (familiarity).

OLPR (in 3-5 seat districts)

This is known as Open List Proportional Representation. Here's a blog post proposing it, by Jack Santucci, an expert in the US history of voting reform. Here's how it works:

  1. As above, the state/province/country is divided up into districts (see above for discussion of terminology), each with enough population to elect 3-5 legislators. The number of seats in a district is known as its "magnitude" M.
  2. Voters each choose one candidate from their district. Each candidate is either associated with a party, or independent. (The risk of wasted or inefficient votes is slightly higher when voting for an independent candidate.)
  3. Parties are assigned a number of seats based on the number of "quotas" of votes for all their candidates combined. As above, one quota is 1/(M+1) of the total votes in the district. Each independent candidates are treated as being in a party by themself.
    1. In case this leaves seats un-assigned, a least-remainder method is used to assign those.
  4. A party's seats go to the candidates in that party in descending order of votes.

Here are the upsides:

Here are the downsides:

MMP (Modified-Bavarian style)

This is known as Mixed Member Proportional. In particular, the method I'll explain is based on the open-list MMP method used in the German region of Bavaria, with some STV-inspired modifications to further equalize voting power. Here's how it works:

  1. Each state/province/country is divided into single-seat districts corresponding to 50%-70% of the total number of seats. The remainder of the seats will be assigned proportionally. (In the US case, existing single-seat districts could be used, and the House simultaneously expanded to 650 or more members to make room for the proportional seats.)
  2. Ballots list all candidates in the state/province/country, with one column of candidates per party for the largest parties, and independents and smaller parties on the back. The local district candidates would be listed at the top of their respective columns, probably in a larger font.
    • Yes, this is a lot to pack onto a ballot, but they manage 17 columns for 17 parties in Bavaria, so it's definitely possible.
    • Note that the largest states/provinces/countries, such as California in the US or England in the UK, would probably divide into a few sub-elections of around 20 seats each, to keep the ballot from being much much too big.
  3. Voters choose one local candidate, and one non-local candidate. These can be from the same or different parties.
  4. In each district, the local candidate with the larga plurality is seated.
  5. The correct number of seats for each party is calculated. Ballots which did not vote for the local winner are counted as one vote each for the party of their chosen local and non-local candidates. Ballots which did vote for the local winner are counted as two votes for that candidate's party (this first modification to the Bavarian method reduces the chances of "overhang" and thus the incentives for strategic clone parties). Ballots which are half-blank (only local or only non-local) are counted as two votes for the candidate and party on the other half.
  6. Parties under a certain threshold (such as 5%) are eliminated and, for any of their votes that came from ballots voting for two different parties, those votes are transferred to the other party (this second modification to the Bavarian method reduces the amount of wasted votes.)
  7. Each party's ideal number of seats is calculated using a least remainders method. The seats they've won locally are subtracted from those numbers, and the leftover seats are filled by any of that party's candidates who haven't won yet, in descending order of their total number of personal votes received.

Here are the upsides:

Here are the downsides:

PLACE voting

PLACE voting stands for "Proportional, Local, Accountable Candidate Endorsement". This is a voting method I designed in 2016 to show off what's possible in terms of mechanism design. I am serious about this in that I have done enough simulations and analysis to be confident that this method would work well, but I do realize that it's highly unlikely that a country of tens or hundreds of millions would adopt an untried voting method solely on my advice.

Here's how it would work:

  1. Candidates run in specific single-seat districts, and are affiliated with parties. Before the election, each candidate may designate some of the other candidates from their party as "faction allies", and some of the candidates from other parties as "coalition allies". These pre-designations are public information and are available for voters to look up in voting locations.
  2. To vote, you either choose one of the local-district candidates, explicitly listed on the ballot; just choose a party, without choosing a candidate from that party; or choose a party and write in the name and/or district of another candidate from that party.
  3. Each vote is converted to a preference order, starting with the chosen candidate, then their faction allies, then other candidates from their party, then their coalition allies. Ties are broken by direct vote tallies. (Thus, the preference order follows the chosen candidate's predeclarations and party affiliation for the general distinctions, but other voters for the fine distinctions; this impedes detailed vote-trading agreements between candidates.)
  4. Any candidates who do not get 25% of the votes from their local district are eliminated (unless that would leave fewer than 2 remaining candidates from that district).
  5. Winners are chosen using an STV-like process, with the guarantee that there will be exactly one winner per district.
  6. Each winner is assigned a "territory" which includes their district and possibly others, so that each district is in the territory of exactly one winner per winning party. Thus, even if your favorite party did not win in your district, you will have a representative from your party who's responsible to you.

Here are the upsides:

Here are the downsides:

TLD̦R ("Too Long, Didn't Read"; Summary)

I believe that politics in the US, Canada, and UK are broken — even more contentious and stuck than politics inevitably is — and that moving to a proportional representation method would help find more positive-sum outcomes. I've discussed four options, which have different relative advantages, both in terms of potential viability and of probable consequences. I've tried to focus this article on the concrete; if you want more abstract or theoretical discussions of voting theory, there's plenty in my other articles.

I have more hope, right now, of seeing at least one of these good options implemented in at least one of the countries I've mentioned in the next decade, than I have had at any time in the last 25 years I've been a voting reform activist. 


Comments sorted by top scores.

comment by ChristianKl · 2021-01-10T13:12:49.230Z · LW(p) · GW(p)

Each winner is assigned a "territory" which includes their district and possibly others, so that each district is in the territory of exactly one winner per winning party. Thus, even if your favorite party did not win in your district, you will have a representative from your party who's responsible to you.

What's the incentive of a representative to care about his assigned territories besides the one in which he's running? 

Replies from: jameson-quinn
comment by Jameson Quinn (jameson-quinn) · 2021-01-10T16:22:38.775Z · LW(p) · GW(p)

Formally speaking, nothing. Indirectly speaking: the candidate is a Schelling point for voters in those districts, especially if they are not excited by the that-party candidate in their own district. So those voters are a potential source of direct votes for that candidate, which help win not just directly, but also by moving the candidate up in the preference order that gets filled in on ballots cast for other candidates.

Replies from: ChristianKl
comment by ChristianKl · 2021-01-10T21:27:06.534Z · LW(p) · GW(p)

Why would voters in those districts want to gather around a Schelling point?

I think both in US elections and in German elections, the primary reason of why it's valuable for a politician to be a good represenative of his districts is not about votes in the main election. In the US a candidate that does badly at local represenation risks getting primaried. 

This is do to the combination of for example a major of a city having a lot of insight into whether the interests of their city are well-represented at the national stage by the local representative while at the same time the major has local political power that can be used to draft primary opponents. 

Given that being primaried is one of the ways a congressman fails to be reelected it's valuable for them to not antagonize their local political power players and represent them in parliament. 

It seems like the assignment has very little power in your system I find it unlikely that representation of a district will work better then in FPTP as you claim. 

Replies from: jameson-quinn
comment by Jameson Quinn (jameson-quinn) · 2021-01-11T14:17:10.712Z · LW(p) · GW(p)

PLACE is compatible with primaries; primaries would still be used in the US.

Thus, PLACE has all the same (weak) incentives for the local winner to represent any nonpartisan interests of the local district, along with strong incentives to represent the interests of their party X district combo. The extra (weaker) incentives for those other winners who have the district in their territory to represent the interests of their different party X district combos, to fill out the matrix, make PLACE's representation strictly better.

comment by ChristianKl · 2021-01-09T23:31:53.915Z · LW(p) · GW(p)

In the US, the Democratic party is poised to have the trifecta they'd need to pass Federal laws

Any voting reform is going to be bad for the career of some congressman. I don't think that they majorities are strong enough to pass substantial voting reform on the federal level in the US.

Replies from: jameson-quinn
comment by Jameson Quinn (jameson-quinn) · 2021-01-10T00:31:42.666Z · LW(p) · GW(p)

You're right, the sentence you quoted is only a small part of the necessary ingredients for reform; finding a proposal that's minimally disruptive to incumbents (unless they owe their seat to gerrymandering) is key to getting something passed; and even then, it's a heavy lift. 

The 4 methods I chose here are the ones I think have the best chances, from exactly those perspectives. It's still a long shot, but IMO realistic enough to be worth talking about. 

Replies from: ChristianKl
comment by ChristianKl · 2021-01-10T13:12:32.443Z · LW(p) · GW(p)

In a senate with one vote majority and a congress with nine votes majority you are not going to get bills passed that are problematic for some of the representatives within your party. You need larger majorities to do that. 

Replies from: jameson-quinn
comment by Jameson Quinn (jameson-quinn) · 2021-01-10T16:18:59.781Z · LW(p) · GW(p)

This is not an article about the specific circumstances in the US. Suffice it to say that, while you make good points, I stand by my assessment that things are more hopeful for electoral reform in the US some time in the next decade, than they have been in my 25 years of engagement with the issue. That doesn't mean hopes are very high in an absolute sense, but they're high enough to be noticeably higher.

comment by iceplant · 2021-01-10T07:31:14.916Z · LW(p) · GW(p)

These are cool strategies! I hope some kind of voting reform is able to pass in these countries for all the reasons you listed. I’m curious if you have any thoughts about the social mechanisms to get voters and politicians on board with alternative voting systems. It seems like RCV has gotten a lot of apathetic pushback, and I’m wondering how you think we should address it. Here are some challenges that come to mind:

  1. All politicians are the beneficiaries of the system that elected them. Getting them to do surgery on the hand that feeds them could be a tough sell.

  2. People are scared of change. People are even more scared of math. For example, Massachusetts voters rejected a proposal to start using RCV at the state level despite almost no organizing or funding on the NO side.

  3. The city of Burlington (yes, the home of Bernie Sanders) repealed RCV after running into the condorcet paradox in 2009. If people were confused about RCV already, just wait till you have to explain to them why they most preferred candidate, the candidate with the most votes didn’t win, or the election algorithm failed to produce a winner and we have to redo it.

Please redirect me if this has been addressed more fully somewhere else. I feel like the people I talk to about alternate voting systems are already into voting systems and agree that FPTP is pretty garbage most of the time. I feel like I have fewer tools for convincing people who aren’t already into voting systems.

Replies from: jameson-quinn
comment by Jameson Quinn (jameson-quinn) · 2021-01-10T16:29:23.154Z · LW(p) · GW(p)

It seems that most of what you're talking about are single-winner reforms (including single-winner pathologies such as center squeeze). In particular, the RCV you're talking about is RCV1, single-winner, while the one I discuss in this article is RCV5, multi-winner; there are important differences. For discussing single-winner, I'd recommend the first two articles linked at the top; this article is about multi-winner reforms.

Personally, I think that the potential benefits of both kinds of reform are huge, but there are some benefits that only multi-winner can give. For instance, no single-winner reform can really fix gerrymandering, while almost any multi-winner one will.

The issue of politicians not wanting to "do surgery on the hand that feeds them" (I like that metaphor) is a real one. The four methods I've chosen to discuss are all chosen partly with an eye to that issue; that is, to being as nondisruptive as possible to incumbents (unless those incumbents owe their seats to gerrymandering, in which case, fixing gerrymandering has to take precedence). Actually, of the four methods, RCV5 is the most disruptive, so if this is your main concern, I'd look more closely at the other three methods I discuss.

comment by Dacourt · 2021-01-10T17:24:30.615Z · LW(p) · GW(p)

It’s worth noting that the U.K. had a referendum on Alternative Vote (AV) and voted 68% to 32% to reject it and maintain the status quo of First Past The Post. Turnout was 42%. 

Replies from: jameson-quinn
comment by Jameson Quinn (jameson-quinn) · 2021-01-11T14:09:53.751Z · LW(p) · GW(p)

Also worth noting that both AV and FPTP are winner-take-all methods, unlike the proportional methods I discuss here. The AV referendum question was essentially "do you want to take a disruptive half-step that lines you up for maybe, sometime in the future, actually fixing the problem?"; I'm not the only one who believes it was intentionally engineered to fail.