Thread by @bettercount_us

In the shadow of the deplorable SCOTUS decision in #Rucho, it's time once again to talk about how to fix gerrymandering. So, here comes a thread.

(Betcha never seen a Twitter thread w/ToC before!)
1/76

Elections twitter peeps have probably heard me make these points before. You may agree with some, disagree with others. Either way, respond! Above all, I want to spark dialogue. I saw in BC how, when we don't get ahead of the curve, we end up behind it.
2

The tweet count on this thread is gonna be high, but I'm gonna make just 4 key points here. Here they are, up front, as a table of contents:
3

ToC:
① The best way to #endgerrymandering is congressional action from a Democratic trifecta: 5
② Redistricting rules can help, but #ProRep is better in both theory and practice: 10
③ Time to talk about #ProRep details: 18
④ Options to discuss: RCVn (STV), MMP, DMP, PLACE: 37

①: The way to #endgerrymandering is Congress.

No hope, now, that SCOTUS will save the day. State legislatures are the problem, not the solution. Diminishing returns on state referendums and lawsuits. So, Congress.
5

Does Congress have the authority to #endgerrymandering via #ProRep? Yes!

Equal protection (14a). Voting equality by race (15a). Time, place and manner for congressional elections (Art I, §4). Republican government for state elections (Art IV, §4).
6

All of these would apply, especially to a law that kicked in only when a partisan districting process led to a partisan-biased gerrymander, and that allowed states to craft their own #ProRep methods but had a fallback option for states that refused to do so.
7

One important consequence of thinking in terms of Congressional action: the key audience is Democratic incumbents in House and Senate. Important, of course, to make a good proposal on its merits and to rally enough grass-roots support to get foot in door. Gotta <ize disruption.
8

On that count, the existing #FairRepAct bill is flawed.

It's a great start, but I don't think it could pass even a Democratic-controlled Congress in its current form. Too disruptive to incumbent House reps; & steps on too many toes at the state level.
9

②: #FairMaps good, #ProRep better.

I'll argue this from both theoretical and practical points of view.
10

Theoretically, my basic point goes to *wasted votes*. It's tough to give a rigorous math definition here, and I won't even try on Twitter, but pretty easy to know it when you see it. Under choose-one (FPTP), at least 50% of votes are wasted: all losing votes, all >50% for a seat.

Gerrymandering weaponizes those wasted votes. But even the best good-faith district map still has them. At best, there's partisan balance.
12

In particular, under choose-one, the goal of minimizing wasted votes is in *direct* conflict with that of ensuring competitive elections and accountable representatives. Either you have 49% losing voters with no true rep, or you have 20% losers and unaccountable 80% reps.
13

Important to note that algorithmic redistricting, or better single-winner voting methods, do *not* fix this problem. It's still a choice: competitive elections mean high variance and thus high mean-squared-error disutility, and uncompetitive ones mean unaccountable, corrupt reps.

(Aside to @UtilaTheEcon, who always shows up in these threads: I'm ready to back up this anti-algorithmic-redistricting stance, but not on Twitter. Tl;dr scorevoting.net links don't save this low-bandwidth platform for detailed debate. Name your forum, but not here.)
15

(Aside to @UtilaTheEcon, who always shows up in these threads: I'm ready to back up this anti-algorithmic-redistricting stance, but not on Twitter. Tl;dr scorevoting.net links don't save this low-bandwidth platform for detailed debate. Name your forum, but not here.)
15

Independent redistricting commissions are *way better* than partisan gerrymandering, but they don't solve wasted votes. As long as >50% of votes are wasted, danger of partisan bias remains high. Leaves urban centers as unexploded cluster munitions. Partisan capture of commission?

Proportional representation (#ProRep), OTOH, fully fixes the problem. Poof: wasted votes go down to a rounding error. With suitable open-list methods, incumbents remain accountable. Outcomes better than very best districting maps, w/o risk of bad map by mistake/enemy action.
17

③: now is the time to have this discussion.

Let me tell two stories about Canada.
18

Nationally, Trudeau was elected with an unequivocal promise to #EndChooseOne. "This will be the last FPTP election," he said. He made the #ERRE commission to look into reform. But when they recommended #ProRep, instead of his favored AV/IRV/RCV1 (gah, terminology), he killed it.

The lessons from this #ERRE and mydemocracy.ca debacle are:

Don't assume anyone who opposes FPTP supports real reform. Backstabbers exist.

Don't use vague terminology to sweep disagreements under the rug. Looking at you, "RCV".

Never to soon to discuss brass tacks.
20

Second story, British Columbia: #PR4BC.

Greens and NDP both had voting reform in their platforms, but obviously existential need was higher w/Greens. 2016 results put them in coalition, with reform a key plank. Conditions were ripe, right?
21

Mistake #1: They gave the job of designing the referendum to a "neutral" party, AG @DaveEby. I'm not criticizing Eby, but "neutral". Imagine if in order to write a legislative bill, you had to be neutral on whether it was a good idea! Crazy!
22

And of course, anti-voter lobbyists like @BillTieleman (I respect your personal ethics, Bill, but that's the job you were doing) jumped on that "neutral" framing to work the refs. "Oh no: the guy structuring the referendum may have sympathy for it passing!"
23

Of course, I'm not saying that there shouldn't be neutral arbiters running the actual voting process. But that was the job of @ElectionsBC. Whoever designed the referendum Q, be it @DaveEby or other, should have been an unashamed partisan for #ProRep.
24

Mistake #2: wasting months with a bloodless "public consultation" #HowWeVote survey process that just recapitulated the false dichotomies from the mydemocracy.ca sabotage job.
25

A few false dichotomies in the #HowWeVote #PR4BC survey:

"Many small parties vs. a few big ones": Well-designed #ProRep can target a happy medium.
26

"Should reps obey parties, or voters?": With fewer wasted votes, the gap between those is narrower.

"Lots of choices, even if it's unclear how votes become seats": Mixing number of choices, with ballot format, with counting process; three separate issues.
27

"Parties cooperate to govern, or one party governs alone": This should be up to the voters in each election, not set in stone by the voting method.

"Ranked ballots, or one candidate?": There are other formats.
28

"Single-seat reps, multi-seat reps, or mixed reps?": Again, not the only options.

"Which PR methods should be on the referendum? List PR; MMP; STV; Japan-style semi-MMP; other": WTF with this list? Nobody asked for closed list / Japan; 2005 citizens' assembly ruled these out.
29

This waste-of-time survey would never have happened if Eby hadn't been fruitlessly chasing an image of neutrality that opponents were never going to grant him. Instead, they should have run a new #CitizensAssembly.
30

So this is where I entered the picture. Instead of watching helplessly from US, I decided to do something about it: help organize the #BCPRsymposium. We got funding @ElectionScience (thanks!), academic sponsorship @SFU_polisci (thanks!), & invited Canada's top experts.
31

Our voice had an impact; Eby's referendum announcement cited us 3 times, more than any other submission. But even so, we were just volunteers stepping forward. With no democratic mandate to make decisions, we were limited to laying out the good options, not recommending one.
32

Shout out to BCPRsymposium folks:
@GraceALore @rgcjohnston3 @rkcarty @1alexhemingway @ruckertgisela @fairvotingBC @youth4pr @MaxwellACameron @sean_gra @jasonmclaren cc: @David_Moscrop @PR4BC. Note: this was an academic, not advocacy, exercise; not imply endorsement of #ProRep.
33

And then the #PR4BC referendum didn't pass. I could cast blame on the "no" campaigns fearmongering and, frankly, superior minority outreach (IMO minority communities like Chinese-Canadians shoulda embraced #ProRep), but opponents are a given. "Yes" campaign was unfocused.
34

The moral of both of these stories: the time for arguing out the details of how #ProRep should work is *before* the political moment comes to pass these reforms. Get your ducks in a row, as much as possible! Vague support for vacuous common principles is not enough!
35

In other words, in the US, the time for debating between #ProRep methods is now.

Not 2021, when (I hope) a Dem trifecta is taking power.

Now.
36

④: So, let's discuss specific #ProRep proposals.

Ignoring stupid closed-list ideas, I see 4 reasonable ideas:

❶ RCVn (STV)
❷ modified Bavarian MMP
❸ DMP
❹ PLACE
37

As I said above in point ①, the goal here is to find a proposal that could pass a Democratically-controlled US congress. So I'll look at these 4 options from a few perspectives: incumbent congressperson; ethnic minority voter; third-party activist; political scientist.
38

❶ RCVn (STV)

Known in academic terms as Single Transferable Vote, and in modern activist branding as Ranked Choice Voting (eliding the differences between n-seat RCVn and non-prop 1-seat RCV1). Used: Ireland, some Australian elections, Malta. And here in Cambridge, MA!
39

In RCVn, voters rank candidates in preference order; the weakest candidates are eliminated and those votes transfer, until a full slate of candidates has one quota of votes each, with less than one quota of wasted votes.
40

Incumbent congressperson perspective: RCVn is appealing because it has a US history since the Progressive era (@JackSantucci's the expert) and currently growing grassroots support. National orgs like @FairVote and energetic local ones like @VoterChoiceMA are on board.
41

Also appealing: in 3-5 member districts ("small-M"; @LaderaFrutal's the expert), RCVn wouldn't cause too much party fracturing. Existing big-tent/coalition parties could stay largely intact.
42

But there are downsides for incumbents. Principally: it's a radical change, and current districts would disappear. Inevitably it would cause more turnover than some of the other options. That's scary to incumbents.
43

Minority voter perspective: like any #ProRep, it would increase numerical representation, especially for minorities with over 17% in some multiseat district(s). But that threshold would be even lower under other #ProRep methods.
44

Also, by weakening links between individual represenpartatives and their parties, RCVn might not be as helpful for substantive representation as for numerical? I'm not sure on this; @GraceALore's the expert.
45

Third-party activist perspective: Unlike the other options I'll discuss, RCVn doesn't take candidates' party into account at all. This makes it appealing.
46

Personally, I think that's largely an illusion. The other 3 methods I discuss all give voters viable ways to support independent candidates and ensure their votes count cross-party if needed. And RCVn, n=3-5, would still substantially favor established larger parties. But, OK.
47

Political scientist perspective: RCVn has some issues with centralized/hard-to-audit counting, and with spoiled ballots/ballot complexity. But all-in-all, it's a well-known, solid method.
48

❷ modified Bavarian MMP

By this I mean open-list mixed-member proportional. The model is the German Land of Bavaria, but I'd include a few modifications to make it a bit more STV-like.
49

Lets say there was a 50/50 split between regional and partisan reps. So, either you'd double the size of the House, or districts would be twice the size. Voters would choose 1 candidate from their own district, and/or 1 from some other district.
50

In each district, the local winner would be seated. Then each party's fair proportion of votes would be calculated, counting both local and nonlocal votes equally.
51

This is where the modifications come in. First off, if your local choice wins, your nonlocal vote counts for that same party, even if you cast it for a different party. Akin to "using up a quota" in STV; stops strategic cross-party voting a la Wales/Italy/etc.
52

Second, if exactly one of your two votes went to a party (or independent) who won no seats, then that's recounted for the party of your other vote. Akin to "elimination transfers" in STV.
53

So now the party's fair proportions are calculated. They each get enough seats to fill up their proportion, assigned to whichever unseated candidates have the most votes (local and nonlocal).
54

Incumbent congressperson perspective: Districts double in size. That's less disruptive than RCVn, but more so than PLACE.

Also, probably leads to more party splitting than other options. Scary to D incumbents.
55

Minority voter perspective: Great for organized, cohesive minorities. Less so for others.
56

Third-party activist perspective: Allocates seats by party, even though lists are open. That leaves a bad taste.

However, in practice, quite open to independent/3rd party candidates.
57

Political scientist perspective: Comparative political studies dream!
58

❸ DMP

Dual-member proprotional, as invented by @sean_gra. 2-seat districts. First seat goes to local winner, second allocated by party in a way that ensures each district ends up with 2 reps.
59

Incumbent congressperson perspective: Second-least disruptive option. But probably party splitting as much as MMP?
60

Minority voter perspective: Relatively little voter choice, so probably worst of these 4 options for minority voters.
61

Third-party activist perspective: Like MMP. Parties have a formal role, "yuck". But in practice, good for independent candidates.
62

Political scientist perspective: Interesting and new in some ways, but under the hood, basically a new biproportional variety of MMP.
63

❹ PLACE Voting

Proportional, Locally-Accountable Candidate Endorsement. Proposed by... me. So yeah, I'm biased. But as I hope you see from this thread, I've given these issues a lot of thought.
64

Uses existing single-seat districts. Voters choose one candidate — either from the list of local candidates, or by writing in one from another district.
65

Seats are allocated by an STV-like sequential, transferable, quota system. If your chosen candidate gets eliminated, your vote transfers to one of the allies (at one of three levels) that they pre-declared — first by faction, then by party, then by coalition.
66

Ensures that it takes at least 25% of local votes to win a seat; and that there's one winner per district. Here's an FAQ which explains it in detail: electowiki.org/wiki/PLACE_FAQ
67

Incumbent congressperson perspective: Least-disruptive to incumbents of any of the methods I've discussed. In fact, with a few light assumptions, you can simulate results from past elections, and show that, except in case of gerrymandering, most incumbents probably woulda won.
68

Also, the 25% local threshold means that it would likely lead to even less party fragmentation than RCV5. Voters are free to support new parties, but incentives still encourage those parties to ally with an existing major party. Again, non-disruptive.
69

(In BC, one of the main anti-PR talking points was "it would elect extremists". Because of the 25% threshold, PLACE wouldn't. Without wasting votes, PLACE would encourage compromise coalitions, not single-issue intransigence.)
70

Minority voter PoV: I think it's the best of all these options. Any minority with >25% in any district would get numerical representation. And smaller minorities would tend to get substantive voice, as other candidates court vote transfers. Unified negotiation=strength.
71

Third-party activist perspective: Substantial third parties — Greens, Libertarians — would win a few seats, though they'd probably naturally become a bit more mainstream as they did so. Fringe groups wouldn't. Is this good? YMMV.
72

Political scientist perspective: This is really weird. Delegated vote transfers? Local vote thresholds? I spent a lot of time and thought getting all these moving parts to work together smoothly, but it's not what people are used to.
73

So, is this just a Time Cube? That is, am I just some Twitter crank ranting about a panacea? Obviously, I think not.

Even with a 75-tweet thread.
74

Aside from organizing the BCPRsymposium, I designed and helped pass the #EPluribusHugo proportional nomination system used successfully for years by thousands of voters to award the well-respected Hugo awards. I'm on the board of @ElectionScience (but here, I speak for myself).

But the point of this thread isn't "hey everyone look at me", it's "we need to be talking details NOW, so we'll be ready in 2021". If you think PLACE sucks, I still want you responding to this thread.

end!
76/76

Bonus gif for anyone who makes it through this elephantine thread: