I feel like this might go a little too far in formalizing the two-party system in a way that would have undesriable consequences. A better approach to drawing Congressional districts would be something like this:
One better is to not have districts at all, and instead use proportional representation, which is what's used in more than 50% of all major democracies around the globe[1].
IMHO, more important than eliminating Gerrymandering, is eliminating the two party system. In truly competitive governments with 10-15 parties holding seats, it's incredibly risky to lie, BS, or ignore arguments. In our two party system most of the arguments are just designed to talk you out of voting at all if you favor the other party.
The US only has a two-party system because our method of voting ensures that it devolves into a two-party system. Our two-party system is a symptom of the problem, not the cause.
Give us ranked voting and proportional representation and we'll see more than two parties emerge.
I really like that a lot, but it needs more awareness of geography as it relates to human travel. In particular, main road networks and major waterways. A district should not count as compact if it is not contiguous via major roads, avoiding toll bridges and ferries whenever possible. And it should have a resistance to crossing major rivers.
For example, Richmond and Sausalito should not be in the same district, and the Sacramento river should be a divide at least as far up as Antioch and Rio Vista.
Likewise, Vashon Island in WA should not be split across 2 congressional districts, but (since too small to be a district on its own) should be entirely in one of the districts it has a ferry connection to. And Puget Sound should be a hard divide.
California had a statewide ballot initiative (Prop 20) which established the California Citizens Redistricting Commission, which drew districts ignoring party affiliation. The resulting districts lean heavily Democratic (because, duh, California) but are also fiercely competitive; there are very few safe seats and incumbents have to hustle. Frankly it's fantastic, a model for the nation.
California has proven out the solution: collect a set of citizens from both parties as well as some unaffiliated, aggressively de-politicize the process, give them broad guidelines on how district lines should be drawn, and let them do their thing. No need for weird cake-cutting machinations.
If you're suggesting that because a system draws from a group of people consisting of both sides of politics (or no side of politics, then it is therefore "functional, fair and apolitical in its outcomes", then I suggest your conclusion is somewhat naive.
The idea of "I cut you choose" is an incredibly inspired one because fairness is inherently built in because no participant wants to risk getting the bad side of the slicing and therefore works very hard to make it fair.
Plenty of people are perfectly willing to draw fair districts. 7.4 billion people on Earth right now, and at least 7 billion of them would be perfectly honest and fair in drawing congressional districts. The fact that the US generally picks people from the tiny slice of committed partisans to draw boundaries is idiotic.
while I like to be optimistic about human nature, the ability to find people that don't have political and/or financial interests involved in those decisions is not trivial. The 7.4 Billon people number is irrelevant. No one is going to someone from Sudan to draw districts in Wisconsin-they won't even go to someone from Illinois or Minnesota for that advice.
They could. Get somebody from out of state, and give them just a raw map with population counts/density/whatever. Heck, do this 3 times, and then get local party affiliates to look them over, and each party can eliminate one of them so the most middle-ground is selected.
Can you elaborate on what guidelines they were given?
There is one surprising bias you have to be aware of. It seems if we have a state that is let's say 35% democrat, 30% republican, 20% libertarian, and 15% green then it would seem to be most fair for districts to have a similar composition. But this would be disastrous as in our district driven first past the post system this would mean that the party with 35% of the public support would end up receiving 100% of the congressional seats!!
Of course the desired result should be one proportional to the makeup of a state. And this is done most naturally by a system where individuals vote for a party (or individual running as a one man party) at a state level with seats distributed according to the total vote. If a state has 10 seats then 10% of the vote guarantees at least one seat. The parties then distribute their representatives to districts weighted on the vote. E.g. a district that voted for 80% democrat would be prioritized to receive a democrat representative more than one that voted 20% democrat.
The 'problem' is that if you do the above all the sudden you'd have socialists, libertarians, pirate party guys, and all other sorts with some minority representation in congress - and we can't have that can we? /s if that wasn't clear. But really, the powers that be would never agree to a system that would meaningfully undermine the establishment and so with that condition, "I Cut You Choose" does seem to be the most reasonable and realistic possibility for fair outcomes.
---
More generally here, I think systems that assume bad behavior and find solutions that acceptably in spite of that tend to be vastly better than ones that assume good behavior and result in collapse when undermined. It's basically just finding game theory optimal solutions to problems.
So yes, if you removed close to 3 million democrats from the state somehow, it would be a tie. Probably an easier method would be for Republicans to make their policies more attractive to more people.
The topic was on gerrymandering districts - there are more republican districts. The part of the comment I was replying to was "DISTRCITS lean heavily Democract" which is not true, more DISTRICTS lean Republican.
I can’t see how you came to that conclusion whatsoever, though I may be biased as an SF-born Democrat.
If you’re looking at the governor race, the votes eyeball to 2.8m Dem, 1.7m Rep.
As well, it was surprising to me that you would choose to remove SD, as a Republican lead that county, though the Democrat leader there was only down by 10k (of ~400k total) votes.
He's talking about counties, which are spread across a diverse geographic region. As per the image there it looks like the clear majority of California's counties leaned republican for governor. But the counties that leaned democrat are very densely populated and so end up choosing for the entire state.
On a national level this is the reason for the existence of both the electoral college and the senate. Imagine we did not have the college or senate and there was one very heavily populated state with a large group of mostly homogenous voters. Suddenly that state would be deciding every single action for the entire nation. Whether this is a good or bad thing probably depends on your views on what representation means, but it's safe to say that people in many states would feel they lacked such - which was the very reason that we rebelled against the British Empire.
It’s not hard to understand - land doesn’t vote, people vote, and measuring party share by the wrong metric is silly. Tango’s explanation was very clear and perhaps helpful to others but didn’t tell me anything I didn’t already know.
Ed: Your comment elsewhere in thread about districts helped the most in realizing we were talking about different things. MB
In effective two party systems, votes cancel each other out. If I vote A and you vote B, then our votes effectively negate each other. This is why 'strategic voting' is a thing - a problem that does not exist in more sane proportional systems with more diverse representation.
The same thing would happen all the way up. Imagine we have two parties, again A and B, that are completely mutually exclusive - one representative always voting exactly the opposite of the other. And now imagine we have a state Massivefornia that has 1 billion voters and a proportional number of representatives, but it's split exactly 50/50. And then there's the other state, Lonely Island, with exactly 1 voter and 1 representative (who... I guess would be one and the same?). In this example, Lonely Island would decide each and every action that goes forward as Massivefornia and their 50/50 split of A/B cancels itself out on everything.
One of the better ways I’ve heard is multi-member districting, where you have a slightly larger district with multiple reps (maybe 3-5 representatives) where seats are partitioned proportionately for that district. Voters vote for a party, and seats are apportioned to each party in accordance with their share (with some rounding). This guarantees a result that is at least as proportional (I’d posit significantly more proportional) as the current representative system, and maintain the ‘theoretical’ locality of representation, meaning constituents still vote for people representing their small localities.
There are different ways to assign representatives in proportional representation systems, but at such small scale you could have district-wide primaries where the top N finishers for that multi-member district are slotted into seats in the order in which they finished in the primaries.
Unfortunately this system is still imperfect, as strategic voting can still occur l, but significantly less so than in our first past the post single-member system. Also it excludes extremely small parties (>10% of the national electorate) from influence. It’s a tough problem.
Get rid of FPTP first, since that is the worst of your problems. Multi-member districts will be an incremental improvement on top, and actually be meaningful once every vote ends up selecting a representative.
With a STV system like Hare-Clarke you will likely see independents standing with very centrist platforms, with the major parties quickly scrabbling to get those centrist votes.
> Get rid of FPTP first, since that is the worst of your problems
No, unequal apportionment and the electoral college are the biggest problems, FPTP and single-member districts in state and federal legislatures are, in combination, just the biggest problem which isn't federally Constitutionally mandated and which can therefore be fixed with legislation at the appropriate level(s) without federal Constitutional Amendment.
> With a STV system like Hare-Clarke you will likely see independents standing with very centrist platforms
Perhaps, but every candidate elected being a compromise candidate that relatively few people feel they would have preferentially votred for first is horrible for representation, and that's what you get with any STV-like system with single member districts in an electorate where political views aren't unimodally distributed (and if views were unimodally distributed, FPTP would get you exactly the battle for the center you propose and you wouldn't need STV to get it.) You want compromise in the legislature on policy where the voices in the legislature reflect those in the populace to the degree feasible; you don't want the compromise to happen mechanically in selecting members of the legislature.
> No, unequal apportionment and the electoral college are the biggest problems
What do you mean?
I think the biggest problem is giving electoral votes based on total state population instead of based on citizenship.
California and Texas both end up with more than they should have.
Edit: I'd like electoral votes to be distributed proportionally instead of winner takes all - it sucks that if you are not in a swing state you have less influence.
There are good reasons why no state awards their electors proportionally.
Electors are people. They each have one vote. The result would be a very inexact whole number proportional system.
Every voter in every state would not be politically relevant or equal in presidential elections.
It would sharply increases the odds of no candidate getting the majority of electoral votes needed, leading to the selection of the president by the U.S. House of Representatives, regardless of the popular vote anywhere.
It would not accurately reflect the nationwide popular vote;
It would reduce the influence of any state, if not all states adopted.
It would not improve upon the current situation in which four out of five states and four out of five voters in the United States are ignored by presidential campaigns, but instead, would create a very small set of states in which only one electoral vote is in play (while making most states politically irrelevant),
It would not make every vote equal.
It would not guarantee the Presidency to the candidate with the most popular votes in the country.
The National Popular Vote bill is the way to make every person's vote equal and matter to their candidate because it guarantees the majority of Electoral College votes to the candidate who gets the most votes among all 50 states and DC.
Article I, Section 2 of the Constitution mandates the U.S. Census count every resident in the United States.
The current system gives "illegal immigrants" a 10 vote advantage in the Electoral College for the Democrats...because they tend to live in safe Democratic states.
An election for President based on the nationwide popular vote would eliminate the Democrat’s advantage in Electoral College members arising from the uneven distribution of non-citizens.
California is 87% citizens compared to 93% for the country. That doesn't change the voting math much. But Wyoming citizens gets 7000% more Senate representation.
> It's a pretty big chunk of EC votes - more than Wyoming gets in total!
No, it's about the same. If you pull non-citizens out of the calculation nationally, California loses between 3 and 4 seats in the House.
> The senate is meant to represent the states - not the people
That's no more true than saying the House is supposed to be scaled to people and not citizens. Both are features of the design of the system. “Is supposed to be X” doesn't stop something from being a problem in terms of a goal other than X.
Mostly the Senate and it's resulting effect on the Electoral College, though granularity effects from the very small number of house seats are also an issue.
> I think the biggest problem is giving electoral votes based on total state population instead of based on citizenship.
The existence of a legal class of permanent resident non-citizens is itself a fundamental problem for the entire idea of popular government.
But even if you accept what you pose as a distortion at all, it's less significant than the disproportionality in federal legislative and executive voting power produced by the Senate and the corresponding Electoral College seats.
I like the idea of Approval Voting, where voters give every candidate a thumbs up or thumbs down. A vote for a third party candidate is not "wasted" and the system would be much easier to explain to voters than STV or IRV.
We've a STV system in Ireland, I'm not sure you'd classify our independents as centrist (by the Irish definition of centrist - by the US definition almost all our politicians are far left).
Overall the system works quite well, and the proportion of representatives a party has in parliament roughly aligns with first preference votes.
In Ireland you have had enough time for parties to trend towards the biggest chunk of votes, which is what I’m labelling “centrist” though perhaps this is really “populist” :(
Not constitutionally, but by using single member districts and our current voting algorithm, the equilibrium is two parties. So we are a de facto two party state.
It’s First Past The Post that forces the two party system. Get rid of that, STV is what will bring lasting change and reduce the the tendency for the two major parties to be exaggerated opposites of each other.
For the presidential elections approval voting would be more appropriate, for local representatives proportional representation is the important part for which you don't really need STV (although I guess voting for political parties with proportional representation is somewhat similar to STV with a predetermined set of preferences).
But then I look at Canada, or the UK. These are single member districts, with first past the post elections, and they have multiple parties, including regional parties. The US doesn't have anything like that.
In practice i’d consider the UK a 2 party state as well. FPTP basically results in either Labour or Conservatives leading government. The other parties are geographically limited. Eg the Scottish Party cannot win outside of Scotland, same for the 2 Irish parties, and although LibDems should not be geographically limited, in reality you can almost consider them the London party.
Go check the Dutch or Belgian systems. They have proportional representation and our (NL) current government is a coalition of 4 parties. (Yes, it moves slower, but also results in a more nuanced government and more gradual change)
The requirement of parliamentary systems to form a government lets some smaller parties have a look-in, since they can get concessions from the large parties to agree to join a coalition. Without that, a smaller party has no leverage.
But the UK is the classic two party system that Washington warned against.
The American two party system is less limiting than it sounds. Bill Clinton shifted the Democratic Party and Trump is shifting the Republican Party. In practice the parties don't define the candidates, the candidates define their party.
While not covered in the article, it's likely that the same process works with more than two participants, as long as one makes the partitioning and some other one (adversary) freezes one district; so A partitions, B freezes one and partitions again, C freezes one and partitions again, A freezes one and partitions and so on while there are districts.
We're also not constitutionally a lot of things that we do all over the country (e.g. discriminating based on race, gender, sexual preference, etc; presidential power to wage war through executive orders; and more!)
How do you evenly split a quantity of something (e.g. drugs) when you have no scales? One person does the splitting, then presents both halves to the other person, who can choose either one as the half they keep. Everyone is incentivized to transact as fairly as possible.
Exactly! Gerrymandering is often seen as one party stealing votes from the other, but really the more important problem is that elected officials are stealing power from the electorate. In some cases officials of the two parties are even happy to collaborate to achieve this.
In each round, I reckon the optimal strategy for the "cut" will likely be the same as partisan cuts are now; densely pack the opposition's supporters into a small number of districts, while slicing the rest up to have a small but stable majority for one's own party.
This gives the opposition the chance to choose/freeze either a district they know they'll win, or a district they're pretty sure they won't win. I would choose the one I know my party would win, and redraw to pack all my opposition's voters into as few districts as possible.
If this is repeated a few times, the map ends up with many safe districts for each party and a small number of "left-overs" which will be contested. This means parties will have an incentive to concentrate on the small number of voters in the few contestable districts, at the expense of probably the majority of voters. It doesn't seem like a recipe for every vote counting and voters' voices being listened to.
If we break up monopoly businesses when the lack of choice increases the cost to consumers, should we also break up monopoly parties when the lack of choice increases the "cost" in time and effort that voters must "spend" to get their issues dealt with?
I believe that cake cutting would improve on the current situation, when it is the incumbents of one party cutting and choosing, but it is not a full solutions to all the problems in the space.
Beyond just incumbents - this would favor whatever the party favors, since it still let's you pack and crack anti-establishment forces in your primaries. Though I think primaries are a real cause of our polarisation.
Same thing can happen in countries where the state funds political parties and outlaws large private/corporate donors. The funding is given according to how many seats each party already has and so favours the status quo.
My older brother's "solution" to being told he must cut the cake and let me choose which half to take was: a sloped cut similar to the one in this picture [1] but which slightly exaggerated the uppermost surface area of the smaller volume.
An analogous tactic against this "I-cut-you-freeze" task might be to provide especially appealing district boundary for your opponent to freeze, while harbouring some knowledge about that district that that your opponent is unaware of (eg suppose that you know the trajectory of demographic changes much better than your opponent)
The competence of the participants to judge the slices is outside the scope of this. If one of the parties doesn’t know what they’re doing, that’s their problem.
What do you mean by "arbitrarily long"? There's this related procedure for N cake eaters, which is straightforward:
1. Arbitrarily choose an order
2. Repeat while not everyone has a slice:
a. First person without a slice cuts a slice
b. Each other person, in order, who's without a slice may either take the slice or pass it
c. If everyone else has passed, the cutter must take that slice
It's quadratic time -- worst case you go fully around the (remaining) circle for every slice -- but that's still not a very big number even for four or five political parties. How many parties are there actually going to be?
(I'm not exactly sure how this maps to the article's scheme, either; the point is that I'm not sure it's all that complex.)
From the relevant wikipedia page: "envy-free divisions of connected intervals to 3 or more people cannot be found by any finite protocol." - I can agree that your algorithm is quadratic, and therefore, that your algorithm is not envy-free.
Your suggested algorithm seems to be unfair because the first party to choose has a clear advantage over the parties who choose later and did not cut: depending on the order of the parties there is also likely to be collusion.
The immediately easiest (and politically realistic) way to game your system is to split yourself into multiple parties and then cut your allies/sock puppets a massive slice.
I cut you choose is not fair. I actually heard an example in real life about a week ago when something was being cut for two people (they weren't following the "I-cut-you-choose" protocol, just trying to be fair). The person doing the cutting was literally saying "It is very hard to cut it so they're both the same."
This is true. Many things don't naturally cut into two equal pieces despite the person's best attempts, and in certain cases it may be impossible. (For example, if there is are "valuable" toppings that are indivisible.)
It is just plain easier to choose than to cut. The worst case choice for the person choosing is that they are indifferent to which piece they get, but the worst case for the person cutting is that they get a worse piece.
I used to joke around with my brother that it'd be funny for the cutter to intentionally cut the cake like 9/10 to 1/10 then get outraged when the chooser took the bigger piece. But, you're absolutely right - choosing is better.
This algorithm is effectively like 2 people dividing N cookies using the cut/choose method. Each side will both cut and choose N/2 times. If N isn’t tiny then worst case it’s odd (e.g 17 for PA).
Thats why this method has the "back and forth". After a choose, the cake gets magically "re-cut" by the previous chooser and so on and so forth until all the pieces are taken.
Here is an idea I had awhile ago - the rules should limit the ability of a boundary to be concave. It isn't reasonable to say "no concave polygons", but you could create a measure of "concavity". If you look at a ridiculously drawn boundary, you'll see it can kind of spiral - it'll hop down a street turn a couple times all in order get the current congressman's house within his existing district.
If there is any point within the boundary that has to cross the boundary more than 3 times to get to a point outside of the boundary, then the proposed congressional district is too complex. if it crosses 4 times (or any even number), the point isn't actually within the boundary to begin with. Crossing the boundary 3 times still allows for some concavity, but 5 is kinda ludicrous.
This further deepens the two-party system. That's no solution at all.
Districts should be compact, as simply shaped as possible, equally populated, and ideally follow as many lines of local government concern as possible.
We already have utility districts, school districts, city precincts, county precincts, townships, drainage and levee districts, cities, towns, villages, counties, wards, neighborhoods, state police districts, state House districts, state Senate districts, and more boundary lines within states. Rather than one county precinct being cut into four parts so each can be a portion of four different Congressional districts that also overlap other county precincts, perhaps as much as possible your part of town should have the same candidates on the Congressional race as your neighbors using the same voting station.
The top percentage of votes on the highest relevant regional area.
If you are voting for the mayor of a city, just count every single vote of the citizens. Why do you first need to split up the people into arbitrary groups, and then count the group outcomes?
> The major problem—one of the major problems, for there are several—one of the many major problems with governing people is that of whom you get to do it; or rather of who manages to get people to let them do it to them.
To summarize: it is a well-known fact that those people who must want to rule people are, ipso facto, those least suited to do it.
To summarize the summary: anyone who is capable of getting themselves made President should on no account be allowed to do the job.
- Douglas Adams (The Restaurant at the End of the Universe)
I think the truth in this humor is quite relevant to those in whom you're placing your hope.
Great quote :) To rephrase I think the problem is that the skill set you need to get into a political position is completely orthogonal to the skill set you need to do good work in said position.
My comment was utter sarcasm, which apparently wasn’t obvious, and that’s sad for us collectively. But as long as that is the case, I think this was a great answer.
What makes you say it wouldn't work? Why would waiting for honorable experience prevail over an algorithm designed to play to the standard participant?
I think it has potential, but I can see where if there are two players, they could "trade horses" and still keep on gerrymandering. Still, I think it potentially could work and would be worth trying out.
There is no "we trade districts and both profit" scenario (apart maybe cutting out 3rd parties that have no chance of winning anyway). If they trade one gets the strictly better deal, the other loses.
The courts exist to provide relief to the people and regardless of whether a solution like this were applied, the courts could continue to provide relief. It would be interesting to see a solution which is more self-organizing at a grassroots level and which might mitigate the whole "from on high" decision-making. So long as a small group of driven individuals is making decisions like this, we will continue to have risks from material non-public knowledge used to game the system, and risks to the successful foothold by minority (of any sort) populations except those large enough to warrant party support. It's not an issue that the courts have had to step in -- without the courts the people would be left without relief.
https://bdistricting.com/
"Impartial Automatic Redistricting... optimized for equal population and compactness only."