Ugly Gerry – Gerrymandering font | Hacker News Reader

Ugly Gerry – Gerrymandering font

fontsarena.com

327 pointssohkamyung177 comments

A, N, M, R, U look particularly egregious.

Looking at M, it seems like the districts don't even have to be a single physically connected area because there are speckles floating above the main shape.

"First past the post" is a terrible way to conduct an election. This kind of map-drawing games wouldn't matter nearly as much if many members were elected from each district as they'd have to be much larger and could more obviously follow existing state and city boundaries.

Yes, exactly. It seems to me that any first past the post system is an accidental democracy (at best). Technically, if every voting district has voted exactly 51:49%, then every single vote of the 49% group will be void, ignored. Winner takes all. Last elections in the USA shoved discrepancy of up to 3 million votes if I remember correctly, meaning majority vote number vs. actual number. And that number can go up to almost half of the population in the most absurd scenario. But even 10-20 million discrepancy, which is much more likely, can possibly cause some huge problems and protests.

> And that number can go up to almost half of the population in the most absurd scenario.

It can get worse. If there are N parrties, the winner could take all seats with a vote percentage of 1/N+epsilon. Or could take the majority with 1/(2N)+epsilon.

For example a party could win the majority of seats with 10%+epsilon of the votes in a five party system. While another party could have a minority with 60%-epsilon.

I wonder if there could be a federal law that requires all districts in all national or state-wide elections to be large enough to send at least ten representatives, leaving it up to the states to choose their preferred mechanism for how voters will select those ten from a longer list.

Won’t change a thing as long as it is winner take all.

How would that work in this case? The state assigns one seat to the person who got the most votes and the other nine seats to whoever regardless of their vote counts? That’s not much of a democracy.

If it’s simply “top 10 candidates with most votes get the seats”, that would be a great improvement over gerrymandered districts already. Other states could opt for ranked voting, d’Hondt’s method for party-proportional representation, etc.

Because the whole point of messing with the districts is to make each district 50.01% party A 49.99% party B. A winner take all system with multiple candidates would work like the presidential election, you vote for them as a group.

If you switch to top 10 then it's not winner take all anymore.

More than half the states don't have 10 House members, so this is impossible without a massive Constitutional change.

Well, ten or the maximum available, in which case the whole state must be a single district. (States with 11-19 seats would also have to be a single district since they wouldn’t be allowed to create districts smaller than ten. The specific number 10 is of course arbitrary, I didn’t think about it at any length.)

that is entirely a self inflicted hurdle, the law used to be there was to be one representative for every 30,000 persons, but then in 1929 congress passed the Permanent Apportionment Act of 1929, fixing the number of representatives at 435. if we had stuck with the original we would have 11,080 representatives. technically speaking the only barrier to this is repealing that act, that and building a much larger capital building

True enough, it's not strictly a Constitutional problem (we don't need an amendment). But going from 435 to >10,000 House members is a massive change.

I've long though that pegging House districts at the population of the smallest state was a reasonable middle ground. That would put us at 571 members instead of 435.

The advocates of multi-member districts in the United States usually prioritize three- or five-member districts, or a mixture. Ten-member districts are unwieldy for voters, and the marginal increase in proportionality drops off quickly.

"M" is the 8th congressional district of New York, and the speckles are islands.

Some states require congressional districts to be contiguous and some do not, but in practice every or nearly every district is contiguous, up to a reasonable interpretation of "contiguous" -- allowing for things like islands or peninsulas. (I could not find any counterexamples.)

https://en.wikipedia.org/wiki/New_York's_8th_congressional_d...

Hilarious and sad.

Makes me wonder, is there a fairer, algorithmic way to carve up a geographical area into election districts so we can apply the same solution to every state?

Perhaps every district should be convex?

I don’t think you can tile convex shapes without leaving gaps.

sure you can, but you might not want to go that small.

(simple proof is to turn everything into triangles.)

Sure you can. For example, regular triangles, squares, and hexagons can all tile the plane, and they're all convex. There are an infinite number of sets of convex shapes which tile the plane.

Unless state boundaries suddenly become convex it's definitely impossible. Ignoring state boundaries you have the issue that you need approximately equal numbers of population. It's possible that you can still tile the plane with convex shapes, but I'm not quite sure how to approach a proof (or disproof).

That sounds overly restrictive (not because it is impossible, though). Maybe we can require the area of the convex hull to be at most xx% larger than the original area. But that will be hard to enforce when a shoreline is involved.

First you need to define the word 'fair'. Noting that you have to really define the word, and not just pick a simple interpretation that also happens to favour one party. Note that the US doesn't automatically do things based on a 51% majority (which, honestly, they picked a sensible policy - really 60%+ should be the norm for political action). So "this favours the group with the 51% majority!" isn't an argument of fairness.

Requiring 60%+ just means unfairly favoring the status quo over the will of the majority? :)

It's not really about favoring anything, but about preventing the degenerate case of democracy - a tyranny of the majority. It's the reason a president can veto congress, a judge can 'veto' a law both congress and the president agreed upon, or why Rhode Island has just as much influence as California in the Senate.

"Fairness" being seen as the same as to majority > all would just result in what we've seen happen with our democracy experiments in places like the Mideast. The majority group just ends up using democracy as a cudgel against minority groups. Those minority groups of course don't just stand idly by and next thing you know the entire system is in meltdown and beelining right towards the next iron fist.

You're right. Far better to have a tyranny of the minority instead.

Minority groups cannot pass laws by themselves. Their one and only "power" is to prevent the majority from being able to arbitrarily impose their will on everybody else. Of course the other side of this coin is that a minority group can prevent "positive" change from being enacted by a thin majority.

But in general this cost pales in comparison to the alternative system where by merit of having a simple majority, a majority group can impose their every whim and desire on the every minority group with absolutely no restraint. If you can solve both problems in some way, do feel free to share - and not only here. You'd have solved a core problem of political science!

> Minority groups cannot pass laws by themselves.

Yes, they can, especially where the combination of FPTP with gerrymandered districts, unequal allocation (by population) of seats (in all three of the Senate and, because of granularity limits, the House and, as a consequence of both of those, in the Electoral College) makes it possible for a minority with the right geographic distribution to have majority control of political institutions.

Right. Say a state has nine districts. Say the state favors one party 55% - 44%. How should the districts be drawn?

Should each district be as competitive as the state, so each district roughly works out to be 55% - 44%, knowing that by definition the state's districts would be 9-0 for one party?

Should you instead have five districts clearly for one party, and four districts clearly for the other, knowing that by definition each district would have a huge incumbency advantage?

A lot of people would say they want the districts to be both representative of the state's population, and competitive, without realizing those two principles are in tension.

Proportional representation solves this problem.

Each state has x districts drawn by an independent third party which must pass a certain threshold for compactness, and y seats not associated with a district.

Via any mechanism you choose (preference, ranked choice, FPP with parties, whatever) votes are assigned. Any voter in one of the x districts that does not get representation from their local member has their vote flow to one of the y seats.

The exact implementations have tradeoffs, but all are far better.

Proportional representation trades one set of problems for another, though.

If you look at the journey from citizen opinion to passed bill, there's a reconciliation process, where those opinions are sorted and filtered to consensus and final vote.

Without proportional representation, more of that process happens earlier. More responsibility given to voters to burden some of that compromise process. When people complain about "lesser of two evils" when thinking of candidates in primaries, or doing calculus on who has a better chance in the general election, that's an expression of that process.

With proportional representation, voters have less of that burden, and have a better chance of electing someone that is closer to exactly what they want. But the burden doesn't go away - instead, that burden increases in the deliberative body itself. More factions and more division in the House.

I'm not trying to represent one approach as better or worse - just saying it's a trade-off, and it isn't the case that proportional representation is "better". From the citizen voter's perspective, maybe, but from the citizen that cares about the final outcome via House bills, it's not so clear.

Those are both upsides. More debate and more different viewpoints means things that pass actually represent the people and that there is some transparency.

"Yeah sure, your representatives will be more aligned, but then the outcomes won't be along strict two party lines and will be more aligned with those representatives".

a pretty clear tenant of fairness is that it should be impossible to win the majority of people and lose the majority of seats

You'd think, right? But "real Americans"! But "coastal elites"!

I saw a paper on this a while back

https://www.math.cmu.edu/CNA/summer_institute/2018/projects/...

Yes! There has been a lot of scholarship on the topic,

https://www.rangevoting.org/GerryExamples.html

https://archive.siam.org/news/news.php?id=1237

https://news.cornell.edu/stories/2020/11/fairmandering-draws...

http://autoredistrict.org/

I've seen a study that a simple grid-based subdivision design that allocates proportionally equal parcels to each electoral division would be unfair... to Republicans. They would lose most elections horribly, so proposed legislation banning Gerrymandering gets shot down every time by them, on the grounds of unfairness.

Just a wierd thought - what if the grid system was done such that one party chooses a square, then the other, alternating? First choice is by coin toss. Maybe the area is required to be contiguous. Maybe not.

How would a party choose a square? They’re just trying to draw the squares to include favorable proportions of voters. If the squares are fixed, there is no choice?

I was thinking that there would be an overlay of squares, each identical, like graph paper. As long as the squares contain a relatively small population as compared to the rest of the state, then the two parties are forced to make real choices. And that's the real goal - for both partoes to make real choiced, and not to simply game the system via gerrymandering.

It's kind of a half-assed idea at this point. But I was thinking that, by forcing identically drawn boundaries, then it would force both parties to choose.

Now that I think about it more, it probably wouldn't work. Both parties would just grab squares that guaranteed their victory. It would end up ultra-gerrymandered.

Oh I think I see. A kind of adversarial map-drawing process. You are given a fine-grained grid ahead of time and then let the parties fight over how the squares will be merged together into districts. Each side tries to draw the map to favor themselves but is countered by the adversary. Neat idea.

What do you mean by "unfair" here? That more Republicans would lose?

I think the simplest definition of "fair" that most people would agree with (at least before being told who would benefit and the tribalism kicks in) would result in a distribution of representatives that roughly matches the distribution of the population at large.

That is, if the whole population of a state votes 60% for one party and 40% for the other, and there are a total of 10 representatives from that state, the fairest distribution would result in 6 representatives for one party and 4 for the other.

If by you saying it's "unfair" to Republicans by just having a final representative distribution that more closely matches the distribution of votes, then I'd argue that is indeed fairer and the current system is the one that's unfair.

I’d say that politicians — who really only truly represent the richest 1% of citizens — losing is perfectly fair. If they want to play to win, they should adopt policies with a broad enough appeal to win the hearts and minds of more voters.

I find it odd that rangevoting.org colors their maps with 5 colors.

Subject to someone always being unhappy, sure; see: Optimization Approaches to Political Redistricting Problems.

There are several commercial GIS packages for this and a wealth of papers on the subject.

The great issue, in the US, as I understand it, is that for some entirely batshit devoid of rational reason ruling political parties oversee new districts on a semi regular basis. This fosters a culture of redistricting for personal party gain and ratcheting inequalities that follow.

Saner countries involve apolitical blind consultants for initials district proposals and have party nased challenges follow "I cut, you choose" type patterns to resolve conflict.

The “I cut, you choose” sounds optimal from a game theoretic perspective and avoid all the complex algorithms that are proposed in the links of other replies here

Sorry, choose what exactly? (in the context of districts)

I'd argue for a "simpler" method: no district should be non-contiguous and it should be "round" (no weirdly shaped or long districts). Something like: area of the circle that fully encompasses the district cannot be more than 30% of the whole district area. (With some exceptions for coastal areas)

Essentially - the opposition has the veto power

The entire idea of a geographical area selecting just one person to represent their views in a legislature is outdated and undemocratic.

It tends to a two party system which inherently can't represent people's views accurately. In the US you see people talk about white and black districts as if all white people think the same.

A proportional system means everyone's views are represented in the legislature and at that point compromises can be brokered.

> The entire idea of a geographical area selecting just one person to represent their views in a legislature is outdated and undemocratic.

I don't think that alone makes it undemocratic. Many systems implementing proportional representation elect a single person per geographic area but also create an overall balance based on the proportional vote. See https://en.m.wikipedia.org/wiki/Leveling_seat for instance.

I feel like a fairer way to do this would be through proportional representation. Most issues that voters care about are not local issues, so why do need a system to carve up representation along geographical boundaries?

https://en.wikipedia.org/wiki/Proportional_representation

While I do think the current system has gone way too far in the other direction, the concern I'd see here is that for issues that _are_ of largely local concern there's no real voice.

> Most issues that voters care about are not local issues

Do you have an data to support this claim?

Indeed, most political science supports the opposite conclusion. Almost all politics is local.

politics is WAN when nuance is harder to find

> Indeed, most political science supports the opposite conclusion. Almost all politics is local.

Interesting. I'd say all the very toxic politics is non-local. The local stuff might be more, but it's mostly mundane.

Which makes sense. My city and county representatives can't get away with passing something wildly unpopular that >> 50% of local people hate. While it's far from perfect, there is a reasonable amount of accountability with local politicians.

Federal-level politicians though, are so far removed from the people they purportedly represent and their election is so heavily gerrymandered that they are untouchable. They can get away with serving only special interests and passing laws which most people hate.

How people vote (or for those who consistently vote one party no matter what, whether get nothet to show up to vote) is consistently due to local issues. Not necessarily local laws or measures mind you, but things impacting them in their local everyday life. This include having a job or the price of gas (incumbents who bring home pork contracts to their district do very well).

Because America is the greatest democracy in the world, which means the democratic system it was built on is the best in the world. Anyone who suggests otherwise hates America.

I'm not being serious, but at least in the UK one of the major defences of the British system whenever electoral reform is brought up is, "it's the oldest so it must be the best".

When NZ adopted MMP (proportional system similar to Germany's) both the major parties (National & Labour) campaigned against it on the "better the devil you know" basis, which the electorate see through. The first MMP election was mildly rubbish, with a populist kingmaker gaining disproportionate leverage, but since then the electorate has been shrewder and in a subsequent referendum MMP remained the preferred system.

> Most issues that voters care about are not local issues

If this is actually true (which I am not sure about), it is a defect, not something we should entertain.

> why do need a system to carve up representation along geographical boundaries

Because people in the same geographic areas tend to have correlated interests

> Most issues that voters care about are not local issues, so why do need a system to carve up representation along geographical boundaries?

That seems more like ratifying a failure mode rather than fixing it. Having geographic representatives mean that area is less likely to get railroaded in various ways, because they have an advocate who's accountable specifically to them, rather than some diffuse ideological block.

Also, a lot of issues that voters care about are actually local issues made national through centralization.

Proportional representation can be done at multiple scales. You can simply add more representatives if districts feel to big.

Because when you do have a local issue you need a local representative who is beholden to you and is more dependent on your vote.

I have an issue with a bullet ridden sign on the Erie canal that needs to be taken down. It's been that way for years and I'm tired of looking at it. I will first contact the canal corporation as it's their responsibility but if they do nothing the local assembly member for that district can make things happen.

It's funny - I only left the United States six years ago and yet this idea is already unimaginably weird to me.

Why is this your representative's job? It just baffles me. Why is this guy dealing with some sign?

The idea that government is broken so you rely on personal relationships to get around is fundamentally bad.

In the Netherlands, you would just take a picture of the problem and post it to the Gemeente (municipality) - in fact, they have an online map you can pin these complaints to, and it's public.

You get a case number, and you get updates until it's fixed, which is usually pretty fast.

If you aren't online, you can just walk in to the office and they'll help you do it.

It's his job because this sign is in a location where outside visitors will frequently see it and it sends a bad message.

> I feel like a fairer way to do this would be through proportional representation.

Very much so, and there is evidence that proportional systems produce, among modern democracies, better satisfaction of the electorate with government, richer political dialogue with more axes of variation.

> Most issues that voters care about are not local issues

That’s very much not true, but in the US system, more local issues tend to be the focus of more local governments (not always, but usually.) At the same time, local impacts of national issues are important and need representation, but proportional representation can be done in ways which respect that.

> Why do need a system to carve up representation along geographical boundaries?

Several reasons: representation of local concerns, accountability of individual representatives to the general electorate and not just a party, and providing individual citizens at least one representative that clearly and concretely represents them particularly, not just their party of preference.

But you can do PR with local district representatives. Two key methods which do this:

Single Transferrable Vote (STV; the multiwinner analog of Instant Runoff Voting/Ranked Choice Voting) provides proportional representation by district in multimember districts, which provides coarse-grained proportionality overall while providing local representation, and mitigating the impact of specific districting decisions, since its very hard for them to have any significant effect on partisan representation overall.

Mixed Member Proportional, in which people have two ballots for a legislative body, one for a district elections (which can be single-winner FPTP, multiwinner STV, or just about anything else), and one for a party. The winners of the district elections are elected, but in the same body enough additional seats are added, from party lists, to achieve proportionality (within the tolerances specified in the system) overall, by reference to the party votes in the second ballot.

STV is the most straightforward and the one that would probably be least problematic for US state or federal legislative elections. (Federally, it would require a change to statute to either allow or mandate states to use it, because there is a statutory single-member-district mandate currently.)

Living in an MMP country, I can say I think the party list system truly sucks, because you get career politicians that you can’t vote out, whom are essentially elected by themselves into the party.

I think one of the most important strengths of democracy is voting out people we don’t like, and MMP badly breaks that ability.

The argument for the list is when 5% of the population would vote for NZ First, and Winston Peters would get a seat because he put himself first on his party list. He would get over-representative power when a coalition government was required: in that specific situation he had a constituency. However in the larger parties (National, Labour) there are situations where people on their list get in on their party vote, even though the person is wildly disliked in the electorates, and would never get a constituency. During a past landslide Labour win, a bunch of nobodies wayyyy down the list (not expected to ever win) got in as members of parliament with zero constituency.

We don't have much gerrymandering in my jurisdiction; boundaries are chosen by an nonpartisan committee. But in theory the legislatures can override it. Sometimes part of the proposed boundary changes do seem illogical. For example, a town closely linked to a nearby city, but districted with the neighbouring rural area, that it has little in common with. Those situations are most likely to result in tweaking the boundaries. (Which sometimes leads to accusations of gerrymandering. That's how it starts, after all!)

It's very hard to quantify that sort of thing algorithmically, like what region a town feels it is part of. And I do think it is important to take into account; part of the whole reason for districts, is for local character to be considered and reflected in the democracy.

Sure but most of the energy goes into algorithms to achieve the most unfair results possible instead.

Probably the best way is to not have the districts in the first place. If a state has five representatives, run a single state-wide election with five winners, using some variant of proportional representation.

Even better: modify the fundamental nature of the House of Representatives such that a representative's voting power is proportional to the share of the vote they got.

Both of those are pretty major reforms, so absent those probably the best thing is to make sure the redistricting process is not controlled by a single party, and maybe imposing some restrictions on how complex the shapes can be.

> modify the fundamental nature of the House of Representatives such that a representative's voting power is proportional to the share of the vote they got.

Germany does that. They have districts, which get to send the winner in each district to parliament. But these only account for a part of the total seats available.

The others are filled up from a party list in such a way that the proportions are equal to the proportions of the votes.

That way you get the best of both worlds, you have representatives that have a connection to each district, but you still have the distribution in parliament that represent the actual distribution of the votes.

We do the same here in New Zealand. It's called MMP (Mixed Member Proportional). https://en.wikipedia.org/wiki/Mixed-member_proportional_repr...

Hungary does this too, and the party list seats are filled in a way that increases the influence of smaller parties. The first seat goes to the party with the most votes, but once that seat's assigned, the winning party's vote total is cut to 1/2. Then the next seat is assigned to whoever has the most votes afterwards, and the winner's total is cut to 1/2 if it's their first time winning, or 1/3 if it's their second time winning. And so on.

The Hungarian system is pretty much useless since the reform in 2013. Before 2013 party candidates that did not win the districts contributed their votes to the list vote totals. This is losers compensation. Since then we have a winners' compensation: winners also contribute their extra district votes (the difference of their votes and the second place candidate's) to the list totals. This exacerbates gerrymandering and pretty much guarantees the 2/3s majority needed for amending the constitution to Viktor Orbán's Fidesz party with their 45%-55% popular vote.

What you're describing seems to be the D'Hondt method which favors larger parties, not smaller parties.

We have a similar system for a Scottish Parliament elections (Additional Member System). But it is still FPTP for the Westminster elections.

Than who advocate for you? Having a local representative s an essential part of getting government by to work for you.

Have a look at Mixed Member Proportional: https://en.wikipedia.org/wiki/Mixed-member_proportional_repr...

I consider it a feature that Congress doesn’t map directly to party voting proportions.

This is smug nonsense. In a PR system there will be more than 2 parties, and Americans that love their "they all suck" badge of honor will find something worth identifying with.

Moving towards pluralism would be the number one way in which the US could rejuvenate their democracy.

Absolutely!

That was unnecessarily dismissive, as I was being quite sincere.

It is a feature, not a bug, that we elect an individual rather than a party. It is likewise a benefit that our distribution of representatives is weighted somewhat by geography and (indirectly) economics rather than purely by population. Should flyover country count less than coastal areas because farming automation has reduced agricultural jobs per acre compared with LA or NYC? Yes, certainly, but to what degree is the question. The industrial and agricultural importance of these areas means that we are better off when we give them larger representation per-capita than they might otherwise expect due to their meagre populations.

The current system is not without faults, such as gerrymandered districts and the various factors that contribute to us having a rigid two-party system. I suspect a lot of people on HN would rather throw it all out and have strict proportional voting and/or equal-population districts. However that would be throwing out the baby with the bathwater.

Are there districts that are low on people, industry, and agriculture that get representation based purely on their size?

The Senate

I really still degree on all of that

> It is a feature, not a bug, that we elect an individual rather than a party

Not it is not. Politics is "team sport". Vetting individuals (look at Sinema's bleeding heart eyeroll to pure cravenness transformation) is hard. Vetting organizations is much easier.

> It is likewise a benefit that our distribution of representatives is weighted somewhat by geography and (indirectly) economics rather than purely by population. Should flyover country count less than coastal areas because farming automation has reduced agricultural jobs per acre compared with LA or NYC? Yes, certainly, but to what degree is the question. The industrial and agricultural importance of these areas means that we are better off when we give them larger representation per-capita than they might otherwise expect due to their meagre populations.

This is ridiculous. This is not the 19th century; Industry and Agriculture are big enterprises that are perfectly capable of being run on a national scale. If you want to promote them, do so based on the national interest with industrial policy, https://en.wikipedia.org/wiki/Industrial_policy. Land-based representation promotes backwards romanticism, not well-engineered modernity.

The Senate giving us stupid things like corn ethanol is not a policy success of unfair representation!

> However that would be throwing out the baby with the bathwater.

Do you really think tons of parliamentary systems around the world are hurting because they through out some sort of "baby" by adoption fair representation?

Why?

I live in a country where I pick a party to represent me in the national government, not a local person, and it seems to work just fine.

Scaling - you live in a country the size of California with a quarter of that states population ( ~10 million ).

Large countries have different issues in areas that for you would be two or three whole European countries distant, and countries with larger populations have many multiples of your entire population to represent.

Those problems should be solved locally and often have no right to be influential on the national level. The fact that some areas are heavily invested in fossil fuel shouldn't cause the nation to stop a transition to renewables, but that's what seems to be happening a lot in the US.

In stead, at a national level, a transition can be started and budget can be created to support the states that will be negatively affected. The way that budget is handled locally will then depend on local politicians who can be voted on and need to show that they're actually doing the right thing locally.

Right now, national level politicians can get reelected locally because they're just obstructionists that want nothing to ever change. They'll never actually have to do anything except fund raise and pretend to care about their constituents by blocking anything those constituents don't like. It's a great way for a couple of people to make lots of money while killing the country.

If I am having some problem with the federal government, I can call up my local representative and their office will help me sort it out. If it's a situation where I'm being treated unfairly or not according to the law, a sternly written letter from the representative usually sorts things out. If it is a case of the law being unjust, they might get the law changed, if it's not a politically difficult problem.

You might find it surprising that this actually happens in a country of 330M people and just 435 representatives, but it does. In part because only a small percentage of people need this service, but everyone benefits from having it as a contingency plan. Helping a single constituent can pay big dividends at election time as they tell their friends, family, and social media about your help. Likewise, not helping can cause great harm as you alienate your voter base.

Then you end up with representatives that represent dense areas, and the rural folks don’t like that.

Almost as if rural folk don't actually care about democracy

Almost as if rural folk have VERY different needs than urban ones and having their needs be unmet due to being the minority.

The problem is, urban areas have grown so massive that they can just screw over rural areas by sheer numbers - in Germany, 77% of voters live in urban or agglomeration areas, that's way blowing past the usual 2/3rds-majority requirements.

At that point, who is to prevent parties from solely catering to urban area voters at the complete expense of rural voters? It's already seen as problematic that most public investments like infrastructure get directed towards urban areas, the term "flyover states" exists for a reason and we've been seeing for the last decade just how damaging that can be for a democracy.

[1] https://www.deutschland.de/de/topic/leben/stadt-und-land-fak...

Don't forget state elections! Tons of state legislatures are affected by gerrymandering, too. It's frustrating that my city's representation is cracked and packed away.

Some constitutional clause prohibiting representatives to rewrite the rules of their own impending elections in bullshit ways sounds like a pretty big oversight.

I'm no constitutional scholar though - maybe such a thing was debated

A rule I'd propose for this would be:

For every administrative region within a state (county, parish, city, village, borough, ward, etc.), a proposed congressional district must meet one of the following criteria:

1. The congressional district is completely contained within that region. 2. The congressional district completely contains that region. or 3. The congressional district contains no part of that region.

Unless all the regions are both simply nested instead of partially overlapping and extremely conveniently sized with regard to population, this will at a minimum pretty consistently violate the Constitutional “one man, one vote” case law, and may often also be intractable.

In most countries you just hire some officials as part of an independent agency and tell them to create reasonable districts based on geographic and social groupings, and it works fine.

I'm not sure if the problem is that the polticians are afraid of giving up the power to control the districts, or if the extremely adversarial nature of American politics means that Americans mistrust their goverment and their fellow citizens so much they don't believe that the possibility of that government making non-partisan decisions even exists.

> In most countries...

How do you calculate that? (not joking: i don't know where one looks for such data)

Well, yes, I don't know, I'm only going off never having heard of gerrymandering being a major problem in any other country.

Plus that in general America seems to involve politics in decisions which are more usually left to officials to do in an, at least theoretically, non-partisan way. I don't know of any other country in which judges are elected by popular vote, or Ambassadors are appointed and approved by politicians instead of being professional government officials(this one I am pretty sure is unique).

it's easy to find it elsewhere (i.e., American politicians aren't the only ones)

The second link below is "fun": seems gerrymandering in France correlates to the French revolution.

https://www.thecanary.co/uk/analysis/2021/10/21/the-tories-a...

https://www.washingtonpost.com/outlook/2019/07/14/what-frenc...

I think it’s two things: 1. What you said, politicians have little incentive to give up power on gerrymandering. 2. The American political system is designed to be adversarial, and adding an “independent” agency with that much power ~feels~ like it invites distrust and corruption.

In Australia Gerrymandering isn't even part of political discourse. An independent committee composed of the Electoral Commissioner, the state or territory Surveyor-General and the state or territory Auditor-General make the decisions. They may invite submissions from the public and political parties but they do not pander to politicians. The boundaries also generally make sense.

This is an active area of research for sure. One idea is basically to come up with a few rules for what the districts must look like (as unions of precincts) and use Markov chain Monte Carlo to sample from a probability distribution on the possible districting plans. Several court cases have featured expert evidence from math researchers asked to quantify how crazy the districting is using this methodology (ish). Wesley Pegden's expert witness report in one such case is readily findable, for example.

One of the main difficulties is that we don't have much rigorous theory telling us how long to run the Markov chains.

Not even a little bit. The reasonable thing to do would be to toss out the old system and have everyone cast a ballot for a representative and weight congressional seats by how many votes you receive, with some kind of thresholding system to keep from having to rebuild the capitol all the time. IE representatives below the threshold can delegate or amalgamate their votes.

Yes, when gerrymandering was challenged in 2019 a bunch of mathematicians filed an amicus brief with a mathematical fairness test. Their work is here: https://mggg.org/

Other tests and models have also been proposed in the challenges to gerrymandering. The supreme court just had to pick one. But what the court ruled in 2019, along strict partisan lines, is this:

The Court ruled that while partisan gerrymandering may be "incompatible with democratic principles", the federal courts cannot review such allegations, as they present nonjusticiable political questions outside the remit of these courts.

https://en.m.wikipedia.org/wiki/Rucho_v._Common_Cause

So just like how the supreme court ruled in citizens united that it wouldn’t put any limits on money in politics, it ruled in this case that it wouldn’t prevent politicians from redistricting in their favor. The supreme court is imho part of the problem of American politics, not part of the solution.

Translation: get the "wrong" unlimited money and the "wrong" party gerrymandering and they might get off their behinds and change their opinions

Now as for the SC being part of the problem, well, maybe it's an issue of assignments and term limits, but it is not as simple as not having one

The correct legal decisions were made in both cases.

In Rucho,

> Roberts made clear that partisan gerrymandering can be distasteful and unjust, but that states and Congress have the ability to pass laws to curb excessive partisan gerrymandering.

In other words, it's up to the legislatures to ban gerrymandering first before SCOTUS can rule on it.

As for Citizens United, the law granted the FEC to pick and choose who was allowed to exercise political speech:

> The Chief Justice wrote separately to explain why the Court went as far as it did. Essentially, Roberts explained what I’ve written in this article: that there was no rhyme or reason anymore to which entities were allowed to influence elections and which weren’t. According to the government, Michael Moore can but Citizens United couldn’t. The New York Times (a corporation) can but other corporations couldn’t. There were so many holes and exceptions to the law organizations had been essentially reduced to asking the government for permission to criticize it, and that is contrary to the essential purpose of the First Amendment. No more narrow exceptions given to one group but not another, the whole thing must go.

https://www.cato.org/commentary/michael-moore-trumpland-migh...

They even went as far as to claim the authority to ban books:

> A close reading of the transcript, however, reveals that General Kagan did not change the government’s answer. She suggests that someone who had their book banned by the FEC would have a “quite good as-applied challenge,” and suggests (wrongly, it turns out) that the FEC has never pursued books in any enforcement issue.

> This is not the same, however, as saying flatly that under no circumstances does the government have the authority, under the Constitution and the First Amendment, to ban books for political content. Instead, Kagan asserts only that the statute in question (203) does not currently cover books, while conceding that another (441b) does cover books — but she promises they won’t use it to go after books. Promise, pinky-swear!

(yes, that Kagan as well in case you were wondering)

https://www.ifs.org/blog/citizens-united-its-all-about-the-b...

> Roberts made clear that partisan gerrymandering can be distasteful and unjust, but that states and Congress have the ability to pass laws to curb excessive partisan gerrymandering.

In other words, it's up to the legislatures to ban gerrymandering first before SCOTUS can rule on it.

...so the very legislatures that practice gerrymandering to make sure their party stays in power should, after a sudden revelation that this practice is distasteful and unjust, decide to forbid themselves from doing it? Yeah, that sounds very likely!

It's admittedly unlikely, but the 17th Amendment which required direct election of senators was ratified by state legislatures who previously had that power.

In this case, Congress would have the power to do so if state legislatures will not (and for that matter, some states have).

> In other words, it's up to the legislatures to ban gerrymandering first before SCOTUS can rule on it.

You're saying that SCOTUS shouldn't act until the people rigging elections in their favour... Voluntarily stop rigging elections in their favour?

I'm not sure if that argument is just nakedly partisan, or is simply one that doesn't make a lick of sense.

It's right up there with 'criminals should preside over their own trials' and 'the watchmen should watch themselves'.

It feels like a similar interpretation of checks and balances as Nixon's belief that the President cannot commit a crime.

I'm saying that, like it or not, there's simply no legal basis for SCOTUS to ban gerrymandering if gerrymandering isn't illegal.

Yeah, the check on the legislature for this kind of thing isn't the courts but the Constitution. Maybe we'll see an amendment in our lifetime.

I don't think an amendment would be necessary. Congress has the power to do so without it.

Why would the legislatures that gerrymandered themselves in amend the constitution to ban this practice?

Waiting for a constitutional amendment on anything in this century is like waiting for Santa Claus. He's not coming.

What you’re saying is true, but it’s not exactly the end of the story in our post Roe world. SCOTUS has proven they are willing to completely reverse their decisions absent a change in fact or law. If nothing changes except the partisan makeup of the court, then their decisions look partisan when they are reversed.

So when SCOTUS says they are bound by the law and Constitution to not make certain decisions, a lot of people now rightly ask: are they really? Because it seems that they can move with great alacrity to twist the law when it suits the political aims of their party.

There was nothing “correct” about either decision.

They may have correctly followed a legally justifiable theory that led to their decision, but there were also other equally, if not more, legally justifiable theories they could have followed to reach the opposite decision.

For example, a theory that would incorporate the idea that a badly acting legislature cannot be the only one to depend on to curb the badly acting legislature when the bad behavior’s goal and effect is to shape the very mechanism that provides the only check on their behavior, ie voting.

You go from "pedantic, but hard to argue against" to "lost the plot and are more wrong" when you say "there were also other equally, if not more, legally justifiable theories" and then list a much harder to arrive at and less legally defensable reason that is also unrelated to the actual facts of the case...

The best criticism of the ruling, imo, is that it should have been decided under the freedom of press clause. The result was correct, but if it was decided under that clause it could have avoided the campaign contribution implications.

> but there were also other equally, if not more, legally justifiable theories they could have followed to reach the opposite decision.

Only what the law actually says is justifiable in law.

> For example, a theory that would incorporate the idea that a badly acting legislature cannot be the only one to depend on to curb the badly acting legislature when the bad behavior’s goal and effect is to shape the very mechanism that provides the only check on their behavior, ie voting.

On what legal basis is this based on?

The fact these cases made it to SCOTUS basically ensures there were multiple legal theories that arguably correct. If there was only one true interpretation, the cases would have been solved by a lower court. Then SCOTUS itself was split (along partisan lines no less).

Flag burning made it to SCOTUS too, and was a 5-4 decision as well.

And there were reasonable arguments on both sides. And there were 2 dissents and 1 concurrence filed, so even within the Court, there were different interpretations, even among Justices who agreed (on each side).

No, I don't believe any argument that burning a flag isn't protected speech is reasonable.

Well then you have 4 supreme court justices you need to argue with. What is the law of the land and what is moral don't always line up. All we get is the hope that our system tends toward morality over time.

I don't need to argue with them because flag burning is already protected speech and the justices are all dead by now.

Nearly 40 years later, it may seem to be a settled question.

But then we have Roe, which was settled, until it wasn't.

> The fact these cases made it to SCOTUS basically ensures there were multiple legal theories that arguably correct.

That doesn't seem to support the idea that

has to be one of those theories.

> > but that states and Congress have the ability to pass laws to curb excessive partisan gerrymandering.

> In other words, it's up to the legislatures to ban gerrymandering first before SCOTUS can rule on it.

The gerrymandered legislatures, you mean?

The idea that the states can cheat in elections how they will, even federal elections, and there is no recourse - do you not see that even if this has the color of law, it is in fact _wildly unjust_?

> As for Citizens United, the law granted the FEC to pick and choose

Given the FEC's extremely limited budget, the fact that they can only enforce a subset of violations is completely necessary.

This is true of all agencies, police organizations and the like.

---

These two are crappy decisions that are bad for the country and bad for the average citizen.

The people who defend them never do so on the basis of justice or fairness, because on that matter they are indefensible. In their hearts, they think, "Our evil side will be able to win and win and win forever if we set the rules to 'Cheat Mode'".

> do you not see that even if this has the color of law, it is in fact _wildly unjust_?

It's not good, and it's certainly not ideal, but it's much worse for the courts to be usurping powers they don't have.

> Given the FEC's extremely limited budget, the fact that they can only enforce a subset of violations is completely necessary.

> This is true of all agencies, police organizations and the like.

The FEC specifically gave exemptions selectively to certain organisations. Not 'they didn't have the resources to pursue this' but 'they said it was ok for Michael Moore to do this but not Citizens United'.

> In their hearts, they think, "Our evil side will be able to win and win and win forever if we set the rules to 'Cheat Mode'".

I'm so glad you're able to glimpse into my heart of hearts and determine what I think for myself.

> The idea that the states can cheat in elections how they will, even federal elections, and there is no recourse

This is false. The recourse, and the correct course of action, is for state citizens to be shown convincing evidence that gerrymandering is unjust and bad (which isn't hard to provide) and elect better representatives. Representatives elected via gerrymandered elections still do actually have to be voted for by members of those districts, and so a sufficiently large grassroots effort will likely succeed.

> Given the FEC's extremely limited budget, the fact that they can only enforce a subset of violations is completely necessary.

This is selective enforcement. I hope you agree that that's bad.

> These two are crappy decisions that are bad for the country and bad for the average citizen.

No, they're decisions that are consistent with the constitutional system of the United States. They have undesirable outcomes because of the laziness of voters, but they are principled and correct.

Subversion of the tripartite system of the executive, legislative, and judicial branches of government to try to obtain a better outcome is obviously unsustainable and extremely abusable - a clearly terrible idea, advocated for only by those with a partisan desire for some specific changes to be made, and not because the justices are actually doing a bad job (so far).

Not a perfect solution, but Michigan created a bipartisan commission to redraw districts in 2020. (Well Michigan didn't create it, voters did in a proposal during the 2018 election.) https://www.michigan.gov/micrc

yes. and we need not even reach for the mathematical perfection readily available.

Growing up a member of large semi feral child population there was one simple rule that kept the peace.

       "DIVIDER GETS LAST CHOICE."

Lots of good responses to this question.

The most plausible solution being pushed right now is the one advocated by Lee Drutman and Fix Our House, proportional representation via large multi-member districts:

https://www.fixourhouse.org/

It's more plausible than others because it doesn't require amending the U.S. Constitution. I'm not convinced you couldn't still gerrymander a bit to tilt this system one way or the other, but it's way more robust than what we have now.

My ideal system has the same basic design, a combination of local and proportional representation, and would look like this:

(1) Redistrict algorithmically using a simple algorithm based on population density alone, minimizing a relatively easy-to-understand objective function (e.g., average distance of people to other people in their district). The goal of this algorithm is not to be "best" or "fairest" in any sense, but to be robust to attacks like John Roberts's frightening line about "sociological gobbledygook." Once you have subjective measures of fairness, you open things up to political maneuvering (and accusations thereof).

(2) Assign multiple members of Congress to each district. An election for a district involves picking n candidates via instant runoff voting.

(3) Members don't get 1 vote. They get a vote proportional to the fraction of people that voted for them in the election.

That way, if you have a 4-member district, with the top 4 vote-getters being a Democrat with 45% of the last-round vote, a Republican with 43%, a Libertarian with 7%, and a Green with 5%, they each get 0.45, 0.43, 0.07, and 0.05 of a vote, respectively.

(4) Do this for the Senate too, ha ha ha. Senate still serves a purpose in this fantasy: it's a longer-time average of public opinion with less time spent running for re-election.

Proportional voting system makes gerrymandering (almost completely) pointless.

The "almost" is there only because there's probably some threshold below which votes are wasted - and the wasted votes need to be reassigned to the other candidates somehow, which makes it not exactly proportional. But in practice it's not a big enough incentive to gerrymander.

The problem with democracy is that any bugs in the process are how people who can fix them are elected, so there's negative incentive to fix them and positive incentive to abuse them even more. I'm of the opinion that a well-functioning democracy needs revolution every century to fix the accumulated trash.

> is there a fairer, algorithmic way to carve up a geographical area into election districts so we can apply the same solution to every state?

I would argue to throw out the constraint that each district must contain the same number of people. There is no reason for that other than the fact that "one man, one vote" sounds good when you hear it. It makes just as much sense as applying the same constraint to senators.

Representatives are supposed to represent a common interest group, not Democracy In The Abstract. If two people are a part of the same community with the same interests, why wouldn't they share a representative? The right approach is to identify natural districts first and then not care whether they're the same size.

There's a really good series of podcasts from fivethirtyeight.com about this:

https://fivethirtyeight.com/tag/the-gerrymandering-project/

The short version: Everyone agrees that gerrymandering is wrong and bad, but it's not 100% clear what the "right" result is. Is the right result proportional representation? Is the right result having more competitive districts? And what about minority influence?

One thing they brought up which used to make the whole thing more complicated is that some civil rights laws also make requirements (or used to make requirements anyway) about the makeup of districts: namely, that if you could, you had to make a district which gave Blacks a significant influence in at least one district. But doing that requires "packing", which is one of the pair of techniques (along with "cracking") used in a partisan gerrymander.

Lots of states have recently implemented a variety of responses to gerrymandering; in a few years it should be clear which approach(es) work best.

The whole system is utterly broken. It benefits a few at the top, while heavily disenfranchising the 99%. It's literally a race to the bottom, chasing a "free market" that will never truly exist, and would cause immense damage to the world (even the highly regulated and captured "free market" we claim exists nowadays is accelerating our collective demise). "Everything not forbidden is compulsory."

Who decided to call it Ugly Gerry and not Sans Democracy?

Sans Sherrif

Hellvotica. Garamondering. Lucida Election.

That is a great question!

IMHO, an important consequence of gerrymandering is that it makes districts less competitive. Instead of a district where neither party has clear grip (so the winning candidate needs a broader appeal), we have one which is either clearly R or clearly D. Now a candidate no longer has to care about the other team, because once your own party approves your candidacy, you are as good as elected. So party loyalty is what matters.

The result is a funnel to select fanatics over moderates, from both sides.

I don't think that's an unexpected consequence. That's the intention. Rope off the folks who won't vote for your party into one district and you get to win the rest of the districts. Or run the lines so that your opposition's support is broken up and sprinkled harmlessly amongst your supporters so that their votes are swallowed up by your supporters' votes. Then you've got no reason to actually work to improve anyone's lives and can spend your time figuring out how to leverage your position to make as much money as possible.

The Wikipedia article has a good graphic that helps visualize the strategy.

For bonus points, devise ways to make it more difficult for the demographics that are likely to vote for your opponent to get to the polls.

Suggest you get involved with local politics. That’s not how it works. Otherwise nearly every gerrymandered seat would be occupied by an extremist.

Exactly, all these one party districts, and states, are not great for democracies. Basically, you get institutionalized corruption, nepotism, inefficiencies, etc. Doesn't matter if it's red or blue. It's bad either way. Swapping out parties once in a while is a good way to shake things up and break that pattern.

Is there any worse gerrymander than the Hudson Yards of NYC that ROBBED the poor to fund the new construction?

Wow, I didn't know about that until I read your comment so googled for more info. That's really just disgusting. More details: https://www.bloomberg.com/news/articles/2019-04-12/the-visa-...

It's fitting, since the word itself is also very, very ugly. It _feels_ like a word for a 50-year-old-something guy walking in the street letting his arms bounce wildly.

It's as ugly as shenanigans, flamboyant, gentrification, astroturfing, tomboy and such.

Jerrymeandering?

I dislike the focus on district shapes in gerrymandering discussions. The thing that matters is how severely the districting process subverts the will of the electorate, and that can be done with squares instead of squiggles.

Squiggly lines do not necessarily imply a gerrymander. California is unfairly on this list even though it’s districts are drawn by and independent bipartisan commission that attempts (generally successfully) to give an even balance to both political parties as well as different ethnic groups.

>to give an even balance to both political parties

Hows that a positive?

I don't think that I am understanding the objection. The idea is that the legislature should roughly match the popular vote. Having a legislature that proportionately represents the demographics and opinions of the populace is essential for a meaningful democracy what is the problem?

"both political parties"

That's a big part of the problem right there. We've shoehorned US politics into two silos that constantly steal all the feed from the much smaller independent silos.

Not necessarily. It would take a pretty tremendous gerrymander to give the Greens (as an example) a seat in the state legislature. At the point where they held a meaningful percent of the vote we could easily incorporate their voters into this framework. The simple truth is most people support one of the two main parties (independents are about 1/3 of voters but these are generally aligned between the two major parties than any alternative). You could take issue with the legislature being allowed to localities instead of parties. The scheme used by the European parliament is nice for representing more fringe political orientations but it neglects the immediate material interests of different regions. Given that most of what the state legislature does is allocate resources to different localities, the ideological issues are probably secondary to the basic questions like who gets water, or funding for roads.

I live in a permanently Democrat state and we've got the worst Gerrymandering I've ever seen, anyhere.

“The documents Stephanie released eventually showed that her father intentionally gerrymandered North Carolina by manipulating voting districts to benefit his political party and disenfranchise the other, and that he advised the Trump Administration to add a citizenship question to the 2020 Census to help discriminate against people of color for redistricting purposes.”

https://ncpolicywatch.com/2019/12/20/the-gerrymanderers-daug...

Can you be more specific?

Gerrymandering seems to be thought of as something Republicans do.

And the name of the state is ....

Illinois is known for extremely gerrymandered pro-democrat districts, that also seem designed to minimize minority impact within the party. They've cleaned up some of the shapes so it's less obvious recently.

Thanks. I was able to look it up. It's quite obvious. Looks like a pinwheel

https://gerrymander.princeton.edu/redistricting-report-card?...

Their other maps are less egregious

It really feels like a crime to do this. It's pretty wild that's legal

Probably Maryland. It's one of the worst in the country.

https://gerrymander.princeton.edu/redistricting-report-card/...

Both the major parties do it to their advantages. Doesn't really matter what it's "thought of", the reality is that they're both equally guilty.

Dan Crenshaw's in Texas is pretty bad - it curves around Houston, staying in predominantly white, upper class areas, avoiding the poorer areas in the middle.

https://res.cloudinary.com/graham-media-group/image/fetch/f_...

You could say the same about Sheila Jackson Lee's district

Yes, Lee's much district actually sits inside of the hole in Houston I was referring to. If Texas districted both of them properly, Crenshaw would get a lot of Lee's district, and Lee would get a lot of his. Would definitely be more equitable.

Same. My district was just gerrymandered this election cycle. Didn't even bother voting, what's the point if the government can just change the lines if they don't like the results. It's evil whether R or D does it.

Yes, like many of the bad things in American politics, they both do it.

Relevant podcast: America's hidden duopoly.

https://freakonomics.com/podcast/americas-hidden-duopoly_rad...

Both the Democrat and Republican parties like the system as it stands, because it guarantees nobody else can join the table.

If only we could make gerrymandering illegal.

It's hard to make it illegal, because under the current system someone needs to draw the lines somewhere, and as long as they can reason why those lines are in those places, it's hard to prove malice.

Attempts to negate gerrymandering typically focus on making sure that those deciding the lines are as unbiased as possible, usually with cross-party committees drawing the lines instead of elected officials.

The best option though is to remove the need to draw the lines at all.

In theory a third party could spring up with the singular goal of fixing democracy by doing away with gerrymandering, instituting proportional representation, and removing overreaching executive powers and probably do well enough to get some movement towards those ends from the other parties if nothing else. They could all promise to resign and hold new elections once their goals are met and to not vote on anything else. Throw federally legalized pot in there if they really want to clinch support.

Lawrence Lessig attempted this, raising over $1M dollars with a kickstarter campaign. So successful that only polysci nerds remember.

https://abcnews.go.com/Politics/harvard-professor-larry-less...

Lessig was going to step down and put his Vice President in power, who would explicitly be a Democrat.

That's not going to get broad appeal even if the whole effort was marketed correctly.

And I'm not talking about just president, but congress as well. Get as many people as possible into as many offices as possible with this goal.

Perfect mix of geek and politics

It is now my system font

I promise

When one finds themselves holding the short end of the demographic stick, it is time for some creative thinking...

Great submission. I was just thinking the other day "you know what needs to be highly politicised and discussed on HN? Fonts!".

Reply on news.ycombinator.com