If we had double or triple the number of representatives in the House, gerrymandering would be a lot tougher.
Would that affect the actual shapes, though?
That’s probably why the new NY-8 in Brooklyn and Queens looks like a bong.
Elections Canada comes pretty damned close.
The elecotral boundaries here don’t seme to be an issue; they are for the most part a pretty honest attempt to group up the right number of people in a geographically logical way. There is nothing even remotely resembling the ridiculous gerrymanders you see in the USA which are so openly, transcendently cheating that I can’t believe anyone with a conscience could be involved in it.
Hmm I never actually thought about that before… Let’s increase it to the constitutionally allowable maximum amount. Doesn’t that work out to be like 30,000 congressmen?
I’d like to see states gerrymander when they have to draw up several thousand districts!
It’s actually pretty simple. You let the majority party or whoever the law requires make up district maps. Then you have a poll of 1000 randomly selected voters - you show them an outline of each district (you don’t show them the map, just a solid color outline) and ask them if it looks gerrymandered. If any district doesn’t pass this test (let’s say 66% polled have to agree it isn’t gerrymandered), the map goes back to revision. Rinse, repeat, until all districts pass.
But what about my plan of having 30,000 congressmen??? How cool would that be!
(Your idea is pretty awesome I suppose)
Why not 100,000. Or a cool 1M. One congressman per 100 voters or so! Gerrymander THAT.
Because, believe it or not, there is a clause in the constitution that sets an upper limit on how many congressmen there can be in the Union. And if I remember correctly, with the current american population that works out to about 30,000.
But I’m cool with amending the constitution to allow for a million congressmen.
“The number of Representatives shall not exceed one for every thirty Thousand”.
That works out to upper limit of about 10,000 representatives. Piddly number.
No, actually, simple geometry is enough. An impartial algorithm might occasionally split up an urban area, but it won’t be often enough to be a real problem unless you’re systematically trying to do that. And if you do systematically try to do that, under the minimum-proposed-boundary system, then some politician from the other party is going to have an easy time in pulling the rug out from under you.
And for that matter, a minimum-proposed-boundary system probably won’t split up urban areas, anyway, except in the obvious case where the urban area is larger than the population of a single district. An urban area is by definition a place where the people are closer together, and so you can cordon off a given number of them with a smaller boundary, contained entirely in that urban area. If you have an urban area that’s about the right population to be its own congressional district, the shortest boundary map will almost always just make it one.
Assume that Liberal City, pop. 300,000, is in the center of the State of Rockrib, pop. 1 million, entitled to four districts. What prevents the Conservatives from creating pie-slice districts—each of them compact squares or triangles—that split the residents of Liberal City up to be small portions (75,000 each) of Conservative districts that are 250,000 each? Now nearly a third of the state’s population is unable to elect one of its own as a representative.
Define “one of its own”.
Someone who represents their views and interests?
… and how do you determine that? By voting? Then you have to group the voters somehow. So - what is the objective criteria you suggest to group “them” to elect someone who represents “them”?
However you grouped them - let’s say the representative is elected with 55% vote. Does he represent the 45%'s “views and interests”? If you say yes, explain why they didn’t vote for him in the first place. And if you say no, then there you go - that 45% is not represented by “one of its own”. Unfair?
I think the problem is nobody can agree on what the goal is.
Suppose you have two political parties: A and B. And you’ve got a state that’s entitled to five congressional districts and in which 40% of the people belong to Party A and 60% belong to Party B.
Our state looks like this:
AAAAA
AAAAA
BBBBB
BBBBB
BBBBB
If you divide the state into north-to-south strips, you get five strips that look like this:
A
A
B
B
B
Each strip has the same proportions of party members as the state has. But when you have the next election, every candidate for Party B will win their elections because each district has a B majority.
Or do you divide the state in east-to-west strips? You’d get two strips that look like this - AAAAA - and three strips that look like this - BBBBB. In the next election, you’d end up with two Congressmen for Party A and three from Party B.
So which system is more “fair”? The system where each district reflects the make-up of the state or the system where the outcome of the election reflects the make-up of the state?
Given a typical (uneven-density, variegated) distribution, no mere slicing-in-strips seems appropriate. Small population centers should be whole in their own districts; large population centers should be split into the minimum number of pieces, wherever possible using the boundaries that tend to actually lead to distinctions in interests: fixed sub-state jurisdictional boundaries (like county lines), and geographic boundaries. The pie-wedging of the OP’s city is obviously going to divide people of similar interests in the vitality of their city, to no purpose except to divide them. Where it is necessary to group rural areas with population centers, territory should be linked to the area it is closest to, with a raw-distance radius modified by those natural or established boundaries.
There may be quibbling about the relative weighting of all these factors, but as a general proposition it should be possible to have CDs of almost precisely equal population drawn by computer, with no regard whatever, in each iteration, for party affiliations.
If you disagree with the principle of compact, geographically-based districts on the grounds that they might disenfranchise some minority group (by party or otherwise), it seems that you should be arguing for an entirely different system–at-large districts, multiple-representative districts, party list voting, proportional representation, something along those lines.
But this doesn’t address the central question. Should districting be aimed at the goal of making each district as representative of the state as possible (with a mixture of different groups that exist in the state present in each district in the same proportions that they’re present in the state) or should districting be aimed at the goal of making each district as homogeneous as possible (with the aim of gathering together like-minded voters so the greatest number will feel their elected representatives represent them)?
So I guess the 10,000 congressmen idea just isn’t working for you guys then?
Don’t the Canadians have a system for drawing up “ridings” that is generally considered fair and nonpartisan? How does that work?