Legislative Districts and the Legacy of Elbridge Gerry

The Sandbox Political Topics
1

I’m weird and have the apparently odd notion that geography should matter less in legislative representation.

For example, I could see House representation being based on superdistricts represented by up to 5 congresspeople, with voters only getting to vote for a single individual (as opposed to voting for “up to 5”).

I’m not a fan of the “one per 30k” because of how ridiculously unweildy an 11,333 member House would be, but I wouldn’t hate formal adoption of the cube root standard (approximately 698 House members).

1 Like
2

Sounds awful. With only 1 vote, why waste it on the candidate sure to be elected instead of helping someone who is close? Unless too many people think that way.

Superdistricts for 5, each voter voting for up to 5 makes much more sense.

And ranked choice voting makes the most sense, super district or single candidate district.

3

1 person can’t adequately represent 500k people

Having a tiny number of representatives doesn’t work well with a 2-party system

Maybe if we had more of a parliamentary system with the 2 parties forced to split into 3 parties each (a left center right version of each), so that everyone is basically a 3rd party

Haven’t decided yet how I feel about ranked choice voting

4

That’s fine.
Given the increasing unity of parties, you could have one vote for a party, then reps are from party lists. That’s “proportional”.

Or, do 80% of the seats by one-member districts, then the remainder from party lists to “correct” to the proportional ratio. “Mixed Member Proportional”, first used in Germany.

5

And, if a rep is only 1 in 11,000, that rep doesn’t have enough clout to “adequately represent” anybody.

At that point, why not go all the way and use referendums?

1 Like
6

“Ssuperdistrict electing n delegates, voters can vote for up to n” style voting was previously extremely common in the south as a means to block non-white (and non-Democratic) city/county council members.

When the US DOJ started supervising voting in the south in the wake of the Voting Rights Act, they disallowed this style of voting because of the effective disenfranchisement of African American voters.

In theory, in a (5-representative superdistrict, vote for 1) model, once the parties learn appropriate strategy, the system should stabilize such that the representation looks somewhat like the voting public as a whole, while still preserving the geographic linkage that so many Americans seem to find important.

Two big problems with single-representative districts in a politically polarized society are:

  • A sizeable minority of the population, approaching 50%, are effectively disenfranchised in the sense that they have no realistic chance of “their” preferred candidate ever being elected, and the elected individuals only need to cater to the interests of members of the majority party in their home district.
  • The system creates an incentive to gerrymander so as to maximize the power of the majority party, increasing their power in a particular state beyond what actual voter mix would suggest.

RCV/IRV reduce the problem…but I don’t think they eliminate it.

There are other systems of legislative representation that would probably work. For example, I hadn’t been aware of how members of the German Bundestag were elected until their recent election (a blend of single-representative constituencies, and then additional members selected from party lists to approximate vote results)…and that feels interesting, although I haven’t had the time/interest to fully game out how it would work / how it could be abused in a climate as polarized as the US’s.

FWIW, my opinions on the subject are informed by my not being a fan of either the Democrats or the Republicans. (I used to use the “small ell libertarian” label before some folks started using the “libertarian” label as cover for some pretty horrible stances.) In any single-representative district system, I am unlikely to ever be represented by a representative that I like.

In an RCV/IRV system, there’d be an improvement in the sense that candidates elected would have an incentive to represent a broader segment of their district’s voting public, and that would be better…but I still would likely not be represented by a good advocate for my views.

But in a superdistrict system, or in some nongeographic system of selecting representatives… I might actually have a shot at being represented.

7

Hopefully we’ll get to change the rules after the violent return to power. As it stands there will never again be a meaningful federal election. All results that would shift control/power away from the R’s will be challenged and suppressed. After the midterms no matter the actual vote counts there will be an increase in R’s in the house and senate and a decrease in D’s. Authoritarians do not give up power willingly.

What or who would stop the R’s from not recognizing elected D’s back into the congress after the election. They don’t have to arrest all of the D’s on trumped up charges, just a few.

1 Like
8

Other countries do this in a non-partisan manner: the US is exceptional in its gerrymandering. For example, Canada has a non-political process:

“Redistricting in Canada, also known as the redistribution of federal electoral districts, is the process of reviewing and adjusting electoral boundaries to reflect population changes after each decennial (10-year) census. This ensures that each electoral district has a roughly equal number of voters, promoting fair representation.

Here’s a breakdown of the process:

1. Initial Seat Allocation:

*** The Chief Electoral Officer first recalculates the number of seats each province is entitled to based on population data from the census.**
*** This calculation involves using a representation formula outlined in the Constitution.**

2. Independent Commissions:

*** Ten independent, provincial electoral boundaries commissions are then established to redraw the electoral district boundaries within their respective provinces.**
*** Each commission consists of three members: a judge who chairs the commission and two members appointed by the Speaker of the House of Commons.**

3. Boundary Review and Adjustment:

*** The commissions review the existing boundaries, taking into account factors like population distribution, community of interest, geographic size, and historical patterns.**
*** They consult with the public and Members of Parliament (MPs) throughout the process.**
*** The commissions ultimately determine the new boundaries for their province.**

4. Implementation:

*** The new electoral boundaries come into effect when a new Parliament is called at least seven months after the boundaries have been confirmed.**
*** Elections Canada provides administrative and technical support to the commissions throughout the process.**

Key Considerations:

*** Voter Equity: Ensuring that each electoral district has a similar number of voters.**
*** Communities of Interest: Recognizing and respecting the unique characteristics and connections within communities.**
*** Geographic Size: Considering the practical challenges of representing large geographic areas.”**

1 Like
9

We can’t help it.

It is in our DNA to politicize everything.

3 Likes
10

Have you seen the stats on the gerrymandering in Illinois and California? The GOP gets 40-44% of the vote in those states and holds 17% of the House seats.

I don’t like gerrymandering. It allows people with extreme views to hold too many seats and is horrible for the country. But I find it incredibly hypocritical that Newsom and Pritzker are acting like its high treason that Texas is trying to move to a map that is the equivalent of their states, just in the opposite direction.

11

Illinois absolutely is a jurisdiction that has embraced gerrymandering.

California, however, with the passage of Prop 20 several years ago, has the “California Citizens Redistricting Commission” draw the Congressional districts. That’s a group of 5 professed D’s, 5 professed R’s, and 4 professed non-partisans, hired through a rather bureaucratic process that’s supposed to ensure they’re qualified and independent of the politicians.

It’s not impossible that the CCRC would gerrymander (in the conversational sense anyway; in some definitions of the term, the use of an independent body precludes the result from being a gerrymander)…but I think in California the disconnect between vote tallies and representation is more an artifact of the problem of having representation by single-representative geographic districts, than it is gerrymandering.

1 Like
12

That is not necessarily an indication of gerrymandering. If you randomly assigned voters to districts (in some kind of electronic voting system), you would end up republicans about 40-44% of each district, and they would likely win less than 17%.

Of course if you tried to make the districts as compact as possible (I doubt there is a clear enough definition that would create a “best” map for compactness), you would get some districts very heavily democratic and some very heavily republican. Whether the republicans would be favored to get more than 17%, I don’t know.

The Texas situation is especially flagrant given that it is an unusual (timing) move and that the stated purpose of the extraordinary redrawing is to change the results of the next election. The major new information available is how precincts voted more recently.

1 Like
13

Optimal compactness is achieved when the aggregate length of the inter-district boundaries is minimized.

1 Like
14

To illustrate this point:

Assume a state with 10 representatives, where the voting population is made up of (60% people who vote Orange Party) and (40% who vote Yellow Party).

In a perfectly fair system, you would expect the state to have 6 Orange representatives, and 4 Yellow representatives.

Assume also:

  • Representatives are elected on a single-representative geographic district basis; and
  • Every household in the state is comprised of 3 Orange-preferring voters and 2 Yellow-preferring voters.

With those two assumptions in place, the state will have 10 Orange representatives no matter how the district boundaries are drawn.

The outcome in this extreme example is not due to gerrymandering; it’s the defect of the underlying architecture.

1 Like
15

Totally agree. Your “first past the post” system produces especially skewed representation in the US since you only have two major parties. If you had a few more popular parties you might get somewhat more representative results but obviously still not as fair as in some form of proportional representation. You could end up with coalition governments but those are often quite productive.

The US will never significantly change its election system as any change would be branded as communistic by the GOP.

1 Like
16

Is that a provable theorem or a general rule?
Note that for redistricting, we have the constraint of equal populations.

17

I’d frame it as: “The US will probably never change its election system as any change would be disadvantageous to the two major parties, and there is no effective mechanism to override the parties at a national level.”

1 Like
18

I’d think that it’s fallout from the definition of “compact” (or maybe “a definition of ‘compact’”?)

Put another way: given multiple options for district boundaries, drawn to meet other criteria (equal population, support of proportionate voting power for protected classes), the most compact option will be the one where the length of the inter-district boundaries is less than the other options.

Or maybe it’s “district boundaries”, not “inter-district boundaries”… I haven’t thought through that complication.

(I have spent WAY too much time thinking about the subject. I suppose that’s the downside of being a map geek, a politics geek, and an algorithm geek.)

EDIT: It is “district boundaries”, rather than “inter-district boundaries”. If you don’t include the external boundaries, you’ll have a greater chance of a circular outer district in the map that wins the algorithm.

19

Yes. For example, another definition of compact is “minimize sum of all the distances between the center of population in each district and each of the residents”.

Given the fact that people aren’t uniformly distributed, I think those two definitions can give different results.

Another possible constraint on the map would be that district borders follow existing features such as roads, city/county borders, and geographic dividers (rivers). That complicates things because boundaries may not be straight lines.

20

Yes. Probably the best way to see/game this would be in an extreme example of:

  • A circular state that is being divided up into n districts, and no need to consider internal geography or protected classes.
  • At the center of the state is a circular city that is densely populated
  • The surrounding area is sparsely populated
  • The residents of the city are uniformly distributed across the space of the city.
  • The residents of the surrounding area are uniformly distributed across the space of the surrounding area.
  • The surrounding area has 1/n-th the population of the state.
  • The surrounding area is comprised of “different” voters than the city’s voters.

In a “minimize sum of all the distances between the center of population in each district and each of the residents” system, I think you end up with the state having been divided into n wedges of equal size.

In a “minimize length of inter-district boundaries” system, I think the city would be divided up into (n-1) wedges, and the surrounding area would be its own district.

I’m not intuitively certain what the result would be in a “minimize the length of the district boundaries” system. I have a bad feeling the question is going to stick in my brain and annoy me until I have time to actually work it out.

I do think that in this example, however, a casual observer would say that the “inter-district” result would be “fairer” than the “sum of the distance” result.