Partisan gain vs. communities of interest

[Senator Napoleon]

Obviously, what is and is not a CoI is debatable, so go by your personal definition.

[muon2]

Neither.

I prefer a map that utilizes geography rather than sociology, but is tempered by political reality. The typical delegation elected from a map should reflect the state as a whole but be subject to change reflecting the will of the voters.

[BigSkyBob]

This is a giant example of the false alternative fallacy. The notion that each state has exactly the same number of "communities of interest," whatever that means, as districts that are equally populated is a joke.

[Senator Napoleon]

This is a giant example of the false alternative fallacy. The notion that each state has exactly the same number of "communities of interest," whatever that means, as districts that are equally populated is a joke.

Genius, that wasn't stated or even implied.

[Torie]

The problem of course is that if your state is going to be fairly drawn, where your party can do a gerry, than the same needs to obtain by an offsetting amount in one or more other states, where the other party can do a gerry. Unilateral disarmament is rarely a good idea. Muon2 is right in his implication at least, that in the end, on a national basis, what one would want is that the share of the pie each party gets to reasonably approximate the popular vote share, with a healthy number of CD's that will switch as the nation swings one way or the other.

[Senator Sbane]

The problem of course is that if your state is going to be fairly drawn, where your party can do a gerry, than the same needs to obtain by an offsetting amount in one or more other states, where the other party can do a gerry. Unilateral disarmament is rarely a good idea. Muon2 is right in his implication at least, that in the end, on a national basis, what one would want is that the share of the pie each party gets to reasonably approximate the popular vote share, with a healthy number of CD's that will switch as the nation swings one way or the other.

Exactly. While a fair map is preferable, you don't want to unilaterally disarm. This is why some sort of a constitutional amendment taking it out of the arms of the legislature across the country would be preferable. I would go even farther and take it out of the hands of elected officials totally. If not, we end up with Washington and Arizona.

[BigSkyBob]

This is a giant example of the false alternative fallacy. The notion that each state has exactly the same number of "communities of interest," whatever that means, as districts that are equally populated is a joke.

Genius, that wasn't stated or even implied.

It was certainly was implied. Absent that implication, you question makes absolutely no sense. You are asserting a dichotomy between objectively fair, and objectively unfair, redistricting. For you to assert objective COI-based redistricting exists there has to be an objectively "fair" way to redistrict any particular state based on "communities of interest," whatever that means.

"Communities of interest" is a buzz word with no particular meaning. It could mean racial gerrymandering: lumping each race into own districts so those voters share the values and outlooks of that race. It could mean bipartisan gerrymandering: create as many uncompetitive districts as possible so that most voters live in districts were their party's candidate wins. It could mean segregation by income. In the real world a "community of interest" is an ex post facto rationalization for taking a favored redistricting choice.

What is particularly galling about your initial query is that I've read numerous theories of what constitutes objectively "fair" redistricting: respect for existing boundaries, county lines, city limits, etc.; continuity of existing districts, aka "least change;" theories about "compactness;" meta-level theories about "fair" partisan distribution of seats;  meta-level theories about "fair" racial distribution of seats; "political choice" theories that claim that voters indirectly vote on maps by electing legislators; and theories about "communities of interest," whatever that means. For you to assert that either you subscribe to the last theory, or you favor "partisan gerrymandering"  is an giant example of the false alternative fallacy.

[krazen1211]

I prefer legislative redistricting subject to rules concerning geography. Specific ones and not bull words like 'compact'.

[BigSkyBob]

The problem of course is that if your state is going to be fairly drawn, where your party can do a gerry, than the same needs to obtain by an offsetting amount in one or more other states, where the other party can do a gerry. Unilateral disarmament is rarely a good idea. Muon2 is right in his implication at least, that in the end, on a national basis, what one would want is that the share of the pie each party gets to reasonably approximate the popular vote share, with a healthy number of CD's that will switch as the nation swings one way or the other.

Exactly. While a fair map is preferable, you don't want to unilaterally disarm.

That makes sense only to elected legislators whom value their party's interest more than the interest of their constituents. If they think some commission scheme is preferable to the current system they ought to have the courage of their convictions to pass it. If "gerrymandering" is bad for the folks in their own state, they should oppose it.

[BigSkyBob]

I prefer legislative redistricting subject to rules concerning geography. Specific ones and not bull words like 'compact'.

"Community of interest" makes  words like "compact" seem precise.

[Vasall des Midas]

The problem of course is that if your state is going to be fairly drawn, where your party can do a gerry, than the same needs to obtain by an offsetting amount in one or more other states, where the other party can do a gerry. Unilateral disarmament is rarely a good idea. Muon2 is right in his implication at least, that in the end, on a national basis, what one would want is that the share of the pie each party gets to reasonably approximate the popular vote share, with a healthy number of CD's that will switch as the nation swings one way or the other.

Exactly. While a fair map is preferable, you don't want to unilaterally disarm.

That makes sense only to elected legislators whom value their party's interest more than the interest of their constituents. If they think some commission scheme is preferable to the current system they ought to have the courage of their convictions to pass it. If "gerrymandering" is bad for the folks in their own state, they should oppose it.

I agree with Bob.

[Senator Napoleon]

This is a giant example of the false alternative fallacy. The notion that each state has exactly the same number of "communities of interest," whatever that means, as districts that are equally populated is a joke.

Genius, that wasn't stated or even implied.

It was certainly was implied. Absent that implication, you question makes absolutely no sense. You are asserting a dichotomy between objectively fair, and objectively unfair, redistricting. For you to assert objective COI-based redistricting exists there has to be an objectively "fair" way to redistrict any particular state based on "communities of interest," whatever that means.

"Communities of interest" is a buzz word with no particular meaning. It could mean racial gerrymandering: lumping each race into own districts so those voters share the values and outlooks of that race. It could mean bipartisan gerrymandering: create as many uncompetitive districts as possible so that most voters live in districts were their party's candidate wins. It could mean segregation by income. In the real world a "community of interest" is an ex post facto rationalization for taking a favored redistricting choice.

Read the first post, genius.

[Vazdul (Formerly Chairman of the Communist Party of Ontario)]

The problem of course is that if your state is going to be fairly drawn, where your party can do a gerry, than the same needs to obtain by an offsetting amount in one or more other states, where the other party can do a gerry. Unilateral disarmament is rarely a good idea. Muon2 is right in his implication at least, that in the end, on a national basis, what one would want is that the share of the pie each party gets to reasonably approximate the popular vote share, with a healthy number of CD's that will switch as the nation swings one way or the other.

Exactly. While a fair map is preferable, you don't want to unilaterally disarm.

That makes sense only to elected legislators whom value their party's interest more than the interest of their constituents. If they think some commission scheme is preferable to the current system they ought to have the courage of their convictions to pass it. If "gerrymandering" is bad for the folks in their own state, they should oppose it.

The problem there is that gerrymandering is bad for the folks in their own state, but it's good for their own self-interest and the interests of their party, so they won't oppose it. This is because the gerrymander allows them to do things that are not in the interests of their constituents without the fear of suffering the consequences at the next election. For this reason, there must be very strict rules regarding how redistricting is performed, preferably federally mandated rules so that what is prohibited in one state isn't allowed in another. "Communities of interest" absolutely must be part of the equation, but the term needs to be well defined.

[lowtech redneck]

This is because the gerrymander allows them to do things that are not in the interests of their constituents without the fear of suffering the consequences at the next election.

No, the gerrymander allows politicians to choose their constituents; they are still accountable to the constituents they have chosen.

[krazen1211]

I prefer legislative redistricting subject to rules concerning geography. Specific ones and not bull words like 'compact'.

"Community of interest" makes  words like "compact" seem precise.

At least by 1 definition, it is not. Current law defines people as a community of interest merely based on race, which is at least well defined.

[Bacon King]

Automatic redistricting by computer. Something like this, perhaps, except based on travel time rather than geographic distance, and maybe using a population-weighted district center rather than a simple geographic center. Or abolish districts entirely and go to PR or something.

From the lesser of two evils presented above, though, I pick the one that isn't partisan gerrymandering.

[Vazdul (Formerly Chairman of the Communist Party of Ontario)]

This is because the gerrymander allows them to do things that are not in the interests of their constituents without the fear of suffering the consequences at the next election.

No, the gerrymander allows politicians to choose their constituents; they are still accountable to the constituents they have chosen.

While this is technically true, in practice the constituents usually don't hold politicians accountable in non-competitive districts.

Automatic redistricting by computer. Something like this, perhaps, except based on travel time rather than geographic distance, and maybe using a population-weighted district center rather than a simple geographic center.

No. Just no. Sure, those districts look much prettier than anything that a state legislature would come up with, but they absolutely destroy communities of interest, just as if the map were gerrymandered.

I'd really like to see the maps on the page you linked with county boundaries overlaid, so we can see exactly how much algorithms such as this screw urban areas. Just a cursory glance at the Kentucky map on the first page is enough to see that Louisville is split right down the middle, with both halves put in districts dominated by rural interests. It's harder to tell, but it appears that Lexington gets the same treatment. The Republicans couldn't do a better job gerrymandering the state if they tried.

In the end, there has to be some human discretion involved with redistricting, otherwise you end up with travesties like the one you linked to.

[freepcrusher]

i think there should be basic rules like no double crossing county lines and whatnot. You can still do gerrymanders but at least its fun that way because it challenges one to do it and it is more subtle. I could probably upload the 80s CA gerrymander, where the districts respect county lines but within the counties (particularly LA) is where a lot of interesting linedrawing is.

[muon2]

i think there should be basic rules like no double crossing county lines and whatnot. You can still do gerrymanders but at least its fun that way because it challenges one to do it and it is more subtle. I could probably upload the 80s CA gerrymander, where the districts respect county lines but within the counties (particularly LA) is where a lot of interesting linedrawing is.

That's why states with populous counties need municipal and even pre-defined community areas (eg wards in OH) to provide delineation beyond constraints of the VRA.

[Bacon King]

Automatic redistricting by computer. Something like this, perhaps, except based on travel time rather than geographic distance, and maybe using a population-weighted district center rather than a simple geographic center.

No. Just no. Sure, those districts look much prettier than anything that a state legislature would come up with, but they absolutely destroy communities of interest, just as if the map were gerrymandered.

I'd really like to see the maps on the page you linked with county boundaries overlaid, so we can see exactly how much algorithms such as this screw urban areas. Just a cursory glance at the Kentucky map on the first page is enough to see that Louisville is split right down the middle, with both halves put in districts dominated by rural interests. It's harder to tell, but it appears that Lexington gets the same treatment. The Republicans couldn't do a better job gerrymandering the state if they tried.

In the end, there has to be some human discretion involved with redistricting, otherwise you end up with travesties like the one you linked to.

I agree with you regarding the shortcomings of that guy's existing process; minimizing travel time to district centers weighted by population is certainly preferable to minimizing geographic distance to geographic district centers, as it would ensure that urban areas aren't split like that (and would also make sure that you don't see stupid results like the New Hampshire districts, or the weird dangle of the red Mississippi district, for example).

And I mean, Iowa manages to use computer redistricting just fine I think an algorithm, finely-tuned enough, can produce a map more fairly than inherently biased humans.

[TheDeadFlagBlues]

Neither, I believe in PR without constituencies.

[Χahar]

Neither, I believe in PR without constituencies.

But then you get legislators who represent nothing but their parties.

[Vasall des Midas]

Neither, I believe in PR without constituencies.

But then you get legislators who represent nothing but their parties.

Not if they're open lists... and the candidates actually campaign independently.

[Χahar]

Neither, I believe in PR without constituencies.

But then you get legislators who represent nothing but their parties.

Not if they're open lists... and the candidates actually campaign independently.

It doesn't feel like that would be possible in a legislature of any significant size.

[TheDeadFlagBlues]

Neither, I believe in PR without constituencies.

But then you get legislators who represent nothing but their parties.

My semi-solution is to create huge regional constituencies to accommodate for cultural variation.