Redistricting Reform:

[hopper]

People complain about gerrymandered districts so what is the best way to do redistricting reform? What rules should be put in place for a result of a non-gerrymandered fair redistricting map?

[Flake]

One that requires the most competitive districts possible, and approved by both chambers, the Governor, and the major state parties (Dems and Reps for every state but Vermont, Dems, Reps, and Progressives for Vermont)

[morgieb]

The only rule is that the maps should be drawn by non-partisan officials. I don't like drawing boundaries for racial reasons, those maps are just as ugly as partisan gerrymander. I'd want some community of interest in maps (so ideally keep cities whole/make several districts out of the metropolis), and the maps ideally should look easy on the eye. Like racial boundaries, competitiveness and applying to the state's partisan lean shouldn't be relevant if it makes the map into an effective gerrymander.

[Southern Senator North Carolina Yankee]

The courts have produced some rather good maps when the legislatures have failed to act. Colorado isn't too bad, New York came out much better than expected and upstate was nearly perfect.

Not all legislatures are horrific though. In some states, particularly the small ones with whole counties requirements the results, like IA and WV, were fine. Whereas at the same time the independent commission flubbed Arizona. I suppose the strength of the criteria is more important than who is doing it, but in most states a commission is better than the legislature.

[Hatman 🍁]

Non partisan boundary commissions with public input. The Canadian model is pretty good.

[muon2]

The courts have produced some rather good maps when the legislatures have failed to act. Colorado isn't too bad, New York came out much better than expected and upstate was nearly perfect.

Not all legislatures are horrific though. In some states, particularly the small ones with whole counties requirements the results, like IA and WV, were fine. Whereas at the same time the independent commission flubbed Arizona. I suppose the strength of the criteria is more important than who is doing it, but in most states a commission is better than the legislature.

These two points are the core of the issue. A balanced set of objective criteria can heavily constrain political gerrymandering. Some legislatures have strong traditions of fair application of criteria like IA, but others have failed. MI is an instructive example of that failure.

From 1970-90 divided government in MI led to a special master, Bernie Apol, overseeing redistricting. In the late 90's he codified his rules and the legislature adopted them as statute. Though his maps were regarded as fair, the legislature failed to realize the loopholes when they passed them into law. The result is that those same neutral criteria in the hands of a partisan legislature in 2001 and 2011 created plans skewed towards one party. The good rules were not robust against those who had clear political goals. That's where an independent commission can do better.

However, I believe that sufficiently robust rules can produce fair maps, and yet allow some participation by the legislature which represents the whole of the state. IA is probably the best model of fair interaction between a legislature and neutral criteria for a map. IA uses a hybrid of a neutral body to draw the map subject to constraints in statute with the ability for the legislature to veto the map based on specific criteria that the legislature finds lacking, again governed by statute.

[hopper]

The courts have produced some rather good maps when the legislatures have failed to act. Colorado isn't too bad, New York came out much better than expected and upstate was nearly perfect.

Not all legislatures are horrific though. In some states, particularly the small ones with whole counties requirements the results, like IA and WV, were fine. Whereas at the same time the independent commission flubbed Arizona. I suppose the strength of the criteria is more important than who is doing it, but in most states a commission is better than the legislature.

These two points are the core of the issue. A balanced set of objective criteria can heavily constrain political gerrymandering. Some legislatures have strong traditions of fair application of criteria like IA, but others have failed. MI is an instructive example of that failure.

From 1970-90 divided government in MI led to a special master, Bernie Apol, overseeing redistricting. In the late 90's he codified his rules and the legislature adopted them as statute. Though his maps were regarded as fair, the legislature failed to realize the loopholes when they passed them into law. The result is that those same neutral criteria in the hands of a partisan legislature in 2001 and 2011 created plans skewed towards one party. The good rules were not robust against those who had clear political goals. That's where an independent commission can do better.

However, I believe that sufficiently robust rules can produce fair maps, and yet allow some participation by the legislature which represents the whole of the state. IA is probably the best model of fair interaction between a legislature and neutral criteria for a map. IA uses a hybrid of a neutral body to draw the map subject to constraints in statute with the ability for the legislature to veto the map based on specific criteria that the legislature finds lacking, again governed by statute.

Yeah I looked at the Michigan US Congressional Map from the decade of the 90's and the district lines were drawn pretty good.

[muon2]

The courts have produced some rather good maps when the legislatures have failed to act. Colorado isn't too bad, New York came out much better than expected and upstate was nearly perfect.

Not all legislatures are horrific though. In some states, particularly the small ones with whole counties requirements the results, like IA and WV, were fine. Whereas at the same time the independent commission flubbed Arizona. I suppose the strength of the criteria is more important than who is doing it, but in most states a commission is better than the legislature.

These two points are the core of the issue. A balanced set of objective criteria can heavily constrain political gerrymandering. Some legislatures have strong traditions of fair application of criteria like IA, but others have failed. MI is an instructive example of that failure.

From 1970-90 divided government in MI led to a special master, Bernie Apol, overseeing redistricting. In the late 90's he codified his rules and the legislature adopted them as statute. Though his maps were regarded as fair, the legislature failed to realize the loopholes when they passed them into law. The result is that those same neutral criteria in the hands of a partisan legislature in 2001 and 2011 created plans skewed towards one party. The good rules were not robust against those who had clear political goals. That's where an independent commission can do better.

However, I believe that sufficiently robust rules can produce fair maps, and yet allow some participation by the legislature which represents the whole of the state. IA is probably the best model of fair interaction between a legislature and neutral criteria for a map. IA uses a hybrid of a neutral body to draw the map subject to constraints in statute with the ability for the legislature to veto the map based on specific criteria that the legislature finds lacking, again governed by statute.

Yeah I looked at the Michigan US Congressional Map from the decade of the 90's and the district lines were drawn pretty good.

It shows that Apol was using implicit criteria to make that 1990 map. His explicit criteria were not enough to prevent the districts in the current map. One challenge is to ferret out implicit criteria used to make good maps and make those criteria explicit. If you have followed some of my many threads over the last couple of years you'll see that is a major effort of mine. Feel free to weigh in on the recent New England thread abut which maps look good and which look gerrymandered.

[politicallefty]

While it isn't perfect, I wouldn't mind seeing the California model (or something very similar) be adopted throughout the country. It requires a commission with 14 members (5D, 5R, 4I) and at least nine votes for action (at least 3D, 3R, and 3I).

Legislatures should have absolutely no part in drawing districts. In my mind, the only thing worse than partisan redistricting is incumbent protection. Any redistricting reform must eliminate partisan bias and incumbent protection.

[muon2]

While it isn't perfect, I wouldn't mind seeing the California model (or something very similar) be adopted throughout the country. It requires a commission with 14 members (5D, 5R, 5I) and at least nine votes for action (at least 3D, 3R, and 3I).

The CA (and AZ) model could be improved by a better set of criteria. Many of the criteria are left as "touchy-feely" items that the commission can define on the fly. If commissioners have strong opinions, particularly on non-political criteria, they can sway votes to reflect their bias. One area of tension in commission criteria is in the use of political data. If one wants competitive districts one must use that data, yet that data can be manipulated to create a partisan bias (see AZ). I don't mind a commission system, as long as they have objective criteria that are defined before they start mapping.

You must be logged in to read this quote.

What makes the IA system work is that the legislature doesn't draw the map, but instead chooses whether to accept a map drawn by a neutral body according to statutory criteria. The result is that there is no incumbent protection, because the mapping body doesn't have addresses or other political data. Yet the legislature can veto a map under narrow circumstances when they see local effects that are unrecognized by the mapping body. It's the advantage a legislature has by having membership from all parts of a state, something that a commission like CA has lacks.

[Orser67]

Seems like what we should maybe do is implement several systems across several states, and see which ones tend to work best. The only thing that's clear to me is that you don't want to give legislatures (and governors) the unrestrained ability to create any kind of map that they want.

[ilikeverin]

The correct answer is to create a muonocracy!

[publicunofficial]

I think the California Commission-style works fine, or court drawn maps in some cases (The Nevada and New York maps are pretty close to perfect)


I'm not a fan of the Arizona-style, which focuses too much on the voting patterns of areas rather than whether communities should be in a district together. I also don't like the Washington/New Jersey system that only serves to shore up incumbents.

[RedSLC]

The best answer would be to have maps drawn by independent commissions or courts. Anything but partisan bodies.

I've thought of some proposed rules, too:
-Counties that cannot house at least one entire district cannot be split between more than two districts.
-Incorporated communities and CDPs cannot be split unless they either cross county lines (in which case they can only be split along the county border), or are too big to fit entirely in one district (in which case, they should have a maximum number of districts that fit inside them). In some cases, this may require precincts to be redrawn as well, to prevent them from splitting communities.
-For states that divide counties into townships, townships cannot be split unless they are too big to fit completely within one district. This applies to New England towns as well.

[muon2]

The best answer would be to have maps drawn by independent commissions or courts. Anything but partisan bodies.

I've thought of some proposed rules, too:
-Counties that cannot house at least one entire district cannot be split between more than two districts.
-Incorporated communities and CDPs cannot be split unless they either cross county lines (in which case they can only be split along the county border), or are too big to fit entirely in one district (in which case, they should have a maximum number of districts that fit inside them). In some cases, this may require precincts to be redrawn as well, to prevent them from splitting communities.
-For states that divide counties into townships, townships cannot be split unless they are too big to fit completely within one district. This applies to New England towns as well.

Great, let me ask some follow up questions, if I may.

In states that have both townships and communities, and they don't match up, which should have preference?

Should a district be permitted to have a strange shape just to keep counties intact?

Would you want to keep metro areas generally as intact as possible?

If a metro area with a large central city can be covered by two districts, is it better to keep the city intact and have a donut shaped district wrapping around it, or should the city be split some to keep the districts more compact?

Should there be an effort to maintain minority voting strength in a district even if it is a little below the threshold where the VRA requires a minority district?

[Sol]

SLValleyman, there's a bit of an issue with your 1st and 3rd criteria, in that sometimes it's necessary to split an extra county, even if it isn't too big for a district. In TN, for example, it's necessary to split one extra county in the Nashville area.

[RedSLC]

The best answer would be to have maps drawn by independent commissions or courts. Anything but partisan bodies.

I've thought of some proposed rules, too:
-Counties that cannot house at least one entire district cannot be split between more than two districts.
-Incorporated communities and CDPs cannot be split unless they either cross county lines (in which case they can only be split along the county border), or are too big to fit entirely in one district (in which case, they should have a maximum number of districts that fit inside them). In some cases, this may require precincts to be redrawn as well, to prevent them from splitting communities.
-For states that divide counties into townships, townships cannot be split unless they are too big to fit completely within one district. This applies to New England towns as well.

Great, let me ask some follow up questions, if I may.

In states that have both townships and communities, and they don't match up, which should have preference?

Should a district be permitted to have a strange shape just to keep counties intact?

Would you want to keep metro areas generally as intact as possible?

If a metro area with a large central city can be covered by two districts, is it better to keep the city intact and have a donut shaped district wrapping around it, or should the city be split some to keep the districts more compact?

Should there be an effort to maintain minority voting strength in a district even if it is a little below the threshold where the VRA requires a minority district?

After giving this some thought, here are my responses to each of your questions:

-Instances of this that I can think of would be if a CDP straddles multiple townships. If this is the case, since the township has local government, and the CDP doesn't, the township should be given priority. I've also considered situations where the arrangement of cities and/or townships makes avoiding splits difficult (i.e. multiple exclaves, or enclave communities surrounded entirely by a larger one, such as St. Louis City and County). If this is the case I would be OK with considering the county/community and its enclaves a single entity.

-I think that districts should reasonably match up with population density patterns (though I concede that it's rather difficult to have a concrete definition), with Statistical Areas anchoring districts (more in the next answer). Whole counties should be examined, but county splits can be permissible if it means preserving metro areas. The whole county Kansas map you posted in the map thread could be seen as an example of how whole counties are not necessarily always the best option (the map splits up Wyandotte and Johnson Counties, for example.)

-I think that Metro areas (as well as other Primary Statistical areas) should serve as "anchors" for districts, containing one or more areas and the surrounding counties. At the very least, districts should avoid splitting the counties of the metro area that contain parts of the actual urban area, unless it is too big to be contained in one district, or if the county does not contain a significant portion of the urbanized area.

-Personally, I think that in that specific situation, a "donut hole" would be a better fit, as it allows both the central city and its suburbs to have increased voting power in picking the representative of their choice. But if the metro area is similar sized, only with a smaller central city, I would be okay with splitting the central county, as long as the communities themselves are not split.

-If the minority group is concentrated enough to have its own compact minority-majority district, then I would say yes.

I hope this can answer some of your questions.

[muon2]

Great, let me ask some follow up questions, if I may.

In states that have both townships and communities, and they don't match up, which should have preference?

-Instances of this that I can think of would be if a CDP straddles multiple townships. If this is the case, since the township has local government, and the CDP doesn't, the township should be given priority. I've also considered situations where the arrangement of cities and/or townships makes avoiding splits difficult (i.e. multiple exclaves, or enclave communities surrounded entirely by a larger one, such as St. Louis City and County). If this is the case I would be OK with considering the county/community and its enclaves a single entity.

I asked this question because in some states like IL, there are both townships with government functions, and municipalities with their own separate government. Elections follow precincts which follow townships, but residents identify with the municipality which provides more services than the township. Municipalities tend to have very irregular boundaries based on unrelated acts of annexation and leave unannexed parcels behind, while townships are regular in shape and cover all the land in a county (with a few exceptions). Given these competing factors, would you give preference to townships, municipalities, or both equally?

You must be logged in to read this quote.

-If the minority group is concentrated enough to have its own compact minority-majority district, then I would say yes.
[/quote]

This question was based on situations where a single minority group is less than 50% of the voting age population in a compact district, so it doesn't qualify for automatic protection. Yet there can be a strong case for a community of interest that might significantly affect elections if kept intact (think of rural blacks in the south). If the population is under 50% but close, a minority can elect a representative of choice by forming a coalition with other minorities or whites who share their political views. Coalition districts are permitted, but not required, so the question is do you think that coalition districts should be encouraged when they don't overly conflict with other redistricting principles?

[YL]

Non partisan boundary commissions with public input. The Canadian model is pretty good.

This, though sometimes the non-partisan commissions have some pretty bizarre ideas...

[muon2]

Non partisan boundary commissions with public input. The Canadian model is pretty good.

This, though sometimes the non-partisan commissions have some pretty bizarre ideas...

Their leap to Hale and Ditton across the river without a bridge is the type of contiguous but non-connected plan that I think should be avoided. This type of leap can addressed by criteria that constrain the commission's choices.

[traininthedistance]

Ideally we would have multi-member districts elected via proportional representation (and, also, would increase the size of the House to something roughly double its current number).  The districts would not all necessarily have to be the same size, but perhaps small states (i.e. less than 4 or 5 reps) would be elected on an at-large basis and large ones could be portioned into large districts that follow metro areas and/or widely-recognized geographic regions (say, the Grand Divisions of Tennessee), and would try to be within 1% of equal population.

As long as single-member FPTP is unavoidable, I favor a holistic approach that tries to come up with a score for various problems to avoid (split counties, townships, or municipalities; keepings metros, CSAs, urbanized areas together; general erosity; VRA concerns; partisan skew), and pushes plans that score well across the whole spectrum of factors.  Muon's pareto optimality approach is, I think, promising, but I might allow for a little bit of extra flexibility to accommodate various local sentiments (for example, double-spanning Nassau and Suffolk to create a North Shore and South Shore district on Long Island).

[hopper]

I think the California Commission-style works fine, or court drawn maps in some cases (The Nevada and New York maps are pretty close to perfect)


I'm not a fan of the Arizona-style, which focuses too much on the voting patterns of areas rather than whether communities should be in a district together. I also don't like the Washington/New Jersey system that only serves to shore up incumbents.

Well the New Jersey one I will go through:

Pallone's district shouldn't even be in Monmouth County. It probably should be up in The Northern Half of Middlesex County(Woodbridge, Iselin, Colonia, Perth Amboy, and maybe even Sayerville) and take in the Union County Suburbs like Roselle, Roselle Park, Winfield, and Clark. Maybe even take in the town of Kenilworth.

Chris Smith's district should take up the Northern Half of the Jersey Shore, make up most of Ocean and Monmouth Counties instead of some of Monmouth County and some of the Mercer County Suburbs like he has now.

[politicallefty]

The CA (and AZ) model could be improved by a better set of criteria. Many of the criteria are left as "touchy-feely" items that the commission can define on the fly. If commissioners have strong opinions, particularly on non-political criteria, they can sway votes to reflect their bias. One area of tension in commission criteria is in the use of political data. If one wants competitive districts one must use that data, yet that data can be manipulated to create a partisan bias (see AZ). I don't mind a commission system, as long as they have objective criteria that are defined before they start mapping.

I wouldn't compare the California system with that of Arizona. I don't recall California using political data when drawing the current lines. I do believe those that draw the lines in California are forbidden from using political data.

You must be logged in to read this quote.

I wouldn't support a legislature vetoing a map wholesale like Iowa. If we're going to allow alterations to the map, perhaps we could have something like Washington (which I believe allows for minimal changes to the commission's map).

[Southern Senator North Carolina Yankee]

In IA, the legislature can only veto a map once or twice and they are stuck with the last map that they are presented with, if I recall correctly that is.

[muon2]

The CA (and AZ) model could be improved by a better set of criteria. Many of the criteria are left as "touchy-feely" items that the commission can define on the fly. If commissioners have strong opinions, particularly on non-political criteria, they can sway votes to reflect their bias. One area of tension in commission criteria is in the use of political data. If one wants competitive districts one must use that data, yet that data can be manipulated to create a partisan bias (see AZ). I don't mind a commission system, as long as they have objective criteria that are defined before they start mapping.

I wouldn't compare the California system with that of Arizona. I don't recall California using political data when drawing the current lines. I do believe those that draw the lines in California are forbidden from using political data.

The use or non-use of political data is driven by the criteria each of those commissions uses. AZ is mandated to create competitive districts, and there's no way to measure competitiveness without political data. Lots of reformers like the idea of more competitive elections, so that brings political data to the process.

CA does have to use political data, but only to determine compliance with the VRA. Since voting differences between the white and minority populations in an area are a required part of the Gingles test, they use political data in those areas with significant minority populations. Also, any commissioner with knowledge of the state's politics will have at least a rough idea of the political data even if it's not explicitly provided. This can create a bias that's difficult to detect without a check of the final districts to see if there's a partisan skew.