CA CD Wine Country Map Quest poll

[muon2]

Splitting a jurisdiction into two relatively even parts is used to dilute their vote if the split jurisdiction can no longer compete against the other parts of the district.

It can also be used when the jurisdiction is large enough or nearly so for its own district but by making a relatively even split the jurisdiction can effectively control two districts.

Either way it is a classic gerrymandering technique. I think it is desirable to limit it by attempting to keep most of a jurisdiction intact. I agree that the 20% number is arbitrary, but that's what we had been batting around. As I just suggested it can be modified to remove the reliance on a specific number yet still having a meaningful impact on a map.

So if a jurisdiction is small, splitting it weakens its power.  If it's large, splitting it enhances its power.  If it's medium-sized, I guess it does neither? 

In any case, my point is not that splits aren't bad (and you still haven't convinced me that a 80-20 split isn't equally bad - doesn't that sort of dilute the 80% part and really dilute the 20% part? - but never mind that), but that having a hard and fast rule regarding them is bound to have perverse consequences. 

Thus the "finding" out for most stuff that goes beyond the Michigan rules.

The question is not whether splits are bad. Every serious neutral mapping model tries to minimize them. The first thing that the public notices is how many districts serve their jurisdiction, and those splits generally get the most public criticism. Gerrymandering is about trying to tilt the vote, and it isn't only about party, it can also be for power or to deny another power. The closer a jurisdiction is to whole, the less manipulation can take place. So that would say that a 80-20 split is preferable to a 50-50 split if everything else is equal.

Torie, are you saying that this should be the rule, or that you want to massage it first? I still fear you will not be satisfied because all good rules will create a bind in some part of the map. The question for me is whether the rule is soundly based and can be agreed to. If that makes a bad spot or two for an observer, does that obviate the rule? Can the good done by a rule elsewhere in a map overcome uncomfortable results in another?

Chop counting rules are explicit in some states and the 0.5% rule has been used in a number of competitions, such as in OH, so I don't see the need to engage in findings before agreement. I can show you the results. As I've noted, rule 3 is one I stuck in to try to complete the discussion on this point. I'm not exactly comfortable with it as is, but I didn't suggest the 20% originally. If you'd rather not have any constraint on deep chops, then you have to be willing for the commission to make them.

[muon2]

My concern about a .5% chop not counting as a chop is that it will force twisted sister maps. I suppose staff could come up with the options using that escape hatch to get a minimum chop count, and the Commission could say that they all suck - in other words the Commission could veto the maps with a supra majority vote, just like is required to pass a map. Then you go back to the minimum chop rules.  That plus the supra majority, with a majority from each party, should preclude anything approaching a Michigan style gerrymander.

I still strongly reject however banning the percentage of CD that chops limitation rule.  The 20% limitation of a county rule that is chopped is sufficient, and the Commission should have the power to override that with a finding as well (e.g., it is OK to have a more than 20% chop of SLO in my signature map on the grounds or protecting the Santa Cruz metro area, with a wall on the Ventura County line, because wall placement elsewhere does lead to a SC chop or a crossing of the Tehachapi 's, or over Luther Pass).

I am pleased that counting the number of chops per county method does not bias where the chops are apparently.  Whew!  

So let me summarize the proposed rules.

1. A plan shall minimize county CD pieces. All pieces count towards the total, including one for the case where a county is entirely in a CD. (BTW this works against minimizing the number of counties chopped, since a tri chop counts the same as two dual shops, but that may be OK in CA).

2. When a county is split, the plan shall minimize the number of census places split within the county (ie cities). Census divisions that span a place boundary can be counted on either side of the line.

3. Pieces of counties less than 1 CD, or pieces of places between 20K and 1 CD, shall not be split leaving less than 80% of the county or place in one piece. (I don't think we settled on a number here, so I'm guessing, but you suggested that some small communities may need a larger split.)

4. Pieces that are less than 0.5% of a CD shall not count in assessing rules 1 through 3.

5. Violations of the aforementioned rules require a finding based in federal or state law approved by a supermajority of the mapping body.

Are we ready? Remember, I don't want to draw and then change the rules to suit a change. That's what the biased groups do, and commissions can have biases that aren't partisan.

We are getting close on the rules, but I want to massage them a bit (and of course, only one of us is a near genius, and neither geniuses), so if something comes up in practice that is unanticipated, either of us have a right to say, oh dear, wait a minute. But hopefully that will not happen.  I wonder how many maps given all the constraints would be theoretically possible (putting aside the details of how the interior of a county are chopped up).  Do you have any idea?

To get back to this question, you need only look at the wide range of maps submitted in some of the competitions. Most of them had stronger rules than what we are talking about here.

For instance, in OH there were quite a few different variations of maps to find groupings of counties that allowed one to take advantage of the 0.5% rule while minimizing chops (they used counting method 1). Those maps were also simultaneously trying to factor in compactness (also used by MI but just to resolving chops) and maximize political fairness and competitiveness, none of which have we even touched on here.

[muon2]

Thus the "finding" out for most stuff that goes beyond the Michigan rules.

I don't need any experimentation to know that there will be perverse consequences.  

For instance, suppose you've got three adjacent counties with equal population.  How do you split them into two districts, while maintaining your rules?  It's impossible.  

To give a case where it's possible, but has ridiculous implications, consider 5 counties.  County A, with a population of 5, is farthest west, bordering only county B, population 1, just to the east.  B borders both C and D which both have population 5.  Finally county E is farthest east and borders both C and D, with population 2.  Divide this up into 2 districts following your rules - the only way you can do it is if you have one district comprised of counties A, B, and E plus a connecting strip in either C or D.  

The use of a connecting strip between whole counties is not unusual in a state with whole county preservation rules. Ohio Senatorial districts often end up with such a strip. Check out current SD 12 and 33 in OH as examples. New HD 102 in MI is another example in a state with county preservation rules.

[muon2]

Thus the "finding" out for most stuff that goes beyond the Michigan rules.

I don't need any experimentation to know that there will be perverse consequences. 

For instance, suppose you've got three adjacent counties with equal population.  How do you split them into two districts, while maintaining your rules?  It's impossible. 

To give a case where it's possible, but has ridiculous implications, consider 5 counties.  County A, with a population of 5, is farthest west, bordering only county B, population 1, just to the east.  B borders both C and D which both have population 5.  Finally county E is farthest east and borders both C and D, with population 2.  Divide this up into 2 districts following your rules - the only way you can do it is if you have one district comprised of counties A, B, and E plus a connecting strip in either C or D. 

The use of a connecting strip between whole counties is not unusual in a state with whole county preservation rules. Ohio Senatorial districts often end up with such a strip. Check out current SD 12 and 33 in OH as examples.

Yes, I've seen those.  They're ugly, and if I were a mapmaker I wouldn't want to be forced into one of them for the sake of an arbitrary 80% split cutoff. 

A percent split isn't what forced those examples to happen. They arose just from the way county splits are counted. Would you like to suggest a rule that would discourage those type of districts?

[muon2]

A percent split isn't what forced those examples to happen. They arose just from the way county splits are counted. Would you like to suggest a rule that would discourage those type of districts?

I'm aware of that, too.  My point was that we shouldn't want to force ourselves into making more of them.  Ohio's rules at least have the virtue of being non-arbitrary; the 80% rule that you are suggesting does not. 

In my 5-county scenario, do you honestly prefer an A/B/strip from C/E district over trying to find a natural division of C or D to make one western district and one eastern district? 

By natural it appears you mean compact. Compactness is a perfectly acceptable standard, but there are many different mathematical definitions. It's been abused in many states where compactness is not specifically defined. IL lists compactness, but absent a definition it has been twisted all sorts of ways by who ever draws the map.

You must be logged in to read this quote.

As far as proposing new rules go, no, I think trying to prevent shenanigans by writing a comprehensive set of convoluted and semi-arbitrary rules will be counterproductive.  C'mon, Muon, you're a Republican - you should know all about unintended consequences of well-meaning regulation. 
[/quote]

I'm for allowing the people to have the most say in who their representatives are. I've seen states where the regulation of the mappers aids the public by encouraging more choice, and I've seen states where lack of restrictions on the mappers results in bias whether intentional or unintentional. I'll choose the path that provides the most freedom for the public, not the mappers.

[muon2]

I haven't seen any other feedback, but if I take into account dpmapper's concern for the shape of some districts due to county preservation, and lack of specific justification for the 20% clause. I would offer the following revised rules. Some simple testing leads me to believe that there is some flexibility, much like the OH contest. I'm also adding a post test to block blatantly partisan maps as one gets in OH or MI on geographic rules alone.

1. A plan shall minimize county CD pieces. All pieces count towards the total, including one for the case where a county is entirely in a CD.

2. Contiguous county pieces require a connection by public road, and two whole counties in a CD shall not be connected if their only connection is through a split piece of a county.

3. When a county is split, the plan shall minimize the number of census places split within the county (ie cities). Census divisions that span a place boundary can be counted on either side of the line.

4. Pieces that are less than 0.5% of a CD shall not count in assessing rules 1 through 3.

5. Violations of the aforementioned rules require a finding based in federal or state law approved by a supermajority of the mapping body.

*****
As an example, here is my plan for the northern inland counties. All CDs are within 100 of the ideal population. County pieces in Siskiyou (12), Glenn (3), Yuba (1), Tuolomme (6), Merced (9) are all under 0.5% of a CD (3515). That leaves a plan with three CD pieces in Alameda and Sacramento,two CD pieces in Fresno, Madera, and San Joaquin, and one in all other counties shown. The excess piece in Madera is due to my drawing of CD 11 to have an HCVAP over 50% (62.0% HVAP).

I also like that the counties including CDs 4-11 are only 28 persons short of the exact population for 8 CDs.

[muon2]

I put my first salvo of thinking on this on the wrong thread.

And what does the bolded bit below mean exactly, and in particular the words "based in?"

You must be logged in to read this quote.

For instance, a split of a community of interest is specific to CA law, so if one wants more chops it requires a supermajority. A federal example would be our difference in the Fresno area. A section 5 map doesn't need Madera, but if one want to preclude any section 2 challenge then the 50% HCVAP district needs to be drawn. That sort of decision would also seem to me to need a supermajority vote of the body.

[muon2]

Here's the coastal counties. I maintained Torie's walls in the south, so the 0.5% rule wasn't applicable within this region. The region is short of 11 CDs by 4197 persons and gains population from Siskiyou and LAC each with a fragment under 0.5% of a CD.

The wine country is almost together as in my version earlier in the thread. It just loses some of Sonoma county south of Santa Rosa in exchange for all of the north coast. This should please some of the other forum "commissioners" concerned about Eureka paired with Redding. About 68K from Marin south of San Rafael is attached to SF for pop equality, otherwise SV and the central coast is not unlike Torie's map.

There is one split more than the minimum. I was unable to do better than the commission for the Monterey CD (20) without a long appendage into SJ. As drawn CD 20 has HVAP 46.3% for section 5. Staying entirely within Santa Clara county I was still able to get CD 17 to AVAP 50.7% while only splitting SJ.

[muon2]

2. Contiguous county pieces require a connection by public road, and two whole counties in a CD shall not be connected if their only connection is through a split piece of a county.

Not a bad rule... but doesn't your proposed CD-12 violate it? 

You are right. But as I mentioned many on the thread preferred a community of interest that linked the whole coast, so this commission might well accept that violation. Also, with that grouping of counties Napa will always link to other counties through a split Sonoma, or there would be a split of both Napa and Sonoma for no reason other than to preserve that rule, and for me the first rule has priority over the second. Another alternative I looked at was some other grouping, such as my initial wine country plan, but that created districts on either the east or west side of the state that were roundly rejected.

I understood the sense of the posters was to have rules that had a certain amount of flex so that there was a way out of a bad map. The way they are constructed, there are times where they will conflict with each other and the mapper can choose how to resolve the conflict.

[muon2]

This is the southern region of CDs. The counties closely match Torie's wall and is only 1019 over population for 31 CDs. The counties can be divided into three groups: Kings-Tulare-Kern-Mono-Inyo-SanB (5 CDs), LA (14 CDs), Riverside-Orange-SanDiego-Imperial (12 CDs), and each of these groups is close enough in population to a whole number of CDs that they could all be drawn within 0.5% of the ideal size without crossing out of their group of counties.

In this map, populations are with 100 of the ideal (except 24 at -119) so the microchop rule is used on four fragments in LAC (22, 24, 26, 27) and one in Riverside (25). There is one extra fragment due to the VRA requirements for CD 23 at 65.2% HVAP and it is over 50% HCVAP.

CD 53 uses the excess in Riverside in Coachella to avoid any part of the city of San Diego and get 65.5% HVAP. That allows all of the city of San Diego to be only in either CD 50 or 52.

The minority CDs are as follows:

CD 23: 65.2% HVAP
CD 27: 63.1% HVAP
CD 29: 63.9% HVAP
CD 32: 61.3% HVAP
CD 34: 64.5% HVAP
CD 35: 51.0% AVAP
CD 36: 68.4% HVAP
CD 37: 43.3% BVAP/44.9% HVAP (BCVAP majority)
CD 38: 67.5% HVAP
CD 39: 72.1% HVAP
CD 42: 51.3% HVAP
CD 46: 65.0% HVAP
CD 53: 65.5% HVAP

[muon2]

I'm not sure I quite understand the purpose of drawing an Asian-majority district in Santa Clara County. I can't think of any reading of the VRA that would require it.

I don't think the VRA under the 9th circuit does require it. But I could do it while only splitting one city that had to be split anyway. I chose that to be my community of interest rather than using some other grouping.

[muon2]

Excellent map, Mike.  I am wondering why you messed with my LA County CD's however.  What bothered you?

Do you have any sense of how many options vis a vis regions, and vis a vis micro chops, the Commission would have to choose from under the micro-chop regime?  That is a sensitive point.

It was faster for me to start with my January map, so it wasn't that I was trying to mess with your CDs. I do like that Brea is just about the right size to join the Asian tiger so that Chino Hills avoids the chop. Have you looked at that as an option?

I don't know the number of permutations, but from my comments about Napa you can see where I've provided for some trade offs in the rules. For instance I use four microchops to eliminate your single chop into Pomona. If a single chop better preserves a city that becomes a legit tradeoff to consider. With this plan I count 106 total pieces, including whole county pieces, and 10 microchops.

[muon2]

Well my last map got rid of the Chino Hills chop, and had that small chop in Pomona. Where are the 4 micro-chops, and how do they avoid my small chop into Pomona?  I didn't consider appending Brea to the Asian CD (it lost Chino Hills, and took most of West Covina instead, and it worked out pretty well), and there is a big empty zone between Diamond Bar and Brea through a long pass, and the two towns are really different worlds. That is an undesirable mix, and should be avoided if possible. I much prefer adding Seal Beach to Long Beach.  That is an excellent fit, with Corona being appended to an OC based CD the second best (strong commuter and economic connections there).

I agree that the Chino Hills chop has to go. Leaving it in costs a chop over all since the excess pop is in OC/SD/Riverside, not in SanB. I noted that since Kern and SanB were inked at Ridgecrest I could consider the whole 6 counties as a single unit with 5 CDs. It was short by 11,774 persons which you grab from Pomona (plus the extra 1K pushed through from Tulare's microchop in your map. If I take the 11,774 and divide it between 4 CDs I can make it into 4 microchops. They are in NW LAC from Kern (24), in Claremont from Upland (26), in Pomona from Chino (27), and in Calimesa from Yucaipa (25). The 1 K that you chopped into Tulare I rotated all the way around to a microchop into Agoura Hills from Oak Park.

I didn't realize the Orange Fwy into Brea was any worse than the connection from Anaheim to Corona which also seems pretty empty between. I'll look at your Seal Beach version again, and see what it does to HVAPs. But there's some advantage to multiple implementations, as it shows that there is limited flexibility within the rules.

[muon2]

2. Contiguous county pieces require a connection by public road, and two whole counties in a CD shall not be connected if their only connection is through a split piece of a county.

Not a bad rule... but doesn't your proposed CD-12 violate it? 

You are right. But as I mentioned many on the thread preferred a community of interest that linked the whole coast, so this commission might well accept that violation. Also, with that grouping of counties Napa will always link to other counties through a split Sonoma, or there would be a split of both Napa and Sonoma for no reason other than to preserve that rule, and for me the first rule has priority over the second. Another alternative I looked at was some other grouping, such as my initial wine country plan, but that created districts on either the east or west side of the state that were roundly rejected.

I understood the sense of the posters was to have rules that had a certain amount of flex so that there was a way out of a bad map. The way they are constructed, there are times where they will conflict with each other and the mapper can choose how to resolve the conflict.

You are preserving exactly what rule?  As to this business of the Commission voting for something that is not legal, do you mean simply per the statute having the power to depart from the otherwise applicable rule, maybe by a supra-majority?

Another issue is compactness. To get compactness might require an override of something else. We should consider one of your definitions of that perhaps. This compactness issue mostly obtains in the CA-01/CA-02 territory, or around the Sierra's potentially. But that may be a mechanism by which Sonoma acting as a bridge could be adopted as the price for compactness (the Commission used it as a bridge too of course, I suspect for that reason actually).

The rule I was preserving was to minimize chops. Consider that I could have split Napa into two pieces shuffled some population in Sonoma and attached the southern piece of Napa to northern Marin through southern Sonoma. At that point there is no rule violation for a bridge since Napa is not whole, but is was a gratuitous chop of Napa that allowed the bypass. It seems silly to chop Napa just to avoid that rule, so I decided that the bridge rule was of lower priority.

Based on comments, I viewed the rules as binding the commission without an override vote. My observation is that not all see the law the same way. I would say the Commission's map violated section 2 of the VRA in the CV, but clearly they agreed that section 5 allowed them to meet the VRA without 50% HCVAP districts in spite of the finding that the Gingles conditions were met.

I agree compactness should count, and there are two ways. One is to look at the district as a whole. The other is to look at how a chop changes the compactness of a district as in MI.

[muon2]

Is this 4 micro-chops in lieu of one large but still small chop really going to sell? How can it be defended without looking silly?  Maybe a micro-chop only counts where it is in lieu of one non micro chop. I think the idea is to drive to your regional groupings (assuming we have some choices), and require a supra majority to dump them.

The basis of the microchop is that SCOTUS has permitted population deviations in some narrow instances creating a range of up to 1% for CDs, which can be achieved by keeping individual deviations to 0.5%. For example MA had a range of 0.29% for its CDs in the 2000. The WV case will probably add some clarity, and the SCOTUS order is generally viewed as favoring that state's position on deviations. With microchops, a plan can have exact equality and if they are removed by swapping the population back to district on the other side of the county line, the districts are all still within a 1% range.

Generally CDs require exact equality under the standard of Wesberry. In followup cases the court ruled that states can deviate but must provide specific justification for deviations showing they could not be reduced. This is from Karcher:

You must be logged in to read this quote.

The CA Constitution provides a possibility for deviations:

You must be logged in to read this quote.

I read all of that as saying if law permits the deviations to meet the constitutional goals properly applied, then there can be population deviations. SCOTUS permits deviations for compelling state interests such as constitutional clauses involving whole counties or towns. Therefore deviations that would exist by swapping microchops back into whole counties are permitted, and if they are not for specific districts, restoring the microchop is the simple remedy. AR did just that in 2001 where they had two maps one with whole counties and one with microchops for equality. The whole county map was not challenged.

In this scenario the commission would have to determine if a certain number of microchops outweighed one regular chop. The size and placement of those microchops could also factor into the commission's determination.

You must be logged in to read this quote.

Reading the report, this is the only conclusion I can draw. If there was some other basis for avoiding a section 2 district when it was possible and the preconditions were met, it is not in the report.

[muon2]

What is the SCOTUS microchop case, and if the law is clear, why then just not have the population deviations? That makes a lot more sense than a host of microchops, on the theory that they can be done away with, but won't be. In any event, when you say the commission can decide on one bigger chop or four microchops, is that by a normal vote, or a supra majority vote in your thinking? If by normal majority, than the microchop regime gives the commissions more options. I assume you think it needs a supra majority if the issue is one normal chop versus one microchop, is that correct? It is only when you have extra microchops, that this issue arises.

There is some lack of clarity in the law, which is why almost all states go to exact equality. The court said that if you don't want exact equality you had better show why in detail. New Jersey lost in the Karcher case, because they couldn't show that they had the best way to achieve state goals with minimum deviation. Yet IA never has exact equality because they have rigidly adhered to their standards. Historically AR and MA also had significant deviations as they held to standards involving counties and towns respectively. However, they both bailed in this cycle and went to exact equality, not because of the courts, but because the legislature didn't like the map those standards would create.

Here's a useful table showing the states and their deviations in the last cycle. One can see a number with deviations beyond exact population. Of note is the 0.6% range (0.3% max deviation) in ID. That came from a commission and includes a county and city split.

As I noted this is the crux of the WV case. Most observers felt that the SCOTUS order overturning the lower court's decision to force a remap for smaller deviations read favorably for the state. I hope this case clarifies the standard required of a state to deviate from equality.