North Carolina Legislative Redistricting

[jimrtex]

Sigh! My plan provides for 121 districts.

The multi-county groups had population for 72.970 districts (just short of 73) and I created groups entitled to 74 districts (an error of -1.4%)

[Kevinstat]

Sigh! My plan provides for 121 districts.

The multi-county groups had population for 72.970 districts (just short of 73) and I created groups entitled to 74 districts (an error of -1.4%)

Did you travel to Vermont recently?  (Do you remember what I'm alluding to here?)

Which of the following from your post where you talked about the number of groupings is incorrect?  The 43.93 districts for the "large" counties that aren't within range of an integer number of districts or the 29.04 districts for the "small" counties?  Because those two numbers add up to 73.97, not 72.97.  Okay, I don't know what I was thinking there.  Something about the 9s that made my mind think the decimals added to almost two rather than almost one, even though one of the 9s was not in the fractional portion.

Was a good portion of the error from the districts and groupings you left as is (because the districts or groupings weren't overturned - not that a court would overturn or uphold a grouping per se but where the simplest solution seemed to be to keep the groupings)?

[jimrtex]

Sigh! My plan provides for 121 districts.

The multi-county groups had population for 72.970 districts (just short of 73) and I created groups entitled to 74 districts (an error of -1.4%)

Did you travel to Vermont recently?  (Do you remember what I'm alluding to here?)

Which of the following from your post where you talked about the number of groupings is incorrect?  The 43.93 districts for the "large" counties that aren't within range of an integer number of districts or the 29.04 districts for the "small" counties?  Because those two numbers add up to 73.97, not 72.97.

Was a good portion of the error from the districts and groupings you left as is (because the districts or groupings weren't overturned - not that a court would overturn or uphold a grouping per se but where the simplest solution seemed to be to keep the groupings)?

43.93 + 29.04 is 72.97. The groups should have had 73 districts. But because they were slightly small on average (0.986 quotas) I ended up with a total of 74 districts. My spreadsheet has sums for the total populations (120.00 quotas) and counties (100), but where I should have totaled the whole number of representatives (and checked for 120), I had simply rounded 120.00 to 120.

I thought I had been careful. I had checked the number of districts for the single-county districts before I even began. It is quite close to a whole number (47.03). Wake and Mecklenburg both have fairly large errors, but they cancel (11.573 + 11.339 = 22.912, and 12 + 11 = 23). So while Mecklenburg is entitled to an extra 1/4 of a representatives, they get a whole 1.

As I worked my way east, I thought I was checking the total number of whole representatives vs. the total number of quotas, but since instead I was simply rounding the number of quotas they always matched.

When I was reviewing Peterson's affidavit, I noticed that when he checked potential covers, generated by random placement of superdistricts, he rejected covers that totaled 119 and 121. I thought, "Hah! If you crafted a map by hand you wouldn't have such problems!"

When I started analyzing Peterson's cover with the most superdistricts, the one with 41, I found it had 9 excess cuts, just as predicted by theory. But then I noticed I had eight excess cuts, despite having 42 districts. It took me a long time to find my mistake.

Where I had most of the problem is with districts with more than one representative. When you can get an area between 0.95 and 1.05 representatives, you are grateful to find something that is in range. But when the target is 2.85 to 3.15 there are lot more possible combinations that work. So it was generally super districts with more representatives that were the problem. I'll show this on another map.

I have a corrected map, but it is down to 41 superdistricts. It might beat other covers on some secondary tests.

[jimrtex]

This shows the error for maps in my original plan.



The percentages are not the relative error for the superdistrict, but the deviation from the ideal total for the superdistrict.

For example for the orange district (Union-Anson) east of Charlotte, has a population equivalent to 2.871 quotas, which is 0.129 quotas below the ideal of 3.000 quotas (0.129 quotas equals 12.9% of a quota).

If you were to total all of the deviations on the map (one per superdistrict, not one per county - it is easier to label every county, rather than one county per superdistrict) they will add to -100.0% or enough to create one extra seat.

Notice there is not too much error in the single-member superdistricts. There is also a cluster of four underpopulated superdistricts south of Greensboro and Raleigh, eastward from Charlotte, and another two along the coast near Wilmington.

Collectively, there is a group of 10 superdistricts with a population equivalent to 27.302 quotas that has 28 districts for a relative error of -2.5%. If we can rearrange these superdistricts so that 27 districts are apportioned, then they will have an average error of 1.1% and the correct total of 120 districts. An important aspect of this is that it preserves the cluster of single-member districts in the northeast, which is one of the key aspects of my map.

This is my revised map, showing the deviation.



It has 120 districts, but unfortunately only 41 superdistricts, as 10 superdistricts in the southeastern part of the state have been rearranged into 8 superdistricts. The other 33 are unchanged from my original map. It preserves six single-member multicounty districts in the east. It rearranges one from Richmond-Scotland to Hoke-Scotland, which is probably preferable from a racial standpoint, since Hoke has a larger minority population than Richmond. The three single-member multi-county districts along the Tennessee line are also kept.

Along with the three single-county single-member districts, these are the only districts that fully comply with the North Carolina Constitution and the the North Carolina Supreme Court's expressed concern about cutting smaller counties. A small county that can not constitute a house district is at somewhat of a disadvantage, and this is exacerbated by being minced with parts attached to other districts. When combined with other small counties, they likely share a common regional interest such as being rural, and there is less risk on one county dominating or being dominated.

The attachment of Jones to Duplin is optional. Jones could be attached to Craven as well, but this division provides better population balance, and it also moves a division of Pender further north, closer to the Duplin County line, keeping a larger share of Pender within a single district.

There is a North Carolina state highway connecting the two counties directly. Jones is somewhat separated from Onslow by a large pososin along the county line.

This is my revised plan in a more conventional form.



This is my analysis of my revised plan:



Green: County groups that are unchanged and that do not contain overturned districts. The districts do not need to be redrawn, and would not be contested if the court orders a special election. The groups contain 42 districts.

Generally, they are in areas with smaller black populations, though there are exceptions.

HD-71 and HD-72 in Forsyth (Winston-Salem) are both 45% BVAP. It was determined that it was impossible to reach 50% BVAP for both of them. One factor that the court used in determining racial predominance was use of a bright line 50%+1 test for its districts. Neither district was challenged by the plaintiffs (A VRA Section 2 challenge requires a district-by-district analysis, though it can be in the context of an overall analysis).

HD-23 (Edgecombe-Martin) and HD-27 (Halifax-Northampton) have 52% and 54% BVAP respectively. But because they are comprised of whole counties, they adhere to traditional neutral redistricting principles of respect for political subdivisions and the whole county provisions of the North Carolina Constitution. Neither district was challenged by the plaintiffs. While the court did not find them constitutional, no litigant suggested that they were unconstitutional.

Blue: New groups, but that don't contain any overturned districts. The districts will have to be redrawn, and in some cases, there may be racial concerns.

Union-Anson is currently a county group with three districts, two suburban districts that are on the Mecklenburg line, and another district that includes the remainder of Union plus all of Anson. Stanly and Montgomery a slightly too large (1.112 quotas) for a single district, but there is a district comprised of Stanly and most of Montgomery.

The four counties were combined in a single group to better balance population. Shifting a bit of the excess population from Stanly to Union-Anson will keep the basis outlines of the existing districts, with tweaking of boundaries to equalize populations. Three of the districts have BVAP of 12%, 12%, and 13%. The Union-Anson district has a BVAP of 24% due to Anson having a 48% BVAP. But combining Anson with Richmond with only reaches 36.3% and not quite enough for a district. Bartlett v Strickland does not require drawing a VRA district unless an area has BVAP%. The crux of the  current litigation appears that you identify the compact area simultaneously with determining whether there it is 50% BVAP, rather seeing if you can construct a connected area that is 50%+ BVAP that can somehow be characterized as being compact.

The Currituck-Dare-Hyde-Pamlico groups is only 11% BVAP. Dare (3%) and Currituck (6%) have particularly low BVAP population.

The Brunswick-Columbia group includes some of the area that contained a once 50% BVAP district that later was at the heart of Bartlett v Strickland The district had a focus in the city of Wilmington and then meandered across rural areas Pender, Brunswick, and Columbus which featured three separate point connections where districts crossed each other. Demographic change (white retirements toward the coast, and statewide growth requiring larger districts) and the elimination of point connections reduced the BVAP below 50%. Stephenson forced the district into New Hanover and Pender, and Strickland resulted in the non-division of Pender.

The current version of the district remains centered in Wilmington, but extends into Brunswick. The district is only 29% BVAP, and the Brunswick portion of the district is only 19% BVAP (overall, the county is 11%  BVAP). Columbus is 30% BVAP, and it should be possible to draw the district that includes Columbus and extends into Brunswick with a BVAP in the mid-to-high 20%'s. The Brunswick district should be quite similar to the existing HD-17, but will requirea small augmentation of its population.

The Cartaret-Onslow group replaces the Onslow-Pender groups. Two groups will remain in Onslow, with a portion along the eastern line being added to Cartaret. Onslow has a 16% BVAP. Replacing the 18% BVAP Pender with the 6% BVAP of Cartaret, reduces the BVAP of the group from 16.4% to 13.1%.

It should be able to maintain HD-15 pretty much as it is. HD-14 could be shifted to take up the area vacated by HD-16, and give up area to HD-13 coming in from Cartaret. There could also be some shifts. The incumbents for HD-14 and HD-15 are both near the edges of their districts. There is no risk of them being moved out of their districts, but it would be easy to move the boundaries away from their homes, and make the two districts more compact.

[Kevinstat]

I've corrected my last post and taken a guess as to where I what caused me to mess up the math.

I see your new plan no longer keeps all 53 districts in groups without any overturned districts the same.  I can see why you rearranged the southern coastal area and the area just east of Charlotte, though, as getting rid of the shortages there was doubtless helpful in getting rid of that extra district.  I know earlier you had critiqued others for not keeping the "unchallenged area" as it was.

I was being silly when I mentioned the Vermont thing.  I know you are more careful and methodical than the Vermont LAB was in their State Senate redistricting.  With their large deviations, they were practically begging for something like that to happen.

[jimrtex]

I see your new plan no longer keeps all 53 districts in groups without any overturned districts the same.  I can see why you rearranged the southern coastal area and the area just east of Charlotte, though, as getting rid of the shortages there was doubtless helpful in getting rid of that extra district.  I know earlier you had critiqued others for not keeping the "unchallenged area" as it was.

What had set me off on this quest was the claim made the legislature was going to have redraw (IIRC) 81 house districts, even though only 19 were found unconstitutional.With 53 retained groups, I thought that  would be reduced to 67 that would be redrawn, and some not materially redrawn, if at all.

The areas that I lost in my revised plan were on the periphery. Is still keep the block of districts in the western part of the state. But I am now down to 42 retained districts, requiring 78 (65%) to potentially be redrawn.

Incidentally, the North Carolina House and Senate redistricting committees will hold a joint meeting next Wednesday, July 26.

[jimrtex]

Back to the Peterson cover with the most groups (Cover 1211)



Green Groups unchanged from current, and with no overturned districts. 16 groups, 31 districts.

Blue Groups changed from current, with no overturned districts. 7 groups, 22 districts. Note the Halifax-Martin group splits up two groups which had unchallenged majority minority districts.

Yellow Groups unchanged from current which contain no overturned districts. 4 groups, 33 districts.

White Groups changed from current which contain overturned districts.  18 groups, 34 districts.

Overall up to 89 groups (74.2%) may need to be redrawn.

Standard deviation assuming perfect splits of whole county groups 3.24%.

Forced county cuts (9 total):

(1) Columbus. One fragment with Pender, one with Robeson.

(2) Rowan. One fragment with Stanly, one with Davie, one whole district in Rowan.

(3) Montgomery. One fragment with Anson-Richmond, one with Randolph.

(4) Haywood. One fragment with Jackson-Swain, one with Madison-Yancey.

(5) Granville. One fragment with Durham-Person, one with Vance.

(6) Sampson. One fragment with Bladen, one with Harnett-Johnston-Wayne.

(7) Harnett. One fragment with Lee, one with Johnston-Sampson-Wayne, oine whole district in Harnett.

(8) Wilkes. One fragment with Alexander, one with Alleghany-Surry.

(9) Surry. One fragment with Stokes-Rockingham. one with Alleghany-Wilkes.

Seven of the extra cuts are in smaller counties (entitled to less than one representative). Splitting such counties is arguably worse since it will be difficult to be elected from the fragments since many fellow county residents will not be able to vote for a candidate, and the fragments will have less influence over the election of "their" representative.

In addition, Wayne will not have one whole district and a fragment, but will have two fragments.

Number of groups per number of counties (ie there are 15 2-county groups)
1: 12
2: 15
3: 6
4: 4
5: 1
6: 2
7: 1

Number of groups per number of districts (ie there are 11 2-district groups)
1: 13
2: 11
3: 6
4: 4
5: 3
6: 1
7: 1
11: 1
12: 1

30 counties are in 10 single-member, multi-county groups. All such groups/districts are consistent with the North Carolina constitution. There are an additional 3 single-member, single-county groups.


For comparison my revised plan:



Green Groups unchanged from current, and with no overturned districts. 20 groups, 42 districts.

Blue Groups changed from current, with no overturned districts. 4 groups, 10 districts.

Yellow Groups unchanged from current which contain no overturned districts. 7 groups, 42 districts.

White Groups changed from current which contain overturned districts.  10 groups, 26 districts.

Overall up to 78 groups (65.0%) may need to be redrawn.

Standard deviation assuming perfect splits of whole county groups 2.62%.

Forced county cuts (9 total):

(1) Haywood. One fragment with Jackson-Swain, one with Madison-Yancey.

(2,3) Wilkes. One fragment with Alexander-Yadkin, one with Alleghany, one with Surry.

(4) Stanly. One fragment with Anson-Union, one with Montgomery.

(5) Lee. One fragment with Chatham, one with Randolph.

(6) Moore. One fragment with Randolph, one with Richmond, one with Lee.

(7) Bladen. One fragment with Robeson, one with Sampson.

(8) Pender. One fragment with New Hanover, one with Duplin-Jones.

(9) Granville. One fragment with Person, one with Vance-Warren.

Eight of the extra cuts are in smaller counties (entitled to less than one representative). Splitting such counties is arguably worse since it will be difficult to be elected from the fragments since many fellow county residents will not be able to vote for a candidate, and the fragments will have less influence over the election of "their" representative. Five of the fragments are relatively small (Wilkes(2), Stanly, Pender, and Granville).

In addition, More will not have one whole district and a fragment, but will have three fragments.

Number of groups per number of counties (ie there are 13 2-county groups)
1: 12
2: 13
3: 6
4: 6
5: 4

Number of groups per number of districts (ie there are 11 2-district groups)
1: 13
2: 11
3: 4
4: 6
5: 4
6: 1
11: 1
12: 1

30 counties are in 10 single-member, multi-county groups. All such groups/districts are consistent with the North Carolina constitution. There are an additional 3 single-member, single-county groups.

[jimrtex]

One claim in the state litigation was that the legislative plans violated the Whole County Provisions of the North Carolina Constitution. This was rejected by the North Carolina district court (see page 49 of decision).

Dickson v Rucho opinion by North Carolina district court (PDF)

The court ruled that the legislature had better followed the dictates of Stephenson I and Stephenson II in drawing the maximum number of 2-county groups before then drawing the maximum number of three-county groups, etc. They make note of the deposition of Peterson where it was his opinion that the North Carolina Supreme Court was wrong. The district court said it was bound by the precedent of the North Carolina Supreme Court.

The North Carolina Stephenson guidelines produce a paradox.

Imagine that that the most groups possible was 41, and under Plan A, it was possible to have 12 one-county groups, and 28 two-county groups, encompassing 68 counties. The remaining group would have 32 counties, 9 of which would be cut, which means that this mega group had at least 10 representatives. If the intent was to complete chop up one area with the state with almost total disregard for county boundaries, this would be the way to do it.

Stephenson also ignores that any whole county group electing a single representative is completely consistent with the North Carolina Constitution. If it is possible to arrange a nine-county area into two groups of three and six counties, or four and five counties, the constitution is agnostic as which is better. Here again Stephenson producing a paradox. By giving a preference for a group of three, it also gives a preference for a group of six.

The North Carolina Supreme court affirmed the lower court decision (see page 43, etc. of the PDF)

Dickson v Rucho-1 North Carolina Supreme Court (PDF)

The North Carolina Supreme Court is preparing to consider Dickson v Rucho yet again. After the Alabama decision, the SCOTUS remanded the case to the North Carolina Supreme Court, which confirmed its earlier decision. After the SCOTUS upheld the federal court decisions, they vacated the North Carolina Supreme Court decision. Since the SCOTUS has now said that the North Carolina Supreme Court was wrong, that court may revisit the Stephenson guidelines which says that you first draw VRA districts, and then draw the whole county groups around them. But that was seen as evidence of a predominance of race.

[jimrtex]

This is an alternative cover with 42 whole county groups, that combines Mecklenburg and Union in a two-county group, which permits rearrangement of the groups in south central North Carolina.



Green Unchanged, no overturned districts: 25 groups with 62 districts.

Blue New Groups, no overturned districts: 1 group with one district.

White New groups, overturned districts: 10 groups with 33 districts.

Yellow Unchanged, overturned districts: 6 groups with 30 districts.

Overall, up to 58 (48.3%) districts may have to be redrawn. If we discount the six single-district groups that are in new groups, up to 52 (43.3%) districts may have to be redrawn.

Standard deviation assuming perfect splits of multi-district groups: 2.64%.

Forced county cuts (8 total):

(1) Haywood. One fragment with Jackson-Swain, one with Madison-Yancey.

(2,3) Wilkes. One fragment with Alexander-Yadkin, one with Alleghany, one with Surry.

(4) Stanly. One fragment with Anson-Richmond, one with Montgomery (there are alternative splits possible in this group).

(5) Lee. One fragment with Chatham, one with Harnett.

(6) Robeson. One fragment with Bladen, one with Columbus, and one district wholly within Robeson (there are alternative divisions).

(7) Wayne. One fragment with Sampson, one with Duplin, and one district wholly within Wayne (there are alternative divisions).

(8) Granville. One fragment with Person, one with Vance-Warren.

Number of groups per number of counties (ie there are 16 2-county groups)
1: 11
2: 16
3: 6
4: 6
5: 3

While eliminating (from my 41-group plan) one single-county group, the plan also eliminates one five-county group, and replaces them with three two-county groups.

Number of groups per number of districts (ie there are 11 2-district groups)
1: 14
2: 11
3: 6
4: 5
5: 3
6: 1
11: 1
14: 1

Compared to my 41-group plan, this eliminates one 4-district and one 5-district group, and creates one additional 1-district group, and two additional 3-district groups. It also converts the 12-district Mecklenburg group to a a 14-district Mecklenburg-Union group.

32 counties are in 11 single-member, multi-county groups. All such groups/districts are consistent with the North Carolina constitution. There are an additional 3 single-member, single-county groups.

[Kevinstat]

These are the whole county groups which are unchanged from the 2011 law and which do not contain VRA districts. These districts do not need to be modified. The groups contain a total of 53 districts, 45 Republican and 8 Democrat.

These are unchanged county groups that contain overturned districts. New maps can be drawn for each group and slipped in the statewide map with no ripple effects.

This is an alternative cover with 42 whole county groups, that combines Mecklenburg and Union in a two-county group, which permits rearrangement of the groups in south central North Carolina.



Green Unchanged, no overturned districts: 25 groups with 62 districts.

Blue New Groups, no overturned districts: 1 group with one district.

White New groups, overturned districts: 10 groups with 33 districts.

Yellow Unchanged, overturned districts: 6 groups with 30 districts.

So the only seemingly keepable groups from the existing plan (those from the first two maps in this post I'm making) that aren't kept are the Meklenburg group and the adjacent Union-Anson group.  Plus one could argue that the group around the Albermarle Sound (which loses Currituck and gains Washington before being divided into two single-district groups; unchanged from your original plan that ended up having one too many districts) could have been kept, but perhaps not while avoiding a county split within and the Currituck-Washington trade makes sense to me.  The blue group in your third map includes territory from the gigantic 20-county group that included three overturned districts.

I like that your new groups with more than one district tend to be "non-stringy" (three counties that all border each other or four counties with all but one pair sharing a border, or a four-corners situation - am I right that that is the case with Stanly, Montgomery, Anson and Richmond?).  I think I prefer this plan to your other plan (that has the correct number of districts).

[jimrtex]

These are the whole county groups which are unchanged from the 2011 law and which do not contain VRA districts. These districts do not need to be modified. The groups contain a total of 53 districts, 45 Republican and 8 Democrat.

These are unchanged county groups that contain overturned districts. New maps can be drawn for each group and slipped in the statewide map with no ripple effects.

This is an alternative cover with 42 whole county groups, that combines Mecklenburg and Union in a two-county group, which permits rearrangement of the groups in south central North Carolina.



Green Unchanged, no overturned districts: 25 groups with 62 districts.

Blue New Groups, no overturned districts: 1 group with one district.

White New groups, overturned districts: 10 groups with 33 districts.

Yellow Unchanged, overturned districts: 6 groups with 30 districts.

So the only seemingly keepable groups from the existing plan (those from the first two maps in this post I'm making) that aren't kept are the Meklenburg group and the adjacent Union-Anson group. Plus one could argue that the group around the Albermarle Sound (which loses Currituck and gains Washington before being divided into two single-district groups; unchanged from your original plan that ended up having one too many districts) could have been kept, but perhaps not while avoiding a county split within and the Currituck-Washington trade makes sense to me.  The blue group in your third map includes territory from the gigantic 20-county group that included three overturned districts.

The Moore-Randolph and Chatham-Lee-Harnett (in green on the 3rd map) groups are the same as the existing plan.

The 20-county snake goes from Stanly to Dare, but by a circuitous route. In the 3rd map, look at the white area. Take, Mecklenburg, Union, and Anson out and the snake begins Stanly-Montgomery-Richmond, and then goes south of Cumberland (Fayetteville). It leaves out Pitt and Wilson which formed an improbable 2-county group (they have a 3-mile wide boundary that has two districts crossing through it, as well as using the middle mile as a boundary jog between the two districts, and continues east to include Pamlico, Hyde, and Dare in the blue group, plus Washington (but not Currituck).

The 9 counties in the northeast in my two groups, were a single two-representative group (with Washington out and Currituck in). Taking Currituck out is what made my split possible. That is only possible by including Pamlico. While the Pamlico-Hyde connection is a tenuous water connection it makes possible the northeastern split, which I think is a feature of my plan, since it produces a third compact whole-county majority BVAP district.

I like that your new groups with more than one district tend to be "non-stringy" (three counties that all border each other or four counties with all but one pair sharing a border, or a four-corners situation - am I right that that is the case with Stanly, Montgomery, Anson and Richmond?).  I think I prefer this plan to your other plan (that has the correct number of districts).

Yes, I noticed that when I realized there were choices when you decide which counties need to have the extra split.

That is a four-corners. The western boundary of Richmond and Montgomery is the center of the Pee Dee River, which including tributaries form county boundaries up to between Yadkin and Surry.

The Montgomery-Richmond boundary is a straight line that crosses the river, where it meets the center-line of the Rocky River which forms the boundary between Stanly and Anson, and Stanly and Union. The straight line forms the boundary between Stanly and Anson, from the center-line of the Pee Dee River to the west bank (about 300 feet).

If you look at smaller scales, it almost looks like Montgomery comes to a point just touching Anson, but there is actually an almost 90 degree intersection of a straight line and the center-line of the river.

One thing I have not mentioned, is that if one is willing to go outside 5% limits slightly, three cuts can be eliminated.

Rowan-Davidson-Stanly-Montgomery is another four corners.

Pitt is 5.8% above two districts. If it were separated from Beaufort-Craven it would eliminate a tiny sliver of Pitt that is attached to those two counties. Those two counties would be -4.8% under.

Johnston is 6.3% above two districts. If it was separated from Duplin-Sampson-Wayne it would eliminate a tiny sliver that is attached to those three counties. The three remaining counties would be 2.6% over.

Surry is -7.3% below a district. If if was separated from Alexander-Alleghany-Wilkes-Yadkin it would eliminate one of the splits of Wilkes. While Wilkes is split twice, about 93% of the county population is in one district (Alleghany-Wilkes could form a two-county group, but a strip is needed to connect Alexander and Yadkin, and a tiny bit is needed to get Surry up to a whole district.

[jimrtex]

The two legislative redistricting committees held a joint hearing on Wednesday. I could not find any audio or video, but there were news accounts, which were somewhat superficial (the editor for the news probably said, "Don't we have a map of North Carolina? Yeah, one with districts would be good. The House and Senate districts are too small. Let's go with the congressional map? Hmm that doesn't look too bad. What about one with the old I-95 district that the Democrats drew?

There are now committee websites. If you click on the 7/26 folder, you will find some maps.

House Select Committee on Redistricting website

Senate Redistricting Committee website

[Brittain33]

The NC legislature is being taken to court (again!) about their slow-walking their constitutional responsibilities. It will be interesting to see what sanctions come down, if any.

https://www.apnews.com/0553c981d6094ac99e9cd2b5859b07cb

[jimrtex]

Thomas Hofeller drew the current maps, and will likely draw the new maps. He has been involved in redistricting in various states since the 1970s. His testimony basically sealed the case against the current maps.

In his explanation about the process he said that he did not attend any hearings, or committee meetings, or read any transcripts. He then backtracked and said that he might have stuck his head in, but had no reason to stay and listen.

As part of the current litigation he presented what he claims is the optimum plan for whole county groups, and it is from this map that the number of counties that must be redrawn comes from.

Hofeller Optimum Plans (PDF), you should be able to rotate this in your browser or (PDF reader such as Adobe).

Here is the map presented in a form comparable to that for the other plans (the color scheme was adopted from Hofeller's affidavit, with the addition of blue for groups that would require drawing new districts in areas where none were overturned.



His strict reading of Stephenson is that the maximum number of one-county groups must be drawn, followed by the maximum number of two-county groups, etc. His plan does achieve that, with 12 one-county groups, and 17 two-county groups. But as a consequence it ends up three larger monster groups: (1) 6-county groups with 4 representatives on the northern boundary; (2) 6-county, 6 representative snake from Davie to Richmond; and (3) the 7-county, 7-district, 3-legged group  southeast of Raleigh.

A consequence of forming the maximum number of groups is to produce leftover regions which can not be divided into small groups. An extreme case is the nine counties in northeastern part of the state which are split 3 and 6, rather the 4 and 5 in my plan.

If we imagine the drawing of a map much as a master stone mason would carefully choose stones to assemble a wall, the Hofeller plan is the equivalent of an apprentice slapping together smaller stones and then shooting massive amounts of caulk or concrete or plastic to fix his mistakes.

Green Unchanged, no overturned districts: 20 groups with 39 districts.

Blue New Groups, no overturned districts: 4 groups with 12 districts.

White New groups, overturned districts: 11 groups with 32 districts.

Yellow Unchanged, overturned districts: 6 groups with 37 districts.

Overall, up to 81 (67.5%) districts may have to be redrawn. If we discount the five single-district groups that are in new groups, up to 76 (63.3%) districts may have to be redrawn.

Standard deviation assuming perfect splits of multi-district groups: 3.29%.

The deviation for the three monster groups are +4.58% (4 representatives, northern), +3.34% (6 representatives, snake), +5.00% (+4.996%) (7 representative 3-legged monster). Because they are large districts in terms of representatives, they produce an overall imbalance. While single member groups may have large individual deviations, they tend not to accumulate. In the Hofeller map, the thirteen single-member groups have an average deviation of -0.41%. Collectively, they are only 4268 persons short of the population needed for 13 representatives.

The three monster groups have a population equivalent to 17.733 representatives, but are only apportioned 17 representatives. They are cheated out of almost a whole representative.

Forced county cuts (9 total):

(1) Haywood. One fragment with Jackson-Swain, one with Madison-Yancey.

(2) Wilkes. One fragment with Alexander, one with Alleghany-Surry.

(3) Surry. One fragment with Wilkes-Alleghany, one with Stokes-Rockingham.

(4) Rowan. One fragment with Davie, one with Cabarrus-Stanly, one wholly within Rowan.

(5) Stanly. One fragment with Cabarrus-Rowan, one with Montgomery-Richmond.

(6) Harnett. One fragment with Lee, one with Sampson-Johnston-Wayne, one entirely in Harnett.

(7) Sampson. One fragment with Bladen, one with Harnett-Johnston-Wayne.

(8) Granville. One fragment with Person, one with Vance-Warren.

(9) Columbus. One fragment with Robeson, one with Pender.

Number of groups per number of counties (ie there are 17 2-county groups)
1: 12
2: 17
3: 4
4: 3
5: 1
6: 3
7: 1

Number of groups per number of districts (ie there are 11 2-district groups)
1: 13
2: 11
3: 6
4: 5
5: 1
6: 2
7: 1
11: 1
12: 1

30 counties are in 10 single-member, multi-county groups. All such groups/districts are consistent with the North Carolina constitution. There are an additional 3 single-member, single-county groups.

[jimrtex]

The NC legislature is being taken to court (again!) about their slow-walking their constitutional responsibilities. It will be interesting to see what sanctions come down, if any.

https://www.apnews.com/0553c981d6094ac99e9cd2b5859b07cb

Your characterization is not accurate.

The plaintiffs did not bother to file suit until 2015!. They may not have been able to find lawyers, or their lawyers may have been busy with other cases. There was similar litigation in the state courts, which upheld the maps drawn by the legislature.

After the Alabama decision, the SCOTUS accepted the state case, and then remanded it back to the North Carolina Court. The North Carolina Supreme Court once again uphold the state plan, which was then appealed to the SCOTUS, which did not take the case.

Eventually, in August of 2016, the federal district court decided the case. This was obviously too proximate to the November 2016 election to draw districts, especially given the fact that the federal court was concerned about how the line-drawing had been in 2011.

The state appealed the federal case to the SCOTUS. The intemperate federal court ordered North Carolina to draw new districts and hold a special election in 2017. This would require representatives who had been elected to be tossed from office in the middle of their terms, which is an extraordinary solution.

The state asked for a stay of this redistricting order, while the underlying case was being considered by the SCOTUS. At this point the decisions of the federal court and state courts were in conflict, and the SCOTUS would be the appropriate body to resolve the difference. In addition, there was the issue as to whether the plaintiff parties in the state and federal cases were independent of each other (they had the same lawyers, which is perhaps why the federal plaintiffs did not file suit until 2015). The SCOTUS granted the stay.

The state offered to do expedited briefing, which the plaintiffs did not, but the SCOTUS did not accept the offer in any case. The SCOTUS decided the federal case in May - but in their decision they vacated the order to immediately draw districts and hold a special election. They spent half of their opinion explaining why, suggesting that the district court really did not understand their past decisions - that the district court had been injudicious.

The state had to explain to the federal district court that SCOTUS decisions are not final until a certified opinion is issued by the SCOTUS, about a month after the opinion is made.

The official reason for having the hearing today was to help the federal court determine the course of action after their original decision had been tossed by the SCOTUS.

[jimrtex]

The Stephenson decisions by the North Carolina Supreme Court in the 2000s cycle were an attempt by the North Carolina Supreme Court to harmonize the Whole County Provisions of the North Carolina Constitution; the One Man, One Vote decisions of the 1960s and 1970s; decisions of the SCOTUS which mandate single-member districts when at-large elections may submerge racial (or possibly political) minorities in a larger electorate; and provisions of the Voting Rights Act which may mandate disregard for traditional districting criteria such as respect for political subdivisions, including counties.

The role of the North Carolina Supreme Court should be to set objective criteria that a harmonized plan should meet. Instead, they set out a procedure that they thought would meet the criteria. It is analogous to the situation where referees were to determine if a play resulted in a touchdown, but instead of checking for infractions such as multiple forward passes, ineligible receivers, holding, etc., they actually drew up a play to score a touchdown. If the play failed, they might feel obligated to thrust both arms into the air, since it was their play. Or if a team ran their own play, they might throw a flag for illegal procedure.

The 1868 North Carolina Constitution provided that each county should have one representative regardless of population, and that the additional representatives should be apportioned on the basis of population. But with 100 counties, and 120 representatives, there were only 20 extra representatives to be apportioned to the larger counties. Counties with much less than 1/120 of the population would have one representative and counties with several times 1/120 might only be apportioned two representatives.

If the United States House of Representatives had 60 members apportioned among the States on the basis of population, California would have 4 representatives; Texas would have 3; New York, Florida, Illinois, Pennsylvania, and Ohio two each; and all of the remaining States, whether Michigan or Wyoming, one a piece.

A similar provision in Alabama, which required one representative for each county was at the crux of Reynolds v Sims. Alabama had compounded the problem by never reapportioning as was required by the Alabama constitution. Alabama has not since amended its constitution, they just ignore it.

The 1971 North Carolina Constitution removed the requirement that each county have a representative, but included a provision that prevents division of a county. Thus some districts will be multi-county with a single representative, others will be single-county with multiple representatives, and some might be multi-county with multiple representatives.

This map complies with the North Carolina Constitution.



Counties with a population greater than 1/120 of the population would be single-county districts, apportioned additional representatives based on population. Counties with less than 1/120 of the population are aggregated into five contiguous districts: (1) Eastern, which wraps around from Caswell to Columbus, 33 counties with 14 representatives; (2) Western, which stretches from Stokes to Cherokee, 24 counties with 10 representative; (3) Southern, between Charlotte and Fayetteville, 6 counties with 3 representatives; (4) Chatham-Lee, two counties with two representatives; and (5) Lincoln one county with one representative. In total, there are 66 small counties, with a population equivalent to 30 representatives. The 1971 constitution freed up 36 additional representatives, added to the 20 available from the 1868 constitution, to distribute a total of 56 representatives to the 34 largest counties. These 34 counties with 3/4 of the state population are apportioned 90 representatives, or 3/4 of the total.

Of course, there is no reason not to group the smaller counties into single-district groups. There is one three-county two-district group of Alleghany-Stokes-Surry. Otherwise, Stokes would have been a single-county district with 58.7% of a quota.

[jimrtex]

I figured out how to fix my original plan with 43-groups and 121 representatives. Apportion one less representative to Mecklenburg.



The North Carolina Constitution requires that the 120 representatives be apportioned among the districts such that the number of persons per representative is a near as equal as may be. When you apportion 120 representatives using the harmonic means for the divisors, Mecklenburg gets 11 representatives. If there were 121 representatives, Mecklenburg would get it.

The North Carolina Stephenson decision says the deviation should be between -5% and +5% in order to comply with SCOTUS OMOV decisions. But that is not precise. The SCOTUS has sanctioned deviations that go outside that range, particularly in order to recognize political subdivisions such as counties.

Since the North Carolina Constitution forbids division of counties, it is reasonable to go outside the 5% limits to avoid splitting counties. The above map sets Pitt and Surry counties off as single-county districts.

Surry has a deviation of -7.3%. Placing Surry in its own group, eliminates a a cut of Wilkes simply to get Surry into range. The remaining 4-county group of Alexander-Alleghany-Wilkes-Yadkin has two nrepresentatives add a deviation of -1.9%.

Pitt has a deviation of +5.8% for two representatives. Placing Pitt in its own group eliminates a tiny part of Pitt being attached to adjacent counties. The remaining group of Craven and Beaufort has a deviation of -4.8% for two representatives.

Mecklenburg with 11 representatives has a deviation of +5.2%.

Green Unchanged, no overturned districts: 23 groups with 50 districts.

Blue New Groups, no overturned districts: 5 groups with 9 districts.

White New groups, overturned districts: 10 groups with 20 districts.

Yellow Unchanged, overturned districts: 7 groups with 41 districts.

Overall, up to 70 (58.3%) districts may have to be redrawn. If we discount the seven single-district groups that are in new groups, up to 63 (52.5%) districts may have to be redrawn.

Standard deviation assuming perfect splits of multi-district groups: 3.22%.

Forced county cuts (7 total):

(1) Haywood. One fragment with Jackson-Swain, one with Madison-Yancey.

(2) Wilkes. One fragment with Alexander-Yadkin, one with Alleghany. Almost all of Wilkes (95%+) will be with Alleghany.

(3) Montgomery. One fragment with Stanly, one with Randolph.

(4) Robeson, One fragment with Hoke, one with Columbus, and one whole district in Robeson.

(5) Harnett. One fragment with Lee, one with Johnston-Wayne, one entirely in Harnett.

(6) Sampson. One fragment with Bladen, one with Duplin.

(7) Granville. One fragment with Person, one with Vance-Warren.

Number of groups per number of counties (ie there are 17 2-county groups)
1: 14
2: 17
3: 6
4: 6
5: 2

Number of groups per number of districts (ie there are 15 2-district groups)
1: 15
2: 15
3: 5
4: 4
5: 2
6: 2
11: 2

32 counties are in 11 single-member, multi-county groups. All such groups/districts are consistent with the North Carolina constitution. There are an additional 4 single-member, single-county groups.

[jimrtex]

The NC legislature is being taken to court (again!) about their slow-walking their constitutional responsibilities. It will be interesting to see what sanctions come down, if any.

https://www.apnews.com/0553c981d6094ac99e9cd2b5859b07cb

The federal district court rejected the plaintiffs demand for a special election. Some of the districts will have new boundaries for the 2018 election. Given that 2018 is an off year in North Carolina with no presidential, no gubernatorial, and no senatorial elections, Republicans are likely to gain legislative seats.

[kph14]

You must be logged in to read this quote.

I disagree with that and have quite a few reasons to support that:
- It's a Trump Midterm and right now Democrats are far more energized than Republicans which is visible in Special Elections all around the country. Turnout drop of Dems will likely be smaller then in past Midterms.
- Money. Gov Cooper efforts have already cashed in three times as much the Reps have got so far. Eric Holder's group is also very interested in investments in NC
- Redistricting. The current map is a racial gerrymander against black voters. Therefore a new constitutional map has to give African Americans more voting power.
- Voting Patterns. There a quite a lot Republican Legislators in districts Clinton carried. So these Republicans could go under if these Democratic trends continue downballot

Edit: And some of the districts will be modified in 2018? Even Republicans admit that they have to alter two thirds of all General Assembly districts to correct the 28 unconstitutional ones.

[JerryArkansas]

The NC legislature is being taken to court (again!) about their slow-walking their constitutional responsibilities. It will be interesting to see what sanctions come down, if any.

https://www.apnews.com/0553c981d6094ac99e9cd2b5859b07cb

The federal district court rejected the plaintiffs demand for a special election. Some of the districts will have new boundaries for the 2018 election. Given that 2018 is an off year in North Carolina with no presidential, no gubernatorial, and no senatorial elections, Republicans are likely to gain legislative seats.

No, just no.  With how gerrymandered these districts were, they will not be gaining seats.  That's partisans dreams.

[Brittain33]

The NC legislature is being taken to court (again!) about their slow-walking their constitutional responsibilities. It will be interesting to see what sanctions come down, if any.

https://www.apnews.com/0553c981d6094ac99e9cd2b5859b07cb

The federal district court rejected the plaintiffs demand for a special election. Some of the districts will have new boundaries for the 2018 election. Given that 2018 is an off year in North Carolina with no presidential, no gubernatorial, and no senatorial elections, Republicans are likely to gain legislative seats.

No, just no.  With how gerrymandered these districts were, they will not be gaining seats.  That's partisans dreams.

Jim knows that, he's just trolling.

[krazen1211]

The NC legislature is being taken to court (again!) about their slow-walking their constitutional responsibilities. It will be interesting to see what sanctions come down, if any.

https://www.apnews.com/0553c981d6094ac99e9cd2b5859b07cb

The federal district court rejected the plaintiffs demand for a special election. Some of the districts will have new boundaries for the 2018 election. Given that 2018 is an off year in North Carolina with no presidential, no gubernatorial, and no senatorial elections, Republicans are likely to gain legislative seats.

The legislature can fortify and consolidate the districts in Mecklenberg County.

[jimrtex]

You must be logged in to read this quote.

I disagree with that and have quite a few reasons to support that:
- It's a Trump Midterm and right now Democrats are far more energized than Republicans which is visible in Special Elections all around the country. Turnout drop of Dems will likely be smaller then in past Midterms.
- Money. Gov Cooper efforts have already cashed in three times as much the Reps have got so far. Eric Holder's group is also very interested in investments in NC
- Redistricting. The current map is a racial gerrymander against black voters. Therefore a new constitutional map has to give African Americans more voting power.
- Voting Patterns. There a quite a lot Republican Legislators in districts Clinton carried. So these Republicans could go under if these Democratic trends continue downballot

Edit: And some of the districts will be modified in 2018? Even Republicans admit that they have to alter two thirds of all General Assembly districts to correct the 28 unconstitutional ones.

Do you have presidential election results by House District? What is your source for the claim of "quite a few".

Democrats House member hold five McCain 2008 districts, while Republicans hold one Obama 2008 districts. I compared county results from 2008 to 2016 (remember Obama carried North Carolina in 2008), just to make sure that whiter areas in western North Carolina had not voted against Obama because he was black. But support for Clinton dropped off the table in those areas. They didn't vote for Clinton because she was a Democrat and because she was Clinton. Some of those McCain districts that are now held by Democrats, might have been carried by Clinton, since some are in more urban areas.

Her support held up much better in urban areas, but the Democrats already hold most of the seats in those areas:

Mecklenburg (Charlotte) 8:4
Wake (Raleigh) 8:3
Cumberland (Fayetteville) 3:1
Durham-Orange (Durham-Chapel Hill) 5:0
Guilford (Greensboro-High Point) 3:3

Democrats gained a few districts in 2016 which might have been drawn as Republican opportunity districts. The districts were not really drawn in order to bleach adjacent Republican districts. They were drawn to meet an arbitrary quota, to get Eric Holder's USDOJ to preclear the map.

You can not assign persons voters on the basis of race. "Vote here", "Why?", "You're Black" is illegal.

"Vote here", Why?", "You're Black" is just as illegal if done to make their votes more efficient.

You can assign voters on the basis of where they live. That is the whole point of districts, no? If there is a community, it has to be based on locality. Areas where there are AME churches, and barber shops and beauty salons that specialize in treating persons with particular hair types, and style preferences, tend to be predominately black. The churches and barber shops are part of what defines the community.

In the more rural areas of the state, the percentage of blacks goes down as you go south. Bladen-Sampson-Duplin is 29% BVAP and the population for two representatives. Just by chance you might get two 29% BVAP districts, which might well be two Republican districts. There might be some differentiation, so it is 34% and 24%, so it is an even split.

But if you start shifting people around you can quickly dislodge the few remaining Democrats in the area.

Voters who are energized by opposition to Trump tend to be concentrated in areas that would vote for Maduro. In 2018, there will be no presidential election, no gubernatorial election, no senatorial election, no council of state election. Turnout will be down, just like it is in every off-year election.

[jimrtex]

No, just no.  With how gerrymandered these districts were, they will not be gaining seats.  That's partisans dreams.

Jim knows that, he's just trolling.

The districts are not as gerrymandered as some fantasize.

In Cumberland, there were two districts that were 47% and 54% BVAP after the 2010 census. They are now both just over 50%. The problem is that they were achieve a numeric target to get preclearance from the USDOJ. Though Eric Holder might want to say that a plan retrogresses because it is not a Democratic gerrymander, he knew that would not stand up in coutt.

You might remember the Ohio redistricting contest, where the NAACP was one of the sponsors. One of their criteria was to draw two black majority senate districts in Cuyahoga County. The only way to do that is to have two districts running east to west, ignoring ward lines, and placing two very black house districts with one quite white house district in western Cleveland. You were literally having to choose between putting Parma or Lakewood in a black district, because at 7% black, they were blacker than the alternatives. Because I couldn't do this, my plan was not even scored.

They also wanted a majority black district in the Cleveland-Akron area. I don't know why they didn't like my solution of using the median of interstates to connect the two cities. I even included the parts of the cloverleafs between the interstates so that you could drive between Cleveland and Akron and stay in the district.

[Nyvin]

No, just no.  With how gerrymandered these districts were, they will not be gaining seats.  That's partisans dreams.

Jim knows that, he's just trolling.

The districts are not as gerrymandered as some fantasize.

Having super-majorities in a state that votes 3% Republican is not normal.   No matter how much you try to spin it, the maps help Republicans massively.