Using County Clusters in LA

[muon2]

A state cannot pass a law which changes a federal law, unless the federal law accommodates that. So federal courts would evaluate whether a minority population was compact under its own standards. I suppose in fashioning those standards up to a point they might at the margins give some cognizance to the state rules, but they need not do so, and will certainly not do so if to the court a minority population is clearly compact, and the majority-minority CD goes away solely because state rules due to anti-chop provisions or whatever. The VRA doesn't give a damn about chops really. 

You are bringing up a new issue here, as to which I am quite confident of the answer, but what I was focusing on is what the state does, where there is some uncertainty as to what the VRA requires, or as per us, some disagreement as to what the VRA requires.

Excellent. We agree.

So I return to my suggested rule, where I said

You must be logged in to read this quote.

Since you would have regions be a tool and nothing more it could be amended to say: "All federal laws shall be followed including the VRA. If a district required by federal law must violate state redistricting criteria, it shall be drawn in such a way as to violate as few of the other criteria as practicable." Of course I still prefer regions.

[Torie]

Fine but that still does not address what to do if the VRA's scope is uncertain, and even I stipulate it is uncertain on this issue. What we disagree about is the odds.

[muon2]

Fine but that still does not address what to do if the VRA's scope is uncertain, and even I stipulate it is uncertain on this issue. What we disagree about is the odds.

What I observed in a number of states in 2011 was the mappers retaining expert lawyers to advise them on the likelihood of a VRA district. CA did this as a report to the commission, though in states like IL the assessment was made in private. Then in IL the expert came forth after the map was drawn to provide testimony about the need and means to satisfy the VRA. For instance they successfully made the case that a VRA district may need to be in an area with 50% VAP, but the district need not if they can show that it would likely for the minority to elect the candidate of choice. For example using a 46% BVAP in a strongly Dem area.

I know that doesn't help much, so we may have to draw districts both ways under the rules.

[muon2]

If I just use the urban county cluster model on LA without respect to the VRA, and I want to avoid a county fragment as a bridge, I get the following plan. There is only one chop in Tangipahoa, and a microchop in Plaquemines. All the districts are within 0.5%.

The highest BVAP is 35.8% in CD 4 which is R+12. The best chance for a black candidate of choice in in CD 1 with BVAP 34.6% and R+4. Could this alternative survive a VRA challenge?

[traininthedistance]

If I just use the urban county cluster model on LA without respect to the VRA, and I want to avoid a county fragment as a bridge, I get the following plan. There is only one chop in Tangipahoa, and a microchop in Plaquemines. All the districts are within 0.5%.

The highest BVAP is 35.8% in CD 4 which is R+12. The best chance for a black candidate of choice in in CD 1 with BVAP 34.6% and R+4. Could this alternative survive a VRA challenge?

According to jimrtex's map on page 3, St. James County (which you have as the southernmost part of the Baton Rouge district) is actually part of the New Orleans urban cluster:

New Orleans-Metairie, LA 1190  1190  (: Jefferson Parish 433 UA; Orleans Parish 344 UA; St. Tammany Parish 234 UA; St. Charles Parish 53 UA; St. John the Baptist Parish 46 UA; St. Bernard Parish 36 UA; Plaquemines Parish 23 UA; and St. James Parish 22 UC.

[jimrtex]

Edit: I messed up the population split for Jefferson Parish, and did not subtract the Grand Isle portion from the main portion.



The Louisiana Redistricting Commission released this map showing county connectivity.   In so doing they adopted all my recommendations.   Geaux figure.

County clusters are in shown as a solid color.   Iberville has been added back to the Baton Rouge urban county cluster.

Adjacent counties that may be joined directly are shown with blue links.  Adjacent counties that may not be joined directly are shown with red links.  These are due to either corner or near-corner connectivity, or lack of easy road connectivity.   Broken lines indicate where it might appear counties may appear connected, but actually aren't.

Specific Interpretations:

St. Martin Parish:  The two parts of the parish are treated as being separate.   This is to prevent a map that might attempt to connect Lafayette to Assumption via St.Martin.   Population for the two parts are 50,717 and 1,443, respectively for the upper and lower parts.

Iberia Parish: The eastern panhandle (east of Lake Fausse Pointe) is uninhabited, as is treated as if it were a large lake.  It may not be used for connectivity, but would be added back in after the districts were drawn.  Marsh Island south of Vermillion Bay is treated in a like fashion.

Jefferson Parish: Grand Isle is treated as a separate area, connected only to Lafourche Parish. There are no residents on the main"land" south of Lafitte.  It is 118 miles by road (2.5 hours) between Lafitte and Grand Isle.  The unpopulated area will be placed in whichever district Lafitte is placed in.  Population for the two parts are 431,256 and 1,296, for the main and Grand Isle parts respectively.  Grand Isle is not included in either the New Orleans or Houma-Thibodaux urban county clusters.

West Feliciana Parish: The detached part of the parish (Turnbull Island) at Old River is treated as non-existent for connectivity purposes.  In particular, Point Coupee and Concordia are connected, while West Feliciana and Concordia are not (the commission rejected the connection via US 61 through Woodville and Natchez, MS).  The detached area of West Feliciana will be placed in the same district as Angola, in northwestern West Feliciana Parish.

Madison Parish: The detached oxbow below Vicksburg, MS is treated as being an offshore island of the rest of the parish, and does not provide a connection to Tensas Parish, independent of the main county boundary.

Lake Pontchartrain: Parish boundaries are treated as ending at the shoreline for purposes of drawing districts.  They will be attached to the adjacent land after the map is drawn.

The Lake Ponchartrain Causeway connects Jefferson and St. Tammany parishes.  This should be used to connect substantive parts of the parishes (eg Metairie and Mandeville), and not as a sneaky way to link for example, Orleans and Tangipahoa parishes).

The I-10 bridge between New Orleans and Slidell connects Orleans and St. Tammany parishes.  If this link is used to indirectly connect St. Tammany and St. Bernard parishes, all of Orleans Parish east of I-510 must be included in the district.


It might be preferable to use a map based on the 5 or 10 foot elevation  above sea level, plus land protected by levees, which would create a bunch of long peninsulas and islands.  Roads between these areas would be treated as causeways.  Using the actual county boundaries may provide a distorted perception of the population being represented.

[jimrtex]

According to jimrtex's map on page 3, St. James County (which you have as the southernmost part of the Baton Rouge district) is actually part of the New Orleans urban cluster:

New Orleans-Metairie, LA 1190  1190  (8): Jefferson Parish 433 UA; Orleans Parish 344 UA; St. Tammany Parish 234 UA; St. Charles Parish 53 UA; St. John the Baptist Parish 46 UA; St. Bernard Parish 36 UA; Plaquemines Parish 23 UA; and St. James Parish 22 UC.

That map was based on Metropolitan Statistical Areas which are based on commuting patterns, and include counties that aren't particularly urban.

The concept of Urban County Clusters is to handle big blobs of population which are difficult if not impossible to divide into equal population districts based on county boundaries, or which may result in perverse outcomes when they do - such as a large suburban county with the population for half a district stretching across the state to pick up the rest of the district.  In addition in urban areas, county boundaries are a weaker source of identification, with residents, workers, and shoppers crossing the county boundary oblivious to its existence.

An Urban County Cluster means "county-based districts don't work as well here, identify a compact area that contains the blob and can be represented by one or more districts, and treat them separately.

The trick is to define what constitutes the blob.  There are arbitrary elements to the census bureau's definition of urbanized areas.  Metropolitan areas are based on commuting patterns.  And thus also provide an economic relationship, permitting grouping of multiple urbanized areas.

But metropolitan areas also include non-urban areas.   Rural areas may actually be more likely to be included than an area with a large town at the same distance from the big city.  The large town will provide more local jobs, such as doctors and dentists and insurance agents.  There might be a small manufacturing plant and a Walmart.  A smaller share of the work force will have to commute.

The counties eliminated in the first pass were shown in a lighter tint.  These included counties with no urban population, or only small towns that aren't even large enough to be recognized as the core of a micropolitan area.  So in essence, the largest town has less than 10,000 persons, or if it does, less than 5000 are in county.

So the map really shows metropolitan areas with the counties clearly don't belong.

St. James barely meets the first pass, as Gramercy-Lutcher Urban Cluster has about 15,000 persons.

St.Charles and St.John the Baptist have a large share of the population in the New Orleans Urbanized Area.   In the case of St.John the Baptist, it is because the census  bureau allows "jumps" and "hops" over rural territory to reach addition  dense settlement.  Gramercy is just a bit too far to qualify as a "hop".   Because St.Charles and St.John the Baptist are in the New Orleans Urbanized Area, they are central "counties".  St.James qualifies for inclusion based on commuting to those counties, because it is too far for easy commuting into Jefferson and Orleans.

Saint Tammany is outside the New Orleans Urbanized Area, but includes two urbanized areas: Mandeville-Covington and Slidell.  But 25,000 workers commute across Lake Ponchartrain.

[traininthedistance]

Here's a Louisiana that I don't particularly like, but might be worth talking about:



Whole counties!  And the only split urban cluster is New Orleans; and Iberville is in the Baton Rouge group (since it seems like small changes in definition can place it in or out, it would be good to have it)  But the NO cluster is divided among three districts rather than two, and the VRA is flagrantly ignored, and the shape of District 3 leaves a lot to be desired.

District 1 is 35.9% BVAP, and 50.8% Obama (so, R+1 or +2?).

I think ultimately I still prefer muon's first map (well, the NO-BR one, not the NO-Lafayette one).

[jimrtex]

Here's a Louisiana that I don't particularly like, but might be worth talking about:



Whole counties!  And the only split urban cluster is New Orleans; and Iberville is in the Baton Rouge group (since it seems like small changes in definition can place it in or out, it would be good to have it)  But the NO cluster is divided among three districts rather than two, and the VRA is flagrantly ignored, and the shape of District 3 leaves a lot to be desired.

District 1 is 35.9% BVAP, and 50.8% Obama (so, R+1 or +2?).

I think ultimately I still prefer muon's first map (well, the NO-BR one, not the NO-Lafayette one).

The inclusion of St. John the Baptist would disqualify the map.

[Torie]

On the issue of minority geographical compactness again, this Brief filed in a NY case attacking a minority district in Nassau because it was not geographically compact, taking in disparate places, and a re-reading of Abrams v Johnson (with the court commenting on the combining of disparate black population nodes syndrome), are elucidative in suggesting that (i) drawing a CD that picks up two disparate minority nodes is not required under Section 2, and (ii) if by picking up those disparate nodes, other appropriate redistricting principles are abandoned to that end, that CD may itself be an illegal racial gerrymander.

The more and more I explore this, the more and more I am persuaded that the scope of Section 2 in forcing the drawing minority CD's that violate other redistricting principles, and would never be drawn absent race as a factor, is very narrow indeed. If a reasonable minority CD can be drawn, that does not violate other principles in play, and is just one reasonable choice among others, than draw it. Otherwise, it's not required, and if too gross, is itself illegal. Indeed after having done this little excursion, I will be more circumspect in drawing minority CD's than I have been heretofore. I believe that while my former interpretation of the scope of the VRA here was narrower than Muon2's, it was not narrow enough. I just won't be breaking many other redistricting rules to draw them, particularly if the only reason to do so, was to pick up geographically disparate minority population nodes, as opposed to a contiguous one, that just happens to spill inconveniently over a number of local jurisdictional lines - particularly, in our chop obsessed minds here, county ones.

So Judge Torie hereby rules that the NO-BR black barbell populated CD is not required under Section 2 of the VRA, but neither is it illegal, because it does reasonably hew to other redistricting principles in play, or at least not totally shred them.

[jimrtex]

Please remind me again, just why it is preferable to ban one extra chop of an urban cluster, while on the other hand, one extra county chop not involving an urban cluster, is just one more chop, a demerit yes, but to be balanced against erosity factors? Thanks.

Our objective is to curb gerrymandering, and produce districts that are representable and have conceptual integrity.

Compactness is a way to increase regional integrity, so that you and your neighbors elect the same representative, increase the chance that your representative will be from your area, and can be more responsive by representing fewer interests, have to maintain fewer offices, be able to not waste advertising on reaching splinters of media markets.  Erosity is a measure of compactness, and in turn improves conceptual integrity.

Similarly, containing districts within a UCC, means that the representatives are likely  to be from the metropolitan area, and perhaps have a single district office.  While they may still have to pay for expensive media, they are at least reaching most, if not all, of their constituents.   And county cuts within a metropolitan area are relatively inexpensive.

In this particular case, the UCC chop was 45,000 persons, which also means that 45,000 persons elsewhere had to be drawn into a New Orleans district.  I am coming up with plans with under 10,000 in county chops in Louisiana.   So the magnitude of the chop is being hidden under a guise of it just being another county.

[jimrtex]

It also has just occurred to me now, because I'm an idiot, that:

The concept of Urban County Clusters is to handle big blobs of population which are difficult if not impossible to divide into equal population districts based on county boundaries, or which may result in perverse outcomes when they do - such as a large suburban county with the population for half a district stretching across the state to pick up the rest of the district.  In addition in urban areas, county boundaries are a weaker source of identification, with residents, workers, and shoppers crossing the county boundary oblivious to its existence.

An Urban County Cluster means "county-based districts don't work as well here, identify a compact area that contains the blob and can be represented by one or more districts, and treat them separately.

is actually saying that county chops should not really be minded so much within the cluster (perhaps intra-cluster chops should be counted at a discount rate, if at all, in any scoring system?), and that perhaps within the state as a whole the UCC is to be treated as a super-county for the purposes of chop identification and minimization.

Which is a stance I wholeheartedly endorse.

Within major metropolitan areas, you will probably have to cut counties, and the magnitude of the county fragments may be large.  I can't really say that splitting County A 300 thousand and 200 thousand is worse than cutting County B 300 thousand and 125 thousand, along with associated grouping of other counties is better or worse.

I see the division of multi-county regions as a separate step, and it can be controlled by limiting the number of cuts to the minimum necessary (N-1), disallowing multi-spanning where two or more districts have parts of the same pair of counties, and emphasizing respect for local communities of interest, particularly where there are local political boundaries available.

Note that I specifically permit forming whole county regions within UCC.  That actually eliminates a (significant) county cut.

In New Orleans, Jefferson+Orleans (1.026), Jefferson+St. Tammany+St Charles+St John the Baptist (1.013), Jefferson+St.Tammany+St.Charles+Plaquemines (0.980), and Jefferson+St.Tammany+Plaquemines+St.Bernard (0.958) are all within a 5% limit.  However, they both  cut off the remnant of the UCC cluster from the rest of the state.

[muon2]

Here's a Louisiana that I don't particularly like, but might be worth talking about:



Whole counties!  And the only split urban cluster is New Orleans; and Iberville is in the Baton Rouge group (since it seems like small changes in definition can place it in or out, it would be good to have it)  But the NO cluster is divided among three districts rather than two, and the VRA is flagrantly ignored, and the shape of District 3 leaves a lot to be desired.

District 1 is 35.9% BVAP, and 50.8% Obama (so, R+1 or +2?).

I think ultimately I still prefer muon's first map (well, the NO-BR one, not the NO-Lafayette one).

This illustrates why it is important to have more constraints than just whole counties. If the only constraint is whole counties I could potentially attack the plan as long as I produce a map that is also whole counties but reduces the population inequality. WV survived its challenge in Tennant because it had the additional constraint that it wanted to shift the fewest number of residents to a new district.

For example, your plan has a range of 0.755% (st dev 0.292%). This erose mess has a range of 0.360% (st dev 0.133%).



As to train's map and chops of clusters. I view that the urban county clusters must sit within one region. More regions correspond to a better plan with the constraint that erosity should not increase while increasing regions. Since the Lafayette CD includes part of the NO cluster, it is in the same region as CD 1 and 2. That leaves a total of 4 regions which matches the best of any other plan including my various offerings. Then erosity kicks in, and the shape of train's Lafayette CD will hurt that plan. As a side note, my erose plan above loses a region since CD 5 is now pulled into the BR cluster, so it loses on that count as well.

BTW, when did Iberville go back into the BR cluster? I thought that was resolved earlier in this thread and it did not meet the 50% UA threshold.

[Torie]

Jimtex's map is not gospel as to where the urban clusters are? There is a modification of what the census says?  I am raising this in the context of Iberville, which I believe was in the NO cluster on Jimtex's map, which was derived from the census definition. I some counties are culled, what precipitates that exactly again? We have so many rules running around, not put in any way place, either comprehensibly, or incomprehensibly, so sometimes the discussion devolves into something rather chaotic.

[Torie]

Please remind me again, just why it is preferable to ban one extra chop of an urban cluster, while on the other hand, one extra county chop not involving an urban cluster, is just one more chop, a demerit yes, but to be balanced against erosity factors? Thanks.

Our objective is to curb gerrymandering, and produce districts that are representable and have conceptual integrity.

Compactness is a way to increase regional integrity, so that you and your neighbors elect the same representative, increase the chance that your representative will be from your area, and can be more responsive by representing fewer interests, have to maintain fewer offices, be able to not waste advertising on reaching splinters of media markets.  Erosity is a measure of compactness, and in turn improves conceptual integrity.

Similarly, containing districts within a UCC, means that the representatives are likely  to be from the metropolitan area, and perhaps have a single district office.  While they may still have to pay for expensive media, they are at least reaching most, if not all, of their constituents.   And county cuts within a metropolitan area are relatively inexpensive.

In this particular case, the UCC chop was 45,000 persons, which also means that 45,000 persons elsewhere had to be drawn into a New Orleans district.  I am coming up with plans with under 10,000 in county chops in Louisiana.   So the magnitude of the chop is being hidden under a guise of it just being another county.

So the argument is that UCC's have more of a community of interest than an array of rural counties, and therefore an excess UCC chop should be banned as opposed to merely penalized? I am not comfortable with that argument. If an excess chop makes for a far better map overall, with less chops and erosity elsewhere, it should be permitted. I am not persuaded such a ban is necessary to make for better maps, and less gaming.

[muon2]

You think SCOTUS would revise its interpretation of Section 2 Muon2, and revise the first prong of the Gingles test, to require a black CD if the percentage population of blacks in that state is high enough, even if not geographically compact?  If so, I don't agree, but in the meantime I would prefer to rely on existing precedents as extrapolated the way a lawyer would extrapolate them to fill in the gaps.

No and yes. I don't think they would revise their interpretation of section 2, but they would revisit the plain text of the statute. This is a hypothetical case where a southern state creates a set of redistricting rules and based on those rules provides for no district where blacks would have the opportunity to elect a representative of choice. I think SCOTUS would refine prong one of Gingles to provide for compactness to conform with generally recognized neutral redistricting principles, which I think we agree my NO-BR district follows. My district just doesn't follow principles as strict as the anti-gerrymandering rules we are entertaining. They have refined Gingles a number of times since it was first decided and I think this hypothetical would cause them to refine it yet again.

[Torie]

You think SCOTUS would revise its interpretation of Section 2 Muon2, and revise the first prong of the Gingles test, to require a black CD if the percentage population of blacks in that state is high enough, even if not geographically compact?  If so, I don't agree, but in the meantime I would prefer to rely on existing precedents as extrapolated the way a lawyer would extrapolate them to fill in the gaps.

No and yes. I don't think they would revise their interpretation of section 2, but they would revisit the plain text of the statute. This is a hypothetical case where a southern state creates a set of redistricting rules and based on those rules provides for no district where blacks would have the opportunity to elect a representative of choice. I think SCOTUS would refine prong one of Gingles to provide for compactness to conform with generally recognized neutral redistricting principles, which I think we agree my NO-BR district follows. My district just doesn't follow principles as strict as the anti-gerrymandering rules we are entertaining. They have refined Gingles a number of times since it was first decided and I think this hypothetical would cause them to refine it yet again.

So as long as a majority minority CD is but a two chop CD, with no traveling chops, and within some erosity limit, than the SCOTUS is going to find that the VRA requires that such a CD be drawn, irrespective of minority compactness or contiguity. Does that about sum it up? Does it matter if such a CD increases the chop count of the map overall?

Returning to the ban versus penalty issue for urban cluster chops, to me that issue turns in the end on high the likelihood is that it would generate ugly maps. I view it as unwise to go there unless and until that specter is put to bed. Univariate rules tend to worry me, when what we are doing in the end is all about finding the Golden Mean, balancing off this against that. I am also unsure how wise a ban would be on policy grounds as well (putting urban clusters and their separation from the rurals, and visa versa, at the top of the heap of considerations, as opposed to but one consideration among others). The rationale is that it further helps to rein in gaming seems weak to me. A penalty would do about the same thing I would think. In the end, the score is the thing.

[jimrtex]

This spin on the compactness test was reiterated by SCOTUS as recently as 2009 in Bartlett v. Strickland, wherein the Court stated the first prong of the Gingles test as:

“a large and geographically compact minority population, is the condition for demonstrating for demonstrating that a dilutive plan injures the Section 2 plaintiffs by failing to draw an available remedial district that would give them a chance to elect their chosen candidate.” (Emphasis added.)

....

In that regard, a pending case wherein a judge upheld an erose black supervisorial district within a county (Fayette County, GA), and is under appeal, the precise issue is whether or not the minority drawn district is sufficiently compact. I strongly suspect that more erosity will be tolerated that is within a given county or municipality, even if the minority population is not contiguous, than a CD that slashes and burns its way from one minority population node to another in a different metro area, with nothing much but alligators and bayou Cajuns in between. Indeed, in the Fayette County case it appears that the black population was more of less contiguous, so the only issue is erosity within the county - to pick up 35 voters somewhere. So in that case, it would not surprise me if the VRA did require that black district in Fayette to be drawn. We shall see (if the appeal is pursued to its conclusion).

The district court in the Fayette County case emphasized that the condition was the compactness of the population, not that of the district that was drawn.

Fayette County currently has at-large elections for school board and county commission, but requires those elected to reside in a district.  What is at issue in this case is not so much the district boundaries, but whether at-large elections can continue to be used.

If Fayette County had district elections, they would probably have a better case.  The districts would simply recognize different areas of the county.  The hybrid system means that if the person who lives in a district is intended to represent the interests of that district, that those who are represented, do not elect their representative.

Both county commissioners and the school board were sued.  The school board settled, and the district court thought that the case had been settled.  But since the commissioners had not, the case continued.  An issue in the case is whether the exemplary district that demonstrated that a majority-minority district could be drawn, that it be used.   The consent agreement provided for a district with a 44.5% district.

The district court granted summary judgment to the plaintiffs, and denied the commissioners an interlocutory appeal.  So currently, the case is in a remedial phase.  The judge did order that the plans submitted by the parties have a 50% BVAP.   The commissioners have submitted a plan, but have also requested that the court delay any order until the legislature has a chance to act.  Apparently in Georgia, the legislature passes local laws for commissioners districts, and the legislature has not met since the courts order.   OTOH after the census, the commissioners were sued for violation of OMOV, and agreed to a consent decree - so it may be the actual practice to sue and then settle, rather than actually legislate.  But it could also be argued that the plaintiffs failed to include a necessary party, the State of Georgia, or the Georgia legislature.

The population of the county is concentrated on the eastern and western sections of the county, with the largest concentration around Peachtree City n the west central part of the county (see concentration of precincts in second map).  So the western arm of the proposed district is crossing over to pick up some blacks inn Tyrone.  It also has the hook into Fayetteville, which has to avoid cutting off the less-black area to the northeast of Fayetteville.

The proposed district rips election precincts apart, so the problem is really not the concentration of the minority population but its numbers.  It is a lot easier to draw 1 of 6 districts when there is a 33% black statewide population, than 1 of 5 with a 20% black countywide population.

The majority-minority district is underpopulated.  So besides the 30 blacks that the article noted provided the majority, the district avoided picking up 600 persons, of which 300 would have had to been black.  This is like in Cleveland when trying to create two majority black senate districts, you had to include Parma or Lakewood to get jurisidictions that had 6/13 of the population and had the most blacks.

The opinion said that people in Tyrone attend a mosque in Fayetteville, which is somehow supposed to establish a community of interest.  Also Fayetteville and Tyrone are 3 miles apart at their nearest point.  Of course the court neglects to mention that Fayetteville has an arm that reaches toward the west, and the portion of Fayetteville that is included in the minority-majority district is on the east side of town, and it is probably 12 miles via the district.  If the school board drew an attendance zone like that they would be ripped apart.

The Louisiana congressional district is most likely a political decision.  With Louisiana losing a district, someone is going to get their feelings hurt no matter what you do.  After the census, LA-2 had a 54.5% BVAP and needed 250,000 added, LA-1 had a 14.5% BVAP.  You couldn't come close to maintaining a minority-majority district when you already had a 40% differential.

But you get black legislators to vote for the district, and you don't have to cobble up so many votes from other members, especially those concerned about the pairing of the two Cajun representatives.

LA-2 is now 58.5% BVAP, LA-1 11.5%, and LA-6 and LA-3 down around 20%.  So you have what amounts to a black pack which the black legislators voted for.  The map has hideous connectivity.  LA-6 lacks road connectivity in 3 places, and probably loops 150 miles to match up the two ends that are around 15 miles apart.  LA-1 and LA-5 are only connected on a political map.

[jimrtex]

Iberville Parish is part of the Baton Rouge urban county cluster.  Grand Isle (Jefferson) and lower St.Martin partishes are treated as separate parishes.

Standard deviation is for the error of the 5 regions from the integer number of districts for the region, expressed as a percentage of the ideal district size.  The errors for the 5 regions are: New Orleans -1.12%; Baton Rouge -0.12%; Lafayette-Lake Charles 0.31%; Alexandria-etc. 0.41%; Shreveport-Monroe 0.52%.

Boundary mileage is measured from county boundary intersection to county boundary intersection. Coastal waters are ignored, as is that shown between Vermillion County and Iberia-Marsh Island.  The Iberville panhandle is treated as a separate county for this purpose.  Region boundaries across Lake Ponchartrain are measured from where the county boundary intersects the shoreline.

Transfers between the regions are expressed as a percentage of the ideal district population.  There must be exactly N-1 such transfers (so for Louisiana 4).  Each transfer must bring one more region into full equality (and the last transfer, two regions).

In this case:

(1) Lafayette-Lake Charles: 0.3%
(2) New Orleans: 0.8%.
(3) Shreveport-Monroe: 0.5%
(4) Alexandria, etc. and Baton Rouge 0.1%

Total Transfer is the total amount of these transfers, expressed as a percentage of the state population, and number of actual persons to be transferred.

If both regions of a shift are within 0.5% of the ideal before the transfer, that shift may be excluded from the Minimum Transfer.

The after equalization standard deviation is that for the regions after all minimum transfers occur.

This map swaps Monroe for Natchitoches.



This map swaps Lake Charles for the area north pf Baton Rouge.



This map creates a Shreveport-Lake Charles district.

[muon2]

jimrtex, What data set are you using for your boundary calculations? Is there a file with all the separate boundary segments?

[jimrtex]

jimrtex, What data set are you using for your boundary calculations? Is there a file with all the separate boundary segments?

I'm measuring using my mapping program.  It uses a polyline (chained line segments), so if I misplace a county junction, it should create too much error.  I started looking at extracting the information from the census shapefiles, but haven't made too much progress.

In a real process, the length of each county-county boundary segment would be provided.

This shows the simplified boundaries that would be used.

[jimrtex]

This is a 60-district apportionment map.   Each multi-county region has an error of less than

min (5%*N, sqrt(N)/sqrt(2)*10%, 20%) from the ideal population.

This process is gerrymander resistant, and ensure districts are largely based on counties.  Since the final districts are of approximately equal population, congressional districts could be formed from 10 of them.  Conceivably this could even be done by agreement among the voters of the 60 sub-districts.



The next map eliminates all multi-county multi-member districts, so that all districts are single-county multi-member, multi-county single-member, or single-county single-member.  Had this level of population equality been achieved in 1960, Reynolds v Sims would have been unnecessary.    

For example, the southern coastal region from Terrebone to Cameron was entitled to 4.03 representatives.  It was divided into 4 districts, Terrebone 1.48, St. Mary 0.72, Iberia 0.97, Vermillion+Cameron 0.86.  In the process, 3 extra districts were created.   For example, the original apportionment was Orleans+St.Bernard (5.03) and Jefferson+Plaquemines (6.01), which was switched to Orleans (4.55), Jefferson (5.72), and St.Bernard+Plaquemines (0.78), resulting in 12 rather than 11 districts.



This shows the BVAP per district or multi-district county.  The underlying principle is that there are real communities of interest that are predominately black, that may be aggregated into a congressional district.  This is not dissimilar to non-race-based congressional districts.  The size and population equality requirements means that districts are actually just aggregates of more local communities that have geographical proximity, and might have a modest regional theme.



The map illustrates the problem with creating a majority-minority district in southeastern Louisiana.  The population is not continuous.   While St.James and St.John the Baptist are majority BVAP, the 4-county Mississippi corridor is only 31% black, or about the state average.  Also notable is the broad swath of Louisiana where the black population is around the statewide average of 30%, with some decline towards the coast, and differentiation around New Orleans and Baton Rouge.

Muon's map includes majority white parishes (Ascension and St.Charles) and splits majority black parishes such as Orleans.  The official map is even more aggressive, but because of the way it cuts off the area to the north, it creates two wrap-around districts which have extremely tenuous connectivity.

Conceivably a majority-minority district could be created by combining smaller districts, perhaps Orleans 4, Jefferson 2, St.John the Baptist+St.James 1, East Baton Rouge 3, and the area north of Baton Rouge 1.   This is 11 districts, which might be necessary because all the districts are underpopulated.

The next map is based on an apportionment of 120 districts, except every county is guaranteed at least one district.   The target size is roughly 5% of a congressional district, or around 35,000.  Large counties would be divided into small areas, which we might designate community areas.   There would be no need for strict equality, and even the number of areas for a county is suggestive.  Smaller counties would have a single community area, and thus could net be divided by a minority-majority district.  What is not obvious from Muon's Alabama map is the population of the counties.  In Louisiana, the counties with the highest black population are usually the least populous (eg St.Helena, Tensas).  



This shows the BVAP per county.  Note there is significant local variation (eg East Carroll and West Carroll)



Musing on the previous map, I created the following, to create two black influence districts.  The apportionment is based on VAP, and there would be a split of Jefferson Parish (coming across on the Lake Ponchartrain Causeway to get about 150,000 persons.   That should put the BVAP for the New Orleans district up over 40%.  In creating the Baton Rouge-Mississippi district, I started with East Baton Rouge which has almost 3/5 of a district and 42% BVAP.  Even selecting the blackest parishes, I was never able to exceed 43% BVAP, because the blackest counties are typically small.



Another thought exercise.  Two black CDs, and 4 white CDs, which affords the black community the maximum opportunity to elect a candidate of choice.  The population in each county is apportioned based on BVAP for the county.  Both maps reflect the Urban County Clusters.

[muon2]

I've pulled this thread out of the larger thread on UCCs in light of the lawsuit over LA-02. It will be interesting to see if the case leads to any clarification on the issues we debated last summer.

Also, after the debate on LA and UCCs we firmed up the definition of UCCs and added the concept of MCCs. MCCs are contiguous counties of substantial minority population, which in this case is 40% or more BVAP. The cluster rules require that chopping a cluster beyond the minimum requirement counts as an additional county chop. In the map below I've identified UCCs with pink circles, using a red number to indicate the minimum number of CDs to cover the UCC. Single counties that meet the standards for a UCC are indicated with a tan circle. MCC counties are shown in green, and four contiguous clusters are present in LA.



The interaction between the UCCs and MCCs create some interesting dynamics. St. James and St. John the Baptist are in the same MCC so splitting St. James off to go with Baton Rouge would count as a chop. Also the Baton Rouge UCC and MCC are each less than one CD in population, but together exceed one CD. The result is that either the UCC or the MCC must be chopped.

[Torie]

Why does chopping a minority county cluster generate a chop if there is not otherwise a 50% BVAP CD in play?  I guess I missed that nuance. What counties are in the Baton Rouge MCC, but not in the UCC? The pink circles on your map Muon2 are not a model of clarity to me.

[Sol]

Some of the maps I've seen earlier in this thread bring to light a concern I've had with the UCC approach from the very beginning- that it can be used to justify drawing problematic districts under the umbrella of a UCC. For instance, upthread there we several maps proposing a district based in Orleans and St. Tammany Counties*. Under an approach that emphasizes UCCs exclusively, that looks great- it's an easy connection, compared to something that snakes around Lake Ponchartrain or splits Orleans. But as a CoI, it's absolutely terrible. It seems like, when it comes to fair maps, the big pitfalls of a UCC-based approach are rural minority areas (which train pointed out earlier- I'm not convinced MCCs really deal with this) and metro areas that are quite diverse in terms of CoI- this is an issue in NC as well.

IMO, the UCC approach needs to have room for local interests.


* I dislike the fact that Louisiana Counties are called "parishes". IMO, all county equivalent units should be called counties.