Using Urban County Clusters To Guide Redistricting

[jimrtex]

Of course, what the SCOTUS should do is admit that the legal basis for Wesberry was flawed (read Justice Clark's concurring opinion).  The reason why the rationale for congressional redistricting standards is so hard to comprehend, is because there is no justification in the Constitution.

And then presuming that SCOTUS followed Clark's opinion, they would by necessity walk it back to Baker v. Carr. As Clark put it "The trial court, however, did not pass upon the merits of the case, ...  I believe that the court erred in so doing. In my view, we should therefore vacate this judgment and remand the case for a hearing on the merits. At that hearing, the court should apply the standards laid down in Baker v. Carr."

(1) It is not necessary that the SCOTUS do this in order to apply the Karcher test and approve the Alabama plan.   It is your claim that they would have to go back to Baker v Carr.  It is not my claim, and it certainly is not my claim that the SCOTUS wants to do so.

While admittedly not adding a portion of an adjacent county to Jefferson is aggressive, it is not indefensible.  If the Alabama map had been in place at the time of Wesberry, it is quite unlikely that subsequent litigation would have overturned it.  It certainly conforms with international norms.   I could get 4 votes from the liberals because of this, I could get Kennedy's vote because it is non-political, and get 4 votes from the conservatives because it respects federalism.   The plan wins 9-0.  Kennedy writes the majority opinion and stresses its conformance with the Karcher test.  Thomas writes a concurring opinion that says we should respect Alabama's plan, and that if Congress wants to do anything about it they have full authority to do so.   Breyer also writes a concurring opinion stresssing international norms.

(2) From a legal standpoint, Wesberry v Sanders is not a strong decision.  There is no reason to read Article I, Section 2 as mandating equal-population districts, let alone random splitting of counties; or of stripping Congress of its authority to override State election regulations.  Interpreted literally, it means that before the 13th Amendment, if elections were at large, every vote should count equally; but if elected by districts, areas with large number of slaves, and relatively few voters (particularly if land ownership was a perquisite) would have few voters per district, while the upcountry areas where most persons owned a small piece of land, and had few slaves, would have more voters per district.

Equal protection would have served just as well as the basis for Wesberry.  After all, it cites Baker v Carr as the basis for justiciability and standing.   A 5% or 10% range as a safe harbor does not follow from equal protection, and the SCOTUS has hinted that it may not always be true.

There could be the same test just like under Karcher.   If the plaintiffs could produce a plan with less deviation, then the defendants would have to justify the variation of their plan.  By now, it would have been established that congressional districts must conform to political boundaries,

(3) Point (2) is irrelevant to Point (1) as far as my argument regarding the Alabama plan.

(4) Something to think about.  Wesberry held (holds):

"The constitutional requirement in Art. I, § 2,that Representatives be chosen "by the People of the several States" means that, as nearly as is practicable, one person's vote in a congressional election is to be worth as much as another's."

Congressional districts with equal population do not have the same number of voters.  While one could argue that population is a proxy for voters, it may not be.   One might argue that it is not practicable to determine the number of voters, but that is not true.   The ACS provides reliable data on the citizen voting age population.   So other than possibly including some felons or those judged to be mentally incompetent shouldn't this be the basis for defining congressional districts?   And shouldn't congressional districts be revised each election?  It is quite clear that the SCOTUS does not consider political practicality the same as practicability.

[muon2]

Of course, what the SCOTUS should do is admit that the legal basis for Wesberry was flawed (read Justice Clark's concurring opinion).  The reason why the rationale for congressional redistricting standards is so hard to comprehend, is because there is no justification in the Constitution.

And then presuming that SCOTUS followed Clark's opinion, they would by necessity walk it back to Baker v. Carr. As Clark put it "The trial court, however, did not pass upon the merits of the case, ...  I believe that the court erred in so doing. In my view, we should therefore vacate this judgment and remand the case for a hearing on the merits. At that hearing, the court should apply the standards laid down in Baker v. Carr."

(1) It is not necessary that the SCOTUS do this in order to apply the Karcher test and approve the Alabama plan.   It is your claim that they would have to go back to Baker v Carr.  It is not my claim, and it certainly is not my claim that the SCOTUS wants to do so.

While admittedly not adding a portion of an adjacent county to Jefferson is aggressive, it is not indefensible.  If the Alabama map had been in place at the time of Wesberry, it is quite unlikely that subsequent litigation would have overturned it.  It certainly conforms with international norms.   I could get 4 votes from the liberals because of this, I could get Kennedy's vote because it is non-political, and get 4 votes from the conservatives because it respects federalism.   The plan wins 9-0.  Kennedy writes the majority opinion and stresses its conformance with the Karcher test.  Thomas writes a concurring opinion that says we should respect Alabama's plan, and that if Congress wants to do anything about it they have full authority to do so.   Breyer also writes a concurring opinion stresssing international norms.

(2) From a legal standpoint, Wesberry v Sanders is not a strong decision.  There is no reason to read Article I, Section 2 as mandating equal-population districts, let alone random splitting of counties; or of stripping Congress of its authority to override State election regulations.  Interpreted literally, it means that before the 13th Amendment, if elections were at large, every vote should count equally; but if elected by districts, areas with large number of slaves, and relatively few voters (particularly if land ownership was a perquisite) would have few voters per district, while the upcountry areas where most persons owned a small piece of land, and had few slaves, would have more voters per district.

Equal protection would have served just as well as the basis for Wesberry.  After all, it cites Baker v Carr as the basis for justiciability and standing.   A 5% or 10% range as a safe harbor does not follow from equal protection, and the SCOTUS has hinted that it may not always be true.

There could be the same test just like under Karcher.   If the plaintiffs could produce a plan with less deviation, then the defendants would have to justify the variation of their plan.  By now, it would have been established that congressional districts must conform to political boundaries,

(3) Point (2) is irrelevant to Point (1) as far as my argument regarding the Alabama plan.

This is where we will have to continue to agree to disagree. I understand your rationale for the "aggressive" AL plan, but legal scholars I have listened to like to note that SCOTUS has not accepted a congressional plan in excess of 1% range. Even within Tennant they note that the error of the district court was to treat the 0.78% range as a major variation due to technology, when 0.79% was minor at the time of Karcher. The question your plan would pose is whether the substantially greater range in your AL plan can be construed as minor in the fashion of Karcher, and thus let the Karcher test proceed.

My citation from Clark's opinion was meant to underscore your statement that SCOTUS should find Wesberry flawed, and then my contention that removing the flaw returns one to Baker. I don't disagree that had a plan substantially like AL here been before the court the direction might have emerged differently. My belief remains that substantially revising Wesberry is the only path by which a range so large as in AL can be accommodated.

[jimrtex]

This is where we will have to continue to agree to disagree. I understand your rationale for the "aggressive" AL plan, but legal scholars I have listened to like to note that SCOTUS has not accepted a congressional plan in excess of 1% range. Even within Tennant they note that the error of the district court was to treat the 0.78% range as a major variation due to technology, when 0.79% was minor at the time of Karcher. The question your plan would pose is whether the substantially greater range in your AL plan can be construed as minor in the fashion of Karcher, and thus let the Karcher test proceed.

The first prong of the Karcher test is for the plaintiff to demonstrate that a more equal plan can be created.   It is not clear how that was interpreted by SCOTUS in West Virginia.  The District Court thought the Perfect Plan was better.   But you can always create a plan with no deviation.   A test that is impossible to fail is not a test.   The SCOTUS was so dismissive of the Perfect Plan that it might be that they didn't even think it satisfied the first prong.

West Virginia conceded that the Cooper whole-county plans had less deviation and that the first prong was satisfied.  In New Jersey, alternatives had actually been available to the legislature.

But under my hypothetical, there were no "Cooper plans" in Alabama.  If there had been, they might have been adopted in stead.

The other way of looking at the SCOTUS decisions is that they have never rejected a plan that was as non-political as my Alabama plan.   Your experts are in essence claiming that there is a quasi de minimis standard of 1%, whereas the SCOTUS has been steadfast in rejecting the notion.  The SCOTUS may look favorably on the use of standard deviation, since it was their observation that if they set a de minimis standard, legislators would aim for the edges, which they have done with legislative plans.  But the record in Alabama clearly shows that range was thoroughly rejected as an inadequate measure of equality.  

In Kirkpatrick and Karcher the SCOTUS was upholding lower court decisions where there had been some irregularities and the justification sounded like excuses.  In Missouri, it was that they were trying to adjust for population change, and resident population, but there wasn't real evidence that they had done this in a systematic fashion.  

If the Karcher test is to have any real meaning, then it has to be open to permitting justification of somewhat larger deviation.

The Alabama plan is somewhat akin to an archery competition in windy or rainy weather.  There is still a winner, even though their score is not very high.

My citation from Clark's opinion was meant to underscore your statement that SCOTUS should find Wesberry flawed, and then my contention that removing the flaw returns one to Baker. I don't disagree that had a plan substantially like AL here been before the court the direction might have emerged differently. My belief remains that substantially revising Wesberry is the only path by which a range so large as in AL can be accommodated.

They could move the legislative decisions towards Wesberry.  There is nothing inherent about equal protection that says 10% is OK, and the SCOTUS have sometimes hinted that it might not be the case.   This is particularly true when you have large congressional-sized districts.  If they were to apply the Karcher test to legislative district, and let legislatures justify deviations due to complying with county boundaries, but not for political reasons, I think you would soon have it sorted out.

[muon2]

But at some point a neutral set of rules could produce a range that is just too large. That will cross a line from a minor to a major deviation, hitherto undefined for CDs. 1%, 5%, 10%, 20% or perhaps only a range so large that they will "know it when they see it"? I agree that it would make more sense for state and federal districts to be treated more alike, but if I'm trying to craft CD rules that could plausibly be adopted by real states, then I'm not going to look at pushing too hard towards larger ranges. The maximum deviation from the quota can be a free parameter, so that if the court moves in the direction of relaxed ranges with strong neutral rules, the parameter adjusts accordingly.

[jimrtex]

But at some point a neutral set of rules could produce a range that is just too large. That will cross a line from a minor to a major deviation, hitherto undefined for CDs. 1%, 5%, 10%, 20% or perhaps only a range so large that they will "know it when they see it"? I agree that it would make more sense for state and federal districts to be treated more alike, but if I'm trying to craft CD rules that could plausibly be adopted by real states, then I'm not going to look at pushing too hard towards larger ranges. The maximum deviation from the quota can be a free parameter, so that if the court moves in the direction of relaxed ranges with strong neutral rules, the parameter adjusts accordingly.

My concern is that if you may pay too little attention to whole counties, you lose any justification in that regard.  The real weakness of Alabama's case is that the process was designed to produce limited county chops:



But looking at the map, the chops of Covington and Bibb were totally gratuitous.  But a court might look at that map and ask why it didn't go even further, and why limit the divisions to whole election precincts?   If 1% is doable, why not go to 0.5%, 0.1%?

But if Bibb and Covington are kept whole, what is the basis for deciding that Jefferson must be augmented?

The difference between my rule where the regions can have a deviation of 5% and yours where the deviation is 0.5%, is that I'm trying to identify small chops to equalize regions, and you are trying to identify regions where major chops can be made within a region.

Can it really be said that the 7 representatives elected from my districts are not elected by the People of the State of Alabama?

[muon2]

Since a number of plans have suggested BVAPs less than 50%, it's worth seeing how far that might go in preserving UCCs. To satisfy the VRA section 2 a district must provide the opportunity for the minority to elect the representative of their choice. One way to do that is by constructing a district where the minority controls the primary and then relies on white crossover voters to support the minority's chosen nominee. In AL, Tuscaloosa and Jefferson both have a number of white Dem voters, so by combining those populations it is possible to make a district with all of Tuscaloosa and part of Jefferson that is 43.4% BVAP and D+4. That should be enough to meet the primary/general crossover option.

I can use that to make a whole county plan (except for Jefferson) that maintains all the UCCs and has a range of 0.5%. The downside is higher erosity and three chops of the black belt MCC. However, since the black belt isn't used to make the section 2 district, are the chops as important? BTW though CD 7 also has BVAP of 43.4%, crossover voting is not as prevalent in that area and it rates as R+3.

[jimrtex]

Since a number of plans have suggested BVAPs less than 50%, it's worth seeing how far that might go in preserving UCCs. To satisfy the VRA section 2 a district must provide the opportunity for the minority to elect the representative of their choice. One way to do that is by constructing a district where the minority controls the primary and then relies on white crossover voters to support the minority's chosen nominee. In AL, Tuscaloosa and Jefferson both have a number of white Dem voters, so by combining those populations it is possible to make a district with all of Tuscaloosa and part of Jefferson that is 43.4% BVAP and D+4. That should be enough to meet the primary/general crossover option.

I can use that to make a whole county plan (except for Jefferson) that maintains all the UCCs and has a range of 0.5%. The downside is higher erosity and three chops of the black belt MCC. However, since the black belt isn't used to make the section 2 district, are the chops as important? BTW though CD 7 also has BVAP of 43.4%, crossover voting is not as prevalent in that area and it rates as R+3.

Using your Mobile district gives me a UCC compliant map with a single chop.



Since you used a two-district region, it is somewhat difficult to calculate its perimeter.  I used three different methods.

1. Take area of Districts 3+6, calculate diameter of circle of the same area (total = 1158).

2. Use Jefferson+Tuscaloosa as a region, which adds perimeter between Jefferson and Blount, St.Clair, Shelby, and Bibb; and Tuscaloosa and Bibb.

3. Modify your map to make Jefferson a region.   This presumes around 24,000 coming from a chop of Blount, St.Clair, Shelby, or Tuscaloosa.

Options 2 and 3 come up about the same.

So the issue is should a plan that does a large chop within a UCC county be treated the same as one where a much smaller chop is done outside the UCC, when a split of the UCC is necessary anyway.

[muon2]

I've picked this up from the Odd and Even thread since the VRA question fits better over here.

Regarding the VRA, it is arguable whether a plaintiff would satisfy the first prong of the Gingles tests, and with this district clearly fails the third prong.

For my version of the Tuscaloosa-Black Belt-Montgomery district:

2012 Obama 53.9% Romney 44.9%
2010 Sparks 53.0% Bentley 46.5%
2010 Barnes 47.6%, Shelby 52.4%
2008 Obama 52.6% McCain 46.8%
2008 Figures 49.8% Sessions 50.1%

I think that there was some consensus that Gingles could be triggered in AL. The key prong is the existence of a compact district with a BVAP majority. The district by train as noted by Torie below seemed to be the strongest proof of principle.

What a court would object to as a racial gerrymander, and what the VRA requires, are two different things. What we are arguing about is what a court would demand, both knowing the answer is uncertain, but with quite different takes.

Anyway, the black CD I think would be most likely to trigger the VRA is the one below, which I think Train already drew. It hews to good redistricting principles, more or less, with but the contiguous stab into Tuscaloosa. Do you disagree?

If I assume that Gingles requires a black district the next challenge is to demonstrate that a district provides sufficient opportunity for the minority to elect a candidate of choice. Drawing districts over 50% BVAP runs into more chops and erosity than is desired. If the district utilizes Montgomery then it splits the UCC to get to 50%.

So, if a district is to be less than 50% BVAP, how secure should the district be to provide the required opportunity under section 2? The argument relies on blacks controlling the Dem primary then getting sufficient white crossover votes in the general election. The critical question is how large a crossover is needed. In northern urban settings, its pretty easy to show a large stable population of white crossover voters. It may not be so easy in the South.

One interesting observation is that the Black Belt swung towards Obama in 2012 compared to 2008. So, for instance jimrtex's district was R+0.8 in 2008 but moved to D+2.5 in 2012. The PVI average would be D+0.9 and would be considered a highly competitive tossup district. That's a marginal pool of white crossover votes, and I'm not convinced that is enough to satisfy section 2.

The largest group of white crossover votes in AL near significant black populations are in Tuscaloosa and Jefferson counties. That led me to construct the plan that created a district that combines Tuscaloosa and Birmingham with a D+4 rating. My first version preserved the UCCs but split the Black Belt MCC three times. This version eliminates those extra splits and improves erosity, but with a slightly higher inequality. CD 6 has 43.6% BVAP and a 2008 PVI of D+3.7 and CD 7 has 46.0% BVAP and a 2008 PVI of R+0.9. Those both become more Dem with the 2012 results and the average PVIs over the two cycles would be D+5 and D+1. The CD 7 may not be sufficient to meet section 2, but I suspect that CD 6 would be.



That brings up the question posed by jimrtex:

So the issue is should a plan that does a large chop within a UCC county be treated the same as one where a much smaller chop is done outside the UCC, when a split of the UCC is necessary anyway.

It actually touches on the larger question of erosity in large population counties in UCCs where there will be substantial chops, including chops entirely within a county. A pure perimeter measure doesn't capture the erosity in a small area district when there are many large districts in the plan. My county connection proxy for an erosity score would also miss ugly shapes embedded within a chopped county.

The best way to resolve this is to look at a chopped county as a collection of county subdivisions. In states with well defined census subdivisions (MCDs as in MI) this is a clear partition of a county. In a state like AL it isn't so clear and one could use the Census CCDs or municipalities when they cover most of a county. Once county subdivisions are established inside a chopped county, one measures connection erosity using those subdivisions except the road connections need only be local roads.

This is a reasonable way to resolve erosity when I chop Jefferson as above. However, this punishes small chops made to reduce erosity. I think this would call for a new category of chop. We already have a microchop category that is less than 0.5% of the population quota . Chops of under 5% of the quota can become minichops. They are used to smooth out district erosity and though they count as chops they don't generate erosity due to county subdivisions. That would up the erosity on my Jefferson split, but not on the plans that split Blount or Walker to add to Jefferson.

[Torie]

"Chops of under 5% of the quota can become minichops."

I have no idea what this is about. 

[muon2]

"Chops of under 5% of the quota can become minichops."

I have no idea what this is about. 

The question jimrtex raised is whether all chops larger than a microchop are equal. In particular my AL plan above chops Jefferson, rather than keeping it whole and adding 24 K from some adjacent county.

The deeper issue this question raises is how do I count erosity in a dense metro like the Detroit cluster in MI. A perimeter based system will be unable to sense gerrymandering of districts entirely within the metro compared to whole county shifts in the large districts outside of the metro area which make large changes in the perimeter. My connectivity proxy, when only viewed at the county level, also would fail to note gerrymandering within a dense metro like Detroit.

The connectivity solution in MI is to take each chopped county and divide it into the minor civil divisions (MCDs: cities and townships). Then count the broken links within the county between MCDs towards erosity, just like broken links between counties for other districts.

This works fine within the metro, but would undermine the type of chop to eliminate erosity that you often identify. Those are typically smaller population chops in counties, but they are bigger than microchops. If those chopped counties are treated like my suggestion for the large urban counties with MCDs measured for erosity, then each of those chops will generate a lot of MCD erosity and the advantage of using those chops to eliminate county erosity is lost.

A system needs to keep the utility of chops to reduce erosity while also allowing meaningful erosity measures in large dense counties. One way to do that is to distinguish between the scale of chops used to smooth irregular districts from the chops that would generally appear in large counties. My suggestion of minichops is one way to draw that distinction. It follows jimrtex's observation that I use a substantially larger chop in Jefferson than appears in other plans around Jefferson, so I set a threshold at 5%.

It would also be possible to make that distinction based on the type of county that is chopped. UCC counties (including 1-county UCCs) could be subject to the subdivision erosity rule while other counties are not. I don't have a strong feeling as to which way is better, but there needs to be a way to scale erosity to dense urban districts.

[Torie]

Given the VRA, and the need to have but one municipal entity chop, just how serious a problem is this issue of gerrymandering by selectively deciding which muni entities go in one CD versus another?  Obviously erosity counts here too, and I guess if you don't measure shapes, but connectivity as a proxy therefor, you do need a connectivity measure to leash erosity, assuming using such a measure within a county really reduces shape erosity by using the connectivity proxy. Bearing in mind shape erosity, it has never arisen as an issue for me. There typically are very few choices how to chop a county, keep muni chops down, and reduce shape erosity.

[muon2]

Given the VRA, and the need to have but one municipal entity chop, just how serious a problem is this issue of gerrymandering by selectively deciding which muni entities go in one CD versus another?  Obviously erosity counts here too, and I guess if you don't measure shapes, but connectivity as a proxy therefor, you do need a connectivity measure to leash erosity, assuming using such a measure within a county really reduces shape erosity by using the connectivity proxy. Bearing in mind shape erosity, it has never arisen as an issue for me. There typically are very few choices how to chop a county, keep muni chops down, and reduce shape erosity.

I don't think it really matters in AL, except for how to deal with the VRA. If the goal is a CD with 50% BVAP then the choices are limited. If one allows BVAP<50% in a sufficiently D district, then that opens up jimrtex's question about whether a severe chop of Jefferson (or Mobile) is equal to any other chop. So far we've said that it is.

His question caused me to consider the situation in areas like the Detroit and Philly UCCs. The VRA doesn't define the whole area and so one gets to how to measure the urban/suburban district shapes. There actually is quite a lot of choice with only population and muni chop constraints. Measuring shape erosity isn't well defined without getting quite mathematically complicated. I believe that connection erosity can provide a simple tool to deal with the urban areas, too.

Erosity measures are going to have transition issues going from large scale districts to small districts. Imagine a district that has 4 whole rural counties that form a nice rectangle with half the population of a district. It is then joined to a small erose piece from an adjacent urban county with the other half of the population. The ragged lines in the urban county look no more severe than a river forming the boundary of a rural county, but there may be a long peninsula in the urban county designed to gerrymander a particular population. When less erose versions within the urban county are possible, how does one measure it?

[Torie]

Presumably one would take the map with the peninsula into a densely populated urban area, and lose it with something less elongated, and no other change, and that map would get a higher score. There clearly does need to be an erosity measure for densely populated areas. And it seems to me the check for games, is that one can just take the map submitted and lose the games, and that map wins the pareto optimality contest. Am I missing something?

[muon2]

Presumably one would take the map with the peninsula into a densely populated urban area, and lose it with something less elongated, and no other change, and that map would get a higher score. There clearly does need to be an erosity measure for densely populated areas. And it seems to me the check for games, is that one can just take the map submitted and lose the games, and that map wins the pareto optimality contest. Am I missing something?

That is exactly the goal. At issue is the mechanism to make that happen in urban areas. One option is to treat each chopped urban county as if it were many separate counties based on munis or towns or CCDs and then measure erosity for the whole district using this expanded set of nodes. Another is to only apply erosity internal to a county as a tie breaker in the same way that population inequality is used to discriminate between plans with equal chops and erosity. Needless to say there are other variations beyond these two.

[Torie]

I assume the tie breaker approach would probably reduce the negative impact on a map's overall score from doing the gerrymandered peninsula, no? Is that the impact? Other than that, we are back to our favorite issue as to how to measure erosity. I no longer reject your approach out of hand on that. Now for me it comes down to whether it meets the pornography test, and whether connectivity is adequate to meet the eye test as to whether or not the level of erosity is pornographic, and whether or not using an explicit shape test generate something shall we say more "artisitc."

In other words, I need to see enough maps using the connectivity test to get satisfied that the proxy is up to its task as advertised. Make sense?

Could you elaborate more on the 5% test, and whether you favor it or not, and maybe put up a couple of maps showing what is in play? I am suspicious of it myself.

[jimrtex]

I've picked this up from the Odd and Even thread since the VRA question fits better over here.

Regarding the VRA, it is arguable whether a plaintiff would satisfy the first prong of the Gingles tests, and with this district clearly fails the third prong.

For my version of the Tuscaloosa-Black Belt-Montgomery district:

2012 Obama 53.9% Romney 44.9%
2010 Sparks 53.0% Bentley 46.5%
2010 Barnes 47.6%, Shelby 52.4%
2008 Obama 52.6% McCain 46.8%
2008 Figures 49.8% Sessions 50.1%

I think that there was some consensus that Gingles could be triggered in AL. The key prong is the existence of a compact district with a BVAP majority. The district by train as noted by Torie below seemed to be the strongest proof of principle.

I don't think that any of the three prongs can be considered more significant than the other.  Is it clear how Gingles test applies to cases like this?

Thornburgh v Gingles was in regard to at-large elections of seats in the North Carolina legislature.   The Gingles test makes sense in such a situation.  You have a minority population that is concentrated in an area of the county, that could elect a representative in a district election, but could not win in at-large election.  It is the practice of using at-large elections that is the violation of the VRA Section 2.

But Alabama is required by federal statute to have district elections for Congress (with certain rare exceptions).

The issue at trial would be whether the Alabama map violates Section 2, not whether it fails the Gingles test - which after all does not require creation of the demonstration district, but rather establishes a requirement to create districts when at-large districts are used.  That is, the Gingles test may not be probative with respect to a Section 2 violation.

And even if the Gingles test is applicable, it does not apply to the particular plan.  Torie's plan does not create a compact district.  Its splits Tuscaloosa, Autauga, and Washington counties segregating the population of these counties solely on the basis of race, and for a marginal increase in the BVAP.

But even if the 1st prong of the Gingles test is passed, and the 2nd prong is conceded; the 3rd prong is not passed, because the Black Belt district does not usually prevent the election of the candidate of choice of the Black population.   For my original UCC district, Obama carried the district by 6.1% in 2008; Sparks carried the district by 12.3% in 2010; and Obama carried the district by 9.4% in 2012.   Only in the 2008 and 2010 senate races was the candidate of choice defeated.  In 2010, the Democratic candidate was a novice, who had nonetheless defeated the incumbent representative for the area, Artur Davis, in the primary, and thus may not have been the overwhelming candidate of choice.   In the 2008 senate race, the candidate of choice of the black population, Vivian Davis Figures was narrowly defeated 50.1% by 49.9% by the 2-term incumbent Jeff Sessions.

What a court would object to as a racial gerrymander, and what the VRA requires, are two different things. What we are arguing about is what a court would demand, both knowing the answer is uncertain, but with quite different takes.

Anyway, the black CD I think would be most likely to trigger the VRA is the one below, which I think Train already drew. It hews to good redistricting principles, more or less, with but the contiguous stab into Tuscaloosa. Do you disagree?

If I assume that Gingles requires a black district the next challenge is to demonstrate that a district provides sufficient opportunity for the minority to elect a candidate of choice. Drawing districts over 50% BVAP runs into more chops and erosity than is desired. If the district utilizes Montgomery then it splits the UCC to get to 50%.

Does Gingles apply in the abstract?  That is, assuming a blank map does it then require certain districts?  Or does it apply in the concrete world?

I assume we agree that a plan must comply with the US Constitution.  And that we agree that the plan must comply with the VRA insofar as the VRA is consonant with the US Constitution (Section 4 was not).  But can there be a presumption that a map that fails the Gingles test, violates Section 2?

You must be logged in to read this quote.

I don't believe that it can be shown that my plan lessens the opportunity for black Alabamians to elect candidates of choice, and to the extent that the current AL-7 is a minority pack, provides a greater opportunity to elect an additional representative of choice.

So, if a district is to be less than 50% BVAP, how secure should the district be to provide the required opportunity under section 2? The argument relies on blacks controlling the Dem primary then getting sufficient white crossover votes in the general election. The critical question is how large a crossover is needed. In northern urban settings, its pretty easy to show a large stable population of white crossover voters. It may not be so easy in the South.

One interesting observation is that the Black Belt swung towards Obama in 2012 compared to 2008. So, for instance jimrtex's district was R+0.8 in 2008 but moved to D+2.5 in 2012. The PVI average would be D+0.9 and would be considered a highly competitive tossup district. That's a marginal pool of white crossover votes, and I'm not convinced that is enough to satisfy section 2.

I don't understand your rationale for using relative measure of support rather than the actual results.  It's as if you are saying that whites in the district voted for Obama, but if there were a white Democrat running, whites would swing back to the Republicans.   Obama ran behind other Democratic candidates in counties like Greene, Dallas, and Choctaw, and particularly behind Sewell.

That brings up the question posed by jimrtex:

So the issue is should a plan that does a large chop within a UCC county be treated the same as one where a much smaller chop is done outside the UCC, when a split of the UCC is necessary anyway.

It actually touches on the larger question of erosity in large population counties in UCCs where there will be substantial chops, including chops entirely within a county. A pure perimeter measure doesn't capture the erosity in a small area district when there are many large districts in the plan. My county connection proxy for an erosity score would also miss ugly shapes embedded within a chopped county.

The best way to resolve this is to look at a chopped county as a collection of county subdivisions. In states with well defined census subdivisions (MCDs as in MI) this is a clear partition of a county. In a state like AL it isn't so clear and one could use the Census CCDs or municipalities when they cover most of a county. Once county subdivisions are established inside a chopped county, one measures connection erosity using those subdivisions except the road connections need only be local roads.

I don't see why townships could not be defined in Alabama.  I don't think CCDs should be regarded as having any merit whatsoever.  They're something that the census bureau defined 60 years ago, and may or may not have had any input from local authorities.

If county chops are kept small, there is little opportunity for gerrymandering, particularly if there is a requirement for whole subdivisions being used, banning double spanning, and approval from the voters.

[muon2]

Does Gingles apply in the abstract?  That is, assuming a blank map does it then require certain districts?  Or does it apply in the concrete world?

In Hastert v State Board of Elections (ND Ill 1991) the court basically did apply Gingles to a blank map to mandate a Hispanic majority district. There was no legislative map and no existing Hispanic district, yet the court agreed with both parties in the suit that such a district was mandated by Gingles. The bizarre shape of IL-4 that resulted was the subject of another suit that went to SCOTUS as King v State Board of Elections, and after being remanded the court affirmed the district as a narrowly tailored remedy to a section 2 violation.

[muon2]

I don't see why townships could not be defined in Alabama.  I don't think CCDs should be regarded as having any merit whatsoever.  They're something that the census bureau defined 60 years ago, and may or may not have had any input from local authorities.

If county chops are kept small, there is little opportunity for gerrymandering, particularly if there is a requirement for whole subdivisions being used, banning double spanning, and approval from the voters.

Townships could be defined in AL but they haven't been. I'd like to avoid creating a new subdivision that presently has no existence. I'm not wild about CCDs, but the Census Bureau claims that they are drawn with input from the local authorities.

There will be cases where county chops can't be kept small because there's too much urban population. There needs to be a way to test for gerrymandering in the form of erosity in an area like the DFW Metroplex as well as in Birmingham. The same standard should apply if both are UCCs.

[jimrtex]

There are two ways to handle the Birmingham UCC.  Either treat Jefferson as a separate region within a 5% limit, and place Shelby in its own region, or place both counties within the same 2-district region.

Jefferson as its own region.

If Jefferson is its own region, and it is made into its own district, the minimum standard deviations is 1.46%, assuming that the other 6 regions equally split the surplus between them (eg 1.006x6 + 0.964 = 7.000).  This error is significantly larger than can be obtained by attaching a small portion of an adjacent county to Jefferson.  If an Alabama whole-county plan was challenged, the SCOTUS might determine that while whole counties are a legitimate state interest, Alabama was elevating it too much above the constitutional requirement of practicable equality.  A second county cut does not improve equality to the same extent, and might be a metric to limit Alabama plans to a single cut.

Scoring of a plan with a Jefferson-only region

The plan should designate an adjacent region from which the chop should come from.  For purposes of calculating the standard deviation, it is presumed that a perfectly-sized chop will be made, so that the population of both resulting districts will be identical to the average of the population of the two regions.

As a practical consideration, the paired region should have a surplus to complement the Jefferson deficit.  Similarly, the remaining 5 regions should attempt to reach perfect equality.  While a deviation of 5% is permissible, and would not cause a plan to be rejected outright, any such plan is certain to be discarded as inferior to other plans.

The interior perimeter would be measured between all adjoining regions, treating Jefferson as a region.  The added population is small (if it were a county, it would rank 44^th among 68 among Alabama counties).  This simplifies the process, permitting regional maps to be compared, prior to finer delineation.

Jimrtex's First Plan



Jimrtex's Second Plan



This a modest change that swaps Washigton for Conecuh, improving both equality and compactness.  This might be sufficient to cause elimination of the first plan, accept for the slight decrease of BVAP in the Black Belt district.



Jefferson and Shelby in a 2-district region.

If Jefferson and Shelby are in the same region, then the cut will be in Jefferson.  The proof is as follows:

If neither Jefferson nor Shelby were cut, then the cut must be made in another county adjacent to Jefferson, which would mean that this is actually a Jefferson region paired with a region including Shelby, a contradiction.  If Shelby were cut, it would be no different than if it were adjacent to a Jefferson-only region.   This is again a contradiction.

That the cut of a two-district region must be in Jefferson is not a general rule, but a consequence of the specifics of the distribution of the Alabama population.  Since the cut is within a UCC, the magnitude of the cut does not matter.   It is a cut within a region that contains a UCC that must be cut.

Scoring of a plan with a two-district region containing Jefferson and Shelby

For purposes of calculating the standard deviation, it is assumed that the region will be perfectly cut, so that the population of each district will be half the region population.

As a practical consideration the two-district region should have a population equivalent to close to two districts.

The interior perimeter would be measured between all adjoining regions, treating the two-district region as a single region.   In addition an estimate of the district cut will be made by measuring the area of the two-district region using simplified boundaries, then calculating the diameter of a circle with the same area.    This may be somewhat longer than a cut needed to split an irregular area, but shorter than an irregular cut.

Muon's First Plan



This has a two-district region, and better equality than either of my plans.  It has worse compactness, and splits up the Black Belt some with inclusion of Coffee (16% BVAP) a particular problem because it is not compact.


Muon's First Plan.  With Jefferson as a separate region.

In Alabama, it should be possible to convert most two-district Birmingham region into into a Jefferson-only region plan.  Even if Muon were not to offer this plan, someone else could take his plan and resubmit it.  This is not a totally implausible plan, with a semi-donut.  While the final districts will be quite different, the only difference is in the scoring of the perimeter.  Jefferson is relative small relative to the area of the region, but the perimeter loops around on three sides.




Muon's Second Plan

This map improves on the Black Belt district, and may make it feasible to create two Section 2 districts in Alabama.  Most of the districts are quite elongated, and compactness suffers.




Muon's Second Plan.  Jefferson-only region version.

As with his first map, it is possible to convert this to a version with a Jefferson-only region.




Jimrtex's Third Plan

This smooths out the the northern part of the southeastern district, while increasing the erosity of the eastern district.  This is somewhat mitigated by the fact that almost 3/4 of the district is in the middle 4 counties (Shelby, St.Clair, Calhoun, and Etowah).  So while geographically it is a bar-bell, population-wise it is an elongated rectangle.




Jimrtex's Fourth Plan

This map substantially modifies the northern part of the map.  It is the most compact map of any so far.




Jimrtex's Fourth Plan, Two-District Birmingham region.

It is possible to convert this into a two-district Birmingham region plan.  A couple of plausible splits of Jefferson have interesting consequences.  If one district were Tuscaloosa-Pickens+some of Jefferson, it might be possible to create a second VRA district, while the other district would connect the northwestern part of the state with Shelby and counties further south, using a bridge across Jefferson.  Another split could place Shelby and 3 other counties with about 60% of Jefferson, while the rest of Jefferson is placed with Tuscaloosa and the other counties to the northeast.  Very compact, but would push the BVAP in both districts down towards 30%.




Jimrtex's Fifth Plan

This is similar to the 4th plan but pairs Jefferson with the northeastern district.    The adjustments around the rest of the state returns the Crenshaw hook, wich adds about 60 miles to the perimeter.




Jimrtex's Sixth Plan

This is a simpler change to the 4th plan, moving Winston to the northeast.  The loop around Winston adds a significant penalty to compactness.




Summary

Summary of Plans showing standard deviation relative to ideal population; internal simplified perimeter in miles; plurality for Black-preferred candidate (Obama 2012 presidential, Sparks 2010 gubernatorial; Obama 2008, and average of 3), and treatment of Jefferson County (single region or part of two-district Birmingham UCC region).


Plan   S.D.    Peri.  2012P  2010G  2008P  Average   Type
JR1    0.50%   1109   9.4%   12.3%   6.1%   10.9%   Jefferson
JR2    0.43%   1079   8.5%   11.8%   5.3%   10.2%   Jefferson
JR3    0.46%   1073   8.5%   11.8%   5.3%   10.2%   Jefferson
JR4    0.48%   1022   8.1%   12.1%   4.9%   10.1%   Jefferson
JR4*   0.48%   1051   8.1%   12.1%   4.9%   10.1%   2-district
JR5    0.51%   1078   8.5%   11.8%   5.3%   10.2%   Jefferson
JR6    0.44%   1080   8.1%   12.1%   4.9%   10.1%   Jefferson
MF1    0.16%   1157   4.5%    9.4%   1.5%    7.0%   2-district
MF1*   0.16%   1154   4.5%    9.4%   1.5%    7.0%   Jefferson
MF2    0.28%   1188   8.8%   12.5%   5.7%   10.7%   2-district
MF2*   0.28%   1179   8.8%   12.5%   5.7%   10.7%   Jefferson

[jimrtex]

Does Gingles apply in the abstract?  That is, assuming a blank map does it then require certain districts?  Or does it apply in the concrete world?

In Hastert v State Board of Elections (ND Ill 1991) the court basically did apply Gingles to a blank map to mandate a Hispanic majority district. There was no legislative map and no existing Hispanic district, yet the court agreed with both parties in the suit that such a district was mandated by Gingles. The bizarre shape of IL-4 that resulted was the subject of another suit that went to SCOTUS as King v State Board of Elections, and after being remanded the court affirmed the district as a narrowly tailored remedy to a section 2 violation.

The defendant SBOE was technically defending the use of the 1980s map which had two too many districts.  As a legal matter, it doesn't matter that they had said that they would go along with whatever the court imposed.  The SCOTUS has told the federal courts not to legislate, so any map they draw should be only remedial.  So the district court in Hastert had to make an affirmative finding of the 1980s map being discriminatory based on the 1990s population.

In an Alabama case, the issue would be whether the map is discriminatory.   I don't think that a district where the candidate preferred by black voters usually wins can be considered to dilute the vote of blacks.

Perhaps the Gingles test could be applied to determine whether two majority-minority districts could be drawn in Alabama.  I suspect that it is impossible to do so without tentacles into Jefferson, through Tuscaloosa, into Mobile, and more precise targeting in Montgomery.  But this is quite different from the case in IL-4 where the two lobes are a couple of miles apart, and the the loop is simply to accommodate IL-7, vs searching out into Birmingham and Mobile, where the "community" is being defined solely on the basis of race, rather than a compact community which is predominately black.

Hastert did determine that the Gingles test did not apply to an attempt to keep Springfield and Decatur in the same district where a very small black population could supposedly influence the election.  It appears that the bright line majority test is based on the conditions of totally polarized voting prevent any opportunity for the minority voters to be successful.

[jimrtex]

Summary

Summary of Plans showing standard deviation relative to ideal population; internal simplified perimeter in miles; plurality for Black-preferred candidate (Obama 2012 presidential, Sparks 2010 gubernatorial; Obama 2008, and average of 3), and treatment of Jefferson County (single region or part of two-district Birmingham UCC region).


Plan   S.D.    Peri.  2012P  2010G  2008P  Average   Type
JR1    0.50%   1109   9.4%   12.3%   6.1%   10.9%   Jefferson
JR2    0.43%   1079   8.5%   11.8%   5.3%   10.2%   Jefferson
JR3    0.46%   1073   8.5%   11.8%   5.3%   10.2%   Jefferson
JR4    0.48%   1022   8.1%   12.1%   4.9%   10.1%   Jefferson
JR4*   0.48%   1051   8.1%   12.1%   4.9%   10.1%   2-district
JR5    0.51%   1078   8.5%   11.8%   5.3%   10.2%   Jefferson
JR6    0.44%   1080   8.1%   12.1%   4.9%   10.1%   Jefferson
MF1    0.16%   1157   4.5%    9.4%   1.5%    7.0%   2-district
MF1*   0.16%   1154   4.5%    9.4%   1.5%    7.0%   Jefferson
MF2    0.28%   1188   8.8%   12.5%   5.7%   10.7%   2-district
MF2*   0.28%   1179   8.8%   12.5%   5.7%   10.7%   Jefferson

[jimrtex]

I don't see why townships could not be defined in Alabama.  I don't think CCDs should be regarded as having any merit whatsoever.  They're something that the census bureau defined 60 years ago, and may or may not have had any input from local authorities.

If county chops are kept small, there is little opportunity for gerrymandering, particularly if there is a requirement for whole subdivisions being used, banning double spanning, and approval from the voters.

Townships could be defined in AL but they haven't been. I'd like to avoid creating a new subdivision that presently has no existence. I'm not wild about CCDs, but the Census Bureau claims that they are drawn with input from the local authorities.

There will be cases where county chops can't be kept small because there's too much urban population. There needs to be a way to test for gerrymandering in the form of erosity in an area like the DFW Metroplex as well as in Birmingham. The same standard should apply if both are UCCs.

CCD were likely created because it permitted the South to be treated like the Midwest and fit in a neat hierarchical scheme, just like the Census Bureau treats counties in New England with greater than actual significance.   If the Census Bureau did consult with local authorities, it would be because they were hiring the enumerators from the local area and it helped organize the work.  They are pretty meaningless in any area that has much growth in the past 1/2 century.

In Jefferson County, Alabama, they don't match city limits very well, so if you use the CCD you will chop cities.  They also don't match voting districts, which indicates they don't match settlement patterns or community of interest either.

Since they would be defined in advance, I don't see a problem with creating them.  The Public Land Survey System is defined in Alabama, so that is one possible starting place.

I don't know that you need an erosity measure if you are requiring that districts be defined by subunits of modest size, limit crossing of county boundaries, and have popular approval of the districts.

[muon2]

Thanks. Does the map change much if it goes to 40% in urbanized areas (still using 25 K as the alternative)? I ask since 40% seems to be the right threshold for counties to count in minority clusters. If urban clusters also used 40% it would establish that as a consistent benchmark to identify a county as part of a larger contiguous community of interest.

The text accompanying the map indicates in green the edge cases (those with urbanized areas with between 25% and 50% of the county population.  Of the 15 edge cases (southeastern states of AL, AR, FL, GA, LA, MS, NC) 9 would fail a 40% threshold.  That is 60% of the 15, which matches (40%-25%)/(50%-25%), which indicates a somewhat linear distribution in that portion (overall, I believe that this represents a saddle).

Edge cases:

Montgomery, AL/Elmore County 27K/34%.   I-65 clips the southeastern corner of Elmore before entering Autauga County.   It appears that the Coosa flood plain is not develop-able, so that Montgomery can't directly spill into the northern counties.

Little Rock, AR/Lonoke County 31K/45%.   US-67 clips the northwestern corner of the county (around Cabot) which is the urbanized area, rather than east along I-40.  US-67 appears to be the traditional direct route to St.Louis (it is not as straight in Missouri, so that I-40 and I-55 via Memphis is just as quick though longer).

Fort Smith, AR-OK/Crawford County 30K/48%.  I suspect that there are enough retirees in the county that holds down the population in the urbanized area a bit.

Atlanta, GA/Walton County 28K/33%.  Walton is east of Gwinnet, and almost to Athens.  Monroe Urban Cluster is still separate from the Atlanta Urbanized Area.  The county doesn't have the best direct highway access to Atlanta proper.

Savannah, GA/Bryan County 9K/31%.  A large chunk of the middle of the county is part of Fort Stewart and non-developable.   The western end is a bit too far for direct connection to Savannah.  Though small, Bryan County has tripled in population in the last 30 years.

Athens, GA/Oconee County 16K/49.7%.  You can see the effect of relatively small county sizes in Georgia, so that even modest cities can have urban growth flopping into the next county.  With larger counties, the urbanized area can be contained within a main county.

Chattanooga, TN-GA/Walker County 31K/45%.  The Lafayette Urban Cluster adds enough population outside the urbanized area to keep it below 50%.

Dalton, GA/Murray County 12K/30%.   Dalton Urbanized Area crosses to include Chatsworth in Murray County.

Baton Rouge, LA/Iberville Parish  11K/34%.  Iberville is the last cross-river parish/county before Hennepin, with the area to the west of West Baton Rouge added to the parish later on.   So Iberville is really south of Baton Rouge, though it gives the appearance of being to the west.   The urbanized area crosses into Iberville in the area east of the Mississippi, and there is another portion linked through West Baton Rouge.

Gulfport-Biloxi-Pascagoula, MS/Hancock County 17K/39.6%.  Hancock is the least populous of the 3 coastal counties, and more than 10 miles from Gulfport, with Biloxi, and Pascagoula even firther way.  The Gulfport Urbanized Area crosses into the county at Bay Saint Louis on US-90 along the coast.  Diamonhead Urban Cluster on I-10 and further inland is separate.

Hattiesburg, MS/Lamar County  28K/49.6%.  Hattiesburg is directly on the county line, with I-59 looping the city to the west on the Lamar County side.  If not for the notch out of the northeast corner of the county, Lamar County would be over 50%.

Raleigh, NC/Johnston County  37K/22%.  Smithfield and Archer Lodge-Clayton urban clusters hold down the urbanized area percentage.  Johnston qualifies under the 25K test rather than the 25% test, and is a central county of the Raleigh MSA.  I would be inclined to keep the 25K test, even if the 25% threshold were increased.

Asheville, NC/Haywood County  26K/45%.  The terrain concentrates the urbanized area west from Asheville.  Haywood County might not qualify as part of the metropolitan area based on commuting.

Rocky Mount, NC/Edgecombe County  17K/31%.  Rocky Mount is on the Nash-Edgecombe county line, and barely qualifies as an urbanized area based on the Nash County portion.  Tarboro Urban Cluster holds down the urbanized area population in Edgecombe County.

Myrtle Beach, SC-NC/Brunswick County 40K/37%.  Brunswick County would qualify under a 25K test.  Brunswick County is a central county of the Myrtle Beach MSA, rather than the Wilmington MSA by a narrow margin.   20279 vs. 19636.   Commuting to New Hanover, NC is 5 times as great as that to Horry, SC.  This illustrates the propensity of urban areas to string out along linear features, like highways, or in this case along the coastline

Of the 190 urban core counties in MSA in AL, AR, FL, GA, LA, MS, and NC, 153 qualify for UCC based on a 50%/50K threshold; 15 fall within 25K-50K/25%-50% region; and 22 fall below the 25% threshold.  78 non-counties were excluded in the first pass.

So of 268 metropolitan counties, 153 (57%) are clearly in, 100 (37%) are clearly out, and 15 (6%) are edge cases.

I've been thinking about this in light of the discussions on chops sizes. 5% of a district seems to be a threshold between smaller and larger chops. Rather than use 25K, it seems that 5% of a CD would be a more rational measure. Below that threshold it would only take a small chop to excise the urbanized area, but beyond that it takes a larger chop. A 5% cut varies from state to state but it would be typically around 35K. I've highlighted the counties in the above list that would survive that or 40% urbanized area. To my eyes this seems like a reasonable separation of the edge counties.

[jimrtex]

I've been thinking about this in light of the discussions on chops sizes. 5% of a district seems to be a threshold between smaller and larger chops. Rather than use 25K, it seems that 5% of a CD would be a more rational measure. Below that threshold it would only take a small chop to excise the urbanized area, but beyond that it takes a larger chop. A 5% cut varies from state to state but it would be typically around 35K. I've highlighted the counties in the above list that would survive that or 40% urbanized area. To my eyes this seems like a reasonable separation of the edge counties.

I'm not sure that chop size is applicable.

What I've tried to do is include counties that are structurally part of the core of a metropolitan area.

The Census Bureau definition of central counties is too inclusive in some cases, and too exclusive in others. 

In 2000, when it switched to to automated identification of urban areas, it split urban areas which had been in different metropolitan areas in 1990.  Yet it then permitted the counties that had been in separate metropolitan areas to be included in other metropolitan areas.  The rules for metropolitan areas were also changed in 2000, so that it is even possible that the counties might not have been in different metropolitan areas based on 1990 data but 2000 rules.

In 2010, existing urbanized areas were grandfathered.  So in effect, the Census Bureau has frozen in place 1990 settlement patterns and 1990 rules.  An example of this is Livingston County in Michigan.  In 1990, Livingston was part of the Ann Arbor metropolitan area.  This meant that in 2000 a new urbanized area was defined in Livingston County, South Lyon-Howell.  Since the division was "near" the Oakland-Livingston line, South Lyon which is in Oakland, ended up as part of the urbanized area.  But since commuting patterns from Livingston into Wayne, Oakland, and Macomb were sufficient, Livingston is an outlying county of the Detroit MSA, and it is possible that some commuting within the South Lyon-Howell urbanized area is counted as commuting into the Detroit core.

The Census Bureau also permits multiple urbanized areas within the same county, but only permits one to be used as the urban core of the an MSA.  In Jefferson County, Texas, Beaumont and Port Arthur are separate urbanized areas,  Because Port Arthur UA includes the mid-county area, it has slighly more people in the county, and is the core of the metropolitan area.  But because Beaumont is the larger city, it is the namesake of the MSA.  In addition, the portion of the Beaumont UA that extends into Hardin does not qualify that county as a central county of the MSA, even though most commuting is into the largest city.

So the first change is to recognize that though the Census Bureau uses a single core urbanized area as the basis for delineating an MSA, there may be multiple urbanized areas that are part of the dense settlement pattern at the core of an MSA.

But the Census Bureau is somewhat over-inclusive in defining central counties of the MSA.  This is because it uses a single standard for delineating micropolitan and metropolitan statistical areas.  Since the urban area that is the core of a micropolitan statistical area only needs 10,000 person, a central county of any CBSA only needs 5,000 persons - which might be almost half of an urban area that spans a county line.   But the core of a metropolitan area must be an urbanized area, that  is, over 50,000 persons.   With urbanized areas extending along highways by hops and jumps, 5,000 persons is a small part of an urbanized area.

When the Census Bureau was considering new definitions for 2000, they had proposed three classes of CBSA, metro-, meso-, and micro-, but would have defined metropolitan statistical areas first.  In the name of simplicity, they ended up using a single standard.   25,000 is half of the size of a minimum urbanized area, and seems like a reasonable standard for defining a core county of an MSA

I don't think that it makes sense to say that such a county is not a core county of a metropolitan statistical area,

[jimrtex]

This is a corrected map, and does not reflect a change in definition.  It is simply showing metropolitan statistical areas, with counties without any significant urban population shown in a lighter tone.



The corrections are San Benito, CA and Tooele, UT which are part of the San Jose and Salt Lake City metropolitan statistical areas.