How Democrats Fooled California’s Redistricting Commission

[muon2]

From what I can tell the GOP was particularly weak at recognizing underlying bias in some neutral mathematical models. In AZ the GOP assumed that the blame was solely on the partisan chair, but even a neutral chair would have had a hard time overcoming the state competitiveness directive given the bias in the chosen elections.

In CA my sense was that the lack of understanding by the GOP of the impact of socioeconomic grouping as a preferred community of interest. The underlying math here works against the GOP as much as a maximally square grid with minimum area districts works against the Dems.

[muon2]

From what I can tell the GOP was particularly weak at recognizing underlying bias in some neutral mathematical models. In AZ the GOP assumed that the blame was solely on the partisan chair, but even a neutral chair would have had a hard time overcoming the state competitiveness directive given the bias in the chosen elections.

In CA my sense was that the lack of understanding by the GOP of the impact of socioeconomic grouping as a preferred community of interest. The underlying math here works against the GOP as much as a maximally square grid with minimum area districts works against the Dems.

Muon,

Of the legal criteria for redistricting, "competiveness" is the least important.

1. Districts shall comply with the United States Constitution and the federal Voting Rights Act. 2. Both legislative and congressional districts shall be equal in population, to the extent practicable. This establishes a new strict population equality standard for legislative districts. 3. Districts shall be geographically compact and contiguous, as much as practical. 4. District boundaries shall respect "communities of interest," as much as practical. 5. District lines shall follow visible geographic features, and city, town and county boundaries and undivided "census tracts" as much as practical. 6. Political party registration, voting history data and residences of incumbents and other candidates may not be used to create district maps. 7. "Competitive districts" are favored if competitive districts do not significantly harm the other goals listed.

What the commission did on the congressional districts is violate rules 3 (compactness) 4 (communities of interest), 5 (geographical/political boundaries) and 6 (party history) in order to create as many Democrat friendly districts as possible.

The entire process from the time Mathis got appointed has been a charade, with the three Democrats (Mathis pretends to be an Independent, but is is fact a dogmatic Democrat) even picked the "Republican" legal counsel!

As I understand it, the above goals 1, 2, 3, 4, 5, and 7 are given equal priority in the AZ Constitution. Federal law elevates goals 1 and 2 as well as the contiguity part of goal 3. The compactness part of goal 3 has historically been given wide discretion by the courts. Item 6 in your list is not a goal, but a requirement during the initial phase of mapping, but is then relaxed in order to test the map against the goals. Specifically voting behavior is necessary for the voting rights act (1) and competitiveness (7).

My point is that any commission is faced after the initial phase with the task of balancing items 3, 4, 5, and 7 and that includes the use of voting data. Competitiveness is easier to objectively measure than communities of interest, but is sensitive to the underlying set of election data. Any commission stuck with slanted election data to test the goals, would be drawn towards districts like the purported swing districts in the AZ plan (see my analysis on the AZ thread).

A truly independent chair might not have been enough given the data used to test the goals. Had the election data been properly vetted, inherent biases in the map would be easier to expose regardless of the chair.

[muon2]

OK, but that deep jut of yours into OC so that a CD that goes from Bellflower (well Lawndale, better) via Long Beach to consume that only all of Huntington Beach, but a slice of Fountain Valley, is just a deal killer sbane. The OC-LA County line is a pretty hard one, trust me (ignored by the Commission when it snapped up Westminster), and any cuts need to have a very compelling reason indeed. There is a very good case for consuming Seal Beach, and in a pinch, even that Los Alamitos gated Seizure World, if need be, since they both go to Long Beach for nearly everything (going south is this huge field of oil wells and strawberry farms and such, and navy stuff, and is probably a toxic waste dump), but that is about it.

Are the county lines generally meaningful in SoCal or is it just that stretch of line you reference. I ask since sometime ago when I was drawing CA maps there seemed to be consensus that a district would have to overlap at some point. Is that better towards the northern end of the LA-OC border?

[muon2]

I'm slowly digesting the Torie map, and it's made harder in that I can't get the 2010 block groups for CA to load in my DRA 2.2 (2000 is fine as are all other 2010 states ). One early observation I would make is that if one is worried about a commission tilted one way as they draw then one needs additional objective measures to constrain them. Communities of interest are very subjective, VRA districts are moderately objective, and political lines like counties and munis are very objective. So my initial test is to look at county splits that are not needed for the VRA.

On that measure I generally like the SD-OC-Riverside districts. Of course I have some caveats.

In a muon world without the VRA I see the following. SD+OC+Imperial is about 46K under the number needed for 9 CDs and Riverside is about 80K over the count for 3 CDs. LA county is 22K short of the population for 14 CDs. I would see some of the Riverside excess go to an OC district, and then from 22K go from OC or the Inland Empire to LA. The remaining bits then can go to San Bernardino with Inyo and Mono and grab 28K from SE Kern. That makes 29 CDs in total.

I also know that VRA concerns will trump counties on occasion. For instance Torie's Asian CD-32 has to make a crossing into San B county and looks good, but without a working App I can't tell how much needs to come across.

I'm also concerned about the lack of an Inland Empire Latino CD. MALDEF drew 2 with 62%+ HVAP - one with Fontana/Ontario and the other linking San Bernardino to Parris. The Commission drew the first as well, but reduced the second to 50% HVAP. As Torie noted, his San B district only gets 57% HVAP, and no other breaks 50%. I'd draw the MALDEF districts and use that to justify any county splits in the Inland Empire and hopefully get rid of the little fragment of Torie CD-41 in Riverside.

More later ...

[muon2]

Muon2, I think the 2010 election districts will load, and I had the same problem and switched to those. Check it out.  

The cut into Riverside County by CA-41 is tiny by the way.  Something has to cut into it, since the only other cut is CA-48 into the middle class white part of Corona, and that cut is also mandatory, since CA-48 is walled in indirectly by the Seal Beach, OC County line, the Asian CD CA-32, and the Saddleback Mountains between OC and Riverside County, with the pass and the Freeway being where I cut into Corona. CA-43 and CA-26 can switch out precincts to make CA-43 more Hispanic, but it means another municipal cut into Ontario, and I wouldn't do it, but maybe it should be done. I will draw an alternative map that way. That will make CA-26 more GOP.  

The way I drew the map, I don't think another Hispanic CD is possible, without cutting into Riverside, and I agree with the commission that that should not be done. It not legally mandated, and enough is enough.

I drew two more CD's by the way. They are drawn about the only way they can be drawn given the VRA and the mountains, unless one does not mind erosity and more chops, and I do mind. CA-20 is 59.8% Hispanic VAP.

Unfortunately the 2010 VTDs don't load for me either.

If you are going to preserve the San B - Riverside county line (and I like that you do). Then I think you will have to beef up your Latino districts. The commission report indicates that there is racially polarized voting in the Inland Empire and they had to go to 64% HVAP in the San B district to get CVAP over 50%. In the Riverside district they brought the HVAP over just 50%, which seems like a reasonable compromise between county integrity and Latino community interests. I still think it should be possible to slide some blocks and get 41 out of Riverside - if only my DRA worked.

I'm also concerned about attaching San Benito to CD-17. The commission didn't seem to check on whether racial polarized voting exists in the San Jose area, but it might, and if so then I think San Benito has to go with San Jose. San Benito isn't that populous and could be replaced by some extra folks from Santa Cruz in CD-17.

[muon2]

I thought I would go into a little more detail on the methods I mentions above.

I divide the state into regions, each roughly equal to a whole number of CDs. In most states I would try to get the regions within 0.5% of the ideal population, but with so many high population counties and natural barriers between regions my tolerance was quite a bit higher.

My goal is to place the districts largely within their region. As I divide each region into districts I look to avoid unnecessary county and municipal splits while complying with the VRA. The purpose of this is to create a starting point that minimizes external influence towards a partisan outcome.

Here are my CA regions, the number of districts, and the deviation from ideal population.

North Coast (2) +4968
Upper Sacramento (2) +4658
Lower Sacramento (6) -51670
San Fransisco Bay (7) -49467
Central Coast (3) +85547
Central Valley (4) +33896
Los Angeles (14) -22065
Inland Empire (6) +40169
South Coast (9) -46072

[muon2]

I would say that Contra Costa County has far more in common with Alameda County than any of the counties to its east; if possible, it would be best to keep the East Bay together.

I would agree if this method were only bout communities of interest. This method relies on a first step of creating regions that are nearly a whole number of CDs. Contra Costa has the size for one and a half districts and shifting it would create regions that were not nearly a whole number of districts. In any case Contra Costa plus Solano is just over two districts in population, so my division of the region would create one district entirely within Contra Costa and one that is split between Contra Costa and Solano.

[muon2]

Does the below precinct exchange between CA-43 and CA-26, a late term "abortion" as it were,  which trashes every reasonable "good-government" parameter on the grounds that doing  racial gerrymandering  via just doing the  VRA deep on steroids, turn anyone on? It gets CA-43 up to 62.2% VAP Hispanic (up from 57.6% Hispanic). Isn't that exciting?

I still think that you would make it better by letting CA-43 jut into Riverside instead of CA-41. It can go pick up some Latino areas and you can rationalize the unavoidable cut, by claiming it is the will of the VRA. For reference 64.7% HVAP gives 51.9% CVAP in that area, so 62.2% would project to 49.4% CVAP, so some additional Latino areas would be useful.

I think your map would be improved if you follow the the lead of the commission and MALDEF and move 43 to the SW. They use Chino, Ontario, Rialto, and a bit of San B to get a solid Latino seat. That would let you preserve Ontario as an intact muni.

[muon2]

Look familiar? This leads to the 38th becoming 65.2% HVAP. The 43rd drops to 45.1% HVAP and it's 50.1% Hispanic for the total population. And this is superior to the commission map since it doesn't split Upland or Highland, and the 43rd doesn't contain Redlands. Another reason to not extend the SBD desert district into the Owens valley!

Another thing I should note about Pomona and the San Gabriel valley Hispanic district. If you don't add Pomona to that district, it becomes 58% HVAP. This is not enough to create a 50% HCVAP district in many areas but in the San Gabriel valley it does seem to be possible (probably due to more established Hispanics living in cities like San Dimas, La Verne, Glendora etc.). Just take a look at the racial stats of the 32nd district as drawn by the commission to see what I am saying. This is why I am comfortable with putting Pomona in the 38th district. That area does need a higher Hispanic population to get to 50% HCVAP. Refer to the 35th as drawn by the commission to see the stats.

Thanks, that's exactly what I was suggesting. I only wish I could load CA 2010 into my computer.

OTOH, I agree with Torie about the cut into Kern. There's a very natural divide that separates the SE corner. I would attach it to Barstow, Inyo and Mono as a Mohave Desert district, but the point is the same. The Central Valley counties in my regional map were over the count of 4 CDs by 34 K, and that's just about exactly the number of people SE of the mountains in Kern.

I thought I would go into a little more detail on the methods I mentions above.

I divide the state into regions, each roughly equal to a whole number of CDs. In most states I would try to get the regions within 0.5% of the ideal population, but with so many high population counties and natural barriers between regions my tolerance was quite a bit higher.

My goal is to place the districts largely within their region. As I divide each region into districts I look to avoid unnecessary county and municipal splits while complying with the VRA. The purpose of this is to create a starting point that minimizes external influence towards a partisan outcome.

Here are my CA regions, the number of districts, and the deviation from ideal population.

North Coast (2) +4968
Upper Sacramento (2) +4658
Lower Sacramento (6) -51670
San Fransisco Bay (7) -49467
Central Coast (3) +85547
Central Valley (4) +33896
Los Angeles (14) -22065
Inland Empire (6) +40169
South Coast (9) -46072

[muon2]

Does this carve-up of the inland "empire" make everyone deliriously happy?  CA-38 is 61.4% Hispanic, CA-26 is 67% Hispanic, and CA-43 is 41.6% Hispanic. And Muon2 gets his f'ing Riverside County cut, which becomes more "convenient" with this map version - all 7,000 or so residents of it.

The key of course is the Pomona chop. We don't like to chop in half big towns like this, but the VRA is a harsh mistress - apparently. We also get a nice erose CA-43,and chop of a couple of more towns in LA County between CA-29 and CA-38.  Life is beautiful.

A win-win-win can be a beautiful thing.

[muon2]

Yeah, not bad. If the Owens valley goes in the 41st then Redlands has to go in the 43rd, otherwise we end up with my map. Keeping Redlands and Yucaipa together is probably a good idea, and the Owens valley can be put in another district. . I Disagree about the Kern cut as well. The 25th to me is an exurban LA district. There are a lot of commuters who commute into either the Santa Clarita valley or into LA from Palmdale and Lancaster. Yes, those areas of Kern are high desert like those two cities but other than that they have very little in common.

That's why I put SE Kern with Barstow and Death Valley. It keeps the high desert together better IMO.

[muon2]

Yeah, not bad. If the Owens valley goes in the 41st then Redlands has to go in the 43rd, otherwise we end up with my map. Keeping Redlands and Yucaipa together is probably a good idea, and the Owens valley can be put in another district. . I Disagree about the Kern cut as well. The 25th to me is an exurban LA district. There are a lot of commuters who commute into either the Santa Clarita valley or into LA from Palmdale and Lancaster. Yes, those areas of Kern are high desert like those two cities but other than that they have very little in common.

That's why I put SE Kern with Barstow and Death Valley. It keeps the high desert together better IMO.

Mike, would you direct me to your "high desert" map?  I think I disagree with your perspective here, but before I comment beyond my tentative thought below, I want to see the impact on adjacent CD's.  The fundamental problem however is connecting a lot of desert to a central valley based CD, and it won't be all of the "high desert" anyway. Victorville is more or less "high desert," and separating it from the balance of the otherwise very lightly populated "high desert," is the consequence.

I'm afraid it is only a conceptual map since I can't load actual 2010 data. The best I have is from 2008, and I used that to sketch the San Bernardino districts. The Ontario district is about 65% HVAP. Rancho Cucamonga is linked to San Bernardino along the north side of I-15, and in principle I would extend up into the foothills more, but that wasn't an option with my data. To the main idea, I linked Victorville and Barstow to SE Kern to put the whole US 395 corridor in one district. I'd love to know what the actual populations are in these districts.

[muon2]

Presumably from where you have an SB district pick up a piece of LA County, no?

No, that cut is into Westlake Village along Hwy 101, a long way from the north LA County CD. As I said, I don't like any of the cuts for the north LA County CD, other than into Kern County. Those cuts represent far less of a community of interest, and cutting into the city of LA to pick up a few people would just be awful.

I understand your reticence to go into the far northern tip of LA city to join with Santa Clarita, etc., but I think that sbane's version is preferable.

I think we all agree that Inyo and Mono should go with eastern SBD, but without using SE Kern, then Inyo only connects by a minor road through Trona or across Death Valley. From that perspective, Inyo and Mono could just as well connect across the Sierras to Fresno. US-395 is hanging right across the line in Kern and the CoI there matches as well.

I don't think that the crossing into LA from the north to balance population is so bad. To me, the connection from Santa Clarita to NW LA along I-5 seems not unlike the OC to Corona connection which is required to balance population there.

Given the choice of attaching two small counties to a large district with minimal road connections or chopping a small part of a large city just over the mountains but following along a freeway, I would strongly prefer with the latter.

[muon2]

Morgan Hill certainly can't be in a Santa Cruz district; there's no usable road through the mountains there, and to get from Morgan Hill or Gilroy to Santa Cruz you need to go through either Watsonville or San Jose.

Cupertino has more in common with Los Altos or Saratoga than it does with San Jose, whereas Campbell would fit better with the San Jose district than with the richer areas to its south. Demographically, Cupertino now has a large Asian majority, but income is probably a better indicator of communities of interest in the South Bay than race would be. It would be nice to simply switch Campbell with Cupertino (and the districts would look cleaner, too), but unfortunately Cupertino is significantly larger.

I'm not sure what Torie's HVAP numbers are for his CD 17, but I was expecting Watsonville to be with the Gilroy/Salinas/Hollister district. Then the question looms as to whether the district needs to extend into SJ and lose Monterrey to break 50% HCVAP.

[muon2]

Morgan Hill certainly can't be in a Santa Cruz district; there's no usable road through the mountains there, and to get from Morgan Hill or Gilroy to Santa Cruz you need to go through either Watsonville or San Jose.

Cupertino has more in common with Los Altos or Saratoga than it does with San Jose, whereas Campbell would fit better with the San Jose district than with the richer areas to its south. Demographically, Cupertino now has a large Asian majority, but income is probably a better indicator of communities of interest in the South Bay than race would be. It would be nice to simply switch Campbell with Cupertino (and the districts would look cleaner, too), but unfortunately Cupertino is significantly larger.

I'm not sure what Torie's HVAP numbers are for his CD 17, but I was expecting Watsonville to be with the Gilroy/Salinas/Hollister district. Then the question looms as to whether the district needs to extend into SJ and lose Monterrey to break 50% HCVAP.

You really are the King of racial gerrymanders, aren't you Mike?  

Anyhoo, one can't get to anywhere near 50% HCVAP for CA-17, no matter how much you just trash the map to try to get there, ignoring every other factor.

I am very sensitive to where VRA attacks on a map might emerge. I was curious to see that the CA Commission didn't ask about racially polarized voting in the Salinas area. The population is certainly enough to get well above 50% HVAP, and in some areas of the state upper 50's translates to a CVAP majority, but I don't know if 50% CVAP is in reach there. In any case the issue of whether 50% HCVAP is needed to elect a candidate of choice if there is sufficient control in the primary is an unanswered question on the national stage.

You must be logged in to read this quote.

My concept was most similar to your map 3. I would leave a corridor along the coast, since I wouldn't want to mess up your nice work down south by creating a barrier between Monterey and Santa Cruz.

[muon2]

I took a peek at MALDEF's version for that area and they claim 45% HCVAP. Their map is pretty similar to mine, so I'm guessing I have about the same.

Here's the legal problem I see. SCOTUS said that you have to meet the Gingles test to claim a section 2 violation. That includes a majority minority in a compact area and racially polarized voting. In Bartlett they said that a majority meant a voting age majority of a single group. They punted on the citizen question. I see a door left open for a voting age majority in the area, yet a sub 50% amount in a specific district as long as the single group had the opportunity to elect a candidate of choice. The Dems made this type of case in their IL legislative map, and they prevailed in court last month. But you're the lawyer, so you tell me if a court would go there in CA.

[muon2]

I took a peek at MALDEF's version for that area and they claim 45% HCVAP. Their map is pretty similar to mine, so I'm guessing I have about the same.

Here's the legal problem I see. SCOTUS said that you have to meet the Gingles test to claim a section 2 violation. That includes a majority minority in a compact area and racially polarized voting. In Bartlett they said that a majority meant a voting age majority of a single group. They punted on the citizen question. I see a door left open for a voting age majority in the area, yet a sub 50% amount in a specific district as long as the single group had the opportunity to elect a candidate of choice. The Dems made this type of case in their IL legislative map, and they prevailed in court last month. But you're the lawyer, so you tell me if a court would go there in CA.

I don't see these conditions met in the area. I'd content myself with drawing a (not VRA-mandated) "Hispanic influence" district... but even that does mean you don't let Watsonville lie just outside it without very very good reason.

I think there is a good possibility they would be met. There is a 50% HVAP population that can be placed in the district, as evidenced by both Torie's map and my own. Polarized voting was identified by teh Commission in the Central Valley as well as in LA county, so I think there is a reasonable expectation that it might be present here as well, though it was not tested. That would satisfy Gingles/Bartlett, and leaves only the question as to whether 45% HCVAP is sufficient elect a candidate of choice in a primary and then general election.

[muon2]

I see the Commission really screwed up the CD numbering system too. So many districts get new numbers - needlessly. Sigh.

The constitution requires that districts be numbered north to south.  The commission interpreted this to mean that districts are numbered based on the latitude of their northernmost point¹.  But if you are searching a map you tend to concentrate on the center of the district. 

There is one district (CD-8) that extends northward through the Owen Valley, but with most of the population in San Bernardino County, including 29 Palms is "north" of San Francisco and Oakland, and almost as far north as Sacramento.

Part of San Francisco County is in the bay almost to the San Rafael-Richmond bridge, so that San Francisco is "north" of Berkeley.

I have been told that someone pointed out that California has a northernmost point and indeed it is all District 1.

I assume that a perfectly valid interpretation would use the geographic center or center of population as the point to determine latitude. If so, the geographic center makes more sense than the northernmost point. As CA-8 illustrates, using an extremum of the district as the point of measurement can result in unexpected outcomes due to long extensions in the preferred direction.

[muon2]

Not all that different than mine, Sbane. As I said, it is cf city. I would post mine, but my Bradlee utility crashed again. It does so, more and more, as I complete more CD's in CA. When it is up again, I will put up my version, and we can discuss it further. Each of our versions has its merits and demerits. My version has a minor little chop into the town in which you grew up.

Are you getting error messages when it crashes? I have been, and I'm curious if yours are the same as mine.

[muon2]

You must be logged in to read this quote.

It isn't Mike, trust me. I drained the swamp dry. Nothing is left. The CVAP for this little VRA monster I bet is closer to something like 40%, maybe a tad higher - no more. You don't really think the courts would require this VRA monster to actually be drawn do you?

Your version does less to trash the map overall, requiring major surgeries all over the place, but it does drop the Hispanic percentage by one or two points. What I drew was the  max pack, saying F it to everything else, just as a masturbatory exercise.

This is the opinion of the DC district court in the Texas preclearance trial as to what the standard for measuring retrogression under Section 5 is:

http://tinyurl.com/74ppdvv

This is a section 5 decision and would be applicable to districts that impact Kings, Merced, Monterey, and Yuba.

A big part of the decision distinguishes between ability districts under section 5 and opportunity districts under section 2. I read this as saying coalitions and crossover voting can be considered in preclearance jurisdictions, though they are not generally applicable for section 2. It also implies that if there is evidence of polarized voting there, one would need a host of statistics to determine if the existing district is an ability district and to determine if the new district retrogresses it in any way.

[muon2]

I'm not sbane, but with Kings and Merced on the section 5 list I put together my version of a split. It turns out there are some very nice county groupings in the southern Central Valley that each equal 2 CDs. I've even used your Kern SE split.

I recognize that there is an ugly bit at the northern part of Tulare into Visalia. I figure you have to shift lines northward anyway to compensate for the cut into LA county from Ventura which I did not assume in my map.

[muon2]

I remain deeply unconvinced of the whole SD map. If it's fine with muon and sbane then it shall stand, of course, but to me the pairing of SD's Hispanic southern suburbs with Imperial, while better than all the alternatives, is bad enough and the Hispanic parts of central SD don't belong in the district unless that were necessary to make a solid HCVAP majority district. Which I'm quite confident it is not, Bob Filner's race notwithstanding. I'd put Coronado and the rest of Chula Vista and the empty parts of SD County in there instead, probably desert Riverside (and its huge penitentiary...) as well even though that is an additional county split. And yeah, I'm aware this has partisan consequences.

I've been thinking about this same issue for a while. I would think that the Imperial Valley and Coachella Valley are a natural pair (minus Palm Springs) and by following I-10 west and then cutting south to Perris one should get a solid Hispanic majority district. That leaves the Chula Vista area as a coalition district with less than 50% HCVAP but less than 20% white.

I'll defer to Torie's expertise on SoCal as to why this wouldn't work to fix the ugly CA-51.

[muon2]

This is the opinion of the DC district court in the Texas preclearance trial as to what the standard for measuring retrogression under Section 5 is:

http://tinyurl.com/74ppdvv

This is a section 5 decision and would be applicable to districts that impact Kings, Merced, Monterey, and Yuba.

A big part of the decision distinguishes between ability districts under section 5 and opportunity districts under section 2. I read this as saying coalitions and crossover voting can be considered in preclearance jurisdictions, though they are not generally applicable for section 2. It also implies that if there is evidence of polarized voting there, one would need a host of statistics to determine if the existing district is an ability district and to determine if the new district retrogresses it in any way.

The more significant part appears to be the DOJ rejecting any sort of percentage test, and instead demanding districts based on the election outcomes.  In Texas, this may mean that you add more Blacks to Hispanic districts since they are more likely to vote for the Hispanic candidate of choice (ie Democrat) than Hispanic voters, so long as there are not so many Blacks as to control the primary.

However, they weren't rejecting it for section 2 cases, only section 5. A percentage test could be OK in and of itself for establishing a section2 district. Even for section 5 the court said that a percentage test was a starting point for considering the outcome, but other factors must be included as well.

That distinction makes a state like CA more complicated since there are only four counties with section 5 coverage, but all with section 2.

[muon2]

Well, it's close to 50%HCVAP. It doesn't matter if it's going to elect an Hispanic, but whether a candidate of their choice is elected. The commission did draw this map, and you would have to do the same to avoid a legal challenge. What the ultimate result of that legal challenge would be I cannot say. But it is clear that there is racially polarized voting in the Central Valley, most especially in the Bakersfield area, so if a district can be drawn that is 50%HCVAP or close to it, I would go ahead and do it.

Is there any evidence that Hispanics vote at a lower rate in the Central Valley, as opposed to just being more illegal or recent immigrants than the California average? It takes a 70% Hispanic district to get to 50%HCVAP here, whereas in most of Socal all you need is about 65-66%, and in the SGV just 62% suffices. That might be why you think Hispanics don't vote here. But if a 50%HCVAP district can be drawn, and being relatively compact like the one I have drawn, I see no reason not to draw it.

It deserves study.  I doubt your Kern based Hispanic CD is anyway near 50% HCVAP, and yes, the turnout levels are lower I strongly suspect. Or is your Kern CD the one that is 65% Hispanic VAP?  Maybe Muon2 has an opinion of the degree of legal risk. If he makes a reasonable case on that that the legal risk is more than remote, then you have your Pub vote anyway, and I guess "our" commission will end up drawing something (although hopefully not as ugly as your map). 

Yes, the Kern CD is 65%HVAP and probably about 48% HCVAP. The commission map is 66%HVAP and 49%HCVAP. They contain roughly the same sort of areas, Bakersfield and rural areas but not Fresno proper. BTW, didn't Muon already draw something that is roughly similar to what I drew?

Yes, but that does not necessarily mean that he thinks it is legally necessary, or that he would draw it as a Pub on the Commission. Muon2 just gets his rocks off drawing these erose little racial gerrys.  Some of them have been just awe inspiring.

You cheated by looking at what the Commission did by the way. I have avoided doing that outside the LA ring, way back when.

I finally have the 2010 CA data working on my computer!

I don't have CVAP data, but I can use MALDEF's plan to estimate the ratio between HVAP and HCVAP in the Central Valley. The draw two districts with over 50% CVAP. The Fresno district was 61.6% HVAP to get 50.1% HCVAP, and the Kings/Kern district was 65.2% HVAP to get 50.0% HCVAP. Their map is proof of principle that they can be drawn.

Kings and Merced are both section 5 so they will have preclearance hurdles for any districts that overlap them. In addition the commission found evidence of racially polarized voting and determined that the VRA would apply there. I think there must be two Central Valley districts that have the opportunity to elect candidates of choice for Hispanics.

I'll update my earlier offerings for that area with new data soon.

[muon2]

This isn't going very well for me is it?    And then the issue is given the HVAP CD's can be drawn, are they legally mandated? What are the legal risks? Can this excrescence be described an hewing together "compact" communities of interest? And suppose given my map design (I am not saying this is the case), we can't reach the 50% HVAP percentage? Do we then redraw the whole state to create one extra 50% HVAP CD? Even that percentage may well not be enough to actually elect an Hispanic. It may still elect an Anglo Pubbie. 

Here's my version of a Kings/Kern/Tulare district. It has 65.3% HVAP and should exceed 50% HCVAP based on MALDEF's calculations. As drawn it should withstand any section 2 or 5 challenge. I think ant HVAP over 65% in this area is likely to survive, and the commission recognized that. Less than 65% HVAP and there is significant legal risk IMO.