New PA Maps In Effect

[muon2]

And failed to persuade Justice Alito?

To be fair, I'm sure Justice Alito would have gone along with any justification (dissenting opinion, Scottish law, Klingon precedent) that preserved 13 Republican butts in seats for 2018, but if he couldn't get 4 of his colleagues to go along there was no point in looking like a loser on the stay.

Another possibility is that Alito knows that SCOTUS has the votes to rule against political gerrymandering and will do so later this year. That means PASC would likely be upheld. Therefore a stay serves no good purpose from SCOTUS' perspective.

[muon2]

http://www.philly.com/philly/news/politics/state/pennsylvania-gerrymander-case-deadline-republicans-strategy-20180208.html

You must be logged in to read this quote.

They told the court they would need around 3 weeks to redraw the map, and then after told to redraw the map, they waste all but <week of that time whining to the US Supreme Court, not even opting to prepare a map (or two) just in case their request for a stay is denied. Now they are left scrambling, looking to cut corners.

Is there any situation where a Republican-controlled legislature would just opt to throw up their hands and say, "fine, you win this round. The decade is almost over, let's just draw a fair map with marginal incumbent considerations and be done with it." Is that possible anywhere in the country? For gods sakes. Stop being so god damn greedy.

Something to note is that the original schedule has been changed a bit, with a map simply needing to be approved by the 15th. Previously, the legislature only had till tomorrow to create a map, now they at least have a few days. The court has also stated that if all parties in this process agree (candidates, governor, legislature, justices) then it would be fine to move the primary a few weeks if incredibly necessary.

But yeah, Pubs are right now in a corner without a real escape plan.

It's not just parties used to the majority; even in a minority position parties have a hard time throwing their members under the bus. In 2011 the Pubs were given a plan with a solid 2nd Latino CD that could have been used to challenge the IL map, but it would likely cost 1 or 2 Pub seats. They went with a weak Latino-influence CD instead that maintained all the Pub seats that had been won in 2010, but lost their claim that the Dems shouldn't reverse so many seats that were affirmed by voters in the previous cycle.

[muon2]

And failed to persuade Justice Alito?

To be fair, I'm sure Justice Alito would have gone along with any justification (dissenting opinion, Scottish law, Klingon precedent) that preserved 13 Republican butts in seats for 2018, but if he couldn't get 4 of his colleagues to go along there was no point in looking like a loser on the stay.

Another possibility is that Alito knows that SCOTUS has the votes to rule against political gerrymandering and will do so later this year. That means PASC would likely be upheld. Therefore a stay serves no good purpose from SCOTUS' perspective.

One can dream....

Well clearly Alito knows the outcome in Gill v Whitford (WI). Like the legislative case in WI, the PA case is based in part on the requirements of the state constitution. That differentiates PA from the gerrymandering cases for MD and NC which are federal questions only. It seems to me that if Gill was going in favor of the state, Alito would almost certainly have gone for a stay for the PA map. But if Gill was going against the state, and the argument in conference was something that would also apply to PA, then then the fate of the PA case is already sealed. In late Mar SCOTUS hears Benisek v. Lamone (MD) and the questions there may shed even more light as to the direction of SCOTUS' thinking on the subject.

[muon2]

And failed to persuade Justice Alito?

To be fair, I'm sure Justice Alito would have gone along with any justification (dissenting opinion, Scottish law, Klingon precedent) that preserved 13 Republican butts in seats for 2018, but if he couldn't get 4 of his colleagues to go along there was no point in looking like a loser on the stay.

Another possibility is that Alito knows that SCOTUS has the votes to rule against political gerrymandering and will do so later this year. That means PASC would likely be upheld. Therefore a stay serves no good purpose from SCOTUS' perspective.

One can dream....

Well clearly Alito knows the outcome in Gill v Whitford (WI). Like the legislative case in WI, the PA case is based in part on the requirements of the state constitution. That differentiates PA from the gerrymandering cases for MD and NC which are federal questions only. It seems to me that if Gill was going in favor of the state, Alito would almost certainly have gone for a stay for the PA map. But if Gill was going against the state, and the argument in conference was something that would also apply to PA, then then the fate of the PA case is already sealed. In late Mar SCOTUS hears Benisek v. Lamone (MD) and the questions there may shed even more light as to the direction of SCOTUS' thinking on the subject.

A complication is that the PA Supreme Court finally explained their decision, and endorsed the efficiency theory, meaning the map needs to be gerrymandered to help the Dems get more proportionality (one map somebody drew at RRH had a seat going from Lackawanna to Centre county, excluding Luzerne in a long erose snake with hideous bridge chops as an example of this concept). It is highly unlikely Justice Kennedy will adopt that concept, so to that extent PA law and federal law will substantially diverge. Things are really a mess now because the efficiency theory and minimizing chops and maximizing compactness, don't go together very well at all in states such as PA.

Plus as jimrtex and I showed last spring, the mathematics of efficiency theory is hogwash.

[muon2]

They're taking the concept of efficiency theory to the extreme.

Yes, but unfortunately the legal minds and partisans on both sides refuse to look at efficiency theory applied to neutral maps in a variety of elections, including typical waves for both sides. Mathematically efficiency theory punishes maps that must last a decade and include a lot of swing districts. The theory also works against creating true section 2 VRA districts, and rewards the creation of minority districts that must rely on white crossover voters.

[muon2]

And failed to persuade Justice Alito?

To be fair, I'm sure Justice Alito would have gone along with any justification (dissenting opinion, Scottish law, Klingon precedent) that preserved 13 Republican butts in seats for 2018, but if he couldn't get 4 of his colleagues to go along there was no point in looking like a loser on the stay.

Another possibility is that Alito knows that SCOTUS has the votes to rule against political gerrymandering and will do so later this year. That means PASC would likely be upheld. Therefore a stay serves no good purpose from SCOTUS' perspective.

One can dream....

Well clearly Alito knows the outcome in Gill v Whitford (WI). Like the legislative case in WI, the PA case is based in part on the requirements of the state constitution. That differentiates PA from the gerrymandering cases for MD and NC which are federal questions only. It seems to me that if Gill was going in favor of the state, Alito would almost certainly have gone for a stay for the PA map. But if Gill was going against the state, and the argument in conference was something that would also apply to PA, then then the fate of the PA case is already sealed. In late Mar SCOTUS hears Benisek v. Lamone (MD) and the questions there may shed even more light as to the direction of SCOTUS' thinking on the subject.

A complication is that the PA Supreme Court finally explained their decision, and endorsed the efficiency theory, meaning the map needs to be gerrymandered to help the Dems get more proportionality (one map somebody drew at RRH had a seat going from Lackawanna to Centre county, excluding Luzerne in a long erose snake with hideous bridge chops as an example of this concept). It is highly unlikely Justice Kennedy will adopt that concept, so to that extent PA law and federal law will substantially diverge. Things are really a mess now because the efficiency theory and minimizing chops and maximizing compactness, don't go together very well at all in states such as PA.

Plus as jimrtex and I showed last spring, the mathematics of efficiency theory is hogwash.

Where is the link to that? Putting aside the internal mathematics, I wonder if there is a metric that more "efficiently" moves towards proportionality (in other words, has a higher number for the ratio for the amount of proportionality gained, divided by the increased erosity and chop count). The VRA will still require performing minority CD's, with crossover white votes or no. That would trump the PA constitution as interpreted by PA's highest court.

I think the map I linked does have equal population districts. The yellow CD is really two CD's with yellows of slightly different shades. The guy who drew the map is a competent map drawer. He thinks this is the kind of map the PA Supreme Court will impose, which yes, would create quite a fire storm if it does, and might move the Pubs to go the impeachment and convict route even (they have the votes to do that).

I don't have the time this morning to collect the links, but they are on this board from last spring. The short form is that I constructed a neutral WI legislative map that comported with the VRA and measured it with both Cook PVI and muon2 SKEW and found it it to be fair and reflective of the electorate from 2004 to 2016. Measured by the efficiency gap test used in WI he same map would fail as a Dem gerrymander in 2008 and a Pub gerrymander in 2016 - the exact same map! The efficiency gap assumes no change in the electorate's behavior over time, so swing districts punish a plan for the party that wins in a wave election year.

[muon2]

Let's just say darn near anything is better than the system in place now...

A system that forces plans that lock in the two parties' seats for a decade (as the efficiency gap would), regardless of swings in the electorate, is not necessarily better than the present system. We wouldn't be talking about the chance of parties changing control of the House with each cycle.Check out the CA maps before their reform. They were bipartisan gerrymanders that would have passed the efficiency test, but the seats didn't change parties except during scandals. That's not representative government either.

[muon2]

What is especially curious is that there is essentially no intersection between this opinion and the Gill v. Whitford case.

The PA Court opinion holds that congressional districts are unconstitutional under a provision of the state constitution that has no equivalent in the US constitution.  Their reasoning is almost entirely on the basis that they are non-compact (as judged by both geometric measures or splitting of existing subdivisions).  They put very little weight on the resulting partisan split of the delegation nor any reliance on constitutional provisions that have a federal equivalent, and they considered no evidence about the partisan motivation of the gerrymander.

The Whitford plaintiffs argue that state assembly districts are unconstitutional under several provisions of the federal constitutions (which have equivalents in the PA constitution, but are ignored by the PA court).  Their argument is based entirely on the claim that the partisan split of the delegation is biased, combined with qualitative evidence of partisan motivation.  Compactness is not at issue at all; indeed the plaintiffs acknowledge that the existing districts are reasonably compact.

Yet Alito rejected the stay without even referring it to the full court. That is what makes me think there is something in the PA ruling that lines up with the direction SCOTUS is taking in Whitford.

[muon2]

What is especially curious is that there is essentially no intersection between this opinion and the Gill v. Whitford case.

The PA Court opinion holds that congressional districts are unconstitutional under a provision of the state constitution that has no equivalent in the US constitution.  Their reasoning is almost entirely on the basis that they are non-compact (as judged by both geometric measures or splitting of existing subdivisions).  They put very little weight on the resulting partisan split of the delegation nor any reliance on constitutional provisions that have a federal equivalent, and they considered no evidence about the partisan motivation of the gerrymander.

The Whitford plaintiffs argue that state assembly districts are unconstitutional under several provisions of the federal constitutions (which have equivalents in the PA constitution, but are ignored by the PA court).  Their argument is based entirely on the claim that the partisan split of the delegation is biased, combined with qualitative evidence of partisan motivation.  Compactness is not at issue at all; indeed the plaintiffs acknowledge that the existing districts are reasonably compact.

Yet Alito rejected the stay without even referring it to the full court. That is what makes me think there is something in the PA ruling that lines up with the direction SCOTUS is taking in Whitford.

We know for a fact it was not referred to the Court?

This is from Scotusblog:

You must be logged in to read this quote.

[muon2]

I count 19 chops in 2 regions:

Western Region - 11 CDs, 11 chops
Crawford
Allegheny (2)
Washington
Fayette
Mifflin
Lycoming
Cumberland
Dauphin
Luzerne
Carbon

Philly Region - 7 CDs, 8 chops
Berks
Chester
Delaware
Montgomery (3)
Philadelphia (2)

In the western region there is only one excess chop. The fact that there are 2 CDs in both Allegheny and Washington shows that there is an excess. In the eastern region there are two excess chops. One is due to 2 CDs in both Chester and Montgomery. The other is due to a loop of CDs connecting chopped counties: Philadelphia-Delaware-Chester-Montgomery.

[muon2]

I agree the Alito's denial should not have been a close call, and I'm glad to see that apparently it wasn't.

But I don't see how this suggests anything about the ultimate disposition of Whitford.  The cases are based on totally different arguments along every relevant dimension.

Perhaps there's another reason and I'm grasping at straws, but this doesn't strike me as something Alito would normally dismiss so readily. Whitford is the only insider info related to redistricting he has that he could use to go straight for a dismissal, given that Roberts stayed NC.

[muon2]

That map is god awful and completely ignores the court order.   19 county splits???  That's absurd.

And the courts wanted more competitive districts and they make PA-1 and PA-2 into even HEAVIER dem vote sinks?

This is a slap in the face

I wouldn't call 19 splits absurd. There are only 3 more chops than the minimum needed to meet their level of pop inequality, and there appear to be relatively few county subunit chops in the chopped counties (perhaps oryxslayer has that exact answer from his reconstruction). It also looks like they can claim bipartisan incumbent residence protection in this plan, keeping those running again in the same CD.

PA-1 and 2 were never going to be competitive, so I doubt the court will care which way they change as long as it doesn't violate the VRA. Competitiveness will likely be measured in the number of CDs that have PVI of 5 or less, and here the new Pub plan actually loses one as PA-16 goes from R+5 to R+7. However, I suspect the Pubs would counter by arguing that the improvements to PA-7 from R+1 to D+2 and PA-15 from R+4 to R+1 are more meaningful than the shift of PA-16. My skew measurement would agree with the Pubs in that regard, but my polarization measure would not.

The Pubs are certainly short of an ideal plan, but they can point to clear improvements in chops, erosity, and skew, at least as I would measure them. That may be their objective if they are faced with another round in court.

[muon2]

That map is god awful and completely ignores the court order.   19 county splits???  That's absurd.

And the courts wanted more competitive districts and they make PA-1 and PA-2 into even HEAVIER dem vote sinks?

This is a slap in the face

I wouldn't call 19 splits absurd. There are only 3 more chops than the minimum needed to meet their level of pop inequality, and there appear to be relatively few county subunit chops in the chopped counties (perhaps oryxslayer has that exact answer from his reconstruction). It also looks like they can claim bipartisan incumbent residence protection in this plan, keeping those running again in the same CD.

PA-1 and 2 were never going to be competitive, so I doubt the court will care which way they change as long as it doesn't violate the VRA. Competitiveness will likely be measured in the number of CDs that have PVI of 5 or less, and here the new Pub plan actually loses one as PA-16 goes from R+5 to R+7. However, I suspect the Pubs would counter by arguing that the improvements to PA-7 from R+1 to D+2 and PA-15 from R+4 to R+1 are more meaningful than the shift of PA-16. My skew measurement would agree with the Pubs in that regard, but my polarization measure would not.

The Pubs are certainly short of an ideal plan, but they can point to clear improvements in chops, erosity, and skew, at least as I would measure them. That may be their objective if they are faced with another round in court.

Your analysis might have some traction if Wolff signed off on this map. Then the issue is whether the map is sufficiently "de-gerrymandered" to be legal. But if Wolff does not (and why on earth would he?), then no map has been adopted, and just why would the court draw a map that is still a GOP gerrymander, but less egregious, just because the Pubs drew it, rather than draw a map that is not gerrymandered, and eliminates the excess chops that are there for partisan reasons? What would you do as a judge? If you were Wolff, in his position, as a Dem, would you sign off on this map? You do agree that this map is still a Pub gerrymander I assume.

Unless there's some push in the court of public opinion I see no reason for Wolf to sign off. His easiest response would seem to note that it hasn't passed the Assembly so he has no formal need to act, but state that he thinks that the Pubs could have done more.

I did read the Feb 9 Pub response to the court that jimrtex linked and I can add that there is one additional chop in Berks that is invisible of the scale of the map. There are 18 subunit chops and that is one more than the minimum, since they are proposing exact equality. They rely on a basket of expert-generated plans that were recognized by the court and shows that the new plan fits into their measures of chops and erosity of that expert basket. They also confirm my guess that they have protected the CDs of all incumbents seeking reelection with 68.3% of the Pub and 69.8% of the Dem prior population retained in the new map.

[muon2]

The document says there are 17 muni chops. Is it 18, because a trip-chopped muni only counts as one chop? Are we in same syndrome as the Ohio contest rules as to tri-chops being the same bi-chops?

To some degree. They count the number of counties that are chopped and the number of munis that are chopped, and separately list how many fragments are in those units that are chopped. What I generally call a fragment, they call a slice. This creates a bias towards multi-chopping units, though they try to show that they didn't let that get out of hand. They also count the number of fragments in each district to show that no district has too many fragments.

[muon2]

I did read the Feb 9 Pub response to the court that jimrtex linked and I can add that there is one additional chop in Berks that is invisible of the scale of the map. There are 18 subunit chops and that is one more than the minimum, since they are proposing exact equality. They rely on a basket of expert-generated plans that were recognized by the court and shows that the new plan fits into their measures of chops and erosity of that expert basket. They also confirm my guess that they have protected the CDs of all incumbents seeking reelection with 68.3% of the Pub and 69.8% of the Dem prior population retained in the new map.

One additional note on population equality that explains the 67 person chop into Berks. The PA Pubs had to redraw their gerrymandered map in 2002 based on a 19 person deviation (Vieth II, an earlier lower court decision related to, but not the same as, the SCOTUS 2004 case). Since they can't point to a rational basis for a deviation they have to have a bunch of chops to get exact equality. I imagine that if they set a specific goal of no subdivision chops, including Philly wards, they could introduce modest deviation based on the WV case. That seems to be a risk they didn't want in their response to the court.

[muon2]

This diagram perfectly illustrates my thoughts on the GOP map:



https://www.washingtonpost.com/amphtml/news/wonk/wp/2018/02/11/pennsylvania-republicans-have-drawn-a-new-congressional-map-that-is-just-as-gerrymandered-as-the-old-one/

It's a partisan gerrymander without the insane lines.   You can't get those results without having partisan intent.

This illustrates one of the more amusing aspects of computer-assisted gerrymandering. Computer map neophytes let the program get every possible decimal point of PVI they can. they gerrymander to improve a map from R+3.04 to R+3.16. Yet a little research would show that a typical CD will move by far more than that just in the 2 1/2 years between the census and the first general election with new districts. So there are all sorts of erose features that really amount to nothing in elections. That's why the PA Pubs can get essentially the same result as they did initially without the contorted shapes.

[muon2]

That ACLU map looks like the naive geographically compact maps on 538. Why would you split Huntingdon and Northumberland county unless they were trying to make some of the middle districts looks round or square?

I agree that it looks like the ACLU elevated compactness above the other criteria. The partisan maps include knowledge of incumbent residences and cores of previous districts.

The Jan 22 order specified only the following criteria, and the Feb 7 order specifically said it did not change them:
Contiguous and compact districts;
Districts as nearly equal in population as practicable;
Avoid division of any county, city, incorporated town, borough, township, or ward, except where necessary to ensure equality of population.
These criteria mirror that criteria in the PA constitution governing the creation of legislative districts. Note, that incumbent protection is not one of the criteria.

Contiguous and compact districts are a frequently used phrase, and in the Feb 7 opinion referred the Reock and Polsby-Popper tests of compactness. When these tests are applied to the entire plan one finds that individual districts may not be compact if it facilitates other districts becoming more compact. Polsby-Popper gives weight to the perimeter length and works against natural irregular boundaries such as county lines in a way that the Reock test does not. Perhaps the ACLU gave more weight to that test.

The as nearly equal as practicable standard for population was the subject of Tennant v Jefferson County (2012). The WV plan had a population range of 0.79% of the quota but survived because of the neutral criteria involved, including division of no counties. In PA it is impossible to divide no counties, but it is possible to minimize the division of counties and divide no county subdivisions. That along with contiguity and compactness should make Tennant the standard for interpreting population equality.

As noted it is impossible to avoid the division of counties. To minimize the division one has to decide is the goal to minimize fragments created by county division, or to minimize the number of counties divided. Looking at the record of plans in PA and their analysis it would seem that the latter is the key factor - the total number of divided counties should be minimized. That factor is modified by the grouping of divided counties into counties split 2 ways, 3 ways, 4 ways, etc. The size of Philadelphia makes a 3-way division inevitable, but nothing forces any greater fragmentation of a county. A reasonable way to meet this requirement would be to avoid plans that have more than 2 chops (3 districts) in a county.

With that in mind I applied those criteria to come up with the following plan, which would be my submission if the public were invited to submit plans. There are only 6 chopped counties and none are chopped more than twice. Chops within a county are drawn to maximize compactness while creating no chops to any county subdivision, including wards within Philadelphia. CD-2 meets the VRA with BVAP 52.6%. The population range is 0.58%, within the Tennant range accepted by SCOTUS. CDs 5 and 10 are arguably not compact, but their shape is largely due to the adjacent compact districts comprised of whole counties, and would not unduly hurt the average compactness of the plan.

[muon2]

That ACLU map looks like the naive geographically compact maps on 538. Why would you split Huntingdon and Northumberland county unless they were trying to make some of the middle districts looks round or square?

I agree that it looks like the ACLU elevated compactness above the other criteria. The partisan maps include knowledge of incumbent residences and cores of previous districts.

The Jan 22 order specified only the following criteria, and the Feb 7 order specifically said it did not change them:
Contiguous and compact districts;
Districts as nearly equal in population as practicable;
Avoid division of any county, city, incorporated town, borough, township, or ward, except where necessary to ensure equality of population.
These criteria mirror that criteria in the PA constitution governing the creation of legislative districts. Note, that incumbent protection is not one of the criteria.

Contiguous and compact districts are a frequently used phrase, and in the Feb 7 opinion referred the Reock and Polsby-Popper tests of compactness. When these tests are applied to the entire plan one finds that individual districts may not be compact if it facilitates other districts becoming more compact. Polsby-Popper gives weight to the perimeter length and works against natural irregular boundaries such as county lines in a way that the Reock test does not. Perhaps the ACLU gave more weight to that test.

The as nearly equal as practicable standard for population was the subject of Tennant v Jefferson County (2012). The WV plan had a population range of 0.79% of the quota but survived because of the neutral criteria involved, including division of no counties. In PA it is impossible to divide no counties, but it is possible to minimize the division of counties and divide no county subdivisions. That along with contiguity and compactness should make Tennant the standard for interpreting population equality.

As noted it is impossible to avoid the division of counties. To minimize the division one has to decide is the goal to minimize fragments created by county division, or to minimize the number of counties divided. Looking at the record of plans in PA and their analysis it would seem that the latter is the key factor - the total number of divided counties should be minimized. That factor is modified by the grouping of divided counties into counties split 2 ways, 3 ways, 4 ways, etc. The size of Philadelphia makes a 3-way division inevitable, but nothing forces any greater fragmentation of a county. A reasonable way to meet this requirement would be to avoid plans that have more than 2 chops (3 districts) in a county.

With that in mind I applied those criteria to come up with the following plan, which would be my submission if the public were invited to submit plans. There are only 6 chopped counties and none are chopped more than twice. Chops within a county are drawn to maximize compactness while creating no chops to any county subdivision, including wards within Philadelphia. CD-2 meets the VRA with BVAP 52.6%. The population range is 0.58%, within the Tennant range accepted by SCOTUS. CDs 5 and 10 are arguably not compact, but their shape is largely due to the adjacent compact districts comprised of whole counties, and would not unduly hurt the average compactness of the plan.

Did you submit your map to the court?

I found no address for public submission or I would have. I have a 5 page doc file with the detailed analysis.

[muon2]

Lamb would probably beat Rothfus.

Saccone is going to be a short-lived congressman. He's not beating Doyle.

Trump won the new 14th by 30 points - if Saccone wins the special, that seat is his. Doyle almost certainly runs in the new 18th.

Saccone's home will be in Doyle's 14. Unless he moves to the new 14 (old 18th) he's BTFO. The new PA17 (old 12th) voted for Trump by just a 3 point margin. Lamb would beat Rothfus.

You don't need to live in a district to represent it.

You kind of have to.

This came up a lot in the GA-6 special election because Jon Ossoff actually lived in GA-5, just outside the boundaries of the 6th.  Some research by the Washington Post (cited here in the Atlanta Journal-Constitution) showed that there were at least 20 current Representatives who lived outside their districts, generally due to cases where a shift of district lines had moved an incumbent to a different district.

Incidentally, if Ossoff had won, it appears that there would have been three sitting Congressmen living in John Lewis's GA-5.  In addition to Lewis, David Scott (GA-13) is also registered to vote in the 5th.

Indeed. In the previous decade at least 3 IL congresscritters lived outside their district and none had trouble with their elections. States typically have residency rules for legislative and local offices, but they can't enforce them on Congress even if they put such laws one the books. The reasoning is the same that blocked term limits when some states passed those. A state may not impose qualifications on congressional officeholders beyond those in the US constitution.