US House Redistricting: Michigan

[Sbane]

It looks like almost 3 districts will fit into Wayne County, save for about 60,000 people. Since county splitting needs to be avoided when possible according to the guidelines for gerrymandering in the state, would the GOP be required to draw 3 Wayne County districts, with a little bit sticking into Oakland or Monroe County? This would make the current 11th district about D+3 or so. On the other hand it becomes easier to make the 9th more Republican. It might also make the 15th more Republican, unless the GOP shores Dingell up by adding Lansing to his district, but I don't see why they would do that.



Something like this. Obviously the GOP won't want to do it, and keep like it is now and give the southern suburbs to Dingell, but will they be allowed to?

[Linus Van Pelt]

I looked this up, and the relevant text is as follows - all emphasis mine. Note that the word "reasonably" appears in the line about county splits, so there's a bit of wiggle room. Also, I don't understand the bolded (iii). Does it mean that if you have to split a county, you have to split the part of the country that's left over from the districts entirely within the county into even fractions?

(link: www.legislature.mi.gov/.../documents/.../mcl-Act-221-of-1999.pdf)

You must be logged in to read this quote.

[muon2]

I looked this up, and the relevant text is as follows - all emphasis mine. Note that the word "reasonably" appears in the line about county splits, so there's a bit of wiggle room. Also, I don't understand the bolded (iii). Does it mean that if you have to split a county, you have to split the part of the country that's left over from the districts entirely within the county into even fractions?

(link: www.legislature.mi.gov/.../documents/.../mcl-Act-221-of-1999.pdf)

You must be logged in to read this quote.

The simple meaning of the section you highlighted is to only split counties between two districts when balancing population. I believe the wording is to permit multiple fractional districts in the large counties like Wayne and Oakland, where the split is not simply to make final adjustments in population between two districts.

In September I posted a neutral version of MI using the statutory rules you cited in a thread about deceptive gerrymanders.

They're not really reasonable, though. They're drawn to dilute Democratic strength.

- MI-03 puts Grand Rapids in the very western corner. A fair district would have Grand Rapids as the center of population.

- MI-04 is specifically drawn to exclude Saginaw, and stretches 2/3rds of the way across the state.

- MI-07 and MI-08 each stretch from the middle of the state to the suburbs of Detroit. Battle Creek and Lansing would more logically be put in the same district, but they're split between the two.

- MI-11 is a bizarre L-shaped district that also attempts to dilute Dem strength as much as possible.

What is interesting about this observation is that MI used fairly rigorous standards to draw districts in 2001. They were based on the standards used by the court-appointed master in 1981 and 1991 and codified into law in the late '90s. The standards rest heavily on minimizing the splitting of counties, townships and municipalities, and the law describes the types of splits that are permissible.

An analysis for the Midwest Democracy Network by Michael McDonald of George Mason U last year showed that the partisan composition of the districts had 5 strong D, 2 strong R and 8 lean R. The fact that the GOP was able to stay within these standards to get an such effective advantage perhaps does rank MI as the most deceptive gerrymander.

To satisfy my own curiosity, I tested the MI standards with the estimated data set. All districts are with 100 persons of the ideal size, and two black-majority districts (58% and 56%) are maintained. The districts were drawn to conform with MI state law as regards to minimization of county, township and municipality splits. Using 2004 presidential votes, I get 5 strong D districts, 4 strong R, 2 lean D and 3 lean R. Based on this I would conclude that the standards were fine, but not tight enough to prevent partisan gerrymandering.


 

[Torie]

That Detroit area map looks like something the GOP would not be remotely happy with, at first glance. It looks just awful. Am I missing something?

[Verily]

That Detroit area map looks like something the GOP would not be remotely happy with, at first glance. It looks just awful. Am I missing something?

They don't have a lot of choice due to the law requiring minimalization of county splits. And it's not bad for the GOP at all; they get to try to dilute Flint by putting it with ultra-GOP parts of Oakland County, and they drop some of the more Democratic parts of Macomb County into a black-majority seat.

I think the way muon deals with Washtenaw County is too risky for the GOP, though. If the GOP has to separate Dingell from Ann Arbor, they should combine Ann Arbor and Lansing to keep their other seats safe.

[Torie]

Where is McCotter supposed to park his hat?  There must be a better way than this.  The idea is to get rid of a Dem, not a Pubbie. Why not have CD-13 stay in Wayne, and pick up Dem precincts from the green district, and have the green district go north of the Wayne County line, and pick up GOP precincts?  At ;east CD-13 should pick up the most Dem precincts in Macomb. It isn't with this map. And do the districts have to be a rectangular as possible?  The Flint district, should reach down and pick up the most Dem parts of Oakland County, not the most GOP parts.

[they don't love you like i love you]

That Lansing district is kind of cutting it close.

[muon2]

Calm down. If you read my post, you'll see it was an attempt to defend the Apol rules that were used for the 1980 and 1990 cycle handled by the court, then ensconsed in statute before the 2000 cycle. I took a blind approach in the map above to show that the rules were not inherently biased, but still left a lot of wiggle room for one party as seen by the map enacted 10 years ago. Needless to say, I assumed that a party in control would act as before, and bend the map within the rules to improve their chances.

[Torie]

Well in my defense, that post was kind of hard to read the way it was formatted.  Anyway, I am pleased that you don't think the law dictates this horrible map. If it does, well the law needs to be changed.   Is there some supra majority requirement to change it, because it is in the Michigan Constitution or something?  

[muon2]

Well in my defense, that post was kind of hard to read the way it was formatted.  Anyway, I am pleased that you don't think the law dictates this horrible map. If it does, well the law needs to be changed.   Is there some supra majority requirement to change it, because it is in the Michigan Constitution or something?  

It's just state law. The constitution provides no guidance on congressional districts.

[Verily]

Here's what I think is the absolute minimum number of county splits (12), with two black majority seats (might be able to cut it down by a split if that requirement were waived). It also happens to be a relatively fair map (6D, 8R, but some of the R seats are winnable for the Dems).

[Torie]

In that map, the GOP gets 6 seats, with the gray district a toss-up.  The Dems have 7 safe seats.

How on earth, do you find 8 GOP seats? The red district is a pretty heavily lean Dem district. It has Eaton, without much in the way of heavy GOP counties to neutralize it, like Livingston.

[Verily]

In that map, the GOP gets 6 seats, with the gray district a toss-up.  

The GOP gets 7 seats on that map, plus the Macomb seat, which would favor them.

The northern seat, the mitt seat, the Ottawa-Muskegon seat, the Grand Rapids seat, the SW seat, the south-center seat and the Livingston seat are all R seats, although the south-center seat and the northern seat in some circumstances are competitive.

The two black seats, the all-Wayne seat, the Ann Arbor-Lansing seat and the Flint-Saginaw-Bay City seats are all D seats. The all-Oakland seat is probably D but is not totally safe.

Eaton County is only 53% Obama. The district overall is R+2 or 3 or so. You're thinking of Ingham County, which is Lansing.

[Torie]

We disagree about the red seat. Oh, my bad. I thought Eaton was Lansing. OK fair enough. This map ain't happening though. It really only has two safe GOP seats, the one with Livingston in it, and the one with Ottawa in it. The rest are somewhat to very marginal.

[Verily]

We disagree about the red seat. Oh, my bad. I thought Eaton was Lansing. OK fair enough. This map ain't happening though. It really only has two safe GOP seats, the one with Livingston in it, and the one with Ottawa in it. The rest are somewhat to very marginal.

I wouldn't really consider the Grand Rapids seat or the mitt seat or the SW seat to be at all winnable for the Dems, certainly not with their current incumbents. And generally speaking this is what has to happen anyway; if there are this many "marginal" R seats on a 6D-8R map, imagine how bad it would be if you started trying to crack Lansing or Ann Arbor. The loss of a seat really puts a squeeze on the current gerrymander since so many R seats are already marginal. I think McCotter is going to end up lost in the shuffle even on an R gerrymander. They certainly can't knock the Dems below their current 5 seats without further endangering incumbents.

[Torie]

Lansing is cracked now. How much population do the two black seats and the Dingell seat need to pick up is part of the question. If they have to pick up a lot, that means a lot more Dems if the map is drawn right, can be shoved into them. That is the key.

[Verily]

They've definitely shrunk, and they will try to extend one of the black districts up to take in Pontiac. But that's not going to be enough to make someone like McCotter safe when he ends up in a district with Peters or in a primary with Mike Rogers.

[Torie]

The trick is whether it is possible to squeeze Levin and Peters together. That is the dance vis a vis the Wayne County CD's.

And here is I think the GOP solution, and will be the solution that the GOP will take. The trick is to have Dingell lose Monroe County (which when added to MI-7 (the gray district that you can barely see a sliver of at the bottom of the screen shot), nicely gives the additional population that district needs), and in exchange, plus having to add population, he picks up white precincts from the southern portion of the black districts, mostly from the one to the west - rather than the white precincts in McCotter's green district (not colored yet, except for a couple of precincts).  Picking up white Dems, not white Pubbies, is what we do for Mr. Dingell. He keeps what he has in Wastenaw (even though the utility is not large enough to show he still has Ann Arbor, but he does). Then the east black district takes from precincts from the west black district, and the west black district rounds up every precinct in Oakland County where there are significant numbers of blacks, plus picks up a few white precincts to equalize population, and effect the link up to Pontiac.

So now the most Dem areas of both the Levin and Peters districts are gone. Candice Miller in the Macomb County red district, has more of a headache, but it can't be helped. The turquoise district (east Oakland north and east of where the black district stops, is not finished. It can expand in northern Macomb, or elsewhere. So  northern Macomb, and the thumb, and the balance of what is left of Peters' MI-9 district in Oakland are open for one district to replace the Peters and Levin districts, and for the other districts to expand into, and the territory is pretty GOP friendly. The GOP districts like dominoes can all be shoved east into or towards that zone, I would think. This map is the only one that creates two black districts that are 60% black. The blacks in Oakland just have to be picked up to hew to a 60% black population target, and if the percentages get down much below 60%, the black representatives are not going to be happy anyway. It should be pretty bullet proof from any judicial standpoint.

The  GOP loser to some extent is Candice Miller, but she will be able to hold her somewhat more Dem district I would think against Levin. It is either that, or get rid of the Michigan law, so more erose and precise gerrymanders can be effected. In exchange though, the GOP has the potential to replace two Dems (Levin and Peters) with one Republican, which is not a bad day's work. It will be very hard for Peters to win in his totally redone district.  At worst, even if Levin wins, that means in effect that Peters is gone, and Miller is replaced by some other Pubbie in a more GOP Oakland district, that can expand into Northern Macomb and the thumb, which it will have to, because the eastern part of his district needs to go to other Republicans. Poor Oakland - the sliced and diced county, time after time.

[map deleted]

[Torie]

And here is the final "solution" for the Detroit metro area.  I suppose actually that the turquoise district should pick up a bit more of Saginaw, so the green district can pick up a few precincts in Oakland from the turquoise district, while losing its four precincts in Wastenaw, so that county split can be avoided. I will do that later, and replace this map.

[they don't love you like i love you]

That is so much harder to view than version 1.0. I'll stick with version 1.0 for most things from here on. The only real advantage 2.0 has is that the partisan data for it now works for California, Texas and New York which didn't in 1.0 for me for some reason, so I'll just use it to compare partisan figures.

[jimrtex]

That is so much harder to view than version 1.0. I'll stick with version 1.0 for most things from here on. The only real advantage 2.0 has is that the partisan data for it now works for California, Texas and New York which didn't in 1.0 for me for some reason, so I'll just use it to compare partisan figures.

It looks like it does real projections.  I don't see how one could avoid a seizure working with the old lat/long data.

[Torie]

Well this is about as nasty a gerrymander of Oakland County as I could manage, and it is probably illegal, but I leave that to Muon2 to tell me. My aim is to try to equalize the PVI's for the pink, turquoise, green and purple districts, with the purple having a bit more, since the Obama swing was more in the purple zone, as far as I can tell, but maybe not given what I have put in it. The goal is to get the PVI's four all four districts to about GOP +3, with the purple district a bit more. I am not sure I have done that (probably not for McCotter's Green district , but I packed as many GOP districts into it as possible in Oakland, while leaving what it has in Wayne County alone), but I did the best I can and it should hopefully get up to about an GOP plus 1-2% PVI anyway. If the turquoise CD  jut down to the Wayne County line is illegal, and the Green district has to fill much of it in, then the green CD's PVI of 0% will not change much, and might even drop a tad.

Leips' application does not have the partisan data for Michigan yet, but looking at the township and city returns, and the black percentages in Oakland, I am pretty sure I have corralled all, or almost all,  of the most GOP precincts for the Pubbies in Oakland, after dumping most if not all of the most Dem precincts in Oakland as possible into the black district. So,  it is just a matter of divvying the more GOP oriented districts in Oakland between the green, turquoise and purple districts. I wanted to chew up Saginaw and Bay City, because I don't think the northern CD's can afford to include either city, and thus they need to be neutralized from the south (and were by the turquoise district), since the Flint district instead of neutralizing Bay City and Saginaw as it did before, now neutralizes Lansing. The turquoise district also has a long thin jut down into south Oakland, which is about even territory for Bush 2004 (maybe a slight Bush 2004 lean of plus 1-2% or so, since I chopped up the townships in the jut area between what appears to be the more Dem and GOP pieces, so it has a somewhat Dem PVI - maybe a Dem PVI +2% or so.  I did that to try to get the GOP PVI for the Green district up as much as possible; the green district  is now  PVI 0%, and it needs as much beefing up as possible. So its expansion was into new more heavily GOP Oakland territory, with a GOP PVI of maybe plus 4% or so, although I am not sure. If the turquoise CD jut down south to the Wayne County border is illegal, then the green CD's PVI of 0% will not change much, and probably go down a fraction to maybe Dem +0.5% or so.

In any event, as long as McCotter, Miller, and Rodgers hang around, Levin and Peters face a near hopeless task of getting elected anywhere, at least absent a big Dem shift in this area, and some new Pubbie has in the purple district pretty ripe hanging fruit to pick up.

[muon2]

Well this is about as nasty a gerrymander of Oakland County as I could manage, and it is probably illegal, but I leave that to Muon2 to tell me. My aim is to try to equalize the PVI's for the pink, turquoise, green and purple districts, with the purple having a bit more, since the Obama swing was more in the purple zone, as far as I can tell, but maybe not given what I have put in it. The goal is to get the PVI's four all four districts to about GOP +3, with the purple district a bit more. I am not sure I have done that (probably not for McCotter's Green district , but I packed as many GOP districts into it as possible in Oakland, while leaving what it has in Wayne County alone), but I did the best I can and it should hopefully get up to about an GOP plus 1-2% PVI anyway. If the turquoise CD  jut down to the Wayne County line is illegal, and the Green district has to fill much of it in, then the green CD's PVI of 0% will not change much, and might even drop a tad.

Leips' application does not have the partisan data for Michigan yet, but looking at the township and city returns, and the black percentages in Oakland, I am pretty sure I have corralled all, or almost all,  of the most GOP precincts for the Pubbies in Oakland, after dumping most if not all of the most Dem precincts in Oakland as possible into the black district. So,  it is just a matter of divvying the more GOP oriented districts in Oakland between the green, turquoise and purple districts. I wanted to chew up Saginaw and Bay City, because I don't think the northern CD's can afford to include either city, and thus they need to be neutralized from the south (and were by the turquoise district), since the Flint district instead of neutralizing Bay City and Saginaw as it did before, now neutralizes Lansing. The turquoise district also has a long thin jut down into south Oakland, which is about even territory for Bush 2004 (maybe a slight Bush 2004 lean of plus 1-2% or so, since I chopped up the townships in the jut area between what appears to be the more Dem and GOP pieces, so it has a somewhat Dem PVI - maybe a Dem PVI +2% or so.  I did that to try to get the GOP PVI for the Green district up as much as possible; the green district  is now  PVI 0%, and it needs as much beefing up as possible. So its expansion was into new more heavily GOP Oakland territory, with a GOP PVI of maybe plus 4% or so, although I am not sure. If the turquoise CD jut down south to the Wayne County border is illegal, then the green CD's PVI of 0% will not change much, and probably go down a fraction to maybe Dem +0.5% or so.

In any event, as long as McCotter, Miller, and Rodgers hang around, Levin and Peters face a near hopeless task of getting elected anywhere, at least absent a big Dem shift in this area, and some new Pubbie has in the purple district pretty ripe hanging fruit to pick up.

The simple application of the Michigan law is that within a split county there can be only one split city or township between a pair of districts. All the districts in your Oakland map seem to share multiple split municipalities. So I'm afraid to advise you that it would be an illegal gerrymander.

For instance, I think that if you are going to have a narrow finger linking Pontiac through West Bloomfield, then you would have to have all of Farmington Hills in the Pontiac - Detroit district. Waterford would have to be out, and Auburn Hills would have to be all in or all out since it looks like it is split three ways.

[Torie]

Thanks Muon2. But can you override all of those annoying little requirements via the VRA  in order to get the black districts up to 60% of the population? You have to slice townships and towns to do that (probably also in Wayne, if the same requirement obtains there), or you are going to fall down to 57% or something, and among VAP, lower still (and among voters, we may then be barely over 50%)? I checked the black percentage of each precinct I appended to the black district in Oakland, in order to maximize its black percentage. And there are no other blacks to be had to speak of in the Detroit metro area. I corralled them all.

By the way, for future reference, here is the text of the Michigan Law. The Pubbies revised it to do their last gerrymander, hewing to some Michigan Supreme Court precedents apparently, based on what I don't know. I suppose it could be tweaked if necessary. For example, my turquoise district is not very compact, but that is just the way it has to be.

[Verily]

But can you override all of those annoying little requirements via the VRA  in order to get the black districts up to 60% of the population. You have to slice townships and towns to do that, or you are going to fall down to 57% or something, and among VAP, lower still, and among voters?

There are VRA districts out there that are only 52-53% black. You certainly don't need 57%, let alone 60%, of the general population to be of the minority group, at least not for blacks, who are not far from whites in proportion in the VAP (compare to Hispanics, who are disproportionately below voting age).

Anyway, the VRA only applies to Michigan insofar as you need to avoid having districts that intentionally dilute the black vote such that their preferred candidate routinely loses. Even 45% black wouldn't be an intentional dilution.