Mid-2014 county population estimates out tomorrow, March 26

[muon2]

Here's the press release from the Census Bureau. The complete set of files aren't up yet.

[muon2]

Why are Illinois and New Mexico losing people? Especially Illinois? I know growth slowed, but with immigration it should be growing?

Net outmigration to other states from 7/13 to 6/14 in IL was almost 100K. Births and foreign immigration kept it closer. The recession was pinning many in their homes, and with its end people are going to where better jobs are. In many ways this data illustrates the reason for the  vote for Gov in 2014 to change the direction of the state.

[muon2]

As an exercise I took the 2014 estimates for IL and projected those growth rates to 2020. Then I assumed that IL would have 17 CDs and divided the state accordingly.

I permitted a variance of 5% of a CD and preserved UCCs, which accounts for the district that wraps around the western exurbs of Chicago. That district plus Lake county (light green) are equal to two CDs if Lake takes up about 20% of McHenry. Kane and DuPage (gold and orange) are together just under 2 CDs with Kane needing about 25% of DuPage to balance the population. Cook county is just under 7 CDs in population.

[muon2]

Here's a map showing the county estimates projected to 2020 for MI then divided into 13 CDs. Except around Detroit they each represent the population of one CD within 5% of the quota. I split the Detroit area into a purple region with 2 CDs and a pink region with 4 CDs to illustrate the relative population.

[muon2]

Here's OH divided into 15 CDs using the projections of the county data out to 2020. Districts preserve UCC covering rules. All districts are within 5% except suburban Cincinnati which is slightly over and is balanced by the adjacent southern OH district. The Columbus and Cleveland areas each contain two districts.

[muon2]

As an exercise I took the 2014 estimates for IL and projected those growth rates to 2020. Then I assumed that IL would have 17 CDs and divided the state accordingly.

I permitted a variance of 5% of a CD and preserved UCCs, which accounts for the district that wraps around the western exurbs of Chicago. That district plus Lake county (light green) are equal to two CDs if Lake takes up about 20% of McHenry. Kane and DuPage (gold and orange) are together just under 2 CDs with Kane needing about 25% of DuPage to balance the population. Cook county is just under 7 CDs in population.

With Illinois losing one CD, what percentage of that lost CD is from 1) the City of Chicago, 2)suburban Cook, 3) the balance of the Chicago UCC ex Cook, and 4) downstate?

The Chicago numbers for 2014 aren't out yet, but I can look at the other three divisions.

Cook: 2010 - 7.29 CDs, 2020 - 6.98 CDs, net loss 0.31 CDs.
Chicago UCC less Cook: 2010 - 4.69 CDs, 2020 - 4.52 CDs, net loss 0.17 CDs.
Downstate: 2010 - 6.02 CDs, 2020 - 5.50 CDs, net loss 0.52 CDs.

There is likely to be at least one CD with substantial population in both the suburbs and downstate area.

[muon2]

IN isn't likely to change in 2020 from 9 CDs and the existing plan largely keeps counties intact. I used that starting point and the projection based on 2014 estimates projected to 2020 to construct this plan. Districts are all within 5% of the quota and preserve the UCC covers. Beyond the UCC, borders follow the existing districts. The Indianapolis area (light gray) has 2 CDs.

[muon2]

Rounding out the Great Lakes states is WI which also is not expected to change from 8 CDs. That's a net loss of 3 from that region (IL, MI, OH).

As with the other states WI is based on projections to 2020 using the 2014 estimates. UCC covers are preserved, meaning that each UCC has no more than the minimum number of districts. The Milwaukee area (gray) has 2 CDs. The map works out very nicely despite the high population counties so that the districts are within 2% of the quota, though I don't expect that to hold up as the decade progresses. The Green Bay district (pink) is nearly identical to the current WI-08 showing that area is growing at about the same rate as the state overall.

[muon2]

As an exercise I took the 2014 estimates for IL and projected those growth rates to 2020. Then I assumed that IL would have 17 CDs and divided the state accordingly.

I permitted a variance of 5% of a CD and preserved UCCs, which accounts for the district that wraps around the western exurbs of Chicago. That district plus Lake county (light green) are equal to two CDs if Lake takes up about 20% of McHenry. Kane and DuPage (gold and orange) are together just under 2 CDs with Kane needing about 25% of DuPage to balance the population. Cook county is just under 7 CDs in population.

With Illinois losing one CD, what percentage of that lost CD is from 1) the City of Chicago, 2)suburban Cook, 3) the balance of the Chicago UCC ex Cook, and 4) downstate?

The Chicago numbers for 2014 aren't out yet, but I can look at the other three divisions.

Cook: 2010 - 7.29 CDs, 2020 - 6.98 CDs, net loss 0.31 CDs.
Chicago UCC less Cook: 2010 - 4.69 CDs, 2020 - 4.52 CDs, net loss 0.17 CDs.
Downstate: 2010 - 6.02 CDs, 2020 - 5.50 CDs, net loss 0.52 CDs.

There is likely to be at least one CD with substantial population in both the suburbs and downstate area.

With the Chicago area draining African Americans, the VRA is going to be tough to nest 7 CD's in Cook County. Southwest suburban Cook has too many whites I suspect for an African American CD to take all of that territory, with the landscape their trapped by the Hispanic CD to the north. It looks like a map drawn to your metrics might end up being an even more natural Pub gerrymander than the 2010 numbers dictate (which suit the Pubs quite well). And all of the Downstate CD's you drew look like what would be the marginal CD's based on a draw using the 2010 numbers all move towards reasonably safe Pub CD's, along with the Will County CD moving substantially in the Pub direction.

Losing a CD makes all the existing districts larger. The problem for the Dems downstate is they don't have a lot of population centers, and to get the 2011 map they had to do a lot of chopping. Making the districts larger makes their problem that much greater. Since this plan preserves UCCs it does about as well as they could hope by linking Rockford and Rock Island in a CD as well as by keeping Metro East intact and not trying to use it in two separate CDs, which didn't work so well in 2014.

It will be hard if not impossible to preserve 3 BVAP-majority CDs in 2020. I assume that there will be only two covered by the VRA. Similarly the case wasn't made for 2 CHVAP-majority CDs in IL in 2011, so I don't see it happening with larger districts in 2020. I'm reasonably confident that 2 black and 1 Latino CD could be nested in Cook in 2020 to comply with the VRA.

[muon2]

Your Cook plan is very similar to what I had in mind. I left the Beverly neighborhood of Chicago with the SW suburbs, which is where it usually is placed. My CD 4 is designed to pick up areas that are anticipated to have additional Hispanic growth by 2020 (58.3% HVAP in 2010, but probably over 60% by 2020) and that leaves CD 5 as an opportunity district (37.2% HVAP in 2010, but probably well over 40% by 2020).

I'll probably do a more detailed analysis at the township level when that data comes out in the summer. I wish the Census provided data based on Chicago community areas.

[muon2]

Muon, what is the partisan breakdown of those districts by 2008 numbers.  Just wondering.

1 & 2 are 91% Obama
3 is 60% Obama
4, 5 & 7 are 75% Obama
6 is 59% Obama

[muon2]

You are correct I didn't approach it with a view to a score. The underlying map is only at 5% variance since it is so early in the decade. As for the split, I looked at it the way a commission might after hearing public testimony. I really wanted to avoid a connection to the north side Hispanics for CD 4, and I think that by 2020 there's a decent chance that the north side Latinos may well be able to control a primary. As I said earlier, I expect I'll revisit this when I have better data at the township and neighborhood level.

[muon2]

Ah, local Illinois politics trumps your set of metrics. I understand.   Myself, I would just say no, but I'm a hardass.

Just out of curiosity, what is your genesis of your squeamishness about linking the two Hispanic nodes?

I have rarely sat through a panel that didn't decry the "earmuff" IL-4. I went to great lengths in 2010-11 to show that the earmuff wasn't necessary to meet the VRA standard established by the 7th Circuit. To the extent possible I hope to show that it isn't necessary for 2020 either, even with larger districts.

[muon2]

Not to beat the drum until it has no sound, but why was the "earmuff" so unpopular? I mean, your Hispanic CD is itself butt ugly erose, in fact to my "artistic" eyes more ugly than my little modest earmuff.

I'm not sure, but C-shaped districts seem to be particularly odious to the good government crowd. The distaste for that shape even colors the IA constitution.

You must be logged in to read this quote.

[muon2]

"Entirely separated" really is not text that accomplishes the prohibition of C shaped districts in my view. Entirely separated means two counties not connected at all to me. Is this some judicial or commission interpretation of the statute. Anyway, putting aside the erosity issue the way you drew the Hispanic CD, you are going to be in trouble with your nesting if the two African American CD's lose as much population as I assumed, because my IL-07 will then either need to take some of the territory used by the Hispanic CD, or grotesquely wrap around it, or chop into another county, or you are back to the earmuffs.

IA redistricting attorneys tell me that was the intent of that text since contiguity was already a requirement. Basically two counties connected by one or more others, but with no part of the district between them: ie the tips of a C. I found it interesting that it became a constitutional issue there.

The fastest growing area in Chicago has been in the areas next to the Loop, and since they are in CD 1 with the parts of the south and west sides losing the most, there should be some offset. I'll have to grind through ACS census tract files to see if that is holding up this decade. Further south I expect that the line between my CD 2 and 3 would just shift west if Orland township outpaces the other townships of south Cook.

The Census estimates for 2013 had Cook growing at about 0.3%/year. Those same estimates had Chicago growing at 0.3%/year and all the townships growing from 0.1% to 0.7% per year and no area stood out as losing population. Given the fall off overall in IL from '13 to '14 and the corresponding downward move in Cook (averaging 0.2% over four years instead of 0.3% over 3 years), I'll be interested to see if there is any notable changes in the townships.

[muon2]

Looking at the trends to 2020 from 2014 estimates in the plains states, I'll start with IA. Interestingly, IA-2 (light green) is growing at a rate so close to that of the entire state that the current projection has it at only -208 persons from the quota, so I left it alone. Most of the growth is in the Des Moines area and I project that it can't be kept with Council Bluffs in the same CD without some unusual shape, so I contracted the Des Moines CD to just 7 counties keeping the UCC together.

This plan has a maximum deviation of 1392 (0.17%) It's certainly possible to get smaller deviations in IA, but given the likely shifts in estimates over the next few years it wouldn't be very meaningful.

[muon2]

Indeed, when I was in Des Moines a couple of years ago I asked about that and I was told that population equality trumped other state provisions, though they could use whole counties as long as the equality was as close as practicable. I showed them a couple of whole county maps I drew that were even closer in equality then their map. My maps were so erose that they hadn't found them, and even if they did I hope they would have invoked their statutory provision on perimeter length to exclude them.

Adopted plan: deviations (-41, +35, +23, -18); average deviation 29.25; range 76.


muon2 alternate A: deviations (+17, +6, -15, -9); average deviation 11.75; range 32.


muon2 alternate B: deviations (+25, +1, -1, -26); average deviation 13.25; range 51.

[muon2]

Growth Trend Map

Nice map. Is this one year or since 2010, and is this growth relative to the national average?

[muon2]

MN was a state on the bubble to lose a seat in 2010, but they dodged a bullet and kept 8. The forecast for 2020 is that MN will lose a seat, though they are still on the bubble. I'll assume they drop to 7 as I use the 2014 data to project county totals in 2020.

The Twin Cities UCC will have about 4 1/4 seats in a seven seat map. This plan preserves the cover count of that UCC and keeps deviations within 5%. The Minneapolis area (orange) has 2 districts.

[muon2]

MO lost a CD in 2010, and it will stay at 9 for 2020 based on 2014 estimates. UCC covers are preserved and all districts are within 5% of the quota. All but the two KC districts are within 1%. The St Louis area (dark blue) contains two districts.

[muon2]

Here's OH divided into 15 CDs using the projections of the county data out to 2020. Districts preserve UCC covering rules. All districts are within 5% except suburban Cincinnati which is slightly over and is balanced by the adjacent southern OH district. The Columbus and Cleveland areas each contain two districts.

Perhaps I made an error, but with my own spreadsheet taking the 7-1-14 numbers and projecting out to 4-1-20 the trends, I find that your OH-04 (the 14 county red CD), is about 82,000 short in population.

The error was mine. The overpopulation of the Cinci suburban district (+5.8%) had to be picked up by an adjacent district or spread over a bunch of districts. Originally I had it picked up by keeping the western district low but then shifted to the southern district on my spreadsheet, but I didn't show it on my map. Madison county should be red and I'll update the post accordingly.

[muon2]

You're also not following the nesting rules for the Columbus UCC.  Seriously, I find that a most important metric myself, more than the coverage metric, which I am willing to fudge if it makes the map work much better. But two CD's need to be contained wholly within Franklin, Delaware, Licking, or Fairfield Counties in my view. You sprawl down to Pickaway, so thus the nesting metric is not followed.

You've got 79K extra folks in the Cincy UCC btw, so maybe the solution is to just have OH-04 cut into Butler County to pick up those extra 79K. That still leaves an extra 25K for the Dayton UCC to dump off somewhere. I guess that is what you were suggesting. Cutting Clermont makes for a prettier map, but maybe the numbers don't work as well. Even losing Madison, OH-07 has 23K extra folks, if it gets Cincy's excess, while the Butler chop gets one into microchop range for OH-04's other cut. To make the map look better, ideally OH-04 should pick up Hancock County.  OH-09 then drops Henry to OH-04 and picks up Defiance, and it becomes chopless, and then OH-04 can cut into Miami deep enough with the 0.5% variance rule, to get both OH-04 and OH-03 within the acceptable population range.

As I said at the outset, I'm not going to try to fit into a 0.5% limit on projections 6 years out. You might sell me on a 3% limit since estimate shifts can amount to about 0.5%/year. So I see this as a guideline to show where the ideal districts are likely to fall, with my projections converging to the rule as the decade progresses.

The problem I have with the pack rule in these projections is that it really requires knowledge of the county subdivisions since it inherently requires partitions. I know that for a final map there are going to be tradeoffs of UCC pack for chop and erosity. Given that, it's usually easier to assess the chop and erosity potential than the pack potential. For example one could create a more erose plan around Columbus that preserves the pack potential, but I don't see a reason to do that so far out from 2020, since the problem may solve itself by then.

[muon2]

Madison, Union, and Pickaway and not growing much in population, so they are not "urbanizing" at the moment.

I think that jimtex's point is that all three counties already have enough urban population to qualify, except that most of the population is in satellite urban clusters, not in the main urbanized area. It would not take a lot of growth for a tendril of Columbus to link to one of those clusters. Then at the 2020 Census the urban cluster would count as urbanized area and the county would meet the UCC criteria.

[muon2]

Madison, Union, and Pickaway and not growing much in population, so they are not "urbanizing" at the moment.

I think that jimtex's point is that all three counties already have enough urban population to qualify, except that most of the population is in satellite urban clusters, not in the main urbanized area. It would not take a lot of growth for a tendril of Columbus to link to one of those clusters. Then at the 2020 Census the urban cluster would count as urbanized area and the county would meet the UCC criteria.

Yes, I more or less understood that, but Madison at current rates is slated to pick up about 500 people, Pickaway 1,100, and Union 1,500.  That seems rather anemic to me.

Consider Pickaway where Ashville and South Bloomfield are in an urban cluster that was only separated by a couple of miles from the Columbus UA in 2010. Derby and Commercial Point are also quite near Columbus and were not yet large enough to form urban clusters in 2010. If the residential areas of those three townships with those communities got pulled into the Columbus UA, then that would place the urbanized population at 38% of Pickaway's total. It wouldn't take much more to get over 40%.

[muon2]

The Columbus UA extends into Pickaway County? What is the definition of an urban cluster?  Anyway, when I look at the precincts, the territory along the Franklin County line, looks like it is not very densely populated at all. Only the precinct north of Commercial Point seems to have more urban like density, which precinct while large in area, has 4,700 people. Whatever. Nobody but us will be paying attention to these definitions anyway in all probability.

There are small parts of Scioto and Harrison townships in the 2010 definition of the Columbus UA that had 2696 people. The Harrison piece extends down from the Rickenbacher airport which serves as a significant cargo and general aviation business airport just north of the county line, but stops north of Ashville. The Scioto piece extends south towards but not all the way to Commercial Point. See http://www2.census.gov/geo/maps/dc10map/UAUC_RefMap/ua/ua19234_columbus_oh/DC10UA19234_002.pdf

Part of projecting the 2020 maps is trying to discern changes the Census will make to its various geographic divisions. Besides our merry band of mapmakers, I can assure you that local governments will watching, too. There are federal grants that can depend on these definitions.